HARVARD UNIVERSITY Library of the Museum of Comparative Zoology dloes not circulate BREVIORA MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE, IN CAMBRIDGE Numbers 1-66 1952-1956 CAMBRIDGE, MASS., U.S.A. 1957 <^ 5° r CONTENTS BREVIORA Museum of Comparative Zoologv Numbers l-6fi 1952 No. 1. A new Panamanian tree frog. By Edward H. Tayi.or. 3 pp., 1 pi. February 8. No. 2. A staurotypine skull from the Oligocene of South Da- kota. (Testudinata, Chelydridae). By Ernest Wil- liams. 14 pp., 2 pis. February 8. Xo. :>. Notes on siphonophores. 3. N ectopyramis spinosa n. sp. By Mary Sears. 4 pp. May 23. No. 4. A unique remopleuridid trilobite. By H. B. Whitting- TON. 10 pp., 1 pi. June 9. No. f), Tricholimnas condificius is probably a synonym of Tri- cholhnnas sylvcstris. (Aves, Rallidae). By James C. Greenway, Jr. 4 pp. August 29. No. (). ( Contributions toward a reclassification of the Formicidae. I. Tribe Platythyreini (Hymenoptera). By William L. Brown, Jr. 6 pp. August 29. No. 7. On the nomenclature of the Pacific gray whale. By Wil- liam E. Schevill. 3 pp. September 29. No. 8. A new species of the cyclostome genus Paramyxine from the Gulf of Mexico. By Henry B. Bigelow and Wil- liam C. ScHROEDER. 10 pp. October 21. No. 9. New species of earthworms from the Arnold Arboretum. Boston. By G. E. Gates. 3 pp. October 21. No. 10. On the earthworms of New Hampshire. By G. E. Gates. 3 pp. December 29. 1953 X.). 11 Cliaracters and synonymies among the genera of ants. Part I. By William L. Brown, Jr. 13 pp. March 20. .\o. 12. A cave faunule from western Puerto Rico with a discus- sion of the genus Isolobodon. By Thomas E. Reyn- olds, Karl F. Koopman and Ernest E. Williams. 7 pp., 1 pi. April 23. No. 13. Fossils and the distribution of chelyid turtles. 1. "Hydraspis" leifhii (Carter) in the Eocene of India is a pelomedusid. By Ernest Williams. 8 pp., 3 pis. April 23. No. 14. fieeord of a hermaphroditic horseshoe crab, Limulus polyphemus L. By John P. Baptist. 2 pp., 2 pis. May 28. Further notes on the earthworms of the Arnold Arbore- lum, Boston. By G. E. Gates. 9 pp. May 15. Notes on the races of Micrurus fronialis (Dumeril, Dumeril and Bibron). By Benjamin Shreve. 6 pp. May 28. A new Ordovician trilobite from Florida. By H. B. Whittington. 6 pp., 1 pi. May 28. Characters and synonymies among the genera of ants. Part IT. By William L. Brown, Jr. 8 pp. Septem- ber 23. Lower Cretaceous nautiloids from Texas. By Bernharp Kummel. 11 pp., 2 pis. September 23. Lower Triassic Salt Range nautiloids. By Bernhard Kummel. 8 pp., 2 pis. September 23. The ancestry of the family Nautilidae. By Bernhard Kummel. 7 pp., 1 pi. September 23. Notes on the ant, Leptothorax ohliquicanthus Cole (Hy- menoptera : Formicidael. By Robert E. Gregg. 3 pp. October 13. No. 23. A new species of Hypognatha from Panama. By Arthur M. Chickering. 8 pp. October 23. No. ^:>. No. l(r No. 17. No. 18. No. 19, No. 20. No. 21. No. 22 1954 No. 24. A new family, a new genus, and tAvo new species of batoid fishes from the Gidf of Mexico. By Henry B. BiGELOw- and William C. Schroeder. 16 pp. Janu- ary 27. No. 25. A new Miocene species of I'clusios and the evolution ol' that genus. Bj'- Ernest Williams. 7 pp., 4 pis. Janu- ary 28. No. 26. A preliminary list of the earthworms of northern New Jersey with notes. By H. Da vies. 13 pp. February 3. No. 27 Anterior regeneration in a sexthecal species of lumbri- cid earthworm. By G. E. Gates. 5 pp. February 5. .\o. 28. Clemmydopsis Boda a valid lineage of emj^dine turtles from the European Tertiary. By Ernest AYilliams. 9 pp. February 8. No. 29. Absence of mesoplastva in a Pelomednsa (Testudines. Pelomedusidae). By Ernest Williams. 4 pp.. 2 pis. February 8. No. 30. Aestivation in a Permian lungfish. By Alfred S. Romer and Everett C. Olson. 8 pp., 1 pi. February 8. No. 31. New freshwater gastropod mollusks of the African genus Lanistes. By T. Pain. 4 pp. March 3. No. 32. Fossils and the distribution of chelyid turtles. 2. Addi- tional reputed chelyid turtles on northern continents : Palaeaspis conyhearii (Owen) — a pelomedusid. By Ernest Williams. 6 pp. March 12. No. 33. Systematic and other notes on some of the smaller species of the ant genus Rhytidoponera Mayr. By William L. Brown, Jr. 10 pp. May 14. No. 34. A review of the coxalis group of the ant genus Sticto- ponera Mayr. B}^ William L. Brow^n, Jr. 10 pp. July 20. No. 35. New or rediscovered pelomedusid skulls from the Tertiary of Africa and Asia (Testudines, Pelomedusidae). 1. Dacquemys paleomorpha, new genus, new species from the Lower Oligocene of the Fayum, Egypt. By Ernest Williams. 8 pp., 1 pi. July 28. No. 36. Present knowledge of the snake EJachistodon wester- manni Reinhardt. By Carl Gans and Ernest E. Williams. 17 pp. August 6. No. 37. On the evolution of an oriental earthworm species, Pheretima anomala Michaelsen 1907. By G. E. Gates. 8 pp. August 18. No. 38. Onnia (Trilobita) from Venezuela. By H. B. Wnrr- TixGTON. 5 pp., 1 pi. November 28. No. 39. New or redescribed pelomedusid skulls from the Terti- ary of Africa and Asia (Testudines, Pelomedusidae). 2. A podocnemide skull from the Miocene of Moghara. Egypt. By Ernest Williams. S pp., 2 pis. Novem- ber 24. No. 40. Some moUusks from the continental slope of northeast- ern North America. By Arthur H. Glarke, Jr. 11 pp. November 29. No. 41. Gomments on the classilication of rodents. By Albert E. Wood. 9 pp. December 17. 1955 No. 42. A new salamander of the genus Parvimolge from Mexico. By George B. Rabb. 9 pp. February 28. No. 43. Speed-induced skin folds in the bottle-nosed porpoise. Tursiops truncatus. By Frank S. Essapian. 4 pp.. 9 pis. April 17. No. 44. A new Murex from Matanzas, Cuba. By William J. Clench. 2 pp., 1 pi. April 8. No. 45. Palaeotaricha oligocenica, new genus and species, an Oligocene salamander from Oregon. By Richard VAN Frank. 12 pp., 3 pis. June 13. No. 46. Cave-fossil vertebrates from Camaguey, Cuba. By Karl F. KooPMAN and Rodolfo Ruibal. 8 pp. June 24. No. 47. A new species of whiptailed lizard (genus Cnemido- phorus) from the Colorado plateau of Arizona, New Mexico, Colorado, and Utah. By Charles H. Lowe. Jn. 9 pp. July 12. No. 48. Notes on Aineric-an earthworms of the family- Luni- bricidae. I-II. By G. E. Gates. 12 pp. October 10. No. 49. Three new shark records from the Gulf of Mexico. By Heney B. Bigelow, "W. C. Schroeder, and Stewart Springer. 12 pp. November 18. No. 50. New fro^s of the genera Asteruphrys and Oreophryne from New Guinea. By Arthi^r Loveridge. 5 pp. November 30. 1956 No. 51. A small mustelid from the Thomas Farm Miocene. By Stanley J. Olsex. 5 pp. January 27. No. 52. Remarks on some Miocene anurans from F^lorida, with a description of a new species of Hyla. By Walter Auffenberg. 11 pp. April 6. No. 5;l Food-finding by a captive porpoise [Tursiops trun- catus). By William E. Schevill and Barbara Law- rence. 15 pp. April 6. No. 54. A revision of the genus Braclupneles (Scincidae), with descriptions of new species and subspecies. By AV alter C. Brown. 19 pp. June 6. Rediscovery otilyla dorsalis and Lechriodus papuanu^ in New Guinea. By Arthur Loveridge. 4 pp. June 6. Notes on the Jamaican and Cajonan Island lizards of the genus Celesius. By Penny Norseen Cousens. 6 pp. June 15. Three new species of ^limetidae lAraneae) from Pan- ama. By Arthur M. Guickering. 14 pp. June 29. Sequence of Passerine families (Aves). By E. Mayr and J. C. Greenway, Jr. 11 pp. June 29. A new subgenus of Chamacleo fi-om Rhodesia and new race of Mahuya from Kenya Colony. By Arthur Loveridge. 4 pp. September 12. No. 60. A new species of Agriognatha from Jamaica, B.W.I. By Arthur M. Chickering. 7 pp. September 12. No. 55. No. 56. No. 57. No. 58. No. 59. No. 61. On regeneration by earthworms of a species of the lum- bricid genus Dendrohaena Eisen 1874. By G. E. Gates. 6 pp. September 14. Xo. 62. A third leaf -nosed species of the lizard genus Anolis from South America. By James A. Peters and Gus- tavo Orces-V. 8 pp. October 3. No. 63. New bathyal Isopoda from the Caribbean with observa- tions on their nutrition. By Robert J. Menzies. 10 pp. October 11. No. 64. Rare species of Copepoda, Calanoidea, taken from the Tzu Region. By Otoiiito Tanaka. 8 pp. October 11. No. 6d. a new species of Agriognatha (Araneae, Argiopidae from Panama. By Artiiitr M. Chickering. 7 pp. November 20. No. 66 The Caninae of the Thomas Farm Miocene. By Stan- ley J. Olsex. 12 pp. November 27. INDEX OF AUTHORS BREVIORA MUSEUM OF COMPARATIVE ZOOLOOY Numbers 1-66 1952-56 No. AUFFENBERG, WALTER 52 Baptist, John P. 14 BiGELOW, Henry B. and William C. Schroedeb 8, 24 BiGELOw, Henry B., W. C. Schroeder and Stewart Springer 4!> Brown, Walter C. 54 Brown, William L., Jr 6, 11, 18, 33, 34 Chickering, Arthur M 23, 57, 60, 65 Clarke, Arthur H., Jr 40 Clench, William J 44 Cousens, Penny Norseen 56 Davies, H 26 EssAPiAN, Frank S 43 Cans, Carl and Ernest B. Williams 36 Gates, G. E. y, 10, 15, 27, 37, 48, 61 Greenway, James C, Jr 5 Greenway, J. C, Jr. and E. Mayr 58 Gregg, Robert E. 22 KoopMAN, Karl F., and Rodolfo Ruibal 46 KooPMAN, Karl F., Thomas E. Reynolds and Ernest E. Williams 12 Kummel, Bernhabd 19, 20, 21 Lawrence, Barbara and Willla.m E. Schevill 53 LovERiDGE, Arthur 50, 55, 59 Lowe, Charles H., Jr 47 Mayr, E. and J. C. Greenway, Jr 58 Menzies, Robert J . 63 Olsen, Stanley J. 51, 66 Olson, Everett C. and Alfred S. Romer 30 Orces-V., Gustavo and James A. Peters 62 Pain, T 31 Peters, James A. and Gustavo Orces-V 62 Rabb, George B 42 Reynolds, Thomas E., Karl P. Koopman and Ernest E. Williams 12 Romer, Alfred S. and Everett C. Olson 30 RuiBAL, RoDOLFO and Karl F. Koopman 46 Schevill, Willlv-m E 7 ScHEViLL, William E. and Barbara Lawrence 53 ScHROEDER, WiLLiAM C. and Henry B. Bigelow 8, 24 Schroeder, W. C, Henry B. Bigelow and Stewart Springer 49 Sears, Mary 3 Shreve, Benjamin 16 Springer, Stewart, Henry B. Bigelow and W. C. Schroeder 49 Tanaka, Otohito 64 Taylor, Edward H 1 VAN Frank, Richard 45 Whittington, H. B. 4, 17, 38 Williams, Ernest 2, 13, 25, 28, 29, 32, 35, 39 Williams, Ernest E. and Carl Gans 36 Williams, Ernest E., Karl F. Koopman and Thomas E. Reynolds 12 Wood, Albert E 41 E V I O R A Miiseiiim of Coimparative Zoology Cambridge, Mass. February 8, 1952 Number 1 A NEW PANAMANIAN TREE FROG By Edward H. Taylor Department of Zoology, rriiversity of Kansas, Lawrence, Kansas While studying ( "entral American frogs in the Museum of ( "om- parative Zoology, Harvard College, a small, presumably undescribed frog was noted, and Mr. Arthur Loveridge, with his characteristic kindness, suggested that I study the form and describe it if it proved new. The most characteristic of the differential features of this di- minutive frog is the very considerable enlargement of the horny spines on the nuptial swelling of the first finger. It is presumed that it is a rivulet species, living in the neighborhood of small streams, rather than a bromeliad species. After my study of the specimen and comparison with other known species of southern ( "entral America, I conclude that the species is new and I present herewith its description. HyLA SHREVEl Sp. nOV. Tiipr. M.C.Z. No. 2(i7()9 cf . La Loma, Chiriquicito, Republica de Panama; K. R. Dunn and ('. Duryea, collectors. Diagnoftis. \ diminutive frog (25.2 mm.) with a br(nid, rather flat head ; skin smooth ; outer fingers with a trace of web, and well-developed discs; canthus rostralis wanting; nostrils nearly terminal; tympanum large directed strongly upward, its diameter at least half of eye length; 2 BREVIORA NO. 1 eye directed obliquely forward; \'onierine teeth in two rounded fasci- culi, lying almost completely behind posterior level of large choanae; no web between two inner toes; others one third, to one half webbed; first finger of male with 19-22 large horny spines. Description of the ti/pe. Head flattened, broader than body, its width (10 mm.) greater than its length (9.1 mm.); tympanum large, semi- transparent, pigmented, facing more upward than outward, the upper edge obscured by a fold arising at eye and curving back to jaw angle, its diameter (1.5 mm.) equal to half length of eye opening (3 mm.); distance between tympanum and eye 2.8 mm.; eye prominent, directed somewhat forward; length of snout less than eye length (2.85 mm.); canthus absent, loreal region concave, sloping obliquely to lip; nostrils nearly terminal, the areas about them strongly swollen with a de- pression between; no notch in upper lip; groove from angle of jaw touches tympanum; choanae subcircular, the distance between them 2 mm.; diameter (transverse) of choana .65 mm.; vomerine teeth in two somewhat circular fasciculi almost completely behind the posterior level of choanae, the teeth in somewhat curving rows; they are ecjually separated from choanae and from each other, a fasciculus as large as a choana; tongue rather thick, broader than long, not or scarcely notched behind; vocal slits very large, lateral; openings of the palatal gland forming a series of separate slitlike openings arranged in a broad \ -shape about the middle of palate. Arm very short, the wrist not reaching tip of snout; first finger short, greatly expanded at base by a very large nuptial swelling and a large metacarpal tubercle; swelling covered with a black, horny plate bearing 19-22 large horny spines; discs of fingers smaller than tympa- num, that on first finger only little smaller than those on outer fingers; a small trace of webbing between outer fingers; subarticular tubercles single; supernumerary tubercles numerous; palmar tubercle partly divided; a slight axillary web present; heel reaches forward to edge of eye; tarsal fold represented by a series of tubercles growing more distinct on the proximal end of tarsus; a large inner metatarsal tubercle and a small distinct outer; no web between first and second toes; one fourth webbed l)etween second and third; approximately half webbed between the third and fourth and two fifths webbed be- tween the fovu-th and fifth, the web forming a slight ridge to discs on sides of digits; subarticular and supernumerary tubercle distinct. Skin on dorsum smooth, slightly wrinkled or corrugated on sides; chin smooth; breast, venter, part of the ventral surface of thighs and region lateral to vent, strongly granular; a rather elongate anal flap, the vent opening midway l)etween upper and lower surface of thighs, followed 1052 NEW TREE FROG 3 l)y a groove bordered by small granules and on ventral surface by a pair of distinctly larger granules; fold above tympanum curves down to jaw angle. Color. In preservative, light l)rown above with a brownish-white shade below; the pigment is in minute chromatophores, that can be clearly seen under a lens; no pigmentation on under surfaces except on edge of jaw, under surface of thigh, part of under surface of tibia and some scattered flecks under hand and foot ; upper eyelids appearing dark from the dark covering of the eyeball; tip of snout somewhat darker than dorsum. Measurements in mm. Snout to vent 25.2; width of head 10; length of head 9.1 ; diameter of tympanum 1.5; length of eye opening 3; snout length 2.85; arm 13; hand 7.2; leg 38; tibia 12; foot and tarsus 15. Remarks. The species is probably related to Hyla zetcki There are numerous similarities in structure, and the measurements are similar but the most striking differences appear to be in the remarkable nuptial asperities on the first finger, the position and direction, and the larger size of the tympanum (^ instead of ^s oi eye); the absence of the spectacle-like markings about eyes. The characteristic anal decoration of zeteki is absent, and the strong granulation on under surface of arm, and on sides is absent. The species is named for Mr. Benjamin Shreve of the Museum of Comparative Zoology, Harvard College. NO. 1 «Ov- Hyla shrevci up. n.uv. AI.C'.Z. Xu. 20709, La Lonia, Chiruiuicitu, Republicu de Panamd. E. R. Dunn and C. Duryea, collectors. (Type d^ X 2.) 1. Dorsal view. 2. Lateral view of head. 3. Enlargement of nuptial spines on inner digit. /Uc'^i- /K r B R E V I Mesemm of Comparsitive Zoology Cambridge, Mass. February 8, 1952 Number 2 A STAUROTYPINE SKULL FROM THE OLIGOCENE OF SOUTH DAKOTA {Testudinata, Chelydridae) By Ernest Williams INTRODUCTION The modern representatives of the Staurotypinae (two genera, Staurotypus and Claudius, and three described species) are restricted to Southern Mexico and Central America. No fossils belonging to these living genera are known. A form known only from a single perfect shell from the Oligocene (Chadron) of South Dakota, Xenochelys formosa Hay, seems to be the only previously described fossil record of this subfamily. Recognition of a staurotypine from shell material as good as that of the unique type of Xenochelys (A.M.N.H. No. 1097) is quite easy. Staurotypines differ from chelydrines in having only 23 instead of 25 marginals and 21 rather than 23 peripherals. The same feature dis- tinguishes them from dermatemydids. They differ from kinosternines in possessing an entoplastron. In these characters Xenochelys is clearly staurotypine. In the elongation of the first vertebral scute and in its precise shape Xenochelys closely resembles Staurotypus. The neural series is much as in Staurotypus. The plastron of Xenochelys has a reduced number of shields, the pectorals and abdominals having ap- parently fused. This again is a staurotypine or chelydrine feature. The general form and height of the shell are very Staurotypus-Wke. There is also some suggestion of the tricarinate condition found in Staurotypus in the carapace of Xenochelys. 2 BREVIORA NO. 2 In other respects the shell of Xcnochclys is more primitive (more dermatemydid-like) than any other member of the chelydrid series. The plastron is relatively large (although the bridge is quite narrow as in chelydrids, not as in dermatemydids). There is a xiphiplastral notch. The nuchal scute is small. There is a trace of nuchal indentation.^ In shell characters, therefore, Xcnochclys is a good structural inter- mediate between Dermatcmys and Staurotypus, and it is also a temporal and phylogenetic intermediate between the complex of fossil forms called dermatemydid and the staurotypine section of the Recent Chelydridae. It should be noted that Hay was not under any misapprehensions as to the affinities of Xcnochclys. He quite explicitly cited Staurotypus and Claudius as "the nearest living relatives of Xcnochclys" (1908, p. 282), and in his assignment of the form to the Dermatemydidae he merely followed Boulenger's 1889 catalog of the Recent turtles in which the Staurotypinae were associated with the Dermatemydidae. When he published his 1908 monograph he had probably not seen Siebenrock's 1907 monograph of the "Cinosternidae" in which the close affinities of the Staurotypinae and Kinosterninae were for the first time made clear; he thus missed an opportunity to point up more clearly the apparent ancestral position of the South Dakota genus. In my 1950 classification of the testudinate order I united the chelydrines, staurotypines and kinosternines in the family Chelydridae as understood in a broad sense. To do so somewhat obscures the ex- treme closeness of relationship of staurotypines and kinosternines, which differ in the simultaneous loss of an entoplastron and acquisition of more or less of a box turtle habitus by the latter. The chelydrid series is surely a natural group, but within that group the staurotypines and kinosternines stand very much closer to one another than to the chelydrines. In the Oligocene this specialized section of the Chelydridae scnsu lato was evidently fully distinct in shell characters, if still somewhat primitive in a few features. It has, however, not previously been known to what extent the skull had evolved concomitantlv with the shell. In the collection of the Department of Geology, Princeton Uni- versity, I have now found a skull (No. 13686), likewise from the Chadronian Oligocene of South Dakota, which surely belongs to a member of the staurotypine-kinosternine section of the Chelydridae. It is distinctly more specialized in a number of ways than any previ- ' Most of these characters might also be counted as kinosternine resemblances. 1952 A STAUROTYPINE SKULL 3 ously known staurotypine or kinosternine skull. Although incomplete it merits extended description and discussion. In the section which follows, the skull is described in detail and compared with Claudius augustatus, Staurotj/pus salvhiii, and Stcrnothenis carinatus. DESCRIPTION OF THE PRINCETON SKULL The skull is complete as far as the postorbital bar. Behind this, however, only the parietals, pterygoids, basisphenoid and basioccipital are retained, all of them somewhat fragmented and incomplete. Sutures are rather difficult to make out because of fractures in critical areas. The profile of the face is strikingly like that of Stauroiypus or Stcrnotherus. The prefrontals project dorsally above the narial opening but laterally are somewhat retracted, so that in lateral view the nostril is seen as a distinct angular indentation. The premaxillae unite in the formation of a distinct median beak, while posterior to this median projection the contours of premaxilla and maxilla form a smoothly sinuous curve, which, however, is doubly incurved in the fossil in contrast to both Staurotypvs and Sternothents. The depth of the premaxilla is markedly greater than in Stavrotypvs. The orbit is rela- tively smaller than in Siavroiypvs and even more distinctly lateral (in contrast to dorsolateral) than in that form. The maxilla below the orbit is marked by a distinct groove running down to the second incurving of the lateral festooning of the jaw. The postorbital bar is rather wide, half the rostro-caudal length of the orbit rather than between V3 and ^-^ as in Staurotypus. In lateral view the prootic is seen to project far foi-ward about the pterygoid in a very exceptionally developed "crista praetemporalis" (Siebenrock 1897). Only a narrow channel separates these two bones. The dorsal margin of the prootic is continued also in a ridge on the parietal. x\ similar forward projection of prootic is seen in Stauroiypus (and in Graptcmys) but is far less extreme. The "crista praetemporalis" which is the feature exaggeratedly de- veloped in the fossil and in Stauroiypus and Graptcmys serves to modify and increase the leverage of the jaw muscles (Zdansky 1924, pp. 101-104). All three genera have widened alveolar surfaces of the maxilla with strong tendency to formation of a secondary palate. In the case of Graptem-ys the wide alveolar surfaces are known to be associated with a malacophagous diet. Probably in the case of the fossil a strengthening of the jaw action and a similar diet are to be inferred. 4 BREVIORA NO. 2 The dorsal planum of the parietal continues the remarkably flat dorsal margin of the face. In anterior view the nostril is very small, much as in Sternotherus, not as in Staurotypus. The prefrontals are anteriorly distinctly convex, posteriorly are very flat. Behind them the frontoparietal area rises somewhat more sharply than in Staurotypus. In palatal view the maxillae have united in the formation of an extensive secondary palate, complete in front with a pronounced median ridge but incomplete in the midline behind. In this secondary palate the palatines share to about the same extent as in Staurotypus and to a significantly greater extent than in Sternotherus. A striking and unique feature is the very impressive dorsomedial slope of the palatal roof. To a very slight extent this condition is prefigured in the other genera, particularly Staurotypus, but the difference is very con- siderable: in this respect no other genus is at all close. There are no ectopterygoid processes and the pterygoids also are bowed dorsally at the midline, so that their lateral flanges are very strong and high, though spreading wide apart. Anteriorly the premaxilla has the deep pit for the tip of the lower jar characteristic of chelydrids. In the specimen it breaks through into the narial region. As in Staurotypus and Claudius (differing in this from all examined kinosternines and chelydrines), foramina incisiva appear to be lacking in the fossil. At the postorbital bar the palate, and thus the whole outline of the skull, is very expanded from the side: this form must have been decidedly brachycephalic. The waist of the pterygoids, however, is only moder- ately broad, as in Stcrnothcrxis, not very broad as in Staurotypus or narrow as in Claudius. The basisphenoidal suture is not clearly dis- tinguished from breaks in this region, but it seems probable that the exposed portion of this bone was very short and broad, not tending to be elongate craniocaudally as in Staurotypus. The infracondylar depression, so marked in Stauroiypiis, is less distinct in this form, as also in Sternotherus, but not so weak as in Claudius.. In dorsal view the great breadth of the skull at the postorbital bar is again evident, along with the considerable incurving of the skull contours just in front of the orbits. The origin of the supraoccipital spine is indicated by a triangular plane surface with well defined lateral margins, as in Staurofj/pus- and Stertlother^ls, not as in Claudius. The sutures bounding the frontals are somewhat obscured by breaks, but it is extremely probable that as in other (helydridae the frontals occupy a very small area and do not enter the orbits. 1952 A STAUROTYPINE SKULL 5 SKULL CHARACTERS IN THE CHELYDRIDAE AND THE ALLOCATION OF THE FOSSIL For the comparisons made in the course of the description just given I have had available the skeletal collections of the Museum of Com- parative Zoology and of the American Museum of Natural History. Included in the M.C.Z. collection are a young and an adult skull of Staurotypus (the young specimen, M.C.Z. No. 4989, is figured), while the A.M.N.H. collection has furnished for study a skull of Claudius (A.M.N.H. No. 65865).^ Because the fossil is a fragment only, comparisons must remain incomplete and portions of the skull which might be diagnostically significant are unavailable. In this circumstance and in the absence of more complete knowledge of variation and difference within the genera Kinosternon and Sternothcrus (valuable information which we may hope to learn in the forthcoming revision of these genera by Dr. Norman Hartweg) I do not attempt to discriminate too finely the affinities of the fossil. However some rather general discussion is possible. We may first consider what characters define a skull as chelydrid in the broad sense, then what features are chelydrine, staurotypine, or kinosternine, and finally what provisional allocation of the fossil skull is possible and expedient. There are six genera of living chelydrids {Chelydra, Macrodeviys, Staurotypus, Claudius, Kinosternon, Stcryiotherus), and of these Sterno- therus might quite properly be relegated to the synonymy of Kino- sternon. The fossil record adds a few more {Acherontemys, Chelydrops, Chelydropsis, Xenochelys). In contrast the Testudinidae has about 30 living genera while the fossil record brings the count up to about 50. It is not surprising, therefore, that the Chelydridae seem a more closely knit group than do the Testudinidae, even if the rather isolated Platysternon is omitted from the latter assemblage. If the skulls of chelydrids (all living forms North America, a few Tertiary fossils European) are compared only with the skulls of North American or European testudinids very clear distinctions are evident. If, however, the comparison is extended to the very rich testudinid fauna of Southeast Asia where, so far as known, chelydrids have never occurred, some of the forms to be found there bridge over the differences which were thought to be significant. Even so astute an observer as Baur, and one so familiar with the testudinate order, found it possible I A (liscimsion of the antitoniy of Claudiut is in preparation. 6 BREVIORA NO. 2 to refer Adelochclys (= Orlitia) to the "Chelydroidea" when he had the skull only, though the shell would have placed the genus without question in the Emydinae. The distinctions between the Chelydridae and Testudinidae are wide enough, when all parts of their anatomy are taken into account and the trends within them are considered, that there is no doubt that, although related, they have long been separated, perhaps since the Cretaceous, certainly since earliest Tertiary. Thus there are differ- ences in cervical formula (Williams 1950), in the form of the eighth cervical vertebra, in degree of development of the costiform processes of the nuchal, in the presence versus absence of inframarginals, in the characteristic reduction of the plastron or its elements in the Chely- dridae, in the proximal end of the femur, and in the absence in chely- drids of gaudy or bright patterns. However, in a way which seems very characteristic of turtles, few of these features hold good with complete fidelity in all cases. In cervical formula there is a striking difference between the two families in that the eighth cervical is biconvex in the Testudinidae, procoelous in the Chelydridae. A few individuals of the Testudinidae (mostly advanced tortoises) vary in the direction of the condition of the Chelydridae, but no chelydrids are known to vary in the direction of the testudinid condition. The eighth cervical vertebra tends to differ in the two families, the Testudinidae generally showing tliree ventral crests on the centrum, though the lateral ones may be barely indicated, the Chelydridae showing a single median crest which may divide into two (some Chdydra, kinosternines). The costiform processes of the nuchal are typically much developed in the Chelydridae, relatively little developed in adult testudinids, but the kinosternine section of the chelydrids approaches the condition of the testudinids, and young emydines have this process rather strongly developed. Inframarginals are never normally present in most testu- dinids, but there is an exception in the case of Platysternon, and inframarginals do occur as individual variations in Chrysemys picta (A.M.N.H. specimens to be reported on by Samuel McDowell). The plastron is never reduced in testudinids; it is strikingly reduced in protective efficiency in chelydrines and staurotypines; in some species of the kinosternines it regains its complete coverage of the ventral surface, but prior to this redevelopment it had lost one of the bones normally present in unreduced plastra. In the testudinid femur a fusion of the trochanters tends to limit the intertrochanteric fossa to a shallow dorsal pit, though a number of emydines (and Platysternon) 1952 A STAUROTYPINE SKULL 7 have the juncture barely suggested. In chelydrids as in most turtles the intertrochanteric fossa is a widely open groove. The lack of bright patterns in the Chelydridae, while characteristic enough of the rela- tively few genera and species involved, is not consistently contrasted with the presence of such patterns in the testudinids (see, for example, the uniform pigmentation of Galapagos tortoises and of some of the Asiatic emydines.) Nevertheless, in spite of these exceptions, these postcranial charac- ters, external and internal, permit in combination a clear discrimination of the two families. Several characters distinguish the skulls of chelydrids and testu- dinids, but here even more than in postcranial characters exceptions reduce the utility of single characters. Chelydrids usually possess at least traces or indications of a sharp median beak or "tooth" on the upper jaw. But this is absent in some kinosternines, and while many testudinids have a notch here, a few (e.g., Terrapcne, Cvora) have a beak quite similar to that of chelydrids. Most testudinids (but not Mcdaycmys) have the temporal bar deeply emarginate from below ; the chelydrids have this bar at most shallowly emarginate. Chelydrids have the nostril, orbits, and otic opening somewhat smaller than is frequently the case in testudinids. The supraoccipital crest is higher or more steepl}^ arched, the premaxillary pit is usually deeper, and in staurotypines and kinosternines there is a more marked festooning of the contours of the upper jaw than occurs in testudinids. The frontals are always small and excluded from the orbit in chelydrids; this feature is variable in testudinids. A combination of most of the characters mentioned defines a chely- drid skull. The absence of all but one or two, most often of all, defines a testudinid as contrasted with a chelydrid skull. Within the Chelydridae determination of chelydrine skulls from staurotypine and kinosternine skulls is at once possible on one key character which offers no difficulty. All chelydrines have the temporal region more fully covered than do any of the more advanced genera. Discrimination of staurotypine as against kinosternine skulls is more difficult on the basis of any characters which have the smallest proba- bility of holding good if more genera are discovered. Perhaps the prema,xillary beak is always more strongly developed in staurotypines and the temporal bar narrower vertically in the same group. This difficulty in finding differences in the skulls of these two groups is akin to the difficulty in distinguishing dorsal shells. The shells of both subfamilies differ from those of chelydrines in the loss of one pair 8 BREVIORA NO. 2 of marginal scutes and one pair of peripheral plates. Shape and height of the shells and scute shape are essentially the same. Keeling is variable. Only in the plastron is a key diflference at once apparent in the absence of an entoplastron. The mobility of anterior and posterior plastral lobes in kinosternines is another differentiating feature. In which of these groups does the fragmentary Oligocene skull find its natural place? It is clearly chelydrid sensu lato. It has the sharp premaxillary beak and deep premaxillary pit of a chelydrid. The temporal bar is broken but there is no suggestion of ventral emargination. The nostrils and the orbits are quite small. The supraoccipital crest is only partially preserved but its root gives evidence of a high arch as in typical chelydrids. The skull is, however, clearly not chelydrine, since the temporal region is fully exposed by posterior emargination as in the two ad- vanced subfamilies, not as in chelydrines. But the postorbital bar is wider than in any staurotypine or kinosternine; this may be a primitive feature. Is it staurotypine or kinosternine? It has resemblances to both groups. The premaxillary beak is very strong as in staurotypines, but the temporal bar is very stout in vertical depth as in kinosternines. The pterygoid waist is moderately broad as in kinosternines, not very broad as in Staurotypus or very narrow as in Claudius. The nostril is very small as in kinosternines and in contrast to the condition in staurotypines. Foramina incisiva are lacking as in staurotypines. Some features, however, are extremely specialized. The degree of development of the secondary palate is greater than in any presently recognized chelydrid species, significantly greater than in Staurotypus. Quite unique (unique in the order) is the extreme obliquity and dorsal arching of the secondary palate. All in all, the skull seems more specialized than that of any living staurotypine, but at the same time more primitive in at least one respect (the strong premaxillary beak) than any living kinosternine, and perhaps more primitive in the width of the postorbital bar than either modern staurotypines or modern kinosternines. The skull is Oligocene in age. The only known shell to which it might belong is Xenochelys, of the same age and not distant in locality. But the shell of Xenochelys is quite primitive for its group. Can so specialized a skull be assigned with any probability of correctness to so primitive a shell? 1952 A STAUROTYPINE SKULL 9 Such an association is by no means impossible. Indeed, in the Princeton skull a few features like the very strong premaxillary beak and the wide postorbital bar may point to a stage of differentiation not very different from that of the Xenochclys shell : advanced in some respects, primitive in others. For the present it seems expedient to refer the Princeton skull — with doubt — to Xenochelys formosa Hay. THE GEOLOGIC RANGE OF THE CHELYDRIDAE The Princeton skull and the American Museum shell of Xenochelys formosa are the oldest known representatives of the Chelydridae (Chadronian Oligocene). An older fossil from the Eocene of Tunis was indeed referred to the family by De Stefano (1903), but his de- scription was based on a few bone fragments associated with the mold of three pleurals. The generic and species names, Gafsachclys phospha- tica, erected on this very insufficient basis may be disregarded as a nomen vanum and need no longer be considered as part of the fossil history of the Chelydridae. It is, of course, surprising that the oldest representatives of the family should be staurotypine rather than chelydrine. It is, however, possible that future more complete knowledge may connect some of the other forms called dermatemydine by Hay (1908) with the Chely- dridae. The type species of Hoplochelys Hay was first called Chelydra crassa by Cope, and this genus, though possessed of a full complement of marginals and peripherals was regarded as possibly related to Staurotypus by Hay himself. Both this genus (Paleocene of North America) and Baptcmys (Lower and Middle Eocene of North America) have the plastron considerably reduced, the bridge narrowed and the posterior lobe pointed, and an arrangement of plastral scutes like that of Recent Chelydra (pectorals meeting femorals). The shape of the shell in both genera is quite like that of staurotypinesandkinosternines, and in Hoplochelys the shell is tricarinate, as it is also in one species of Baptemys. The hiunerus of Baptemys tricarinata is very like that of Chelydra. However, the first vertebral is never elongate as in stauro- typines and kinosternines. The other vertebrals are never as wide as in chelydrines, though wider in later (Torrejon) than in earlier (Puerco) Hoplochelys. The costiform process of the nuchal is said to be short in Baptemys (Hay, 1908). The skull of Baptcmys wyomingensis is known and is quite un-chelydrid in its major characters: the temporal bar is deeply emarginate from below, and the temporal region widely open above, there being a much greater caudal emargination than in 10 BREVIORA NO. 2 chelydrines. The postorbital bar is in consequence relatively narrow. The orbit also is rather large. None of the conditions just mentioned in which Baptemys and Hoplochelys differ from chelydrids positively debars them from ancestry. For the present, however, and until they are better known and tran- sitional forms are discovered, it seems convenient to retain them in the Dermatemydidae, merely calling attention to their possible special relation to the Chelydridae. Unless these forms, perfectly suitable in age, are ancestral chelydrids, there is no record of the family until the early Oligocene, and it is then first represented, as the shell and the referred skull of Xenochelys formosa show, by an advanced subfamily. The first occurrence of apparent Chelydrinae is in Europe and later in the Oligocene. Fragments of doubtful value from the middle Oligocene of Germany have been assigned to "Chelydra sp." by Reinach (1900), and in the later Oligocene of Germany rather good remains are found of an undoubted chelydrine, ''Chelydra" dccheni v. Meyer. Reference of the latter form to the Recent genus Chelydra is, as Zangerl (1945) has already pointed out, extremely doubtful: though the shell shape is that of a chelydrine, there are curious resemblances to Staurotypus and to Macroclemys rather than to Chelydra. Indeed, H. V. Meyer himself in 1852 expressly admitted that the inclusion of this form in Chelydra depended upon a very wide generic concept, and his idea of the genus was very definitely much wider than that current today. In the Miocene of both Europe and North America there is a flowering out — real or apparent — of chelydrine types. In Europe "Chelydra" miirchisoni Bell and four other named species of "Chelydra" and Chelydropsis earinata Peters record the rather widespread oc- currence of the subfamily. The remarks above for "C." dccheni apply also to "C." murchisoni and the other European forms referred to "Chelydra." The carapace of Chelydropsis earinata has been excellently figured by Peters (1869). Unfortunately it does not seem possible to verify in his figure the features upon which Peters relied in distinguishing this form generically. I am unable to interpret the plate as showing the presence of supramarginals, and while a division of the nuchal bone into two parts is clearly shown, I (as also Boulenger in 1889) doubt that this reflects the normal condition of the animal. Never- theless, I consider it probable, if only on the grounds of zoogeography and age, that the genus will stand, though requiring redefinition. It 1952 A STAUROTYPINE SKULL II may at least be pointed out that if the eventual much needed restudy of the European chelydrines should reveal that they all belong to one genus, the name Chdydropsis is available. From the Miocene of North America tliree chelydrines have been de- scribed. Two are known from skulls only: Chelydrops strida Matthew and Macrodemys schmidti Zangerl. Both of these are from Nebraska, M. schmidti from the Middle Miocene, Chelydrops strida from the Upper Miocene. Both are certainly related to Recent M. temminddi, but they are distinct from that form and from one another. Chelydrops is unique among known chelydrids in having a ridged alveolar surface of the maxilla.^ M. schmidti differs from Chelydrops and from M. tcm- minckii by the considerably shorter antorbital portion of the face. The other Miocene North American form (from the Roslyn Miocene of Washington) is known from the carapace only, no portion of the plastron nor any skeletal parts having been recovered. Hay described this form as a new genus and species, Acherontemys hcckmani, because of the close articulation of pleurals and peripherals and because the vertebrals were even broader than in living chelydrines. This shell may belong to either or neither of the forms represented by the skulls before mentioned. Zangerl (1945) has described a skull fragment from the Pliocene (Clarendonian) of South Dakota, which is indistinguishable from Recent M. temminckii. Gilmore (1923) has described from the San Pedro Valley of Arizona, either Pliocene or Pleistocene, a Kinosternon which is said to differ from Recent A', flavcscens mostly in size. Two species of Chdydra and one of Macrodemys have been described by Hay from the Pleistocene of Florida. The value of these forms, based on fragmentary material, will be difficult to determine. There are also scattered Pleistocene records for the Recent species Chelydra serpentina and Macrodemys temminckii. Acknowledgments. I am indebted to Dr. Glenn L. Jepsen for the privilege of studying and describing the Princeton skull. Mr. Arthur Loveridge and Mr. C. M. Bogert have generously made available the comparative Recent material under their care in the Museum of Comparative Zoology and the American Museum of Natural History respectively. Dr. E. H. Colbert permitted examination of the types of Xenochdys formosa and Chelydrops strida. Dr. A. S. Romer and IVIr. L. I. Price have read the manuscript. Mr. Sam McDowell is to be credited with the drawings and a number of useful suggestions. 1 The type and figured adult skull fragment has this ridge. The young specimen referred by Matthew to this form lacks the ridge. 12 BREVIORA ^fO. 2 TABLE 1. FOSSIL SPECIES OF THE CHELYDRIDAE Oligocene: Xenochelys formosa Hay Chadronian Oligocene (S. Dakota) North America "Chelydra" decheni v. Meyer Upper Oligocene (Siebengebirge) Europe Miocene: "Chelydra" murchisoni Bell Miocene (Oeningen) Europe Chelydropsis carinata Peters Miocene (Eibiswald) Europe "Chelydra" meilheuratiae Pomel Miocene (Allier) Europe "Chelydra" lorettana (v. Meyer) Glaessner Miocene (Leithagebirge) Europe "Chelydra" argillarum Laube Miocene (Preschen) Europe "Chelydra" allinghensis E. Fuchs Miocene (Viehhausen) Europe Macroclemys schmidti Zangerl Middle Miocene (Nebraska) North America Chelydrops stricta Matthew Upper Miocene (Nebraska) North America Acherontemys heckmani Hay Miocene (Washington) North America and additional European records for "Chelydra sp." and "Macroclemys sp." Pliocene: Macroclemys lemminckii (Holbrook) Zangerl Early Pliocene (S. Dakota) North America Pleistocene: Macroclemys floridana Hay Pleistocene (Florida) North America Chelydra laticarinata Haj- Pleistocene (Florida) North America Chelydra sculpt a Hay Pleistocene (Florida) North America Kinosternon arizonerise Gilmore Pleistocene (Arizona) North America and additional North American records for M. temtninckii and C. serpentina 1952 A STAUROTYPINE SKULL 13 LITERATURE CITED BOULENGER, G. A. 1889. Catalogue of the chelonians, rhynchocephalians and crocodiles in the British Museum (Natural HistorjO- London. 311 pp. FucHS, Erika 1938. Die Schildkrotenreste aus dem oberpfalzer Braunkohlentertiar. Palaeontographica, Abt. A, Vol. 89, pp. 57-104. GiLMORE, C. W. 1923. A new fossil turtle, Kinosternon arizonense. Proc. U. S. Nat. Mas., Vol. 62, pp. 1-8. Glaessner, M. F. 1933. Die Tertiarschildkroten Niederosterreichs. Neuea Jahrb. Min. Geol. Pal., Abt. B, Vol. 69, pp. 353-387. Hat, O. p. 1908. Fossil turtles of North America. Carnegie Institution of Washing- ton Publication No. 75. 568 pp. 1916. Description of some Floridian fossil vertebrates belonging mostly to the Pleistocene. Rept. Florida Geol. Survey, Vol. 8, pp. 41-76. Hoffman, C. K. 1890. Schildkroten in Bronn's Klassen und Ordnungen des Tierreicha. Leipzig. 442 pp. Laube, G. C. 1 900. Neue Schildkroten und Fische aus der bohm Braunkohlenf ormation : Abhandl. deutsch. naturwiss.-med. Ver. Bohmen "Lotos", Vol. 2. no. 2, pp. 37-56. Matthew, W. D. 1924. Third contribution to the Snake Creek fauna. Bull. Amer. Mus. Nat. Hist., Vol. 50, pp. 59-210. Meyer, H. von 1852. Ueber Chelydra Murchisoni und Chelydra Decheni. Palaeonto- graphica, Vol. 2, pp. 237-247. Peters, K. F. 1855. Schildkrotenreste aus den osterreichischen Tertiar-ablagerungen. Denkschr. math-naturwiss. Kl. Akad. Wiss. Wien, Abt. 2, Vol. 9, pp. 1-22. 1869. Zur Kenntniss der Wirbelthiere aus den Miocanschichten von Eibiswald in Steiermark. I. Die Schildkrotenreste: Denkschr. math-naturwiss. Kl. Akad. Wiss. Wien, Vol. 29, pp. 111-124. 14 BREVIORA NO. 2 POMEL, A. 1854. Catalogue m^thodique et descriptif des vert^br^s fossiles d^cou- verts dans le bassin hydrographique superieur de la Loire et surtout dans la valine de son affluent principal I'Allier. Paris. 193 pp. Reinach, a. von 1900. Schildkrotenreste in Mainzer Tertiarbecken und in benachbarten ungefahr gleichalterigen Ablagerungen. Abhandl. Senckenberg. naturf. Ges., Vol. 24, pp. .3-135. SlEBENROCK, F. 1897. Das Kopfskelet der Schildkroten. Sitz.-Ber. Akad. Wiss. Wien, Abt. 1, Vol. 106, pp. 24.5-328. 1907. Die Schildkroten Familie Cinosternidae m. monographisch bearbeitet. Sitz.-Ber. Akad. Wiss. Wien, Abt. 1, Vol. 116, pp. 527-599. 1909. Synopsis der rezenten Schildkroten mit Berucksichtigung der in historischer Zeit ausgestorbenen Arten. Zool. Jahrb., Suppl., Vol. 10, pp. 427-618. StEFANO, G. DE 1903. Nuovi Rettili degli strati a fosfato della Tunisia. Bol. Soc. Ital. Ecol., Vol. 22, pp. 51-80. Teppner, W. 1915. Ein Chelydra-Resi von Goriach. Mitt, naturwiss. Ver. Steiermark, Vol. 51, pp. 474-475. Williams, E. E. 1950. Variation and selection in the cervical central articulations of living turtles. Bull. Amer. Mus. Nat. Hist., Vol. 94, pp. 511-561. Zangerl, R. 1945. Fossil specimens of Macrochelys from the Tertiary of the plains. Fieldiana, Geol. Ser., Vol. 10, pp. .5-12. Zdansky, O. 1924. Ueber die Temporalregion des Schildkrotenschadels. Bull. Geol. Inst. Univ. Upsala, Vol. 19, pp. 89-114. 1952 A STAUROTYPINE SKULL 15 PLATE 1 Staurotypus salvinii M.C.Z. 4989: Dorsal, ventral, anterior and lateral views of skull. X about 1 }4- 16 BREVIORA NO. 2 PLATE 2 Xenochelys formosa Princeton 13686: Dorsal, ventral, anterior and lateral views of referred skull fragment. About natural size. BREVIORA Miiseiuinni of Comparative Zoology Cambridge, Mass. May 23, 1952 Number 3 NOTES ON SIPHONOPHORES 3. Nectopyramis spinosa n. sp. By Mary Sp^ars^ In a number of samples collected by the "Dana" in tows with 1000 to 3000 meters of wire out, there were several nectophores and bracts with rows of peculiar jagged spines (Fig. C). These spines are quite unlike those on other deep-dwelling species, such as Vogtia pentacantha Kolliker, I', spinosa Keferstein and Ehlers, and Athorybia rosacea Forskal. In fact, the specimens have little in common with either the Hippopodidae or the Anthophysidae. However, one nectophore was described by Moser (1925, p. 425, PI. 25, figs. 5-7) as Hippopodius (?) cuspitaia, although it actually closely resembled the specimens found in the "Dana" collection. The chief difference is that hers had the peculiar jagged spines scattered irregularly over its surface, whereas on the "Dana" specimens these are arranged in discrete rows. Never- theless, the spines themselves seem to be identical in structure. Like the "Dana" nectophores, most of the characteristics described and figured by Moser (1925) make it appear almost certain that her speci- men was also a Nectopyramis, not a Hippopodius. In all examples, the shape of the nectosac, its relation to the hydroecium, and the fact that the nectophores have all been taken singly and at considerable depths are all peculiarities of Nectopyramis. On the other hand, the soma- tocyst is tubular and not "represented by a series of divergent canals" (Bigelow, 1911a, p. 338) as defined in the brief description of the genus. The original account was based on one species, Nectopyramis thetis. Soon thereafter a second species, Nectopyramis diomedeae, with many of the same characteristics was recorded (Bigelow, 1911b). Since then no additional species have been reported, and little more has been 1 Papers from the "Dana" Collection No. 38. and Contribution No. 607 of the Woods Hole Oceanographic Institution. NO. 8 learned about the genus. Hence, there has been little reason for modifying the original description. Nevertheless, our ideas concerning the relationships of this genus have changed considerably. Thus, the family of the Monophyidae has been broken up and the individual genera transferred to other well-established families (Totton, 1932, p. 328) with which they had obvious affinities. Nectopyraviis, there- fore, is now referred to the Prayidae. In this family, the somatocyst and radial canals are structurally quite varied; in some, they are simple, in others branched. At first, these were used as a criterion for separating genera (Bigelow, 1911b), but more recently, a transitional form with branched radial canals and a simple somatocyst has been GG PASLEY DEL V. %^^ 1 V '.-.J^*;;^; •ct^- B discovered (Leloup, 1934, p. 11). This suggests that when more is known about this family, differences in the structure of somatocyst and canals may prove to be of specific rather than generic value. If, then, we consider that the shape of the somatocyst is in all probability a specific character in Nedo-pyramis, as may be the case among other prayids, we have two species originally described by Bigelow (1911a; 1911b), N. theiis and A'', diomcdeae, both of which had branched somatocysts, but simple radial canals and two species, N. cuspitata Moser and A^ spinosa n. sp. (described below) with simple tubular somatocysts. The two latter are both spiny, but the two are readily distinguished for Moser's species is irregularly covered by the spines, but in spinosa these are arranged in regular rows along more or less distinct ridges. 1952 NOTES ON 8IPHONOPHORES 3 Nectopyramis spinosa n. sp. The type specimen is a nectophore about 5 mm. in length taken at "Dana" St. 3933^^ at 11°18'S., 50°13'E. on 20 December, 1929, in a stramin net 150 cm. in diameter towing with 2000 meters of wire out. The type specimen will be deposited in the University Museum, Copenhagen, Denmark. Other specimens were taken as follows : St. 3556". 2°52'N., 87°38'W. 14. IX. 1928. SI 50. 2000 m. wire. 1 nectophore. St. 3677^^. 5°28'S., 130°39'E. 23. III. 1929. SI 50. 2000 m. wire. 2 bracts. St. 3920^\ 1°12'N., 62°19'E. 9. XII. 1929. SI 50. 1000 m. wire. 1 bract. St. 3964'''. 25°19'S., 36°13'E. 15. I. 1930. E300. 3000 m. wire. 1 bract. St.3998"^ 7°34'S., 8°48'W. 1. III. 1930. S150. 1000 m. wire. 1 bract. Nectophore (Fig. A). The unique rows of jagged spines make it obvious that the nectophore is new. Although probably flattened on preservation, in dorsal or ventral view, it is roughly triangular much as are the nectophores of the other known species of N c dopy r amis . The nectosac opens into the base of the triangle and like A' . thetis "is comparatively small, shallow, and saucer-shaped" (Bigelow, 1911a, p. 338) with four primary radial canals along its sides. It is uncertain whether or not these eventually branch because of preservation. A large deep hydroecium lies just above it with a slit-like opening on the ventral surface. This extends the entire length of the hydroecium. The somatocyst insofar as can be determined appears to be a tubular rod extending from the base of the hydroecium to slightly above its apex. The stem and appendages are entirely missing. Bract (Fig. B). Although bracts have not been found attached to the nectophore, there seems to be little doubt that the bracts with rows of the same jagged spines belong to this species. These are globular, about 2.5 mm. high and slightly more than 3 mm. wide. As in the nectophore, the hydroecium is so deep that it extends somewhat into the upper half of the bract. Closely associated with the hydroe- cium is a long tubular somatocyst which follows the outline of the hydroecium. Oa two of the specimens, two threadlike branches are given off dorsally and appear distally to have a globular connection. 4 BREVIORA NO. 3 BIBLIOGRAPHY BiGELOW, H. B. 1911a. Biscayan plankton collected during a cruise of H.M.S. "Research" 1900. Pt. XIII. The Siphonophora. Trans. Linn. See, London, Zool., 2nd ser., 10(10): 337-357, pi. 28. 1911b. XXIII. The Siphonophorae. Reports on the Scientific Results of the Expedition to the Eastern Tropical Pacific, 1904-1905 .... Mem. Mus. Comp. Zool. 38: 171-402, 32 pis. 1931. Siphonophorae from the Arcturus Oceanographic Expedition. Zoologica. N. Y., 8(1): 525-592, figs. 185-220. BiGELOw, H. B., and M. Sears 1937. H. 2. Siphonophorae. Rept. Danish Ocean. Exped., 1908-10, to the Mediterranean and Adj. Seas, Biol. 2: 1-144, 83 text figs. Leloup, E. 1932. L'eudoxie d'un siphonophore calycophoride rare, le Nectopyramis thetis Bigelow. Bull. Mus. Hist. Nat. Belg. 8(3): 1-8, 5 text figs. 1933. Siphonophores calycophorides provenant des campagnes du Prince Albert 1®' de Monaco. Res. Camp. Sci., Monaco, 87: 1-64, 1 double plate. 1934. Siphonophores calycophorides de I'Oc^an Atlantique tropical el austral. Bull. Mus. Hist. Nat. Belg. 10(6): 1-87, 15 figs. 1936. Siphonophores calycophorides (suite) et physophorides provenant des campagnes du Prince Albert 1®' de Monaco. R^s. Camp. Sci.. Monaco, 93: 1-35, 2 double plates. Leloup, E., and E. Hentschel 1935. Die Verbreitung der calycophoren Siphonophoren im Slidatlan- tischen Ozean. Wiss. Ergeb. Deutschen Atlantischen Exped. . . . "Meteor" 1925-1927. 12(2): 1-31, 17 text figs. MOSER, F. 1925. Die Siphonophoren der Deutschen Siidpolar-Expedition. Deut- schen Siidpolar Exped. 17, Zool. 9: 1-.541. 33 pis., 61 text figs. TOTTON, A. K. 1932. Siphonophora. Great Barrier Reef Exped., 1928-29, Brit. Mus. (N. H.), Sci. Repts. 4(10): 317-374, 36 text figs. 1936. Plankton of the Bermuda Oceanographic Expeditions. VII. Siphonophora taken during the year 1931 . Zoologica, N. Y., 21(4) : 231-240. BREVIORA Museem of Comparative Zoology Cambridge, Mass. June 9, 1952 Number 4 A UNIQUE REMOPLEURIDID TRILOBITE By H. B. Whittington LNTRODUC'TION The species described below, and made the type of a new genus, has been known for more than 75 years. Previous illustrations are both few and inadequate, and no attempt had been made to reconstruct the exoskeleton. When such an attempt is made (text-fig. 1) this unique remopleuridid is shown to have been a most unusual trilobite, in which the long, forked hypostome reached back to the anterior margin of the pygidium. It could not enroll, as it would seem that many (or most) trilobites could. If the hypostome was rigidly attached to the cephalic doublure, then only very restricted movements of the thorax and pygidium, relative to the cephalon, were possible. I am indebted to Dr. G. Arthur Cooper, U. S. National Museum, Washington, D. C, for allowing me to study and photograph material in his charge. SYSTEMATK " DESCRIPTION Family REMOPLEURIDIDAE Hawle and Corda, 1847 Genus HyPODICRANOTUS gen. nov. Type species. Rcmopleuridcs striatulus Walcott, 1875, pp. 347-349. fig. 27, from C. D. Walcott's locality in the Trenton limestone of Trenton Falls, Oneida Co., New York. Discussion. The cephalon of Ili/podicranofu.s stridtulu.^- ditfers from that of the type species of Rnuophurids (\\ hittington, 1950, pp. 540- BREVIORA NO. 4 543, PI. 70, figs. 1, 2, 4, 5) in that: (a) the anterior part of the glabella, the tongue, projects for a greater distance in front of the eyes ; (b) three pairs of glabellar furrows are present, and the second is relatively farther back than that of Rr m o pie u rides; (c) the broad spine that curves back beside the thorax is not the genal spine, but a lateral cephalic spine; (d) the hypostome is long and deeply forked, not a Text-figure 1. H ijpodicranotus striatuliis (Walcott, 1875). Reconstructions of ventral (left) and right lateral (right) appearance of outline of exoskeleton, approx. X 3, based upon M. C. Z. Nos. 1616, 1617. The positions of the raised lines on the cephalic and thoracic doublure and the hypostome are indicated in the ventral view. ?h = median suture; p - pit in doublure. In both draw- ings a general outline of the pygidium is indicated by dotted lines. sub-rectangular plate wider than long. The thorax of Ili/podicranottts probably consists of eleven segments, and is like that of Remopleurides, with the wide axis, prominent articulating processes and sockets, diagonal pleural furrows, and backwardly curved and pointed pleurae. 1952 ('NU^rK RE.MOPI.KnUDIl) THILOHITK 3 The axial and pleural furrows in Ili/ijodirranofu.s are shallow, there are no axial spines, and no segment has the pleural spines exceptionally elongated. The pygidium of Ili/podirranotus is poorly known, but appears to be rectangular in outline, longer than wide, with the axis much shorter than that of Rcmophiiridcs. The dorsal exoskeletons of other remopleuridid genera may readily be distinguished from that of Hypodicranotus or Rcnwplcuridr.s. Hypodicranohis is known from the Trenton Group of New York and Ontario, from rocks of the same age in Wisconsin, the Prosser limestone of Minnesota, the Kimmswick limestone of Missom-i and Illinois, the Viola limestone of Oklahoma, and from the Ordovician strata at Silliman's Fossil Mount, 15afhn Island. HvPODRRAXoTrs sTRiATi'Hs (Walcott, bS75) Plate 1, figs. 1-10; text-figure 1. Lrctofi/pc (selected Raymond, 1925, explanation of Plate 3). Mus. Comp. Zool. No. 1()16, original of Walcott, 1875, p. 347, fig. 27A, from the dark-gray colored Trenton limestone of Trenton Falls, Oneida County, New York. The large collection from this locality in the Museum of Comparative Zoology includes the trilobites Ccrauriis plcutrxaufhrmus' Green, 1832; Calliops ntU'urphalus (Hall, 1847); Lconaspi.s ? froifoncnsis (Hall, 1847); Diiicdiithtt.sjji.s ? parnila (Walcott, 1S79); Illnrinis cf. a iiirrlcfuiiis liillings, 1859; Isotdus gigas DeKay, 1824; Flr.ricdli/iinni' .sciutrid (Conrad, 1841), as well as bryozoans, brachiopods, crinoids and asteroids. This locality, from which both W. P. Rust and Walcott collected, is probably in the Denmark member of the Sherman Fall formation of Kay (1937, pp. 267-268). Parah/pes. Mus. Comp. Zool. Nos. 1617, 1618, 1537, same locality and horizon. Additional material, cranidia and hypostomes, is included under Mus. Comp. Zool. Nos. 3267-3269. No. 1617 is original of Walcott, 1875, p. 347, fig. 27a. Description. Dimensions of lectotype, Mus. Comp. Zool. No. 1616, in millimetres: ( 'ephalon : Length (sagittal) 10.4 (exsagittal) 18.2 Width at genal angle 16.4 Height at second glabellar furrow 6.5 Maximum width across palpebral lobes 11.2 Length of eye lobe 7.2 4 BREVIORA .\'0. 4 Thorax : Width axis first segment 10.2 " " ninth " ,approx 4.3 " first segment (across tips of pleurae) 13. G " seventh segment (across tips of pleurae) 7.7 Length (sagittal) complete thorax, approx 13.0 Dimensions of hypostome, Mus. C/omp. Zool. Xo. 1()17, in milli- metres : Length (sagittal) 5.2 (exsagittal) 20.2 Maximum width (at about half the length) 7.4 Width of middle bofly 3.7 The cranidium includes the occipital ring, glabella and tongue, and palpebral lobes. The occipital ring is longitudinally flat, trans- versely moderately convex, bounded laterally by the posterior branches of the facial sutures, which are straight but run diagonally back from the posterior corner of the eye lobe, then curve to run into the posterior margin at right angles. Inside and parallel to the suture is a faint depression, the axial furrow, and just inside the point where the suture cuts the posterior margin is a rounded notch, the articulating socket. The occipital furrow is shallow. Between the eye lobes the glabella and palpebral lobes are gently and evenly convex both longitudinally and transversely. In the mid-line the tongue of the glabella of the lectotype projects forwards 1.8 mm. in front of the eye lobe, and is convex and bent down so that the anterior part overhangs the sutural margin. There are three pairs of glabellar furrows, visible because they are both unornamented and slightly depressed (PI. 1, figs. 2, 4). Each furrow runs in a curve convex forwards, and they are equally spaced from each other, the distance (exs.) between them slightly less than the equal distance (exs.) of the outer end of the first from a point opposite the posterior end of the eye lobe, and the outer end of the third from a point opposite the anterior end of the eye lobe. The furrows commence a short distance inside the palpebral furrow, and do not extend to the mid-line. The first is the most strongly convex, the second commences farthest out, and the third farthest in. The palpebral rim is flat, depressed slightly below the lobe, outwardly sloping, widest posteriorly, narrowing forwards and disappearing at the anterior end of the eye lobe. Posteriorly the palpebral furrow runs into the junction of the extremity of the occipital furrow and the axial furrow. The eye lobe is long, in dorsal aspect curved, most strongly in the posterior part. The eye surface is gently convex 1952 TNigTE REMOFLErRiDii) tkilobitp: 5 transversely, and slopes steeply. The arrangement of the minute, convex facets is shown in Plate 1, figure 10. The outer margin of the eye lobe is defined by a narrow, convex border which commences at the posterior corner and runs forwards to merge anteriorly with the border of the cheek. The anterior branches of the facial suture curve to run at first forwards and inwards from the outer, anterior, corner of the eye lobe, and then run inwards and upwards to the mid-line. Thus in anterior aspect (PI. 1 , fig. 5) the margin of the tongue is a curve convex upwards. The free cheek is triangular in outline, outwardly sloping, widest behind the eye lobe, the genal spine short and pointed. The cheek narrows forwards to disappear opposite the anterior part of the eye lobe. From the lateral border opposite the median part of the eye lobe a broad spine curves back, narrowing and reaching to a point opposite the axis of the fifth thoracic segment. The inner margin of the proximal part of this spine is parallel to, and just outside of, the outer margin of the cheek and genal spine. The doublure of the cephalon is broad anteriorly, gently convex, crossed by a median suture. Antero-laterally and laterally the doublure is narrower and bent upwards. A small pit is situated on the exsagittal line passing through the anterior end of the eye lobe, and about mid-way across the doul)lure (PI. 1, fig. S; text-fig. 1). The hypostome is long and forked, and its position relative to the rest of the exoskeleton is shown in text-figure 1. It is gently convex, in both longitudinal and transverse directions, the transverse convexity greatest at the curved anterior margin. The convexity (in ventral view) is interrupted only by the shallow depression in the median region just in front of the crotch of the fork. The middle body is circular in outline, faintly elevated, defined postero-laterally by a distinct furrow. It is divided into three sub-equal sections by shallow radial depressions, the strong- est leading back to the crotch of the fork, the other two directed antero-laterally (PI. 1, fig. 7). On the inside of the exoskeleton of a smaller specimen (PI. 1, fig. 9) these sections appear more pronounced, and on both sides the postero-lateral pair are the more prominent. Anteriorly the middle body merges with the border, and the anterior wings appear to be short, triangular, and upwardly directed. Each postero-lateral border is extended as a prong of the fork, the inner margin of the prong straight, the outer margin curved, so that the maximum width of the hypostome is in front of half the length, and the prong tapers to a sharp point. The doublure of the hypostome (PI. 1, figs. 4, 9) is narrow along the lateral border, but extends in under the fork to a line almost under the margin of the middle body. 6 BREVIORA NO. 4 The inner edge of the doubhn-e is flexed up sharply. Mus. Comp. Zool. No. 1618 (PI. 1, fig. 4) shows that the doubkn-e of the prong is convex dorsally, with a low, median, longitudinal ridge. The thorax is moderately convex transversely, the axis broad, narrowing rapidly backwards. It seems to consist of eleven segments. On the left side of the lectotype 10 segments can clearly be seen (PI. 1, fig. 2). Behind this only what may be part of the left pleura of the eleventh segment, and part of the pygidium, are preserved. Another specimen (PI. 1, fig. G) shows ten partly disarticulated seg- ments and an incomplete pygidium The last segment has the pleura pointing almost directly posteriorly, whereas that next in front has the pleura directed back and slightly outwards, and the lateral margins are curved. This penultimate segment is very like the tenth of the lectotype. The axial ring is moderately convex transversely, the articulating furrow shallow, the articulating half-ring of length (sag. = sagittal) about half that of the ring (PI. 1, fig. 6). There is no axial furrow as such, but inside the articulating sockets anfl processes a triangular area (broadest anteriorly) of the outermost part of each ring is slightly depressed. The inner margin of this area is a diagonal curved line sub-parallel to the pleural furrow. The articulating process, on the anterior margin of the segment, is slightly raised as well as forwardly projecting. Postero-laterally the inflation is con- tinued across the pleura as a low, curved ridge, dying out at about two-thirds (exs. = exsagittal) the length. This ridge, and a depression on the inner side, define the pleural furrow. The pleurae are gently convex (tr. and exs.), outwardly sloping, the narrow (tr. = transverse) inner part directed transversely, the outer part curved back and pointed. The shape and inclination of the pleurae are shown in Plate 1, figures 1, 2, G. Only a small part of the pleural lobes of the pygidium is preserved in the lectotype. The only other specimen known (PI. 1, fig. 6) is also incomplete. The outline was evidently rectangular, longer than wide, the ill-defined axis convex, short, and wide. Behind and beside the axis the pleural lobes slope down in a curve concave in longitudinal profile. The lateral and posterior margins, and the doublure, are unknown. Scattered over the cranidium, but largest and most closely spaced on, and adjacent to, the palpebral lobes (though absent from the palpebral rim), are small crescentic depressions, the points of the crescent facing forwards, and the concave, anterior margin raised (PI. 1, figs. 1, 2, 4). Similar but larger structures are present on the 1952 T'NIQI'E REMOPLETRIDin TRILOBITK 7 free cheeks, the genal spine, and the base and outer part of the long lateral spine (Pi. 1, fig. 1). There is a tiny median tubercle on the occipital ring, closer to the fui-row than the posterior margin. Faint, small crescentic depressions are scattered on the median part of the axis of the thorax. Strong, well-spaced, raised lines run approximately longitudinally on the inner part of the lateral cephalic spine, the thoracic pleurae, and on the doublure of these areas and the cheeks (PI. 1, figs. 1-4, 6, 8). The lines tend to run in curves convex inwards on the dorsal surface of the pleurae, and transversely in curves convex forwards on the dorsal surface of the pygidium. On the inner part of the doublure of the thoracic pleiu^ae, however, they run in curves concave inwards. These lines die out on the median part of the cephalic doublure, and are replaced by faint, anastomosing, transverse grooves. On the lateral borders and fork of the hypostome (PI. 1, fig. 7) the longitudinal lines are strongest, and that which runs just inside the inner margin of the prongs is a prominent ridge. The middle body shows a faint pattern of narrow, anastomosing ridges running longitudinally on the postero-lateral sectors. On the anterior sector of the middle l)ody and the anterior border, fine, well-spaced, anasto- mosing grooves, like those on the median part of the cephalic doublure, run transversely. Discussion. This species was first described by ^^'alcott (1875, pp. 347-349, fig. 27), and the original material later redescribed by Raymond (1925, pp. 57-58, PI. 3, figs. 4, 5). An incomplete cran- idium from the same general locality as the type (U. S. Nat. Mus. 92528) was also figured by P^oerste (1920, p. 222, PI. 22, figs. 18 A-C). Raymond (1921, p. 31, PI. 9, fig. 7) described a hypostome from the Middle Trenton of Trenton, Ontario, and referred it to H. striahdus, and stated that a cranidium had also been found at Governor Bay, near Ottawa. Specimens kindly loaned to me by Professor G. Winston Sinclair, from the Middle Trenton of Lakefield, Ontario, contain cranidia and hypostomes like those of the type material. A second species of Hi/podicranotus is //. missouriensis (Foerste, 1920, pp. 220-222, PI. 21, fig. 17; PI. 22, figs. 17A, 17B; Bradley, 1930, pp. 246- 247, PI. 30, figs. 4-9) from the Kimmswick limestone of Missouri and Illinois. The type material (U. S. Nat. Mus. No. 78438) includes cranidia and a hypostome. The outline of the latter, and the prom- inence of the postero-lateral areas of the middle body, distinguish it from H. striatidus. The characteristic hypostome of Hypodicranotvs also occurs in the Trenton of Duck Greek Quarry, near Green Bay, Wisconsin (U. S. Nat. Mus. Nos. 72181, 87087), the Prosser limestone 8 BREVIORA \0. 4 of St. Paul, Minnesota (U. S. Nat. Mus.), and I have collected one from about 100 ft. above the base of the Viola limestone, in the road cut on U. S. highway 77, in Carter County, 21^ miles north of Springer, Oklahoma. The matrix at this locality was a finely-granular, light grey-brown limestone, which yielded abundant graptolites and the trilobites Cri/ptolifhoidcs ulrichi Whittington, 1941, Trinodus sp., Robergia sp., and an asaphid. The specimen figured by Roy (1941, p. 155, fig. 114), as Rcmopleurides sp., from Silliman's Fossil Mount, Baffin Island, seems also to be the hypostome of this genus The genus thus occurs in central and eastern North America in rocks of Trenton age, and the Baffin Island beds may be of a similar age. The exoskeleton of Ili/podicranotus striatulus, apart from the hypostome, is like that of other remopleuridids of Middle Ordovician age, e.g. the Irish specimens (\^hittington, 1950, pp. 540-543, PI. 69, figs. 5-10, PI. 70, figs. 1-6) and undescribed species from the Edinburg limestone of Virginia. The latter have a sub-rectangular hypostome, not forked, like that of the type species (Whittington, 1950, PI. 70, fig. 2). The posterior margin reaches back to a point lying no farther back than the occipital furrow. The long, forked hypostome of Hypo- dicranotus is not only unique, so far as is known, among remopleuridids, but its relatively great length is in excess of that of any other trilobite known to me. It is evident that here is one genus of trilobites that could not enroll. The pit in the antero- lateral cephalic doublure (PI. 1, fig. 8, text- fig. 1 ) is a feature which I have observed in several difi'erent species of Middle Ordovician remopleuridids. Silicified specimens show the pit to be the opening of an upwardly-directed tube, which narrows inwards and terminates near the lower, anterior corner of the eye surface. The position and nature of this opening do not suggest that it is homologous with the Panderian opening, which, if present in the cephalon, is situated in the postero-lateral cephalic doublure. Not all remopleuridids appear to show this opening (e. g. those described by Ross, 1951, pp. 84-91, PI. 20, do not), and I have not observed a similar opening in any other group of trilobites. It is tempting to speculate on the mode of life of the holaspid Ili/podirrduotits-. It occurs in company with shallow-water marine forms, many of which are indisputably benthonic. The long hypostome prevented enrollment, but did provide some protection for the ventral surface. If no movement was possible at the hypostomal suture, then the amount of passible movement of the thorax and pygidium in the vertical plane, relative to the cephalon, must have been severely 1952 rXUilK KKMOPLKIKIDII) IKII.OHITK 9 limited. The mode of artieulutiou of the thorax prechides any eoii- siderable movement relative to the cephalon in the horizontal plane. There is ample room for the appendage.s to project downwards and outwards between the hypostome and the thoracic pleurae. But we know nothing of the type of appendage possessed by remopleuridid trilobites, and without this information have little basis for speculation on the mode of locomotion, manner of feeding, etc. Was HiiporUcran- ofus a burrowing, crawling, floating, or swimming form? No definite answer is possible but I am inclined to think of it as either floating or swimming. Referexces Bradley, J. H. 1930. P'auiia of the Kimmswick Limestone of Missouri and Illinois. Contrib. Walker Mus., vol. 2, no. 6, pp. 219-290, pis. 2.3-30. FOER.STE, A. F. 1920. The Kimmswick and Plattin Limestones of Northeastern Missouri. Denison Univ. Bull., J. Sei. Lab., vol. 19. pp. 17.5-224, pis. 21-23. K.^y. G. M. 1937. Stratigraphy of the Trenton Group, l^ull. Geol. Soe. Am., vol. 48, pp. 233-302, pis. 1-10. Raymond, P. PI 1921. A Contribution to the Description of the Fauna of the Trenton Group. Geol. Surv. Canada, Mus. Bull. 31, Geol. Ser. 3S, pp. 1-64. pis. 1-11. 1925. Some Trilol)ites of the Lower Middle Ordovician of PJastern North America. liull. Mus. Comp. Zool., vol. (i7, no. 1, pp. 1-1 SO, pis. 1-10. Ross, R..J. 1951. Stratigraphy of the Garden City Formation in Northeastern rtah, and Its Trilobite Faunas. Peabody Mus. Nat. Hist., Vale Fniv., Bull. 0, pp. 1-161, pis. 1-36. Roy, S. K. 1941. The Fijper Ordovician Fauna of Frobisher Bay, Baffin Land. Mem. Field Mus. Nat. Hist., Geo)., vol. 2, pp. 1-212, 14() fijis. Walcott, C. D. 1875. New Species of Trilobite from the Trenton Limestone at Trenton Falls, N. Y. Cincinnati Quart. .J. Sci., vol. 2, j^p. 347-349, fig. 27. Whittington, H. B. 1950. Sixteen Ordovician Genotype Trilol^ites. ,J. Paleont., vol. 24, no. 5, pp. .531-565, pis. 68-75. EXPLAXATIOX OF PLATE I H ypodicranotus striatulus (Walcott, 1S7.5). Trenton limestone, probably Denmark member of the Sherman Fall formation, Trenton Falls, Oneida County, New York, all X 3, except figure 10, X 17,V2, Mus. Comp. Zool. collections. Figs. 1-3, 5, 8. Lectotype, Xo. 1616, left lateral, dorsal, right lateral, anterior, and antero-ventral views, presumed original of Walcott, 1875, p. 347, fig. 27A, and original of Raymond, 1925, PI. 3, fig. 4. Figs. 4, 10. Paratype, Xo. 1618. 4, dorsal view. The right half of the dorsal exoskeleton of the cephalon and three thoracic segments are preserved; on the left is seen the lateral cephalic spine, and external moulds of the doublure of several pleurae. The left, and tip of the right, prongs of the fork of the hypostome are exposed from the inner side. The hypostome has been slightly displaced. 10, the anterior portion of the right eye surface, showing the arrangement of the facets. Fig. 6. Paratype, Xo. 1537, dorsal view. Fig. 7. Paratype, No. 1617, ventral view, presumed original of Walcott, 1875, p. 347, fig. 27a, and Raymond, 1925, PI. 3, fig. 5. Fig. 9. Xo. 3269, dorsal view of interior of hypostome. BREVIORA Unique Remopleuridid Trilobite. Plate 1 E V I Museuim of Comparative Zoology Cambridge, Mass. August 29, 1952 Number 5 TRICHOLIMNAS CONDITICIUS IS PROBABLY A SYNONYM OF TRICHOLIMNAS SYLVESTRIS (Aves, Rallidae) By James C. Greenway, Jr. Comparison of the type of Tricholimnas conditicius Peters and Griscom (Proe. New England Zool. Club, vol. 10, 1928, pp. 99-103) with a series of Tricholimnas sylvestris (Sclater) makes it appear most probable that the type specimen is an immature female of sylvestris. Furthermore, some recently discovered notes in the handwriting of collector Andrew Garrett cast grave doubt upon the theory that Garrett shot the bird on Apaiang Atoll, between Makin and Tarawa, in the Gilbert Islands, as Thomas Barbour postulated in the original description of the supposed species. Both Mr. Peters and Mr. Griscom agree to this. Naturalists have wondered since then whether a habitat, such as this low sandy island affords, could have supported a population of these rails. The very distance from congeners on New Caledonia and Lord Howe Island, off the coast of Australia, made this seem unlikely, quite apart from the disparity of habitats. However, the arguments brought forward in the original description were difficult to circumvent. This type and only known specimen of conditicins was found pre- served in alcohol in the Museum of ( 'omparative Zoology after a period of long oblivion. With it was a label: "Kingsmill Islands, 1861, Andrew Garrett, Collector". Since it could be proved without question that Garrett was actually on Apaiang Atoll of the Kingsmill, or what is now called the Gilbert group, in the autumn of 1859, it seemed likely that he did collect the bird there and that the date "1861" represented the year of acquisition by the museum. However, it now seems quite 2 BREVIORA NO. 5 as probable that the specimen has been confused with a consignment of specimens that came from the Kingsmills and that the label wa& written in the museum under a misapprehension. Andrew Garrett was a dealer who had in his stock specimens from all over the world. In a letter to (^harles E. Hamlin at the Museum of Comparative Zoology, written on April 10, 1878, from Huahine, Society Ids., he says: "In return I shall be glad to receive any land or fresh water shells from the East Indies and marine shells from any part of the world except Eastern States, California, and Polynesia." And again, "You can send me in return for this box cvcnjthing you can spare from your duplicates no matter from what part of the world . . . don't forget the museum publications for which I will send the ferns in compensation." From Hilo, Hawaii, he writes on Oct. 10, 1856, that he plans a voyage on the whaler 'Lydia', the captain of which was an acquaintance of his. In a letter to Jas. M. Barnard dated April 29, 1857, and written at Hilo, he gives a short history of his life. He had been a sailor and had visited in past years the "Ladrone, Bonin, Loochoo, some of the Australian and several of the East India Islands ... my taste for Natural History, which I dearly love, was acquired while visiting- foreign places." The most important bit of evidence has recently been found. A short note in Andrew Garrett's handwriting and on the same paper used by him in his correspondence is entitled "Remarks upon the Birds of Apaiang one of the Kingsmills Is." and is reproduced here in its entirety. It bears no date but was probably written in 1859. "During my short residence at the above location [iVpaiang Atoll, Sept.-Oct. 1859] I collected one or two specimens of every bird I could get. As I had not time to prepare their skins, I packed them in alcohol. There is only one species of land bird, and this I never saw hut obtained one of its tail feathers which I send you. I think it is a species of Hawk." (Italics are mine.) Unfortunately this feather cannot be found. Two species of migrant hawks have been reported from the Palau Island; possibly a stray bird reached Apaiang. Andrew Garrett was an excellent naturalist; Barbour has called him "an extraordinary genius." Although his interests were centered upon fishes and shells, he knew birds. In a contemporary list of birds, mammals of Hawaii, sent to the Museum of Comparative Zoology, he says, (no.) "7 — Phaethon — This fine bird builds its nests and roosts among the precipices bordering the sea in Hamakua and Hilo." He 1952 TRICHOLIMNAS CONDITICIUS 3 then describes the bird accurately. And again, (no.) "3 — Fuhca — common about our fresh water ponds." It is quite unhkely that he would consider Tricholimnas to be a sea bird. Nor is it likely that he would mistake the short, decomposed tail feather of TricJwUmnas for that of a hawk. Unfortunately there is no record of the actual provenance of the specimen but the weight of evidence points to its having come from Lord Howe Id. It is prol)able that it found its way into Andrew Garrett's collection by an exchange, or perhaps through the good offices of a whaling friend. Further questions occur. Assuming that the type was a member of a population which had for some time inhabited Apaiang Atoll, why was it not found by the contemporaries of Garrett? The answer is that none of them visited the island. However, Kubary, Finsch and others did land upon islands of neighboring groups and no such speci- men was recorded. No bird even faintly resembling this has ever been recorded except from Lord Howe Id. and New Caledonia. If the assumption is correct, we must presuppose that wind born or ship born individuals reached Apaiang and no other neighboring islands and that the population was extirpated a very short time after Garrett visited there. It would be strange if such a population had established itself on this flat, almost desert, island in view of the fact that its un- mistakable relatives prefer forests and hills and that the intervening islands afford more suitable habitat. That it became extinct in a number of island groups during the twenty years 1860-1880 is not at all probable. Granted that no proof can be offered and that at best we are dealing with probabilities, the best calculated guess would appear to be that the type of Tricholimnas conditicius came from Lord Howe Id. No specimens of sylvestris were to be found in the United States in 1928 when conditicius was described. Comparison of the type with a series of twelve in the Rothschild collection, now in the American Museum of Natural History, New York, cause most of the characters ascribed to geographical variation to disappear. The paler crown, throat and underparts, the browner head and tlu-oat are doubtless the result of long immersion in alcohol and it is surprising that more color change has not taken place. The only differences are then the length of wing and bill, which are two millimeters shorter than those of the smallest female of syhestris available. Since G. M. Allen, in the original description indicated that the skeleton was obviously that of 4 BREVIORA NO. 5 an immatiire bird, perhaps it may reasonably be assumed that this character is due to age. Measurements, in millimeters, are as follows: "conditicius" sylvestris lafresnayanus cf 9 Wing 132 135-139 133-142 185 Tail 58 58-63 61-65 109 Bill 45 54-56 47-50 52 Tarsus 47 47-50 43-48 54 In the original description the length of the tail of the type of conditicius is said to be 68 mm. Mr. Peters agrees with me that it should read 58 mm. E V I Mmsemnn of Coinnparative Zoology Cambridge, Mass. August 29, 1952 Number 6 CONTRIBUTIONS TOWARD A RECLASSIFICATION OF THE FORMICIDAE I. Tribe Platythyreini (Hymenoptera) By William L. Brown, Jr. Museum of Comparative Zoology Harvard University The tribe Platythyreini has included the sole genus Platythyrea Roger as treated by former authors. My own investigations show that three additional genera {Proholomyrmex, Escherichia and Eubothro- ponera) must be included. Proholomyrmex Mayr and Escherichia Forel have heretofore been placed among the Proceratiini because of their depigmented condition, atrophied eyes (workers), and especially their vertical, fused and approximated frontal carinae and the closeness of the antennal insertions to the median line and to one another, accompanied by fusion of frontal carinae with the greatly crowded clypeus. Also, these two genera have, according to the describers of the included species, only one tibial spur to each of the two posterior pairs of tibiae. I believe that all of the characters just mentioned are correlated with adaptation of the insects to hypogaeic or other crypto- biotic conditions of life; they appear in widely separated genera of ants and other hymenopters, such as Proceratium, Discothyrea and others in the Formicidae, Psilohethylus, etc. in the Bethylidae, and so on, as rather consistent combinations. The similar modifications of doryline and some other ants may be partly due to hypogaeic or subhypogaeic adaptation, but it would seem that the legionary habit may somehow be more important in accounting for this particular structural modi- fication. For our present purposes, it will be sufficient merely to recognize two facts: (1) the characters combining to produce the "proceratiine 2 BKEVIORA NO. 6 habitus" can and do act to form similar-appearing groups of genera, even though these genera may have the most diverse ancestry; (2) the number and condition of the spurs of the two posterior pairs of tibiae have been found to possess much less taxonomic value than Emery and Wheeler granted in their widely-used keys; in fact, the spurs are now known to be present as pairs, single units, or even to be absent in different species of the same genera in tribes where once the spurs were considered universally constant and tribally diagnostic. With these facts in mind, one can proceed at once to a more natural classification of the genera and tribes of Ponerinae. Proholomyrmex and Escherichia have always seemed anomalous as members of the Proceratiini, and they are not, in fact, at all closely related to any other proceratiine genus. Following a slight clue dropped by Mann in 1923, when he mentioned the similarity of the pilosity and sculpture between his Proholomyrmex boliviensis and the species of Plaiythyrea, I have carefully compared the two last-named genera and find that they agree in an astoundingly complete way. In fact, the point-by- point agreement is so close that I must consider Proholomyrmex to represent a direct derivative of Plaiythyrea modified for a highly cryptobiotic existence in the same way, as I also believe, that the proceratiines are only modified ectatommines. If Plaiythyrea and Proholomyrmex are to be associated, then Escherichia should probably go along with them. While I have never seen an example of the latter genus, it appears from all accounts to be very close to Proholomyrmex, and may even be synonymous and representative of an ergatoid or other intermediate female caste of one of the Ethiopian Proholomyrmex species. A fourth genus that must be included in the Platythyreini is Evbo- throponera, in spite of its unfortunate name. In his original diagnosis of this Australian group, Clark stated, "... certainly close to Boihro- ponera Mayr;" he called the tarsal claws "simple" and gave the palpal formula as maxillary 4, labial 2 segments. I have examined types (Museum of Comparative Zoology) of all the Eiihoihroponera species described to date except E. hrunnipes Clark, and find that all species, at least in the worker caste, possess single, small but distinct median teeth on all tarsal claws. The palpi in this genus are unusually long for a ponerine group, and the maxillary pair may reach the foraminal border behind if fully outstretched. Most of the few specimens available were inconveniently situated or else had the maxillae re- tracted, but in spite of this it is clear that all six species seen have more of both maxillary and labial palpal segments than Clark claims 1952 PLATYTHYREINI (hYMENOPTERa) 3 for them. In E. reticulata Clark and E. tasmaniensis (Forel), a full count was made under good conditions, and in these species the formula was maxillary palpi 6, labial palpi 4 segments. The basal segment, especially in the maxillary palpus, is short, fairly broad, and the segment most likely to be overlooked. Comparing the proportions of the visible segments of the palpi of the less favorably situated speci- mens with those upon which a direct count could be made, I feel quite safe in assuming that the six species I have seen all possess a 6, 4 formula in the worker. Now this formula is the primitive one in Platythyrca, as I can con- firm (vide infra), and since all other characters save one combine to link the two genera, they must at least be placed in the same tribe. In fact, the characters separating them are relatively minor ones, though constant and consistent in combination in the series before me. The body in Eubothroponera is smaller and generally more compact (less slender and elongate), and the characteristic sculpture of Platy- thyrea is less well developed in Eubothroponera, but still basically the same on head and alitrunk. The pilosity must be used to separate the two genera until a study can be made including all castes of a large number of species belonging to both (see key, below). In examining the palpi of some representative species of Platythyrea for this work, it was confirmed, as has been long known, that certain of them have a palpal formula of 6, 4. \Miat is not so generally known, or perhaps even unknown until now, is that certain species of the Old World tropics possess lesser numbers of segments. The New World species P. angusta Forel and P. strenua Wheeler and Mann have elongate palpi, with the 6, 4 segmentation; the basal segment, par- ticularly of the maxillary palpi, is very short, and the total similarity of the palpi to those of Eubothroponera is very striking. Among the Old World forms, P. crihrinodis (Gerstaecker) very definitely has short palpi, formula 3, 2. In P. schuUzei Forel, the maxillary palpi have 2 clearly visible segments each, and there may possibly be a third small basal one, though no third segment could be seen in the single specimen examined; the labial palpi are definitely 2-segmented. Other species, such as P. sagei Forel, were not advantageously placed for a definite palpal count, but had short palpi with formulae almost certainly under those of angusta for both pairs. Also seen in one doubtfully determined species from the Old World were palpi with definite counts of 6 and 4, but very short in overall proportions. It seems that Platythyrea species vary considerably from one to the next in palpal characters, and anyone possessing the material needed for a thorough survey of 4 BREVIORA NO. 6 the group by means of dissection may be able to correlate this vari- ation with other characters, and thereby be able to recognize more than one group or genus among the array of Platythyrea species. Furthermore, the males of Platythyrea seem, from external exami- nation only, to have very distinctive terminalia worthy of further study. Key to the genera of the tribe Platythyreini, based on the workers and probably applicable to the females 1. Frontal carinae and clypeus fused and projecting anteriorly over the mandibles; antennae somewhat incrassate apically, inserted close together on the anterior part of the clypeo-carinal process. (Small, usually pale- colored forms, under 4 mm. in length. Ethiopian, Neotropical, Indo- Australian) Probolomyrmex Mayr^ Escherichia Forel Frontal carinae and clypeus not projecting anteriorly over mandibles; antennae inserted distinctly posterior to the clypeus and apart from one another, the insertions covered by the broad lobes of the frontal carinae, funiculus not or scarcely incrassate in most forms. (Larger, usually pigmented forms, mostly over 4 mm. in length) 2 2. Distinct erect pilosity present and widely distributed (widespread in Australia) Eubothroponera Clark Distinct erect pilosity limited to mouthparts and gastric ape.x. (Tropico- politan; occurring naturally or as tramp species in some warm temperate regions) Platythyrea Roger Below are listed the platythyreine genera as known to me at present, and also a citation of the genotypes and the principal references to the literature. The references are not intended to be complete, but are designed to give the interested worker a start toward the gathering of pertinent titles. Platythyrea Roger 1863, Berlin. Ent. Zeitschr., 7: 172. Genotype: Pachycondyla punctata Fred. Smith, 1858, soldier (recte worker), male; designated by Bingham, 1903. Besides the characters cited in the key (above), it should be men- tioned that this genus, and also Eubothroponera, are further marked by 1 Probolomyrmex and Escherichia are supposed to differ in that the latter has small compound eyes and a discernible postmesonotal groove or line, but it is important to note that Forel himself, in the original description of Escherichia, compared the novelty with several proceratiine genera, but made no mention of Probolomyrmex'. Certain Probolomyrmex (P. boliviensis Mann, P. parvus Weber) are known only from the female caste. 1952 PLATYTHYREINI (hYMENOPTERa) 5 having a complete and apparently functionally flexible suture sepa- rating the pro- and mesonotum, and also by having a large and a smaller spur on each one of the two posterior pairs of tibiae. Emery's survey of the world species (1911, Gen. Ins., Fasc. 118, pp. 28-30) is comprehensive for its time, but is now far out of date due to the addition of many forms since it was written. Wheeler's "Ants of the Belgian Congo" (1922, Bull. Amer. Mus. Nat. Hist., 45: 57-60, 758-761, 1007) gives references to African and Malagasy species. EuBOTHROPONERA Clark 1930, Proc. R. Soc. Victoria, Melbourne, (n.s.) 43: 8-9. Genotype: Eubothro- ponera dentinodis Clark, 1930, worker; original designation, nee E. tasmani- ensis (Forel) designated by Clark in 1934. With the original generic description, Clark characterized and figured three species (pp. 9-11, fig. 1): E. dentinodis Clark (p. 9,, fig. 1, nos. 6, 6a); E. micans Clark (p. 10, fig. 1, nos. 7, 7a); E. bicolor Clark (p. 11, fig. 1, nos. 8, 8a); he also included in the genus Forel's Pachy- condyla (Bothroponcra) tasmaniensis (p. 11), and gave a "key" (p. 9) to the three Western Australian species newly described. In 1934 (Mem. Nat. Mus., Melbourne, No. 8, pp. 32-34, pi. 2, figs. 15-17) Clark redescribed the worker of E. tasmaniensis (Forel) (p. 32, fig. 15) from Tasmania and described the new species E. reticu- lata (p. 33, fig. 16) from New South Wales and E. septentrionalis (p. 34, fig. 17) from Queensland. A seventh and last species was added by Clark when he described E. brunnipes (1938, Proc. R. Soc. Victoria, Melbourne, 50: 361-362, fig. 3) from Reevesby Island, Sir Joseph Banks Group, South Australia (biology, loc. cit., p. 356). The above, I think, includes all the recognized species of Eubothro- ponera, each reported only from the worker caste. The species are quite uniform in structure, and it is felt that Clark has failed in some instances to properly differentiate them; consequently, a review of the genus is necessary. NO. 6 Probolomyrmex Mayr 1901, Ann. Naturhist. Hofmus. Wien, 16:2-3. Genotype: Probolomyrmex filiformis Mayr, 1901. worker; monobasic. M. R. Smith (1949, Proc. Ent. Soc. Washington, 51:38-40) has reviewed briefly the five species known to that date. A sixth species was described at about the same time by Weber (P. parvus Weber, 1949, Amer. Mus. Novit., No. 1398, pp. 3-4, fig. 2, female) from Africa. A description with a good habitus drawing of the female type of P. holiviensis is given in the original reference by Mann (1923, Psyche, 30: 16-18, fig. 2). A seventh species, apparently remaining undescribed, has been taken in the Canberra region of eastern Australia by Mr. Tom Greaves. Escherichia Forel 1910, Zool. Jahrb. Syst., 29: 245-246. Genotype: Escherichia brevirostris Forel, 1910, op. ciL, pp. 246-247, worker; monobasic. Although there are entries in the catalogs of Emery and Wheeler, this genus and species seems not to have been reported a second time. It should be noted, however, that Weber's description of Probolomyrmex parvus, cited above, may just possibly apply to the female of Esche- richia bremrostris when the latter becomes known in association with workers, as the characters and type localities given for the two species do not, to my mind, exclude the possibility of conspecificity. E V I O R A Museiim of Comparative Zoology Cambridge, Mass. September 29, 1952 Number 7 ON THE NOMENCLATURE OF THE PACIFIC GRAY WHALE By William E. Schevill^ The gray whale of the Pacific has until recently been called RhachianccU's glaucus (Cope). In 1937 van Deinse and Junge published an important paper of which the main part is a thorough comparative study of subfossil Kschrichtius rohusius bones from the Netherlands, comprising one adult cranium and two juvenile crania, one of which was associated with the mandibles, scapulae, and other parts of the skeleton. These they compared with an adult Rhachiaiuctcs glaucu.s skeleton in the British Museum, as well as with the published material on both forms. Although the subfossil bones are naturally worn and fragmentary, as well as being for the most part juvenile, they present ample characters to indicate that they and the gray whale are conge- neric. At this point I would like to emphasize that the following criticisms of van Deinse and Junge's conclusions are directed solely at their taxonomy; this involves no disparagement of their important antl valuable osteological work, which has demonstrated the generic identity of Eschrichtius Gray 1864 and Rhachianrctrs Cope 1869 (= Agapliclus Cope 1868). But they also argue for specific identity, and thus unite RJiacliianccics glaucus (Cope) 1868 with Eschrichtius rohusfus (Lilljeborg) 1861. Moreover, they argue that Dudley's scrag whale (1725), named Balacna gihbosa by Erxleben (1777, p. 610), is the same form, and therefore say (1937, p. 181), "After this it is clear that the name of this whale must be changed into Eschrichtius gibbosvs (Erxleben). We must emphasize that with the bones we have now at hand we cannot give any argument that the Pacific and Atlantic representatives of Eschrichtius should be specifically different." But by the same token these bones cannot prove that these whales are specifically identical. 1 Contribution No. 626 of the Woods Hole Oceanographir Institution. 2 BREVIORA NO. 7 Additional and better material of E. robustus may present more specific characters, but until such material turns up, it seems unfortu- nate to take a well-grounded name of a well-characterized living form, known from complete specimens, and replace it with that of a form so imperfectly known from such incomplete material that specific characters cannot be entirely estaVjlished; instead it seems better not to change the name of the well-known form. It is even less desirable to use the name (gibbosa Erxleben) of a form which from the beginning has been an unrecognizable literary curiosity. Considering the available information, which is scanty though oft repeated, the most that can be said for Balacna gibbosa Erxleben is that it may be an Eschrichtius, but since it is not specifically well characterized and since nothing sufficiently like it has ever again been reported from anywhere near the type locality ("the Coast of New England") or elsewhere in the Atlantic, it seems to be stretching probability to submerge in it known forms from other regions. De- tailed consideration of the uncertainties in attempting to identify Dudley's scrag whale (Balacna gibbosa Erxleben) are omitted here. From time to time cetologists have tried to identify this beast, but have always been frustrated by various irreconcilable data, the chief of which is perhaps the utter lack of specimens. Cope's attempt to embody gibbosa Erxleben was undermined by his confusion over a dilapidated specimen of Balarnoptrra acuto-rostrata, so that he formally withdrew in 1884 (p. 1124). Although he did not publish details, it is probable that Cope used much the same reasoning as van Deinse and Junge in associating Dudley's scrag and the Pacific gray whale in the same genus. Thus it appears best to continue to use robusta Lilljeborg for the subfossil European form. As to the nomenclature of the Pacific gray whale, it seems reasonable to retain the trivial name glancus until better material of Eschrichtius /•o6t/5/i/5 may demonstrate specific identity (although Cederlund (1939, p. 282) believes the holotype sufficient for this). It is therefore sug- gested that the living Pacific gray whale be called Eschrichtius glancus (Cope). This status may be summarized thus: Genus Eschrichtius Gray 1864 (p. 350). Type species (by subsequent designation by Lilljeborg 1865, p. 494) : Balacnoptera robusta Lilljeborg 1861. Species Eschrichtius robustus (Lilljeborg) Balacnoptera robusta Lilljeborg 1861, p. 602. The Graso whale. Subfossil in northwestern F.urope (English Channel to Baltic Sea). 1952 PACIFIC GRAY WHALE 3 Eficliric lit ills glaucus (Cope) Acjapfuius glaucus Cope 1868a, p. 160, 1868b, p. 225. Rhnchiancdcs glaucus Cope 1869, p. 15. The California or Pacific gray whale; devilfish. Recent in North Pacific Ocean (Baja California to Korea). REFERENCES Ccderlund, B. A. 1939. A subfossil gray whale discovered in Sweden in 1859. Zjologiska Bidrag fraan Uppsala, 18, pp. 269 - 286, 5 pis. Cope, E. D. 1868a. [Remarks on Cetacea at meeting of 23 June.] Pioc. Acad. Nat. Sci. Phila., 1868, pp. 159-160. 1868b. On Agaphelus, a genus of toothless Cetacea. Proc. Acad. Nat. Sci. Phila., 1868, pp. 221-227. 1869. Systematic synopsis of the species of the cetaceans of the west coast of North America. Proc. Acad. Nat. Sci. Phila., 1869, pp. 14-32. 1884. Catalogue of aquatic mammals of the United States, by F. W. True. Review. American Naturalist 18, no. 11 (Nov.), pp. 1123-1124. VAN Deinse, a. B., and G. C. A. Junge 1937. Recent and older finds of the California gray whale in the Atlantic. Temminckia, 2, pp. 161-188, pis. 4-11. Dudley, Paul 1725. An essay upon the natural history of whales with a particular account of the ambergris found in the sperma ceti whale. Philos. Trans. Roy. Soc. London, 33, no. 387, pp. 256-269. F]rxleben, J. C. P. 1777. Systema regni animalis . . . Classis I. Lipsiae, pp. xlviii + 636 + [66). Gray, J. E. 1864. Notes on the whalebone-whales; with a synopsis of the species. Ann. Mag. Nat. Hist. (3) 14, no. 83 (Nov.), pp. 345-353. LiLLJEBORG, W. 1861 . Hvalben, funna i jorden paa Grason i Roslagen i Sverige. Forhandl. Skandinav. Naturf., 8de M0de, 1860, Ki0benhavn, pp. 599-616. 1865. [Note quoted in] Gray, J. E., Notice of a new whalebone whale from the coast of Devonshire, proposed to be called Eschrich- tius robustus. Ann. Mag. Nat. Hist. (3) 15, no. 90 (June) pp. 492 - 495. BREVIORA MiuseTuinn of Coumparative Zoology Cambridge, Mass. October 21, 1952 Number 8 A NEW SPECIES OF THE (TCLOSTOME GENUS PARAMYXINE FROM THE GULF OF MEXICO By Henry B. Bigelow and William C. Schroederi Up to the present time the myxinids, represented in the North Atlantic by the unibranchiate genus Myxine alone, have been unknown in the Gulf of Mexico. During 1951, however, the experimental trawlings of the U. S. Fish and Wildlife Service vessel "Oregon", in the northern part of the Gulf, yielded three specimens of a myxinid with six pairs of external gill openings that agree closely with the genus Paramyxine Dean 1904, reported previously from Japan only.^ Paramyxine, originally described from a single specimen, as the new species P. atami Dean 1904, is characterized, by its author, by having the outer "branchial ducts of distinctly unequal length, the most anterior several times the length of the posterior" (Dean, p. 22), so that the most anterior pair of external gill openings is opposite the 4th or 5th pair of interna,l gill sacs; and by having the successive pairs of gill openings close together. In other known myxinids with more than one pair of gill openings the successive pairs of external branchial ducts are of approximately equal lengths, or the 1st pair is only a little longer than the rearmost pair, and the successive pairs of gill openings are more widely spaced. On the type specimen the gill openings them- selves were oval, with their long axes transverse (except for the 6th opening on the left-hand side, which — receiving the oesophageal duct — is larger than the others), not round as they are in other polybranchiate species. But the gill openings were of various shapes in other specimens described subsequently by Matsubara (1937), showing that their shape is not a generic character as Dean supposed, 1 Contribution No. 610, from the Woods Hole Oceanographic Institution. 2 We owe to Mr. Stewart Springer the opportunity of studying these interesting specimens. or even a specific character, but is, likely, dependent to a considerable extent on the degree of muscular contraction. This leaves only the relative lengths of the external branchial ducts and perhaps the close spacing of the gill openings, as compared with other myxinids in which they are few in number, as bases for separating the genus Paramyxine from other polybranchiate myxinids. We accept the genus provision- ally, awaiting the opportunity of examining other myxinids with corresponding numbers of gill openings. Characters that Dean (1904, p. 22) included as specific for P. atami were the precise number (6) of pairs of gill openings; 1st pair of ex- ternal branchial ducts about 3 times as long as the 6th pair; 12 and 13 lingual teeth in the outer rows with 13 in each of the inner rows; anterior 3 lingual teeth of each row fused with one another at their base; white gill openings and conspicuously dark mucous pores. His illustration also shows a rather pointed snout, resembling that of Myxine, but a somewhat broader and more paddle-shaped caudal fin fold. The gill openings of the type specimen of atami occupied a space about 9 per cent as long as the distance from the tip of the snout to the most posterior pair of gill openings,^ with those on each side arranged close together in a regular row, and they are arranged simi- larly in an excellent specimen of atami, from Suruga Gulf, Japan, received recently by the U. S. National Museum (U. S. Nat. Mus. No. 161442). But Matsubara's (1937) observations have shown that the serial arrangement of the gill openings, like their shapes, may vary widely from individual to individual, and even between the two sides of the body of a given specimen, for they were irregularly arranged in 13 of his 14 specimens. Muscular contraction may play a role here. Matsubara also found that the dental formula is not a precise one, for his specimens had only 10 or 11 lingual teeth in each row, outer and inner, with only the first two fused at the base instead of three as was true of the type specimen. The Gulf of Mexico specimens agree in general with the accounts and illustrations of the Japanese atami and with the National Museum specimen, in body form, with rather pointed snout and a caudal fin fold somewhat wider than that of Myxinr. Their barbels are similar to those of atami in length and in arrangement; they have the same number (6 pairs) of external gill openings; the external branchial ducts (as measured by probing) are about as long relative to the 6th pair; ' As scaled from Dean's (1904, pi. 1. fig. 3) ilhistration. 1952 NEW SPECIES OF CYCLOSTOME the foremost pair is situated about as far rearward while the gill openings cover about as long a space relatively; the distribution of mucous pores is about the same; also the linear dimensions in general, as appears from the following tabulations: Table 1 Distances, in Per Cent of Total Length Gulf of Mexico atarni, atami, U. S. specimens, 338, Japan type Nat. Mus. 055 and 590 mm. 550 mm. specimen 500 mm. Snout to foremost mucous pore 7.5- 8.7 6.8 9.0 Snout to foremost gill openings 22.5-23.5 26.0 27.0 1st gill opening to 6th gill opening 2.4-5.7 2.5 2.4 6th gill opening to tip of caudal fin fold 72.0-74.3 71.4 73.0 Snout to origin of anal fin fold 37.1-50.0 41.4 43.4 Cloacal pocket (center) to tip of caudal fin fold 13.4-17.5 12.8 13.2 Table 2 Numbers of Mucous Pores on Each Side Gulf of Mexico atami, type' atami, U. S. specimens Nat. Mus. specimen Snout to foremost pair of gill openings 15-15 to 19-19 19-20 16-16 6th pair of gill openings to cloacal pocket 44-45 to 57-57 58-59 47-47 Center of cloacal pocket to tip of caudal fin fold 11-11 to 13-14 ? 10-10 The number of lingual teeth, too, is about the same in a Gulf of Mexico specimen of 505 mm. (14 and 13 in the outer rows, 11 and 12 in the inner rows) as it was in the type specimen of atami, with the same number (3) of anterior teeth fused basally in each of the outer two rows. And while two only, of the anterior teeth in each of the inner rows are fused in the Gulf of Mexico specimen, three were fused in the type, and Matsubara (1937, p. 15, text fig. 3) has already reported and figured Japanese specimens that agreed with the Gulf of I From Dean 1904, p. 11, text fig. 2, D. 4 BREVIORA NO. 8 Mexico one in this respect. The series of mucous pores on each side is interrupted near the anterior end of the cloacal pocket on the U. S. National Museum specimen of atami (Fig. 6), just as it is in the Gulf of Mexico specimens (Fig. 5) — in Mi/xine too, for that matter — to continue rearward at a higher level. Representation of it in Dean's (1904,pl. 1, fig. 3) illustration as continuing unbroken past the cloacal pocket seems to have been an error by the artist. But the rearward pores are situated higher above the ventral edge of the caudal fin fold in the Japanese atami (Fig. 6) than in the Gulf of Mexico specimens (Fig. 5). The depth, too, of the body relative to the distance from tip of snout to 1st pair of external gill openings is greater on all three of the Gulf of Mexico specimens (34 per cent to 37 per cent) than is pictured for atami either by Dean (30 per cent) or by Matsubara (22 per cent), or than it is on the U. S. National Museum specimen of atami (29 per cent). And while the hagfishes are so soft of body that measure- ments of different specimens cannot be accepted as precisely com- parative unless taken under equal tension, the difference in this respect between the two geographic populations may be large enough to be significant. The shape of the caudal fin fold also differs rather notice- ably, its lower margin being nearly straight in all three of the Gulf of Mexico specimens while it is pictured as moderately convex in the two Japanese specimens of atami which have been illustrated (Dean 1904, pi. 1, fig. 3; Matsubara 1937, pi. 1, fig. A), likewise in the National Museum specimen of atami (compare Fig. 5 with Fig. 6). Furthermore, neither of the two larger Gulf of Mexico specimens shows any trace of a dorsal fin fold anterior to the level of the cloacal pocket, whereas Matsubara's (1937, pi. 1, fig. A) illustration of atami clearly shows such a fold, extending forward for one third to one half the distance from the level of the cloacal pocket toward the level of the 6th pair of gill openings while the dorsal fin fold on the National Museum specimen of atami is continued forward as a definite though low dermal ridge to abreast of the 6th pair of gill openings.^ These differences between the Gulf of Mexican and the Japanese representatives of the genus seem sufficient, in combination, for recog- nition in nomenclature. We therefore propose for the former the new specific name springeri, as defined below. What is most interesting, however, in this case, is not that the Gulf of Mexican population 1 The anal fin fold also, of the National Museum specimen of atami is continued forward nearly to the 6th gill opening as a prominent dermal ridge. But the appearance of the latter, in alcohol, suggests that it was the result of aulmuscr contraction, not a normal characteristic. 1952 NEW SPECIES OF CYCLOSTOME 5 differs in minor ways from the Japanese — it would be astonishing if such were not the case — but that two populations so widely separated should resemble each other so very closely, and that myxinids with five to seven pairs of external gill openings should have been found at localities as widely separated as the Gulf of Mexico, Japan and China, New Zealand and South Africa, but nowhere else in the oceans.^ This instance does not stand alone as it concerns the Gulf, for the chimaeroid Hydrolagus alberti Bigelow and Schroeder 1951a, recently discovered in deep water there, seems more closely allied to the Japa- nese H. mitsukurii (Dean) 1904, than it is to either of the two species of its genus that are known from the North Atlantic, i.e., affinis (Brito Capello) 1868, and mirahiUs (CoUett) 1904, or than it is to H. collei (Lay and Bennett) 1839, of the Pacific coast of North America. Equally unexpected was the recent discovery, in the Gulf, in moder- ately deep water, of a representative {Springeria folirostris Bigelow and Schroeder 1951), of the rajoid family Anacanthobatidae, for the latter had been known before only off the Natal coast of southeast Africa; also of two representatives of the skate genus Cruriraja in Cuban-Floridan waters (C. atlaniis and C poeyi Bigelow and Schroeder 1948), for this genus, too, had been known only off Natal, southeast Africa, previously. Altogether, the bottom zone in the Gulf and around Cuba at 75 to 600 fathoms has yielded three new species of sharks, eleven new batoids, and one new chimaeroid since 1938, a foretaste of the rich harvest of unknown fishes to be expected from a more thorough ex- ploration of the deep slopes and of the sea floors of the Gulf and of the Caribbean. Paramyxine springeri, new species^ Type. A specimen, 590 mm. long, Oregon Station No. 489, Lat. 27°44'N., Long. 85°09'W., 254 fathoms, Sept. 29, 1951 (U. S. Nat. Mus. No. 161512). Additional material. Another specimen 505 mm. long from the same trawl haul, and a third, 338 mm. long, from Oregon Station No. 321, Lat. 27°27'N., Long. 87°19'W., 220 fathoms, April 28, 1951. The sex has not been determined for any of these specimens. Description. Dimensions, in per cent of total length, of type (590 mm.), 505 mm. specimen, and 338 mm. specimen. Depth of body, at 1st gill openings, 9, 7.8, 8.9. 1 For a synopsis of the species of myxinids, see Holly, 1933. ' Named after Stewart Springer who discovered this interesting cyclostome. 6 BREVIORA NO. 8 Distance, snout to 1st gill openings, 23.5, 23.2, 22.5. Distance, 1st gill openings to 6th gill openings, 2.4, 2.6, 5.7. Distance, snout to origin of anal fin fold, 50.0, 37.1, 39.8. Distance, center of cloacal pocket to tip of tail, 14.5, 13.4, 17.5. Depth, in per cent of distance, snout to 1st gill openings, 38, 34, 37. Number of mucous pores on each side, type (590 mm.), 505 mm. specimen and 338 mm. specimen: Snout to 1st gill openings, 19-19, 18-18, 15-15. 1st gill openings to 6th gill openings, 4-3, 3-3, 6-6. 6th gill openings to anterior end of cloacal pocket, 55-56, 57-57, 44-45. Anterior end of cloacal pocket to tip of caudal fin fold, 13-14, 11-11, 12-12. Snout obtuse-conical, much as in Myxine; body cylindrical an- teriorly, about 80 to 100 per cent as thick (transversely) at region of gill openings as deep there, but increasingly compressed posteriorly, until only about 38 to 54 per cent as thick as deep at level of cloacal pocket, and flat sided thence rearward to paper-thin margin of caudal fin fold. Depth of body about 5.4 to 6 per cent as great as total length at level of most anterior mucous pores, about 6.3 per cent at anterior end of cloacal pocket on type, and 6.6 to 8.6 per cent on smaller specimens. Fleshy rostrum, anterior to nasal opening, broadly rounded an- teriorly and wider than long,^ the nasal opening conspicuous below it. Barbels in the typical myxinid arrangement, the 1st pair about two- thirds as long as the 2nd pair, the 3rd pair a little the longest. Mouth, when closed, with the usual puckered or irregularly stellate appearance, the conical projection that flanks it on either side about as long as the 2nd pair of barbels. Thirteen and fourteen lingual teeth in the outer rows, eleven and twelve in the inner rows of 505 mm. specimen^, individual teeth conical with slightly blunted tips, curving a little rearward, the most posterior of each inner row strongly so; the 5th and 6th tooth longest in each row; the 7th and the following teeth successively shorter along each row; the rearmost only about one-third as long as the 5th and 6th. The first three teeth in each outer row, and the first two in each inner row fused at their bases; the outer row in each side about 1.2 times as long as the inner row; the longest teeth in the outer rows about 1 .3 times as long as the longest in the inner rows. 1 Longer than wide in Myxine. 2 The teeth of the type specimen cannot be counted without dissection. 1952 NEW SPECIES OF CYCLOSTOME 7 The six pairs of gill openings are low down on the sides, their fore-aft position as given in Table 1 (p. 3), occupying a shorter space on larger specimens (2.4 per cent of total length on the type) than on smaller, the series on the two sides converging rearward, with the 6th pair only about three-fourths as far apart as the 1st pair on the type, about four-fifths as far apart on the smallest specimen; the 6th opening on the left-hand side (receiving the oesophageal duct) larger than the others as is the rule among myxinids, and irregular in shape, the other gill openings ranging in shape from round to oval with the long axis transverse or slightly oblique, the precise shape seemingly dependent on muscular contraction; the largest (apart from the 6th left-hand) a little less than one-half as long as the 3rd (longest) pair of barbels. The gill openings are either in regular serial arrangement on each side as on the type (Fig. 2) and on the smallest specimen, or more or less irregular at least on one side, as on the 505 mm. specimen (Fig. 3), a difference perhaps associated with muscular contraction. The 1st pair of external branchial ducts (as sounded by probing, on one side of the body) are about three times as long as the 6th pair, both on the type and on the 505 mm. specimen. The mucous pores range in size from easily visible to so minute as to be discernible only on close examination. Their location on the body and in serial arrangement is as in the other myxinids; 19 pores on each side from snout to 1st gill openings on the type, 4 on one side and 3 on the other abreast the gill openings, 55 and 56 between gill openings and cloacal pocket and 13 and 14 rearward from the anterior end of the cloacal pocket on the type specimen; their numbers on the other specimens as given in Table 2 (p. 3). Totals of 77 pores on the 338 mm. specimen, of 88 on the 505 mm. specimen, and 92 on the 590 mm. specimen (type) show that the pores increase in number with growth. The pores along the section occupied by the gill openings are ventral from the latter, which it seems is characteristic of the poly- branchiate myxinids in general. The pores rearward from the gill openings are nearly in line with the latter, are low down on the sides nearly to the anterior end of the slit-like cloacal pocket, where the series is interrupted, with the nine posterior pores situated at a definitely higher level, a distinctive character of this species as contrasted with the Japanese atami (p. 4). On the type, those rearward from the cloacal pocket are also noticeably larger (hence more conspicuous) than those farther forward, and they are somewhat more conspicuous than the more anterior pores on the BREVIORA 505 mm. specimen as well. But this regional difference in size of mucous pores is not a specific character; rather it is likely a phase of muscular contraction, for the pores of the caudal series on the smallest specimen are no larger than those anterior to the cloacal pocket. The point of origin of the anal fin fold, about midway of the body on the type specimen, is considerably farther forward on the other two specimens (see Table 1, p. 3) showing that its precise situation is variable within rather wide limits. On the type, the fold itself is so narrow that its width is hardly measurable, and it is traceable rearward only about half the way from its point of origin toward the cloacal pocket. But it is wider on the other specimens (8 to 13 per cent as wide as the body is deep), and traceable rearward nearly or quite to the cloaca. To what degree these differences are associated with different stages in growth, or in the sexual cycle, is a question for the future. Neither the type, nor the 505 mm. specimen shows any trace of a dorsal fin fold anterior to the level of the cloacal pocket, their backs being evenly rounded, in which respect they contrast with P. atami as pictured by Matsubara (1937, pi. 1, fig. A). And while the smallest (338 mm.) specimen has a low fleshy ridge all along the back nearly to the level of the mouth, the fact that it is flanked on either side by a shallow and indefinite furrow suggests that it is the result of muscular contraction, rather than a definite and persisting fin fold. The caudal fin fold fringes the rear section of the trunk, from about opposite the anterior end of the cloacal pocket on the dorsal side, around to the posterior end of the cloaca on the ventral side. Its lower outline is nearly straight as remarked above (p. 4). The irregular fluting of its broadly rounded, posterior-dorsal margin on the smallest specimen (Fig. 5) seems to represent its normal outline better than its margin on the type, (Fig. 1) where it seems to have been injured. One of the most interesting features of P. springeri, reported to us by its discoverer, is that no one of the specimens discharged any mucus at all when brought on board, or so little that none was noticed. Color. All three specimens are greyish brown, faintly tinged with reddish, below as well as above, the edges of the caudal fin fold and of the anal fin paler; the gill openings encircled with white, rendering them conspicuous against the dark background; the mucous pores appearing as dark brown dots, each of them at the bottom of a shallow depression of the skin. Size. The maximum size is not known. 1952 NEW SPECIES OF CYCLOSTOME 9 Range and habits. So far known only from the northern side of the Gulf of Mexico, at the localities listed on page 5. The depths of capture, added to the fact that fishermen have never reported it, show that it is a deep-water species. Nothing else is known of its habits. REFERENCES BiGELOw, Henry B., and William C. Schroeder 1948. New genera and species of batoid fishes. Journ. Marine Research, vol. 7, no. 3, pp. 543-566. 1951. A new genus and species of anacanthobatid skate from the Gulf of Mexico. Journ. Washington Acad. Sci., vol. 41, no. 3, pp. 110-113. 1951a. Three new skates and a new chimaerid fish from the Gulf of Mexico. Journ. Washington Acad. Sci., vol. 41, no. 12, pp. 383-392. Brito Capello, Felix de 1868. Catalogo dos peixes de Portugal. Jorn. Sci. Math. Phys. Nat. Acad. Lisboa, vol. 1, no. 3, pi. 3, fig. lA. 1868a. Descripcaao de dois peixes novos de Portugal. Jorn. Sci. Math. Phys. Nat. Acad. Lisboa, vol. 1, no. 4, pp. 314-317. Collett, R. 1904. Diagnoses of four hitherto undescribed fishes. . . . Selsk. For., 1904, no. 9, 7 pp. Dean, Bashford 1904. Notes on Japanese myxinoids. Journ. Coll. Sci. Imp. Univ. Tokyo, vol. 19, art. 2, 23 pp., 1 pi. Holly, Maximilian 1933. Cyclostomata, in Schultze and Kukenthal, Das Tierreich, lief 59, 62 pp. Lay, G. T., and E. T. Bennett 1839. Fishes, in Richardson and others. Zoology of Captain Beechey's voyage, pp. 43-75. Matsubara, Kiyomatsu 1937. Studies on deep sea fishes of Japan. IIL On some remarkable variations found in Paramyxine atami Dean. Journ. Imper. Fisheries Inst. Tokyo, vol. 22, no. 1, pp. 13-15, 1 pi. 10 NO. 8 ^^^'^^^'-^i^n^'.^. Ai^A.iJ&d,^iS,:.».:.«^.m^S^ Fig. 1. Type specimen, x 34; the mucous pores somewhat exaggerated. Fig. 2. Ventral view of anterior part of type specimen, x }4', the mucous pores somewhat exaggerated. Fig. 3. Ventral view of gill region of specimen 505 mm. long; natural size; the mucous pores somewhat exaggerated. Fig. 4. Lingual teeth of specimen 505 mm. long, from the right-hand side; outer row above, inner row below, anterior ends to the left, x 2. Fig. 5. Side view of caudal region of specimen 338 mm. long, x J-^; the mucous pores slightly emphasized. Fig. 6. Side view of caudal region of P. atami, 500 mm. long, from Suruga Gulf, Japan, U.S. Nat. Mus.,No. 161442;the mucous pores slightly emphasized. E V I O R A MeseeitTii of Coimparative Zoology Cambridge, Mass. October 21, 1952 Number 9 NEW SPECIES OF EARTHWORMS FROM THE ARNOLD ARBORETUM, BOSTON By G. E. Gates I. From Allolobophora caliginosa (Savigny) 1826 there have recently been split off: yl. nociurna Evans 1946, and A. iowana Evans 1948. The new species were thought to be distinguished by differences in location of first dorsal pore, extent of male porophores, etc., and es- pecially by number of segments. These characteristics had not previ- ously been given such weight in lumbricid taxonomy, may be subject to variation and, as in the case of segment number, may be tedious to use when dealing with large numbers of worms. While working on collections from the Arnold Arboretum and other localities near Boston, it was found that material with characteristic caliginosa tubercula pubertates could be rapidly sorted into two groups according to the presence or absence of genital tumescences on segment xxxiii. Further study showed that those specimens with tumescences on xxxiii were referable to caliginosa as apparently restricted by Evans after breeding the worms in the laboratory. The worms with no tumescences on xxxiii were found to be consistently distinguishable from caliginosa by the same sort of characteristics as in the case of nocturna and iowana and, as well as those species, worthy of specific status. Allolobophora arnoldi n. sp. Type. Museum of Comparative Zoology, cat. no. 4441. Length, 55-100 mm., diameter, 4-6 mm. Segments, 152-194 (normal specimens). Unpigmented and almost white, or of greyish appearance or, especially in older worms of second season, with a brownish coloration. First dorsal pore, on 10/11, 11/12, or 12/13. Clitellum, on xxvii, J/^xxvii, or xxviii to xxxiv or V2XXXV. Tubercula 2 BREVIORA NO. 9 pubertates of double origin, on xxxi-xxxiii. Genital tumescences, in- cluding a and b setae, on ix-xi, xxx, xxxii, xxxiv, and occasionally also on XX vi. Spermathecal pores, on 9/10-10/11 on r lines, setae, male poro- phores, calciferous sacs, calciferous glands, typhlosole and last hearts, etc., as in caliginosa. Type locality. Arnold Arl)oretum, Boston, Mass. Distrihution. Known at present only from Boston and vicinity. Remarks. A. arnoldi is distinguished from caliginosa by the constant absence of genital tumescences on xxxiii, occasional presence of tumes- cences on xxvi (instead of xxvii), slightly greater anterior extent of the clitellum (xxvii or xxviii instead of 3^xxviii or xxix), greater number of segments, and perhaps by a slightly more anterior first dorsal pore. From iowana, arnoldi is distinguished by the constant presence of genital tumescences on xxxiv, constant absence of those tumescences on xxxiii, by the two-part tuberculum about as in caliginosa, possibly also by a slightly more posterior first dorsal pore, greater number of segments and absence or slighter development of pigmentation (not dark reddish brown) . From noctnrna, arnoldi is distinguished by the constant absence of tumescences on xxxiii and xii, smaller size, fewer segments, less re- stricted male porophores, and possibly by a more posterior first dorsal pore. Data as to variation in both caliginosa and arnoldi in the Boston area will be presented, if possible, in another contribution. II. In lumbricid taxonomy considerable weight has been given in the past to rather small differences in the tubercula pubertates. Thus, for instance, the species in the following pairs have been distinguished from each other by a one-segment homoeosis of the tubercula: Octo- lasium cyaneum (Savigny) 1826 and 0. lacieum (Orley) 1881, xxx-xxxiii and xxxi-xxxiv; Lumbricus rnhellvs Hoffmeister 1843 and L. castaneus (Savigny) 1826, xxviii-xxxi and xxix-xxxii. Further, Dendrobaena rubida (Savigny) 1826 has been distinguished from D. subrubicunda (Eisen) 1874, by a one segment difference in length of the tubercula, xxix-xxx and xxviii-xxx. Genital tumescences, as indicated above, may also be of considerable value in lumbricid taxonomy. Accordingly, it seems advisable to treat a worm distinguishable at present from the fa%mo*'a-complex only by differences in tubercula 1952 NEW SPECIES OF EARTHWORMS 3 and tumescences as specifically distinct. Allolobophora molita n. sp? Type. Museum of Comparative Zoology, cat. no. 4442. Length, 81 (+ ?) mm., diameter, 5 mm. Segments, 140 (+ ? pos- terior amputee?). Pigmentation lacking. First dorsal pore in ?12/13. Clitellum on xxviii-xxxiv, possibly reaching slightly onto xxvii. Tubercula pubertates, longitudinally placed bands in he, extending across entire lengths of xxx-xxxiii; with lateral margins nearly smooth, but median margins (quite definitely lateral to h lines) slightly indented by 30/31, 31/32, 32/33. Genital tumescences, including a and h setae, present on ix-xi, xxxii and xxxiii (feebly developed). Spermathecal pores, male porophores, setae, calciferous sacs and glands, typhlosole, last hearts, as in ealiginosa. In the clitellum intersegmental furrows are obliterated and dorsal pores are occluded but the epidermal thickening is slight. Sperma- thecae (in x and xi) are iridescent and presumably contain spermatozoa obtained from a copulatory partner. Male funnels are iridescent, also indicating sexual maturity in spite of the feeble development of the clitellum. Remarks. A. molita is distinguished from all of the ealiginosa- complex: caligmosa, noetvrna, iowana and arnoldi, by the four-segment tubercula pubertates. Further distinction may be provided by an apparent restriction of genital tumescences in clitellar region to xxxii- xxxiii. From all of the complex, except possibly iowana, molita is also distinguished by the quadripartite (instead of double) origin of the tubercula. In iowana, tubercula are band-like but nothing was said as to origin. Type locality. Arnold Arboretum, Boston, Mass. Distribution. At present known only from the type locality. REFERENCES Evans, A C. 1946. A new species of earthworm of the genus Allolobophora. Ann. Mag. Nat. Hist. (11) 13: 98-101. 1948. On some earthworms from Iowa, including a description of a new species. Ann. Mag. Nat. Hist. (11) 14: 514-516. E V I O R A Musemmti of Coiraiparative Zoology Cambridge, Mass. December 29, 1952 Number 10 ON THE EARTHWORMS OF NEW HAMPSHIRE By O. E. Gates So far only one record of earthworms in New Hampshire has been encountered in the literature. The worms were found "under the bark of trees" at elevations of 2,000-3,000 feet in the White Moun- tains by a German delegate to the International Zoological Congress at Boston in 1907. The specimens went to an Irish Museum where they were identified as Dendrobaena rubida (Savigny) 1826, by South- ern (1910). The species had not previously been reported from North America and since 1910 there has been no further record from this continent. The record was given in a short article on another kind of worm and, though the article was published in Philadelphia, the record was not included in Smith's review C1917') of the Lumbricidae of North America. Through the kindness of Dr. P. J. Darlington, Jr. who donated the bait remaining from a fishing trip, it is now possible to list a second species for New Hampshire. Southern (1910, p. 18) quoted his collector who stated that "earth- worms were very rarely met with in the forests at elevations of 2,000- 3,000 feet." Similar statements have been made in the past with ref- erence to forests of northern New England as well as of northern New York. Accordingly, Dr. Darlington's observations appear to be worthy of record : "My experience has been that it is always possible to find earth- worms in fair numbers even in the heavy woods in the more remote parts of New England, or at least in the White Mts. and the Connecti- cut lakes districts of New Hampshire. I have looked for them many times when I wanted them for bait for trout and I have always been able to find them under cover along the banks of brooks. I am fairly 2 BREVIORA NO. 10 sure that there are at least two species. One, . . . (see below) ... al- though it occurs on the stream banks, is usually well above the water level under stones or logs or in the roots of vegetation in sandy places that are damp but not wet. The other species is darker and much smaller, in fact hardly large enough to use for bait and it seems to oc- cur in much wetter places, under stones on gravel bars or sandy banks very close to the water level. . . . Practically every brook in New England is fished these days, and many of the fishermen bring worms with them and throw out what may be left at the end of their fishing trips, so that diflferent species must have been introduced many times, even in the deep woods." fDarlington, in lit.^ Allolobophora arnoldi Gates 1952 In steep bank of East Inlet Brook, about 2,000 ft. above sea level, near spruce, in heavy but not virgin forest, above Second Connecticut Lake, May 30, 1952. Four clitellate specimens. Dr. P. T. Darling- ton, Jr. collector. Length, 95-105 mm. Diameter, 5.5 mm. Segments, 91 (amputee), 106 (amputee), 156, 158. First dorsal pore, ?12/13 (1), ??13/14 (2). Clitellum, -Ixxvii-xxxiv. Tubercula pubertates bipartite and. as in A. caliginosa (Savigny) 1826. Genital tumescences, on ix-xi (4), xxx, xxxii and xxxiv (4), xxix (1). Remarks. Alive, these worms appeared to be quite without pigmient and they looked in that respect much as Orfolasiuni ci/avpvm (Savigny) 1826 usually does. The epidermis of segments xxvi, anterior half of xxvii, and of xxxv, is somewhat thickened and of a translucent appearance quite distinct from that of normal epidermis as well as from the yellow or white opacity of the clitellar epidermis, on three specimens. The demarca- tion of opacity and translucence on xxvii is indistinct, one gradually passing into the other, on the fourth worm. On the last segment of the amputees, rudiments of setal follicles and of nephropores are still recognizable. In one of those specimens the new anal region is small and scarcely distinguishable from the ninety- first segment but probably is demarcated by a quite fine, greyish trans- lucent furrow (presumably a rudimentary intersegmental furrow). In the other amputee no special anal region is recognizably demarcated. A. arnoldi was split off from A. caliginosa when it was found in Boston collections that all specimens having genital tumescences on 1952 EARTHWORMS OF NEW HAMPSHIRE 3 XXX, xxxii and xxxiv (with none on xxxiii) could also be distinguished from calighma hy greater rlitellar length nnd by larger segment number. The presence of worms with arnoldi characteristics in northern New Hampshire shows that those characteristics by which the species was recognized and defined are not merely those of a local variant in the Boston area. The occurrence of arvnJrJi in many other states, as well as in Canada, may be anticipated. REFERENCES CRATES, G. E. 1952. New species of earthworms from the Arnold .\rborptum. Boston. Breviora, No. 9: 1-3. Smith, F. 1 01 7. North American earthworms of the family Lumbricidae in the col- lections of the United States National Museum. Proc. U. S. Nat. Mus. 52: 157-182. Southern, R. 1910. A new species of enchytraeid worm from the White Mountains. Proo. Acad. Nat. Sci. Philadelphia, 62: 18-20. E V I O R A Mmseiuiiii of Cormparative Zoology Cambridge, Mass. IMarch 20, 1953 Number 11 CHARACTERS AND SYNONYMIES AMONG THE GENERA OF ANTS Part I By William L. Brown, Jr. Museum of Comparative Zoology Harvard University Work now under way is aimed at a revision of tlie basic classification of the Formicidae down to generic and subgeneric levels. The present classifications are considered both unnatural and impractical, con- cealing as they do large numbers of generic and specific synonyms. I propose to differentiate genera only on the basis of discontinuities of a largely morphological kind, and it is hoped that by this method a useful and accurate key to the genera will become possible. Syn- onymies given here require discussion that is best set forth separately, so as not to clutter more nearly final, synoptic classifications and keys now being prepared. Myrmecia Fabricius Myrmecia Fabricius, 1S04, Sj'st. Piez.: 423. Genotj-pe: Myrmecia gulosa Fabricius, by designation of Emery, 1911. Myrmecia (Promyrmecia) Emery, 1911, Gen. Ins., Fasc. 118: 18-19. Sub- genotype: Myrmecia aherrans Forel, by original designation. NEW SYNONYMY. Myrmecia Clark, 1952, Formic. Australia, Melbourne, 1: 21. Promyrmecia Clark, 19.52, Ibid., p. 119. The references of Clark should be consulted for further synonymy. Clark has stoutly defended the separation of Myrmecia and Pro- viyrmccia as distinct genera, first on the basis of ability or inability to jump of the species concerned, and later, when saltation proved to have numerous exceptions in both directions, on morphological grounds. 2 BREVIORA NO. 11 Years of study of the myrmeciines by this author have produced in the 1952 reference (p. 20) a couplet summarizing the characters supposed to separate Myrmecia and Promyrmccia. The three charac- ters contrasted are: (1) fraction of length by which the antennal scape surpasses the occipital border, (2) distinctness of metanotum, and (3) size, as mirrored in total length. In the first character, it may be observed from the descriptions of species of Myrmecia given by Clark himself in the pages following the couplet, that no less than one-third of the 59 forms recognized disagree with the couplet, and disagree in the direction of Promyrmccia. In at least some cases, the scape pro- portions are exactly as stated in the couplet for the upper limit of Promyrmecia. The distinctness of the metanotum is subject to varying interpretation, but from any reasonable point of view, the generality of the indistinctness of the metanotum in Promyrmecia cannot be defended. In fact, Clark's figures (op. cit.) in rapid survey are the best contradiction of his couplet that I can offer anyone without a large representation of material at hand for direct study. The "size" figures, even as cited in the couplet, are broadly overlapping and useless for the purpose to which they are put. Study of a good repre- sentative series of the species is convincing proof for me that size, as based on any measurable dimension or proportion of the worker or female so far utilized, is graded from the largest Myrmecia to the smallest Promyrmecia without a break. It is entirely possible that study of the male genitalia, when a sufficient number of males becomes available, will demonstrate a set of cleavages that along with other characters will serve as disconti- nuities for the proper splitting of Myrmecia. Clark's and other di- visions are here rejected for lack of evidence. It should be pointed out, in order to save possible future confusion, that Clark has syno- nymized Halmamyrmecia Wheeler with Promyrmccia, although he wrongly maintains Myrmecia nigrocincia Fr. Smith, its type, in the genus Myrmecia as he has defined it. Rhytidoponera Mayr Ectatomma (Rhijtidoponera) MajT, 1862, Verh. zool.-bot. Ges. Wien 12: 731. Genotype: Ponera araneoides Le Guillou, bj- designation of Emery, 1911. Rhrjtidoponera {Chalcoponera) Emery, 1897, Ann. Mus. Stor. Nat. Geneva 38: 548. [Sub] genotype: Ponera metallica Fr. Smith, by designation of Emery, 1911. NEW" SYNONYMY. Rhytidoponera and Cfialcoponera treated as distinct genera, Wheeler, 1922, 1953 ANT CHARACTERS AND SYNONYMIES 3 Bull. Amer. Mus. Nat. Hist. 45: 643-644. Clark, 1936, Mem. Nat. Mus., Melbourne, 9: 14-15. The large genus Rhytidoponera in the present broad sense includes species of typical ectatommine characteristics, distinguished in having the inferior pronotal margins just in front of the fore coxae armed on each side with an acute tooth; the hind coxae are completely unarmed above. Distribution is primarily Australian and Papuasian, with outliers in the "Wallacia" region and in the southern Philippines to the west, and in New Caledonia in the east. The separation of Rhylidoponcra from Chalcoponera, either generic or subgeneric, has been accepted from the time of Emery's first di- vision of the few species then known. Rapid accretion of species in both groups has frequently led to puzzlement of authors trying to place new species in one group or the other, and this puzzlement has resulted in several published expressions of doubt accompanying specific descriptions. The separation has been based on (a) proportions of certain antenna! segments in worker and male, (b) development and pectination of the spurs of the middle and hind tibiae in the worker, (c) development of notaulices in the male, and (d) presence or absence of a normal winged female caste (Wheeler, loc. cit. 1922). If one applies these Emery- Wheeler key characters rigorously to a wide variety of species in Rhytidoponera and Chalcoponera, discrepancies are not long in ap- pearing. The tibial spurs of the middle pair of legs are reduced and narrow (rarely absent) in all species of both genera examined for this work, and may be safely disregarded. The extremes of development of breadth and pectination of the posterior spurs are largely in accord with the conventional separation of the two groups, but exceptions are glaring, and gradual transition from one type to the other is evident among a restricted selection of eastern Australian species. In the species tenuis Forel, which has Chalcoponera antennal funiculi and which has been placed in Chalcoponera by all authors, the posterior tibial spurs are minute and show the strongly reduced pectination supposed to be characteristic of Rhytidoponera s. str. Conversely, certain Rhytidoponera s. str. species, such as R. malandensis Forel, have the posterior spurs large, broad, and strongly pectinate; in fact, R. malandensis has the spurs considerably more broadly pectinate than in any Chalcoponera species I have seen, including the genotype. In the worker funiculi, the proportions of the "critical" segments 4 BREVIORA NO. 11 are intergradient and fully ambiguous in a number of species, among them R. reticulata Forel, C. lamellinodis Santschi (paratypes), C. duhia Crawley, and all the species of the C. impressa Ma\T complex. More ambiguous examples or outright contradictory species could be cited, but the above cases should serve to prove the point so far as the worker is concerned. Concerning the male characters, a case similar to that for the workers can be drawn, but it will serve merely to focus on the New Caledonian fauna, consisting of a handful of forms in which the workers have been placed in Chalcoponera on the usual characters. The males of these species are, unfortunately for the generic division, endowed with the characters of Rhytidojjonrra s. sir. Emery (1914, in Sarasin and Roux: Nova Caledonia, Zool. 1: 397) covers this situation adequately when he states: "The males of the New Caledonian species of Chalcoponera {Rhytid. fulgcns, numcensis, atropurpurca and acupuncta) make an exception to the character that I have attributed to the subgenus concerning the structure of the antennae (Genera Insectorum, fasc. 118, p. 39). For the present, I am incapable of distinguishing the males of the two subgenera of Rhytidoponera." Apparently Wheeler overlooked this paragraph when he raised Chalco- ponera to generic rank in 1922, citing in his key the same discredited male characters. I have dissected the male genitalia of several species representing Chalcoponera and Rhytidoponera s. sir., and the prepa- rations fail to show differences except minor ones among the various species, without regard to the old division. The volsellae are uniformly much like those of Myrmica and the Dacetini in the subfamily Myrmi- cinae, and the other parts are on the usual formicid pattern. It seems fairly obvious that most describers have pretty consistently ignored the formal characters in assigning new species to one group or the other; rather, assignment seems to have been made chiefly on the basis of size and habitus. On such a basis, I can make out not two, but several, species-groups of fairly distinctive relationships, but broadly intergradient one to the next. Of these, the impressa complex seems to have a combination of characters approaching nearest the hypothetical generalized Rhytidoponera. Disregarding certain aberrant species and complexes, such as the tnrneri group and the New Cale- donian stock, the development of the majority of species seems to have followed two lines: one toward large species of the type of viayrt Emery, araneoidcs Le Guillou, and punctata Fr. Smith, and one toward the type of metallica Fr. Smith and victoriae Andre. In the farfiung 1953 ANT CHARACTERS AND SYNONYMIES 5 regions penetrated by Rhytidoponcra s. lat. on the Australian continent, these two broad adaptive types have met with relative success, while the more generalized connecting types have been restricted to favorable forested areas, largely in the mountains of eastern Australia. Thus, while the Rhytidoponcra species inhabiting the deserts, scrubs, heaths and eucalypt woodlands of the great, arid, open regions of Australia may seem to be amenable to assortment into two distinct groups, it must be borne in mind that the more generalized forest-loving species, while less familiar to the obser\'er, form an effective bridge joining these two groups. Before leaving Rhytidoponcra, a word is required on the females. It has been assumed that Chalcoponcra species always produce normal winged females, and that Rhytidoponcra s. sir. species do not. In 1950, Clark {in litt.) indicated to me that he had found females of some sort, presumably differentiable from the workers, in the latter group. To date, he has published nothing concerning this find, so that it is fitting that the old assumption be continued until definite information appears in print. My own field investigations in Australia (1950-51) indicate that the R. imprcssa group {imprcssa, chalyhaea, splcndida) normally produces a regular yearly crop of males and females in most mature nests during the first part of the dry season in each of the regions inhabited. Both sexes are winged before the nuptial flight, and ferti- lization and nest-founding are presumed to follow conditions more or less normal among the Ponerinae, as isolated females have been found dealated and inhabiting small cells in various situations in normal habitats. Outside the season for production of winged forms, only dealate females have been found in established nests in addition to the workers and such brood as occurs. When winged forms are found in a nest, these are always of both sexes so far as my observations go. In species of the mctaUica group, a situation of a cjuite different sort is apparent. Among the common and widespread members of this group, such as mctaUica, tasmanicnsis, and their very close allies, winged or obviously dealate females are very rare or unknown. The victoriac group is similar in this respect, although I have found a dealate female in a mature nest of the species common at Kuranda, in North Queensland, and have also found a lone winged female of victoriac floundering in the damp sand at the surf edge at Seaford, on Port Phillip Bay, in Victoria. Of several thousand nests of victoriae, mctaUica, tasmairicnsis and very close allies of these species that I have opened, not one (with the single noted exception above) has yielded 6 BREVIORA NO. 11 a female of the normal type, either alate or dealate. Yet there are a few isolated specimens of normal females in these groups in the Mu- seum of Comparative Zoology collection, most of which seem to have been taken singly on the wing or otherwise. While it is conceivable that the rarity of females in the nests of such species may reflect the methods of investigation, and not true absence of this caste, another circumstance makes this seem unlikely. I lefer to the very widespread occuiTence of winged males in nests of the same species at all seasons of the year, a phenomenon that stands out among my Australian observations even though I failed to take quantitative notes that would have made it much more convincing. These males are usually present in small numbers, and may be found in the upper as well as the lower chambers of a given nest. They are very active, and either run to hide quickly or else take to ready flight. A more limited number of observations on the large deserticolous members of Rhytidoponera s. sir. indicates that males are to be found in their nests on a somewhat similar basis, though observations through several seasons will be needed to confirm this. Among some other, more rare and distributionally restricted species, R. Croesus Emery appears to have normal females most or all of the time, while R. aspcra (Roger) is recorded as having this caste at least some of the time. For most of the species of the "inter-subgeneric" and aberrant groups, workers only are known at present, but these forms are so rare and so little-investigated that presence or absence of winged females cannot be assumed on any reasonable grounds. Under the circumstances as outlined, utilization of the presence or absence of a true female caste as a taxonomic character would seem entirely premature. The situation in the female-less or female-rare species that have been fairly extensively investigated is, however, of considerable interest apart from taxonomic considerations. In two such species, I have seen a male seeking to enter an alien nest, apparently of the same form. In the spring, in southeastern Australia and the dry Northern Flinders Ranges of South Australia, at least, males of these forms are often seen hawking in rapid flight low over the ground in the manner of certain mutillid and thynnine males during the sexual search. Observations in 1951: On the 12th of September, in a strip of open eucalypt woodland at Burwood, near Melbourne, Victoria, I noticed such a male in flight on a cool, sunny afternoon. While following him, I suddenly saw him 1953 ANT CHARACTERS AND SYNONYMIES 7 alight and disappear without hesitation into a small hole in the ground. This hole, when attacked with a trowel, proved to be the entrance to a nest of R. tasmaniensis, a nest containing, so far as I could dig on this occasion, workers and two males, the latter indistinguishable from males taken in nests of this species on other occasions. One of these males, found just below the entrance, must have been the male I saw entering just before I started to dig. At Wilpena Pound, a sort of natural oasis in the arid Flinders Ranges of South Australia, the evening of November 25 brought an abundant flight of a large brown Rhytidoponcra species, entirely males, to our camp pressure lanterns. Later that night, some rain fell and there were electrical storms on the adjacent peaks. On the next day, in an area of Triodia grass within the Pound, I found a low, gravel-studded mound with wide, slit-like entrance, of the type made by certain arid- land species of Rhytidoponcra and Camponohis. The nest was seen in the path taken on the way out, and was not excavated until returning toward camp, late in the afternoon when the sun had fallen very low. While I was bending over to deal the hard clay mound a first blow, a large male of Rhytidoponern came, flying in low over the ground from a distance, and landed directly on the lip of the entrance slit, which it immediately entered. This was surprising, as no sign of life had been detected around the mound at either time of inspection (the species concerned, probably R. viayri Emery, is, with its close relatives, a primal ily crepuscular and nocturnal forager). The nest was immedi- ately attacked, and the male recovered quickly several inches down along the main entrance passage. Further digging secured only a few workers of R. mayri Emery.^ While these two incidents, observed by chance, are scarcely to be considered definite proof of an established behavior pattern, the obser- vations made to date on various Rhyiidoponcra species suggest the following hypothesis, now being tested by Haskins {in Hit.). Rhytidoponera {s. lat.) varies by species and species-groups in presence, absence or rarity of a normal female caste, and varies corre- spondingly in behavior of and toward the males. In the species in which normal females are rare or absent, it may be assumed that deposition of fertile eggs is wholly or largely taken over by workers ' The species described by Clark as R. stridulator and R. dixoni. along with some other forms' are supposedly distinguished from R. mayri by details of petiolar structure and body sculpture. Single nests of what I take to be mayri, however, show a wide range of variation in these same characters, and Clark has not made the differences sufficientlj' clear to convince me that his species are really distinct. 8 BREVIORA NO. 11 or highly modified ergatoids. It is possible that apparatus for the retention of sperm is strongly impaired in such hypothetical individuals; if so, then the need for constant refertilization would seem to call for the constant presence of consort males in the nest. IMeager obser- vations seem to indicate that males sometimes, perhaps normally, leave the parent nest and enter another nest of the same species, where they may remain as guests for long periods. If observation and controlled experiment corroborates the above hypothesis, a new and extremely interesting kind of behavior will be added to the multiplicity of patterns known among the ants, and the origin of certain similar male habits among the Dorylinae may be easier to understand. Centromyrmex Mayr Centroinynnex Mayr, 1866, Verh. Zool.-bot. Ges. Wien 16: S94. Ckniotype: Cenlromyrmex bohemani Mayr, monobasic. Typ/i'o/eros Kara wajew, 1925, Konowia 4: 128. Genotype: Tijphloteras hamu- latum Karawajew, monobasic. NEW SYNONYMY. Karawajew based his genus on a specimen with only a single, large pectinate spur on the posterior tibiae, and thereby carried it out to Emery's Group III of the tribe Ponerini. This placement is only another example of the great faith formerly held by many authors in the constancy and taxonomic importance of the number and condition of the spurs of the middle and hind tibiae. In fact, the character in question is a very poor one upon which to base a classification, es- pecially in the Ponerini. If one reviews the situation in various Centromyrmex species, including Centromyrmex hamidatus NEW COMBINATION, it is at once apparent that the number and con- dition of the spurs in question differ by species, and possibly even within species. Thus, either the middle, or the posterior, or even both pairs of tibiae may possess the extra lateral spur in a more or less rudimentary condition, while the degree of de\elopment of the medial spurs and their pectination is also variable from one species to the next. In other characters, including the striking general habitus and the probably general termite-eating proclivities, the species here included in Centromyrmex form a very homogeneous and natural-seeming group. To break up this combination on the basis of spur characters, one must, as did Karawajew, go to extraordinary lengths in invoking "convergence" as a possible explanation of the common resemblances; to be thorough in applying the same logic, one would have to distribute the few species involved among three or four genera. In any case, the 1953 ANT CHARACTERS AND SYNONYMIES 9 tibial apices of several of the species bear such a dense growth of heavy, spine-like setae that the identification of a spur rudiment of similar size becomes an academic exercise. In addition to the synonymy of Ti/phlotcras with Ccniromyrmcx, it becomes necessary to point out that the subtribe Centromyrmicini of Emery is a NEW SYNONYM of tribe Ponerini if the spur characters wnll not hold. Pristomyrmex Mayr Pristomyrmex Mayr, 1866, Verh. Zool.-bot. Ges. Wieu 16: 903. Genotype: Pristomyrmex pungeiis Mayr, monobasic. Odoniomyrmex Andre, 1905, Rev. Ent. Caen 24: 207. Hylidris Weber, 1941, Ann. Ent. See. Amer. 34: 184, 190. Genotype: Hylidris rnyersi Weber, monobasic. NEW SYNONYMY. Hylidris defended, Weber, 1952, Amer. Mus. Novit. 1584: 15-22. When, in 1941, Weber first described Hylidris, he did so without realizing that it might be closely related to another genus like Pristo- viyrmex. By 1952, he has realized this relationship and is at some pains to mark out its boundaries. His new approach to the situation lies in segregating the African species of Pristomyrmex from the Indo- Australian ones ; the former group, rallied about H. mycrsi as genotype, is given the name Hylidris. He is vague about what genus the Indo- Australian species are to belong to, but these are presumably to remain in Pristomyrmex. The characters of Hylidris are cited briefly, but are not contrasted with those of Pristomyrmex s. str. in any direct state- ment. Instead, Weber gives briefly his views on generic limits, w4iich he feels should be arbitrarily drawn in some cases. (For my contrasting opinion, see the introduction to this paper.) Summing up, it may be said that Weber's division of Pristomyrmex is based upon characterization of only one of the resultant groups, and no assurance is given that the characters are exclusive to that group. In point of fact, these characters are 7iot exclusive to the African group. Since Weber claims to have examined the Indo-Australian species, at least in part, it may be wondered that he did not note this fact for himself; perhaps he did note it, and oftered the generic-limit discussion as a defense against protests he felt would be lodged against Hylidris. I cannot follow Weber's generic split, which I regard as wholly arbitrary. Significantly, Weber fails to mention the old synon^Tn Odontomyrmex, a name put forward by Andre for a species that appears to have the chief definitive characters of Hylidris. IVIann (1919, Bull. Mus. Comp. Zool. 63: 341) delivered the coup de grace to Odonto- 10 BREVIORA NO. 11 myrmex when he noted two specimens of his Pristomyrmex obesus melanoticus having a prominent tooth on one side of the pronotum, but none on the other side. We do not need to appeal to abnormal specimens in this case, however, as the known, normal specimens of the Indo-Australian and African regions form a tightly intergradient series with respect to the development of the posterior propodeal and anterior alitruncal pairs of teeth and the smooth to foveate-reticulate sculpture. This series, in my opinion, does not even split into re- spectable species-groups on the basis of the known characters, let alone genera or subgenera. Series of several undescribed species in the Museiun of Comparative Zoology and J. W. Chapman collections serve only to fortify this opinion. Gauromyrmex Menozzi Gauromyrmex Menozzi, 1933, Natuurhist. Maandblad 22: 146. Genotype: Gauromyrmex bengkalisi Menozzi, monobasic. Solenomijrma Karawajew, 1935, Treubia 15: 103. Genotj'pe: Solenomyrma acanlhina Karawajew, monobasic. NEW SYNONYMY. Acalama M. R. Smith, 1948, Jour. N. Y. Ent. Soc. 56: 20.5-207. Genotype: Acalama donisthorpei M. R. Smith, monobasic. NEW SYNONYMY. This genus is very doubtfully distinct from Vollcnhovia Mayr, from which it may at present be distinguished by means of the 11-segmented antennae and the bidentate propodeum, vs. 12-segmented antennae and unarmed propodeum in Vollenhovia. The characters are very weak ones, and may be compromised in species in this complex that are presently not available to me. Vollenhovia emeryi Wheeler, with 12-segmented antennae, has minute propodeal teeth, and the median funicular segments are so reduced in length as to be virtually obsolete; the step to Gauromyrmex from this species is a very short one. The amber species Vollenhovia beyrichi (Mayr), if properly placed by Wheeler, would be intermediate in the critical characters. Types of Gauromyrmex bengkalisi and Acalama donisthorpei were examined, and are considered specifically distinct, but not generically so. Specific synonymy is formalized as follows: Gauromyrmex acanthinus (Karawajew) new combination Solenomyrma acanlhina Karawajew, 193.5, Treubia 15: 103-104, fig. 23, worker. Acalama donisthorpei M. R. Smith, 1948, Journ. N. Y. Ent. Soc. 56: 207-208, figs. 1, 2, worker. NEW SYNONYMY. Types of Smith's species and Gauromyrmex bengkalisi have been examined through the kindness of Dr. Smith. The former species has 1953 ANT CHARACTERS AND SYNONYMIES 11 been compared with Karawajew's description and figure, and also with the West Chinese series taken by myseh', and mentioned by Smith {loc. cit., p. 206). The comparison gives the distinct impression that one is deahng with a single variable species. Variation is chiefly in size (slight), distinctness and acuteness of propodeal teeth, and depth of pigmentation, but the present evidence does not warrant subspecific distinction in my opinion. This ant appears to be rather common in India and China, and I believe that I have seen it on several occasions in Bengal Province nesting under loose bark, though the actual specimens were lost in a wartimic shipment. In life, the distinctive point of recognition lies in the extremely depressed ap- pearance of the ant, even when moving about. It appears to the naked eye somewhat like a flattened Lcj^totliorax, and its attachment to tree trunks and plant cavities may be correlated with the habitus. At the time of Dr. Smith's investigation of this insect, he sent specimens to me, and I failed to recognize it as any described species. The published descriptions of Menozzi and Karawajew came to light later. The entire case of these synonymous genera should prove my frequently-held point that the classification of the Myrmicinae is long overdue for complete revision. Tribe AMBLYOPONINI Onychomyrmicini Ashmead, 1905, Canad. Ent. 37: 382. Examblyoponini Donisthorpe, 1949, Ann. Mag. Nat. Hist. (11) 15: 401. Reneini Donisthorpe, 1947, Ann. Mag. Nat. Hist. (11) 14: 183. NEW SYNONYMY. Prionopelta Mayr Prionopelta Mayr, 1866, Sitzb. Akad. Wiss. Wien 53: 503. Genotype: Pri- onopelta punctulata Mayr, monobasic. Ponera Fr. Smith (partim), 1860, Journ. Proc. Linn. Soc. London, Zoo). 4 (suppl.): 105, nee Latreille. Examblyopone Donisthorpe, 1949, Ann. Mag. Nat. Hist. (11) 15: 401. Geno- type: Exambhjopone churchilli Donisthorpe, monobasic: vide infra. Renea Donisthorpe, 1947, Ann. Mag. Nat. Hist. (11) 14: 183. Genotype: Renea testacea Donisthorpe, monobasic. {Nee Renea Nevill, 1880, in Mollusca.) NEW SYNONYMY. Vide infra. Incredible as it may seem, Donisthorpe based his two new genera and two new tribes on two new species, both of which are synonymous with Prionopelta majuscula Emery. The same author had already proposed a new name for a synonym of the same species, so he has 12 BREVIORA NO. 11 achieved the unequalled feat of proposing for the same species three new specific, two new generic and two new tribal names. The syn- onymy of Prlonopclta majuscula follows : Prionopelta majuscula Emery Ponera similliina Fr. Smith, 1860, Journ. Proc. Linn. Soc. London, Zool. 4 (suppl.): 105, "worker," 9, nee Fr. Smith, 1860, op. cit., p. 104. NEW SYNONYMY. IRhopalopone simillima, Emery, 1900, Term. Fiizetek 23: 311. 1911, Gen. Ins. 118: 35. Prionopelta majuscula Emery, 1897, Term. Fuzetek 20: 595-596, worker, 9. Brown, 1951, Bull. Brooklyn Ent. Soc. 46: 102, Examblyopone churchilli synonymized. Prionopelta poultoni Donisthorpe, 1932, Ann. Mag. Nat. Hist. (10) 10: 462, nam. pro Ponera simillima Fr. Smith, IL NEW SYNONYMY. Examblyopone churchilli Donisthorpe, 1949, Ann. Mag. Nat. Hist. (11) 15: 401-402, 9. Renea testacea Donisthorpe, 1947, Ann. Mag. Nat. Hist. (11) 14: 18.3-186, fig., worker, 6"; pp. 590-591, 9. NEW SYNONYMY. In his original description of Ponera simillima II, Frederick Smith describes a "worker" and also gives characters purporting to be those of a winged female. In his treatment of 1932, Donisthorpe mentions as Smith's type "1 dealated 9 " from "Dor." [Dory, A. R. Wallace]. It is by now fairly well known that some of the Dory insect material, at least among the coleopterous collections in the British Museum, is suspected to be from other East Indian localities. Smith's description fits Prionopelta majuscula better than it does either of the other two Prionopelta species known from the Indo-Papuasian area (P. kraepelini Forel and P. opaca Emery), so there is no reason to question this particular record for a widespread New Guinea species. The as- sumption must be made, of course, that Donisthorpe's 1932 assignment to Prionopelta was correct; there is every reason to believe that it was so. In 1951, I determined that a paratype female of Examblyopone churchilli was synonymous with P. majuscula and returned the type to Dr. E. S. Ross at the California Academy of Sciences. Dr. Ross was the original collector (at Mafiin Bay, Dutch New Guinea) of both E. churchilli and Renea testacea, and he has recently sent me workers and males from the type series of the latter species. The workers fit Emery's diagnosis of P. majuscula very neatly, except for the usual small size difference resulting from Emery's habitual under- 1953 ANT CHARACTERS AND SYNONYMIES 13 measurement. This being the case, I asked Dr. Ross to compare the female specimens assigned by Donisthorpe to R. testacea with the female holotype (defective) of E. churchilli. He has replied {in litt.) that the correspondence of form, etc. is as good as can be expected, allowing for the missing parts of the E. churchUli type. The R. testacea workers differ as expected from authentic workers of PrionopcUa opaca and P. kraepelini in the Museum of Comparative Zoology and match the differences from P. opaca listed by Emery in the original de- scription of P. majuscula. P. majuscula is larger than the other two Indo-Papuasian species, and its worker has the dorsum of the head and alitrunk very definitely shining, with minute, spaced punctulation. The female is considerably larger and darker than the worker, and has stronger punctulation; the head is darker than the rest of the body. Both P. kraepelini and P. opaca are not only smaller, but also there is only a slight difference between the worker and female stature in these species; the dorsum of the head is densely and more coarsely punctulate and opaque, especially in opaca. The worker and female of kraepelini and the worker of majuscula are pale to bright yellow in color, while the females of majuscula and opaca and the worker of opaca are darker, ranging from ferrugineous brown to blackish-brown. P. opaca and P. majuscula appear to be widespread on New Guinea and neighboring islands, while P. kraepelini is a more westerly, Indomalayan species that has spread into the Pacific as a tramp and has reached the Philippines in the north at Dumaguete, Negros Oriental: several series (J. W. Chapman and D. Empeso). B R E V I Miiseuim of Coimparative Zoology Cambridge, Mass. April 23, 1953 Number 12 A CAVE FAUNULE FROM WESTERN PUERTO RICO WITH A DISCUSSION OF THE GENUS ISOWBODON By Thomas E. Reynolds Zoology Department, Massachusetts College of Optometry Karl F. Koopman Biology Department, Queens College Ernest E. Williams Harvard University In spite of the diligent v^^ork of H. E. Anthony nearly 40 years ago. the vertebrate paleontology of Puerto Rico must still be described as incompletely known. Anthony's labors have not been followed up in any systematic fashion by other investigators, and much therefore remains to be done. The present paper places on record a collection from an area of the island relatively neglected by Anthony — the extreme western portion of Puerto Rico. The collection, though small and apparently repre- senting a fauna of no high antiquity, permits new locality records for certain forms and includes a very interesting specimen of the genus Isolohodon which has compelled us to inquire into the variability of that genus and into the propriety of taxonomic subdivisions of it. The materials here recorded were obtained by the senior author during the fall and winter of 1948-49 at the mouth of the C'ueva Monte Grande between Mayagijez and Cabo Rojo (Distrito de Mayagiiez, Pueblo de San German, Barrio de Monte Grande). The specimens have been deposited in the Museum of Comparative Zoology. There were several levels in the deposit, and part of the deposit was kitchen midden, but no record was kept of the level at which specific material was found. 2 BREVIORA NO. 12 We cite below only the mammalian component of the faunule. There are, however, also fish, bird, turtle, snake and lizard bones. The lizard jaws have been identified by Rodolfo Ruibal and Max K. Hecht as all belonging to Anolis cuvicri, the living giant anole of Puerto Rico. No attempt was made to identify limb bones in any group. In the case of the mammals we take this opportunity to cite also localities not previously published but represented in the col- lections of the Museum of Comparative Zoology (MCZ) or the American Museum of Natural History (AMNH). The Monte Grande Faunule INSECTIVORA Nesophontes edithae Two mandibular halves. This large extinct insectivore of Puerto Rico has previously been reported from Hacienda Jobo, Morovis and Utuado. The MCZ col- lections show it also from Ciales, Manati and Rio Piedras. CHIROPTERA Noctilio leporinus mastivus One complete skull. Two mandibles. Though recorded by Anthony (1926) only from Old Loiza, this form was collected by Anthony and Goodwin at Vega Baja (AMNH speci- men) during a 1926 West Indian expedition which has never been fully reported in the literature. The present is therefore the third record for this large fish-eating bat in Puerto Rico. Chilonycteris parnellii portoricensis One skull. The specimen so referred seems somewhat small for the species, but it is much too large to be the small Puerto Rican species of the genus {C. fuUginosa inflata). It is probably a young specimen. This subspecies seems to be relatively rare in collections, only 11 specimens having been previously obtained, representing four locali- ties: Cayey, Morovis, Pueblo Viejo and Trujillo Alto. MONOPHYLLUS PORTORICENSIS One skull. 1953 A PUERTO RICAN CAVE FAUNULE 3 This species was recorded by Anthony (1926) from five localities: Bayamon, Cayey, Morovis, Pueblo Viejo and Trujillo Alto. Brachyphylla cavernarum Four more or less complete rostra. Six complete mandibles. Seven half-raandibles. Recorded by Anthony from six localities: Cayey, Comerio, Corozal, Morovis, Pueblo Viejo and Trujillo Alto. Artibeus jamaicensis jamaicensis One skull. Six mandibles. This very common form was obtained or reported by Anthony from 14 localities. Eptesicus fuscus wetmorei One mandible. Anthony (1926) reported this species from four localities (Maricao, Morovis, Pueblo Viejo and San German). He had collected it also at Trujillo Alto (AMNH specimen) during the 1916 expedition but somehow overlooked this record in his publication. RODENTIA Rattus sp. Two mandibles. One half-skull. Rattus, as always in the Western Hemisphere, indicates that at least part of the Monte Grande faunule was post-Columbian in age. Elasmodontomys obliquus A fragment of an upper incisor and a third right upper molar appear to represent this species. This large extinct rodent has previously been reported only from Ciales, Morovis and Utuado. The MCZ collections add it also from Rio Piedras. ISOLOBODON PORTORICENSIS Nineteen skull fragments. Fifty-five half- or partial mandibles. The bulk of the material from Cueva Monte Grande is of a medium sized hypsodont rodent. The enamel folds of the molars are nowhere completely separated to form laminae. The lateral surfaces of these 4 BREVIORA NO. 12 teeth have the pattern of striations characteristic of the genera or subgenera Isolohodon and Aphadrcvs. Even a cursory inspection of the molar pattern, however, rules out Aphadrcvs (Miller 1929a, pi. 2). The specimens then seem referable to Isolohodon. The only species of that genus which has previously been recognized on Puerto Rico is Isolohodon portoriccnsis, and almost all of the specimens agree well with this form in both size and pattern and may immediately be placed as representatives of this species Specimens of Isolohodon portoriccnsis from Monte Grande afford a new record of locality but less clearly than the other forms recorded above, since Anthony has already reported this species from nearby Cabo Rojo as well as from Ciales, Manati, Salina, San German and Utuado. (The MCZ adds Saliche and Aguirri.) Four of the Monte Grande mandibles, however, are not immediately classifiable as /. portoriccnsis and require special attention. All are much smaller in size than fully adult /. portoriccnsis. Two are obvi- ously immature. One of these is edentulous; the other has completely unworn crowns to the teeth, which are, therefore, not readily compared with the worn adult molar pattern. However, by breaking the ventral surface of the mandible, it has been possible to discover from under- neath the pattern of the enamel ridges which will be eventually re- vealed by growth. The enamel pattern so revealed is exactly that of typical /. portoriccnsis. The third small mandible is larger than the other two and although considerably smaller than adult /. portoriccnsis shows no morphological evidences of immaturity (last molar not fully erupted, unworn molar pattern, absence of bony shelf behind the last molar). The fourth mandible is similar to the third in size but is edentulous. Subadult mandibles equivalent in size to these last two Monte Grande specimens are known from many localities elsewhere in Puerto Rico and in Mona and the Virgin Islands and are always indis- tinguishable from the adults except in size. It is, therefore, clear that small size is in itself no bar to considering the four cjuestionable Monte Grande mandibles as /. portoriccnsis, and we refer the two obviously immature specimens to that species without further discussion. The fourth mandible, lacking teeth, manifests no distinctive characters and may tentatively be referred to the same species. The third questionable specimen differs, however, in molar pattern from all the numerous specimens of Isolohodon with which it has been 1953 A PUERTO RICAN CAVE KAUNULE 5 compared. In any Isolohodon three important enamel folds are evident, a single main fold from the labial side and two counterfolds from the lingual side. In the Monte Grande specimen in question the anterior counterfold of the first molar shows a distinct constriction about midway along its length and, correlated with this, there is a marked broadening of the lingual lobe anterior to the counterfold. No similar constriction is present on the second and third molars; the anteriormost lingual fold may have been broadened on both these teeth, but fractures in the critical regions make it impossible to confirm this. No con- striction comparable to that on the first molar, nor any broadening of the anterior lingual lobe at all similar to that seen on the first molar and probable on the second and third molars of the Monte Grande specimen, have been found in the more than 250 Isolohodon portori- ceTisis mandibles examined. This third, small, Monte Grande mandible may, therefore, represent an exceedingly rare species, possibly reaching a smaller adult size, but closely related to Isolohodon portoricensis . On the other hand, it may represent a rare mutant condition in a subadult individual of /. portoricensis. Of these two alternatives we consider the second the more economical hypothesis, since we have been unable to find any other consistent differences between the third, small, Monte Grande mandible and the mass of compared /. porto- ricensis. We feel that the counterfold constriction might have been caused by a single mutant gene. It is interesting that this apparently aberrant individual probably never reached full adulthood. In the course of this investigation, and while small size as a possible species character was still in question, the Monte Grande mandible and other Isolohodon material from Puerto Rico was carefully com- pared with material identified as Isolohodon levir from Hispaniola. It is worthwhile to consider here the status of the latter species, but before doing so a brief resume of its taxonomic history seems in order. /. levir was first described by Miller (1922) as a distinct genus and species, Ithyodontia levir, from two molars found fossil at St. Michel. He did this believing that these isolated teeth were lower molars. Twelve mandibles from the same locality were identified as Isolohodon portoricensis without comment. In 1929, after studying much more material from the same locality, he realized that the two molars of ''Ithyodontia" were actually upper molars of Isolohodon. In an attempt to save /. levir from complete synonymy, he noted that all St. Michel specimens were smaller than Isolohodon from other Hispaniolan 6 BREVIORA NO. 12 (kitchen midden) localities. The latter, in turn, were indistinguishable from Puerto Rican /. portoricensis. Although he could find no morpho- logical characters to distinguish them, he nevertheless separated the St. Michel specimens as a distinct species on size alone. He pointed out, however, that Hispaniolan /. portoricensis came from kitchen middens, whereas /. Icvir came from owl pellet deposits. As he indi- cated, smaller species and individuals are usually found more fre- quently in owl deposits than in kitchen middens, but he believed that the presence in St. Michel caves of relatively large mandibles of the related genus or subgenus Aphaetreus ruled out the suggestion that in the two types of Hispaniolan Isolobodon he was dealing with large- and small-sized samples from the same population. In later papers (Miller 1929b, 1930) he records /. Icvir from several other Hispaniolan locali- ties (San Gabriel, Monte Cristi, Constanza, Trujin), in Monte Cristi actually in association with /. portoricensis. After extensive comparisons of /. portoricensis from Puerto Rico and surrounding islands (AMNH material) with /. levir from Hispa- niola (MCZ material from Fort Liberte as well as material from the United States National Museum, especially mandibles from St. Michel, Monte Cristi and Anadel, but also skulls from these and other localities), we agree with Miller that no morphological characters are to be found to separate the two species. We fail, however, to see any clearcut size difference, since the smaller species ''levir" appears to us to grade into the larger species portoricensis. "I. levir" mandibles from Hispaniola can be matched by equally small mandibles from Puerto Rico. We believe also that in some cases Miller introduced a false dichotomy into his measurements by comparing the largest speci- mens identified as /. levir with the largest /. portoricensis: it is unlikely that a bimodal distribution would have resulted in the one case of actual association of the two forms if measurements of all measurable specimens had been used. For these reasons we consider Ithyodontia levir Miller a complete synonym of Isolobodon portoricensis J. A. Allen. Isolobodon thus emerges as a monotypic genus or subgenus, undiffer- entiated from Hispaniola through Mona and Puerto Rico to the Virgin Islands. As has already been pointed out (Miller, 1918) this wide uniform distribution may well have been brought about by human transport. If that be true just what was the original range of the genus before the coming of man to the Antilles. This would seem impossible to determine now. 1953 A PUERTO RICAN CAVE FAUNULE 7 Acknoiiiedgmcnts. We are indebted to the Departments of Mammals of the Museum of Comparative Zoology, the American Museum of Natural History, and the United States National Museum for the privilege of examining and comparing specimens. REFERENCES Allen, J. A. 1916. An e.xtinct octodont from the island of Porto Rico, West Indies. Ann. N. Y. Acad. Sci., vol. 27, pp. 17-22. Anthony, H. E. 1926. Mammals of Porto Rico, living and extinct. N. Y. Acad. Sci., Scientific Survey of Porto Rico and the Virgin Islands, vol. 9, pp. 1-238. Miller, G. S. 1918. Mammals and reptiles collected by Theodoor de Booy in the Virgin Islands. Proc. U. S. Nat. Mas., vol. 5-4, pp. 507-508. 1922. Remains of mammals from caves in the Republic of Haiti. Smithsonian Misc. Coll., vol. 74, no. 3, pp. 3-5. 1929a. A second collection of mammals from caves near St. Michel, Haiti. Smithsonian Misc. Coll., vol. 81, no. 9. pp. 14-18. 1929b. Mammals eaten by Indians, Owls, and Spaniards in the coast region of the Dominican Republic. Smithsonian Misc. Coll., vol. 82, no. 5, pp. 6-8. 1930. Three small collections of mammals from Hispaniola. Smithsonian Misc. Coll., vol. 82, no. 15, pp. 4-8. BREVIORA NO. 12 PLATI'] Crown views of mandibles of Isolobodon. A. "Isolohodon kvir". U.S.N.M. No. 255874 from Monte Cristi, Dominican Republic. B. Isolobodon portori- censis. A.M.X.H. "0.1" Utuado, Pvierto Rico. C. Aberrant Monte Grande mandible. M.C.Z. Cueva Monte Grande, Puerto Rico. (These specimen.s, all of about the same size, have been selected to show the extremes of variabilit.>- of molar pattern.) 5 x natural size. E V I O R A Mii]is(eii]im of Cooipsirative Zoology Cambridge, Mass. Ai'kil 23, 1953 Number 13 FOSSILS AND THE DISTRIBUTION OF CHELYID TURTLES 1. ''Hydraspis'^ leithii (Carter) in the Eocene of India i.s a Pelomedusid By Ernkst Williams For SO yetirs a fossil turtle from the Eocene of India has heen referred to the Recent South American chelyid genus Ili/drdspis (not of Bell 1S2S, correctly called Phrj/uopft according to Stejneger 1909. and Lindholm, 1929). In spite of the zoogeographical interest of this record the evidence for this assignment lias not ])rc\i()usly heen r(>- examined critically. In view of the prevalent idea that the family reached Austi-alia from .Vsia, the occurrence of a chelyid in the Eocene of India would he neither surprising nor unwelcome. Reference, however, of an Eocene Indian fossil to a modern South American genus is more suspect and obviously calls for re-investigation. Examination of the original de- scription leads to a verdict unfavorable to both the generic Jiufl the family assignment. The fossil in question was brought to scientific attention just on (m- 100 years ago. In 1852 H. J. ( "arter in a study of the geology of the is- land of Homljay described as Tcsiudo Irithii* the remains of a small turtle from the Intertrappean beds. Carter was not deceived as to the affinities of the form; he was using the generic name Tcsiudo in a Linnaean sense and explicitly stated that he regarded his fossil .is close to "Sfrrnotharrus" (= Prlusius). He published two good plates giving ;i reconstruction of the fossil made from nine partial specimens. These Tt. Ti.-iudu Intlm Guuther I'-gvpt. The :iv:iilable Miui i-nrreet name of the hitter then appears t I.ortel 18S:j 2 BREVIORA NO. 13 plates and his text description are the basis of all subsequent discussion. The type material cannot now be located in India and has never been restudied. On the basis of the 1852 description, Gray in 1871 reassigned the Indian fossil. He remarked: "The description and figure of the cara- pace induce me to believe that the fossil is nearly allied to some of our existing South American species of the restricted genus Hi/drasph-; and the remains of the head, which are unfortunately imperfect, lead to the same conclusion. ..." This determination has been very generally copied, in spite of the zoogeographical anomaly upon which Gray himself commented. I find it necessary to disagree with Gray on the basis of the figured morphology of both shell and skull. Plates 1 and 2 are reproductions of Carter's plates X and XI with a few inessential modifications for clarity. According to Carter's plate X (though the area is given in dotted lines only) and according to his express statement in the text there is no nuchal scute in the Indian fossil. Lydekker (1889b, p. 170) stated: "The omission of a nuchal shield in the restoration of the anterior l)order of the carapace is probably incorrect." Perhaps, however, Lydekker made this statement only on the ground that if the form were Ili/draspi^' it should possess a nuchal scute. Absence of a nuchal scute would rule out all Recent genera of Chelyidae except Chrlodina and Emydura (in both of which the scute may be present or absent) and Eheya (in which it is regularly absent). The latter genera are all natives of the Australian region. In Carter's fossil the first vertebral is much smaller than the second vertebral. This precise condition is not met with in the living ( helyi- dae. In most Recent South American forn s including most cf tie species of Hiidraspis! ( = Phriinops) the first vertebral is on the contrary much larger than the second. There is, it is true, an apprcach to the condition of the fossil in the Australian genera Enu;dnra, Ps( iidcnndvia and Khrya anfl in the South American Ilydraspis fiibrw.sa (specirrers in the British ^Museum) and perhaps in some specimens of the South .Vmerican genus Ilydromdusn, but in the latter only if the anterior median scute is interpreted as a nuchal withdrawn from the margin rather than as a transversely divided first vertebral. In none of these, however, is the first vertebral as much smaller than the second as it is in the fossil. In the Indian fossil the gulars are small, and the very Ijrcad inter- 1953 FOSSILS AND DISTRIBUTION OF CHELYID TURTLES 3 gular extends from the anterior margin to tlie humero-pectora! sulcus, separating the humerals in the midhne. This pattern of scutes on the anterior phistral lohe is different from any which is known in the modern ( 'helyidae. In all the genera except Chdudina, Psrudoni/dura, and rarely in Kini/dura (Siebenrock, 1907) the humerals meet in the midline for a significant distance behind the intergular. The intergular also is rarely as broad as in the fossil. In Cliflodina the intergular is very large and separates the humerals, indeed extending deep into the area of the pectorals, and, except in C. iutrrgularis P'ry, it does not reach the anterior phistral margin, the gulars meeting in front of it. In Pscudemydura (Siebenrock, 1907) the intergular is like that of the Indian fossil in its breadth but as in Chdodinn dips deeply between the pectorals. The gular-intergular pattern in the exceptional Kmi/dura siiln/lohosa in which Siebenrock found the intergular separating the humerals is also quite unlike that of the Indian fossil, the intergular being narrow and of quite different shape. The feeble xiphiplastral notch is another feature in which the fossil differs from Ihjdrnspis and other chelyids except the forms of the Australian region and Batrachrmys. The shell, therefore, is not a gooil match for that of any known genus of chelyid. It is perhaps most like those of the Australian genera but differs from all of these in significant details, for example, in the presence of neurals, which are lacking in all the Australian genera. The skull and mandible in their turn provide conclusive evidence against chelyid affinities. The mandible, although incomplete, is stouter and broader than in any known chelyid. The symphysis must ha\'e been long, in strong contrast to the condition in chelyids. The skull, shown l)y Carter only in dorsal view, is radically different from that of any chelyid. The skull roof has undergone emargination from behind as in the Pelomedusidae or most Cryptodira. As a result, the parieto-sciuamosal arch is absent, but a jugal-quadratojugid bar is present. In the ("helyidae and in no other turtles the skull is emargi- nated from the ventral margin only, and a parieto-sciuamosal con- nection is (except in Chclodina) always preserved. In chelyids the quatlratojugal and the bar of which it was a part are always absent. These are as crucial and clearcut differences as it is possible to obtain between skulls of turtles. Carter's fossil cannot be a chelyid. It is most probable that it is a pelomedusid. Reference to this family wou'd, it will be recalled, be a return to the opinion of the original desc.-iber, wh > thouT;ht the fossil was closest to the African 4 BREVIORA NO. 13 peloniedusid genus Prlusios. The characters of the shell fit such a reference extremely well. The nuchal scute is almost always absent in pelomedusids. The first vertebral is almost always smaller than the second (exception in Palaraspis Gray, and sometimes in Pclusios). An intergular separating the humerals is found in Stercogciiys podocne- moidcs and in FJochrh/s prrftcfa as well as in occasional individuals of Podocncmis cxpansa. The xiphiplastral notch is very variable in pelomedusids. Vertebral shields 2 to 4 of the Indian fossil have strikingly convex anterolateral borders, conspicuously concave postero- lateral l)orders, as in some Recent Podocnemis and a nimiber of fossil pelomedusids. A final feature which, as described and figured by Carter, is anoma- lous, may be clinching proof of the pelomedusid affinity of this form. Gray mentioned that Carter's form was "pecidiar also for the under- side of the marginal opposite the (pectoral-humeral) suture being rather l)roader than the rest and angular on the inner edge, which I have not seen in any of the Recent species." If we have to do here with sulci between scutes, the situation is indeed peculiar and unique, but it is noteworthy that the lateral marginal scute boundaries are represented by dotted lines in Carter's original reconstruction (solid lines in plate 2 here) except for the anomalous "marginal" in question. It is noteworthy also that the ventral view of the lateral marginals in Carter's plates does not match the dorsal view of the same marginals. It is evident that Carter was not certain of the exact scute boundaries here, and it is possible that he has figured as the anomalous "mar- ginal" the sutures between bones rather than the sulci between scutes. The lines drawn solidly by Carter in this region are in nearly the right position and have the right aspect to represent in their lateral portions the sutures boimding small mesoplastra and medially the hyo-hypoplastral suture. This interpretation is the more probable because Carter's figure is a reconstruction from nine specimens, one of which may have shown the sutures in this critical region and not the sulci. If small laterally placed mesoplastra were present, this fact woukl definitely place the Indian form in the Pelomedusinae of Zan- gerl (1948) with which, on the basis of other resemblances in its shell, it is most plausibly linked. Carter's turtle is, therefore, most probably a new genus of pelome- dusine. I cannot distinguish it from all previously proposed genera of pelomedusines because not all of these are themselves well-delimited. Thus I cannot distinguish it from Hosasia (Carrington da Costa 1940) 19.53 P^OSSILS AND DISTKIBITION OF CHELYID TITRTLES 5 because at present that genus does not seem to be definable. (On its known characters Rosasia might be a synonym of any one of several genera. The carapacial shield, which alone is known, is not sufficiently diagnostic.) I cannot distinguish the Indian fossil from DacochcJys (Lydekker, 1889a) because there are no comparable parts, that genus having been founded on a mandibular symphysis, a part that is missing in (barter's fossil. There is also no evidence that Dacochclys is a pelome- dusid.* I distinguish Carter's form, with some hesitation, from EJochelys Nopcsa 19.31 because the gular-intergular pattern is not quite that of E. prrfrcta, the type of the genus, and I am not persuaded that the other species referred by Nopcsa to that genus {K. major) belongs there. The critical feature of Elochcli/s also, the absence of a suprapygal, is not determinable in Carter's form. A possibly trivial feature, the feeble xiphiplastral notch, distinguishes the Indian fossil from Sfcrcogrni'.'i podocvrnioidc.s (Reinach, 1903), but with Schmidt (1940) I do not believe that podocnrmoidts belongs to the genus St(rco(ir)ii/.s. Podocncmoidcs and Icifhii may indeed belong to the same genus, but that genus is then imnamed. From the better defined genera of the Pelomedusidae the Indian form is distinguished by the following combination of characters: Carteremys, new genu^ Type. Trstudo Irlthii Carter 18.52. Diagnosis. Skull roof much emarginate from l)ehind; opisthotics prolonged backwards in sharp crests; mandilile with a moderately long symphysis; nuchal absent; first vertebral not divided transversely; intergular large and very broad, separating humerals; xiphiplastral notch narrow and shallow; pubic and ischial scars distinctive in shape and position. * .\n iiieenious device by which Lydekker avoided the possibility of a change in the trivial name of Dacochelys has been the source of confusion in regard to this point. The type of the genus is Dacochelys delabechei Lydekker 1889, but Lydekker considered his form probably synonymous (largely on size alone) with Emys conybearii Owen and therefore (according to Lydekker and Boulencer 1K87) with Emys ddahichei Bell. No name change is necessary if this synonymy is correct and, since the shell of Emys conybearii shows small lateral mesoplastra (Lydekker and Boulengtr 1887), Dacochelys delabechei is then a pelomedusine. But the only valid physical type of Dacochelys (the type of the species upon which the genus is based) is the very peculiar mandibular syn physis, which I regard as quite impossible to assign to family. Lydekker's device has most unfortunately brought the name Dacochelys into the literature (for e.xample in Zangerl, 1948) as a pelomedusine — which it may be but which it certainly cannot at present be proved to be. It should be mentioned that if Dacochelys delabechei were in fact a synonym of E. delabechei Bell as Lydekker assumed, and if E. conybearii Owen were a synonym of E. delabechei Bell as Lydekker also assumed, Dacochelys Lydekker 1889 would be a straight synonym of Palaeaspis Gray 1870, type Emys conyhenrii Owen. 6 BREVIORA NO. 13 Horizon. Intertrappean Eocene of Bombay. Comment may now he made on other records and alleged records of this form. In 1890 Lydekker reported an entoplastron showing half of an intergular scute from the Intertrappean beds in the Xagpur district as a second occurrence of ''Hi/draspis Irlthii." The new fragment was much larger than typical Cartrrrmi/s Irithil and differed also in the much narrower intergular. Lydekker passed over the differences casually with a mention of variability in Hjidraspin ( = Phrynops) hilarii. It is unlikely that he had in hand any member of the genus Hydraspis or Cartcrcmys Irithil. It is probable that he had some other pelomedusid, and he had himself previously (1887) described a Podoc- nrmis Indico in the Eocene of India. Sukheswala (1947. an abstract only) has reported a find of a shell of Tc.siudn (— Cnrtrrrmy.s) IrithH in the Intertrappean of Worli Fill. Bombay. Here, as in the case of the specimens described by Carter, remains of frogs (Indohafrnchus piisillvs) were associated. In this instance there is no doubt of the identification. Dr. Sukheswala has kindly sent me a photograph of this specimen (plate 3). The outline of the shell (somewhat dift'erent from that figured by Carter) is clearly shown, as is also the characteristic sculpture of the surface (mentioned by Carter), which while somewhat like that of some chelyids also resembles that of, for example, the American pelomedusine genus Taphrosphys. More important is the evident presence of several neurals, the first vertebral clearly much smaller than the second, and (less certainly) the absence of a nuchal shield. The plastron and the skull are unfortunately missing. As with Carter's specimens the new shell is small, eight inches long by six wide. The specimen is noM- in the possession of the Geological Survey of India. Two other specimens have since been foimd by Dr. Sukheswala and have been sent by him to the Geological Survey of India. These are recorded in the general report of the Survey for 1948 (West, 1950). These specimens, while recognizable, afford no additional information. Also recorded by the SurNcy in the same report was a possible young shell of Cartcrcmys Icithll from the carbonaceous shales of an Inter- trappean band near Raibasa in the Chhindwara district, Central Provinces, India. The Survey has generously sent me a photograph of this specimen. The identification must be pronounced doubtful as, indeed, the ^Survey has regarded it. While this new locality may be 1953 FOSSILS AND DISTRIBUTION OF CHELYID TURTLES 7 thought of hopefully, it would appear that Caiitrciin/.s Icit/iii is at present known with certainty only from the Intertrappean of the island of Bombay. Acknowlcdgxicnts. Dr. A. S. Romer, Dr. P. .1. Darlington, Dr. P. E. Vanzolini, Dr. Karl ¥. Koopman, Dr. W. E. Swinton and Mr. L. I. Price have read the manuscript and given critical advice. The allo- cation of the Carter fossil has been discu.ssed with Dr. Rainer Zangerl. Mr. Arthur Loveridge. Mr. (\ M. Bogert and Dr. H. \V. Parker have permitted the examination of Recent skeletal material in their de- partments. I wish to thank also the authorities of the Bombay Natural History Society for endeavoring to locate Carter's types, the Geological Survey of India for furnishing me all the information in their possession in regard to the new specimens of Carternnys and for photographs of certain of them, and especially Dr. R. N. Sukheswala of St. Xavier's College, Bombay, for his prompt and courteous replies to my insistent questions and for the gift of the photograph reproduced as plate 3. REFERENCES Carrington da Costa, .). 1940. Um novo queionio fossil. Portugal Com. Serv. Geol., vol. 21, pp. 107-123. Carter, H. .J. 1852. Geology of the island of Bombay. Jour. Bombay Branch Roy. Asiatic Soc, vol. 21, pp. 161-215. Gray, J. E. 1S70. Supplement to the Catalogue of Shield Reptiles in the Collection of the British Museum. Part I. Testudinata. London. 120 pp. 1871. Notice of a fossil hydraspide {Testudo leithii Carter) from Bombay. Ann. Mag. Nat. Hist., ser. 4, vol. 8, pp. 339-340. LiNDHOLM, W. A. 1929. Revidiertes Verzeichnis der Gattungen der rezenten Schildkroten nebst Notizen zur Nomenklatur der Arten. Zool. Anz., vol. 81, pp. 275-295. LORTET, L. 1883. Poissons et reptiles du lac de Tiberiade et de quelques autres parties de la Syrie. Arch. Mus. Hist. Nat. Lyon, vol. 3, pp. 99-194. Lydekker, R. 1887. Eocene chelonians from the Salt Range. Paleontographica Indica, ser. 10, vol. 4, pp. 59-65. 8 BREVIORA NO. 13 1889a. On the remains of Eocene and Me.sozoic Chelonia and a tooth of ?(Ornithopsis). Quart. Jour. Geol. Soo. London, vol. 4.5, pp. 227-246. 1889b. Catalogue of the fos.sil Reptilia and Amphibia in the British Museum (Natural History). Part III. London. 239 pp. 1890. Notes on certain vertebrate remains from the Nagpur district. IL Part of a chelonian plastron from Phisdura. Rec. Geol. Surv. India, vol. 23, pp. 22-23. Lydekker, R. and G. A. Boulenger 1887. Notes on Chelonia from the Purbeck, Wealden and London Clay. Geol. Mag., ser. 3, vol. 4, pp. 270-27.5. NopcsA, F. 1931. Sur des nouveau.x restes de tortues du Danien du Midi de la France. Bull. Soc. Geol. France, ser. 5. vol. 1, pp. 223-234. 1934. The influence of geological and climatological factors on the distri- bution of non-marine fossil reptiles and Stegocephalia. Quart. Jour. Geol. Soc. London, vol. 90, pp. 76-140. Reinach, a. von 1903. Schildkrotenreste aus dem aegyptischen Tertiar. Abhandl. Senckenberg. naturf. Ges., vol. 29, pp. 1-64. Schmidt, K. P. 1940. A new turtle of the genus Podocnemis from the Cretaceous of Arkansas. Geol. Ser. Field Mus. Nat. Hist., vol. 8, pp. 1-12. SlEBENROCK, F. 1907. Beschreibung und Abbildung von Pseiidemydurauinbrina Sicbenr. und fiber ihre systematische Stellung in der Familie Chelydidae. Sitz-Ber. Akad. Wiss. Wien. Math-naturw. Kl. vol. 116, pp. 120.5-1211. Stejneger, L. 1909. Generic names of some chelyid turtles. Proc. Biol. Soc. Wash- ington, vol. 22, pj). 125-127. SUKHESWALA, R. N. 1947. A fossil tortoise (Testudo leithii) from the Intertrapp?an of the Worli Hill, Bombay. (Abstract.) Proc. Indian Sci. Congress, 33rd Session, vol. 3, p. 97. WE.ST, W. D. 1950. General report of the Geological Survey of India for the year 1948. Rec. Geol. Surv. India, vol. 82, pp. 1-253. Z ANGER L, R. 1948. The vertebrate fauna of the Selma Formation of Alabama. Part II. The pleurodiran turtles. Fieldiana, Geology Memoirs, vol. 3, pp. 23-56. Plate 1. Carteremys leithii, dorsal view of shell and skull, slightly modified for clarity from the original reconstruction (Plate X of Carter, 1852). Skull restored from the more perfect half. Plate 2. Carteremys leithii, ventral view of shell and mandible, slightly modified for (darity from the original reconstruction (Plate XI of Carter, 1852). Plate 3. Cartcrenujs leithii, photogra])h of dorsal surface of shell. (Courtesy Dr. R. X. Sukheswala.) E V I O R A Miuseeim of Co:mpsirsitive Zoology Cambridge, Mass. May _'s. 1953 Xtmher 14 RECORD (3F A HERAIAPHRODITIC^ HORSE^^HOE (^RAR, LIMriJ'S POLYPHEMUS L. By John P. Baptist Fishery Research Biologist U. S. Fish & Wildlife Service Xewburyp( )rt . Mass . A hermaphroditic horseshoe crab, Liinulus puh/phciiiu.s- L., was picked up by a small scallop dredge from the channel of Plum Island Sound, Massachusetts on May 16, 1952. As far as can be determined, this is the first such Li mill us to be recorded. The general appearance was that of a mature male. Its width (125 mm.) was about average for males in this area. (The widths of 600 mature males measured this summer ranged from 8S mm. to 140 mm., with an average of 117 mm., while 489 mature females ranged from 130 mm. to 199 mm. in width, with an average of 155 mm.) It had mating claspers, and the anterior edge of the prosoma was curved upward in the middle, which also is a male characteristic. The left genital aperture was typically male, round and located on a definite papilla. The right genital aperture was typically female, a horizontal slit and not on a papilla (Lochhead, 1950). The genital operculum, showing both male and female genital apertures, may be seen in the upper illustration of Plate 1 . On the dorsal surface of the opisthosoma near the telson were darkened areas normally present on mated females. These may be barely discernible in Plate 2. They are pro- du "ed by the abrasive action of the anterior edge of the male as he clings tenaciously to the female during the mating period. Females which have mated for prolonged periods exhibit deep scars on the last pair of immobile spines, caused by the powerful claspers of the male. 2 BREVIORA NO. 14 Such scars were not evident on the hermaphrodite, suggesting that it had been mated as a female for a comparatively short time. Gross dissection of the prosoma and histological sections revealed both male and female gonads. The right and anterior portions were full of eggs in various stages of development (Plate 2), comparable to to those of mature females. Histological sections of the tissue on the left side demonstrated sperm sacs full of tail-less sperm, diverticula of the hepatopancreas and connective tissue (Plate 1, lower illustration). The same structures Avere visible in sections of gonads from norma! males. According to Benham (1885) the tails are apparently produced as the sperm approach the aperture. The presence of both eggs and sperm confirms the external evidence that this specimen is a true hermaphrodite. REFERENCES Benham, W. B. S. 1885. Ou the testis of Limulus. Trans. Linn. Soc. Lond. (Zoo).), (2) 2: 363-366. LOCHHEAD, J. H. 1950. Xiphosura polyphemus. In Selected Invertebrate Tj^pes, edited by F. A. Brown, Jr. pp. 360-381. John Wiley & Sons, Inc., N. Y. 1953 A IIERMAPHKOniTIC HORSESHOE CRAB PLATE 1 Upper — Ventral view of popterijr surface of genital operculum (enlaiKcd), showing right genital aperture (9 ) and left genital aperture (o^). Lower — Photomicrograph of section through testis (X 120). S, spei'm sacs containing tail-less sperm. H. hepatopancreas. C, connective tissue. (Photographs by Alden P. Stickney) NO. 14 platp: 2 Dorsal view of Liiiiiiliis pali/phouxs with carapace of the prosoma removed, showing mass of eggs on the right side, but only a few scattered eggs on the left side. (Telson has been cut off). (Photograph by Alden P. Stickney.) E V I O R A MuseiuitM of Comparaitive Zoology Cambridge, Mass. May 15, 1953 Number 15 FURTHER NOTES ON THE EARTHWORMS OF THE ARNOLD ARBORETUM, BOSTON By G. E. Gates An opportunity to search once more for mature specimens of the Pheretima sp., represented in collections of the previous spring by a single, specifically unidentifiable juvenile, was unexpectedly provided on September 25, 1952. After securing the desired material, as many of the previous sites as time permitted were re-examined, to permit comparison of fall with spring populations. In the natural woods, during the summer, the leaf pile of previous collections had been grown over and had become rather dry. Some fifty feet away, and at about the same distance from the site where the juvenile of Pheretima sp. had been found, another large leaf pile, possibly older and with more gravel, had remained damp and without any plant overgrowth. In the peat bog, where vegetation had been mowed and removed, water was no longer running in the brook though still standing in small, scattered puddles, but previously disturbed soil still seemed moist. In the ponds, water level had fallen several inches exposing a marginal strip of bottom muck that was still saturated. The Newton garden had been planted and could not conveniently be re-examined. The species obtained in the September collections are listed in the following table which also shows sexual stages and number of specimens secured. BREVIORA NO. 15 Earthworms Obtained in the Arboretum in September, 1952 Natural Woods Species Original Second Leaf Pile Peat Bog Leaf Pile Near Side Far Side Allolobophora t * arnoldi 0-2-2 0-1-0 0-0-1 0-0-1 caliginosa 0-0-2 0-0-1 limicola 0-2-0 sp. (1) 4-0-0-0 Dendrobaena suhruhicunda 3-5-32 1-0-1 Eisenia foetida 12-5-6 7-1-7 2-0-1 Lumbricus castaneus 0-0-2 rubellus 3-1-7 2-2-3-16 6-11-17 terrestris 4-1-0-4 1-0-1 Octolasium lacteum 0-0-1 Pheretima agrestis 0-0-35 0-0-21 Figures indicate juvenile-aclitellate-clitellate or small juvenile-late juvenile-aclitellate- clitellate. t Postsexual aclitellate. * Specimen in poor condition. (1) Probably of arnoldi and /or caliginosa. A. arnoldi. Two specimens from original leaf pile listed as clitellate had spermatophores externally but the clitellum had almost completely re- gressed, the site being indicated only by a brownish discoloration. In an amputee from the bog, after enteroparietal healing at 106/107, cvi had been much elongated and at the same time narrowed so as to taper terminally and its setae had been lost though follicle apertures were still visible. One worm from the leaf pile showed the same changes in its last segment. L. terrestris. A small juvenile from the leaf pile, 27 x 2.5 mm. has 149 segments. 0. lacteum. Segments in front of the anus were packed with coelomic corpuscles and parasitic bodies many of which were discharged through dorsal pores during preservation. 1953 EARTHWORMS OF THE ARNOLD ARBORETUM 3 Dendrobaena subrubicunda (Eisen) 1874 Clitelluin, on xxvi-xxxi but occasionally with slight extension onto xxxii (32 + 26), xxvi-xxxii (1). Epidermis of xxv (several + 8) and xxxii (several) slightly modified. Tubercula pubertates, on xxviii-xxx (35 + 27), occasionally reaching onto xxxi. Each tuberculum nearly always is bounded laterally by a deep groove. In two clitellate speci- mens in which tubercula were quite indistinguishable the groove was obvious, on both sides of the body. The tuberculum is longitudinally elliptical in outline. A central depressed portion of similar outline may be more or less clearly demarcated from a band-like marginal area by a slight groove, or, instead, there may be recognizable only a single longitudinal groove at the center. The lateral groove, in some specimens, apparently passes, though less clearly, around ends of tuberculum to become completely circumferential. Genital tume- scences, including ab, on viii (5), ix (14 -\- 11), x (5), xvi (33 + 27), xvii (2 -|- 1), xviii (1), xx (3), xxi (1), xxii (-|-1), xxiii (-1-2), xxiv (12 -f- 12), xxv (1 + 3), xxvi (-}-l), xxvii (26 -f- 25), xxviii (2), xxxi (19 -f 23), xxxii (1), xxxiii (-f 1), xxxiv (-}-l),xxxv (+1). Tumescences on xvi may extend well into be and aa or even be united mesially. Those on xxxii to xxxv are especially well developed. Homoeosis. Male pores, on xvi. Female pores, on xiv. Clitellum, on xxvi-xxxii and on xxxiii-xxxiv of left side only. Tubercula pu- bertates, on xxviii-xxx (right side), xxix-xxxiii (left side). Genital tumescences, on left sides of xxiv and xxxiv. No metameric abnor- mality recognized. Remarks. In live worms a short terminal region had a brilliant yellow appearance that disappeared during preservation and which, presumably, was due to accumulations of coelomocytes in the coelomic cavities of a region with rather transparent parietes. The anal region was not marked off by an intersegmental furrow from the last seti- gerous segment in 19 specimens. Spermatophores were present ex- ternally on eight specimens. Data from a series of 27 specimens (collected at Bangor, Maine) that has been available for comparison with arboretum worms, are included in parentheses above, preceded by the -|- sign. Very little of such information has been reported. 4 BREVIOEA NO. 15 LuMBRicus RUBELLus Hoffmeister 1845 Number of segments: 67, 73, 75, 79, 85, 88, 91, 96, 100, 101, 109 (2), 110, 111, 112 (4), 113 (3), 114 (3), 115 (2), 116 (3), 117 (4), 118 (3), 119 (4). Clitellum, on xxvii-xxxii (40). Tubercula pubertates, on xxviii-xxxi (5), but extending onto xxvii (40). Genital tumescences, including ab, on vi? (1), vii? (1), viii (2), x (6), xi (42), xii (50), xxvi (56), xxvii (3). Abnormality. No. 1. One segment near the hind end spUt, on opposite side, into three segments each with setae. No. 2. Two metameric abnormalities in region shortly in front of anus. Homoeosis. No. 1. Male pores, as well as female, on xiv. No. 2. Male pores on xvi, female pores on xiv. Clitellum on xxvii-xxxii, tubercula on xxviii-xxxi but extending onto xxvii. Genital tume- scences, on xxvi, and possibly also (?) on vii-xiii. No. 3. Female pores on right side of xiii and left side of xiv. Male pores on right side of xiv and left side of xv. Aclitellate, tubercula on right side of xxvi-xxx and left side of xxvii-xxxi. Genital tumescences, on right sides of ix and XXV, left sides of x and xxvi. Remarks. Worms with 79, 88, and 100 segments have large nephro- pores, pit-like depressions at sites of apertures of setal follicles but no setae on the last segment, and probably were amputees. Two speci- mens, of 109 and 119 segments, had setae in the last segment and presumably were also amputees. If any of the other worms having 67 to 101 segments are amputees, no indications thereof were recognized and the terminal region must have been subsequently reorganized so as to look like that of a normal worm. In the clitellar region, intersegmental furrows are always more or less clearly indicated. The worms can be arranged in a finely graded series of stages showing gradual deepening of the furrows, decrease in thickness of the epidermis, reappearance of dorsal pores, and gradual reddening of the parietes. In the final stage, boundaries of the clitellum were not certainly determinable but in much of the clitellar region the red color has a sort of slight whitening that is lacking on other segments. No spermatophores were found. No brownish discoloration such as is present in the clitellar region of postsexual specimens of some species, was recognized in any of the worms. Smith (1917) and Olson (repeatedly) have given the number of 1953 EARTHWORMS OF THE ARNOLD ARBORETUM 5 segments in this species as 95 to 150 or 90 to 145, but with no data as to individual counts in American specimens. Segment numbers in all of the arboretum collections are in better agreement with those of English worms, 95 to 120, as reported by Cernosvitov and Evans (1947). Pheretima agrestis Goto and Hatai 1899 External characteristics. Length, 70-150 mm. Diameter, 5-8 mm. Number of segments: 63 (3***), 04 (*), 66 (2**), 67, 69 (*), 71, 76, 77, 78 (2**), 82, 84, 85 (*), 86, 90, 92 (2**), 95, 97 (2), 98 (2), 99 (3), 100 (7), 101 (6), 102 (7), 103 (5), 104 (1). First dorsal pore, on 12/13 (56); a dark and somewhat pore-like marking shortly behind 11/12 on several specimens, definitely on 11/12 though not a functional pore, in one worm. Setae: 52, 55, 56, 62, 62, 64, 64, 65, 69/xii; vii/2, 4, 4, 3, 4, 2, 1, 1, 1, 2, 2, 4, 2, 4, 4, 1, 2, 1, 3, 2, 2; viii/5, 3, 5, 1, 2, 2, 3, 0, 2, 2, 2, 5, 3, 4, 3, 0, 0, 0, 1, 1. (Clitellum, on xiv-xvi; annular.) Sexthecal, spermathecal pores minute and superficial, on 5/6-7/8, about \C apart. Each pore may be at center of a definitely demarcated circular porophore or the latter may have the appearance of two semi- circles, one in front of and the other behind the pore. (Female pore, single, median, on xiv.) Male porophores, male pores, as well as any rudiments thereof, completely lacking. Definite genital markings also are lacking. The setal circles of vii-viii are uninterrupted (several specimens), or have a small gap on one side (several) or one large median gap which may extend nearly to spermathecal pore lines (25), or a pair of gaps leaving a few setae midventrally (as indicated above) . In the latter case, in ten specimens, in region of each setal gap there is recognizable, in good optical con- ditions, an area where the epidermis is thin or especially cross-hatched with fine furrows. These areas, which have no definite boundary though appearing to be of transversely elliptical outline, are sym- metrically placed across the setal equator. No special pigment, brown or otherwise, was recognized in any of those areas. Internal anatomy. A low collar of iridescent tissue is present on the oesophagus just behind the gizzard. The intestine begins in xv (6) and has a double row of sacculations on each side from about xvi to about XXV, the ventral row on each side usually more definite than the dorsal and with one or more of the sacculations so constricted as to have an appearance of dorsally directed pockets. The caeca are mani- 6 BREVIORA NO. 15 cate, with seven or eight secondary caeca, the dorsalmost the longest. The typhlosole, which begins in the region of the caecal segment, is low, simply lamelliform and from about xxxix gradually decreases in height, becoming more irregular in shape and finally unrecognizable behind Ixvi (worm of 92 segments), Ixxiv (100, 104 segments), Ixxvii (102 segments). A very small but definite and double ridge runs posteriorly on the floor of the gut from the caecal metamere for fifteen or twenty segments. Last hearts in xiii (5), heart of ix on left side (3) or right side (2), hearts of x lacking (5). Subneural trunk much smaller anterior to xiii but continued into ii. The dorsal trunk of one specimen passes superficially through tissues of the pharyngeal bulb and on emerging anteriorly bifurcates under the brain, the branches passing around the gut median to the nervous commissures and reuniting midventrally to become the ventral trunk. Testis sacs unpaired, apparently above the nerve cord and with ventral blood vessel in, on or just above the roof. Testes present, male funnels large and nearly filling the sacs which have very little coagulum. Seminal vesicles laterally flattened, filling coelomic cavities of xi and xii, reaching into contact with dorsal vessel, in xii with a primary ampulla marked off distinctly from the lamina. Vasa de- ferentia, in region of xvi to xxiv, slightly swollen for a short distance and ending blindly without passing into parietes. No trace of male terminalia. Ovaries appear to be unusually large. Spermathecal ducts shorter than contracted ampullae, with thick wall (having a muscular sheen), abruptly narrowed at parietes, lumen in coelomic portion large, irregularly slit-like in cross section as a result of presence of high longitudinal ridges. Diverticulum longer than combined lengths of duct and ampulla, into median face of duct close to parietes, with slender stalk (having muscular sheen) longer than the duct and a longer, more or less sausage-shaped seminal chamber. Distribution. Japan : Takahashi, Tokorosawa, Oarai, Sapporo, Hokkaido, Sendai, Oshima Island, Matsuyama, Tomitaka, Kago- shima. Said to be widely distributed throughout the Aomori and lawati prefectures. P. agrestis has been reported once before outside of Japan and then from Baltimore but several attempts to secure specimens for confirmation of the identification have been fruitless. Abnormality. Specimen of 82 segments with one metameric abnor- 1953 EARTHWORMS OF THE ARNOLD ARBORETUM 7 mality in region of Hi. First dorsal pore on 12/13 but a definitely pore-like though obviously non-functional marking exactly on 11/12. Left spermathecal pore of 7/8 and clitellum on right side of xiv lacking. Female pores: on left side of xiii, in setal circle, and on right side of xv. Heart present in xiv on right side. A testis on posterior face of left rudiment of 8/9, a male funnel on anterior face of left rudiment of 9/10 (no testis sac but sacs and vesicles of x-xii as usual. Left ovary lacking but a rudimentary female funnel present in left side of xii; right ovary and female funnel in xiv. Vasa deferentia end blindly in xxi v. Autotomy and Regeneraiion. On being touched three worms in- stantly autotomized a posterior portion (of 31, 34, 35 segments). In one case autotomy was either through Ixiv or at 64/65. In a specimen of 85 segments, Ixii-lxiii were deeply constricted all around the body but apparently without recognizable break in the musculature. Pos- terior amputees are indicated above by asterisks after setal numbers (one for each amputee) . Specimens with 63, 69, 77, 78, 86, 90 and 92 segments each have an anal region that is not marked ofl' by an inter- segmental furrow from a terminal segment with a complete circle of setae. The last metamere in a 78-segment worm (a) is unusually long, without setae, but with minute pits still representing sites of apertures of original follicles. A similar long terminal metamere but without setal pits characterizes a worm (b) of 76 segments but in this case setae are recognizable, in the penultimate metamere, only within the parietes of the ventrum. Tail regenerates, in spite of the high incidence of posterior amputation, were not found. In the (a) worm, loss of setae with resorption of setal follicles, and elongation of the last segment, are regarded as external indications of a process of reorganization, as one result of which a new growth region, in addition to an anus, is reestablished in place of that which had been lost by amputation. In the (b) worm the penultimate segment is regarded as one that had already been marked off (by an intersegmental furrow from a reestablished growth region), but in which equatorial setae, one of the stigmata of a metamere, had not yet been sufficiently developed to be externally visible except in the ventrum. Differentiation of the new metamere had not yet been fully completed. Further, the length of the terminal (growth) region is such as to indicate possibility of production of yet another segment. If these interpretations are correct, P. agrestis presumably does not, at least ordinarily replace lost posterior portions at levels behind 62/63 8 BREVIORA NO. 15 by the supposedly usual process of tail regeneration. Instead, a growth zone is reconstituted that may produce new segments, but only slowly and one at a time. In new segments so produced, the usual differential stigmata of regeneration, so easily recognizable in long tail regenerates that have been produced by a much more rapid process, may well be lacking. The penultimate segment (counting anal region as a segment), in several specimens that presumably had not undergone posterior ampu- tation, had no externally visible setae and no vestiges of former setal pits were recognizable. Both ante- and pen-ultimate segments in another specimen (of 100 segments) were similarly incompletely differ- entiated. In the rather small anal region of two specimens (of 100 segments each), a short rudiment of an intersegmental furrow is present, in the ventrum (1) or dorsum and then with a rudiment of a dorsal pore. These conditions are regarded as indicating that new segments may be produced, even in unamputated adults, by a slow, one-at-a-time method. Remarks. The clitellum seems to be fully developed in every speci- men. Nevertheless no spermatozoal iridescence was noticed on any male funnel of the six dissected specimens. Nor was spermatozoal iridescence recognizable in any of the thirty-five spermathecae. If sperm are produced they cannot, in absence of male terminalia, be transferred to another individual in copulation. Every specimen in a fair-sized sample being anarsenosomphic, reproduction must take place by self fertilization or parthenogenesis, with the latter more likely because of apparent absence of mature sperm in the probable breeding season. In either case, introduction of a single worm, even a single cocoon, theoretically could have enabled establishment of P. agrestis in the arboretum. Presence of a full set of seminal vesicles, and of a complete battery of spermathecae, all now of no apparent use in reproduction, on a principle of economy, indicates that the anarsenosomphic development in this species has been recent. DISCUSSION During the summer the population of the pond periphery quite obviously had increased considerably. Exposed but still saturated margins were almost covered with castings of Sparganophilus eiseni, 1953 EARTHWORMS OF THE ARNOLD ARBORETUM 9 the only species found (the rubellus sites, just at the high water level not re-examined) . In the original leaf pile, E. rosea apparently had disappeared, and two species, E. foetida and D. subntbicunda, absent in the spring, had appeared. One of them, D. subruhicunda, had become dominant nu- merically. L. terrestris, however, still remained dominant by sheer volume even though not as common as in the spring. From the earth of the original peat bog site all worms had been removed in the spring. Several weeks later more than fifty specimens had moved into the disturbed soil from below or round about. Although that same soil still appeared moist in September only seven worms were found in it. Three species, E. rosea, E. tetraedra and L. rubellus were no longer represented, but two, 0. lacteum and L. terrestris, that had been lacking at time of second collecting were again present. In the newly examined leaf pile, Pheretima agrestis, an exotic from Japan (and possibly brought directly from there) was clearly dominant by weight though equalled in numbers by L. rvbelhis. No cocoons or copulating individuals of any species were found in fall collecting. A clitellum sufficiently mature for cocoon deposition apparently characterized specimens of L. terrestris, 0. lacteum, D. subrubieunda, P. agrestis and E. foetida but appearance of the clitellar region indicated that the reproductive period had ended in A. arnoldi and L. rubellus. Breeding period in P. agrestis is probably in the fall. ACKNOWLEDGMENTS The collecting and identifying of the material considered in this contribution constituted spare time recreation during part of a fellow- ship period. The author's thanks are extended to the John Simon Guggenheim Foundation for making this contribution possible and to the Director of the Arnold Arboretum for permission to collect therein. REFERENCES Cernosvitov, L. and A. C. Evans 1947. Lumbricidae. No. 6 in Sj'nopses of the British fauna. London, Linnean Society. Gates, G. E. 1953. On the earthworms of the Arnold Arboretum, Boston. Bull. Mus. Comp. Zool. Harvard, 107, No. 10: 497-53-1. Smith, F. 1917. North American eartli worms of the family Lumbricidae in the collections of the United States National Museum. Proc. U. S. Nat. Mus. 52: 157-182. E V I Miiseuim of Coimpsirsitive Zoology Cambridge, Mass. May 28, 1953 Number 16 NOTES ON THE RACES OF Micrurus frontalis (Dumeril, Dumeril and Bibron) By Benjamin Shreve As Micrurus frontalis and its associated forms appeared in need of revision, 46 specimens were assembled for the purpose. This total consisted of all the material in the Museum of Comparative Zoology (M.C.Z.), two specimens from the American Museum of Natural History (A.M.N.H.), and all those in the Chicago Natural History Museum (C.N.H.M.). It was found that the material fell into three groups : M. /. frontalis (21 ex.), M. f. altirostris (21 ex.), and M. f. pyrrhocrxjptus (4 ex.).. I have retained altirostris as a race o{ frontalis as evidence is lacking that the two do not have separate ranges joined by an area where both occur together and interbreed, although the supposed area of hybrid- ization is decidedly large and there is doubt how much, if any, inter- breeding occurs within it. Further discussion of this matter is presented below. I regard -pyrrhocryptus as a race owing to its close relationship to frontalis and the fact that it apparently has a range separate from either altirostris or frontalis. This opportunity is taken to thank the cm-ators, Messrs. C. M. Bogert, C. H. Pope and K. P. Schmidt, for these loans, also Mr. K. P. Schmidt for furnishing data on a pyrrhocryptus belonging to the American Museum of Natural History. I also wish to thank Mr. Arthur Loveridge for fm-nishing data on the head coloration of the Uruguay series of altirostris in the British Museum (N.H.). 2 BREVIORA NO. 16 Key to the Subspecies 1. Parietals without pale area except sometimes along the anterior border; triads on body 7-15; ventrals 216-2'42 2 Parietals exhibit a pale area at least anteriorly; triads on body 13-17 (13 only rarely)'; ventrals (197) 203-220 altirostris 2. Light nuchal area between posterior end of parietal suture and first black ring from M to 3J^ scales in width; body triads 9-15 (9-10 only rarely), black median ring of triad usually not broader or but little broader than the black outer rings frontalis Light nuchal area between posterior end of parietal suture and first black ring from 5^2 to S scales in width; body triads 7-9, black median ring of triad usually twice as broad as the black outer rings . . pijrrhocryptus MiCRURUS FRONTALIS ALTIROSTRIS (Cope) Elaps altirostris Cope, 1859, Proc. Acad. Nat. Sci. Philadelphia, p. 345: Type locality unknown. Schmidt (1936, p. 199) gives the type locaUty as South America, but there is nothing to suggest this in the description or title of Cope's article. Range. Uruguay (where /ro7i/a/?;? is apparently unknown) . Outside of Uruguay, except where otherwise noted, it occurs alongside /ro/ito/w with which it may possibly hybridize. It also occurs in the Brasilian states of Rio Grande do Sul, Parana, Sao Paulo (probably also in Santa Catarina and southern IVIato Grosso). In Argentina it occurs in the Gobernacion of Misiones (probably also in Corrientes Province and possibly those of Entrerios and Buenos Aires, as well as Paraguay). Material examined. Argentina: Misiones: Iguazu Falls (C.N.H.M. 9260-2, 9383^), Monte Carlo (C.N.H.M. 9357). Br.\sil: Parana: eastern part (C.N.H.M. 37736-7, 37743-4), Pinheiro Machado (M.C.Z. 51495); Rio Grande do Sul: Itaqui (M.C.Z. 17853), Nova Teutonia (C.N.H.M. 16121-2, 18475), Pelotas (M.C.Z. 27651), Santa Maria (M.C.Z. 43337-9); State of Sao Paulo (M.C.Z. 20848), eastern part (C.N.H.M. 37742). Ventrals (197) 203-219 in cfd', (206) 208-220 in 9 9 ; subcaudals (16) 19-25 in cfd^, (16) 18-23 in 9 9 ; also ventrals 215-220, sub- caudals 19-22 in two unsexed juveniles; body triads 13-17; total length: largest d" (M.C.Z. 20848) 918 (859 + 59) mm.; largest 9 (C.N.H.M. 37737) 881 (834 + 47) mm. 1 12-15 body triads according to Boulenger (1896, p. 428) on Uruguay snakes he assigned to Elaps frontalis, but which are probably altirostris. 1953 RACES OF MICRURUS FRONTALIS 3 As regards the possibility of hybridization between althostris and frontalis, it has not reached a point where doubts are raised as to the assignment of any individual. For example, an altirostris 9 (C.N. H.M. 37736) from eastern Parana has the pale area on the anterior part of the parietals reduced, though otherwise the dark head markings, except for being rather brownish, are those of altirostris. The body triads are 16, which is higher than in a.ny frontalis ; ventrals 217. Only one frontalis, a Nueva Italia cf (M.C.Z. 47028) with 216 ventrals has so low a count, possibly resulting from an admixture of altirostris "blood". Admixture with fro7italis is further suggested by a 9 (C.N.H.M. 37737) and juvenile (C.N.H.M. 9261) altirostris with 220 ventrals, as well as by other material. Further support for the idea of hybridization is afforded by the ventral and subcaudal counts of Uruguayan altirostris which, being presumably free from association with frontalis, average lower than those from elsewhere. It seemed likely that all Uruguayan altirostris might have almost entirely white parietals such as is found in an apparently pure bred d^ (C.N.H.M. 16121) from Brasil, whose ventrals number 204, subcaudals 21, and body triads 15, but an examination of the Uruguay series in the British Museum by Arthur Loveridge reveals that they are as variable in this character as many from elsewhere. The tendency for one form to show one or more characters of the other may be due to individual variation in the nature of a "throw back" to more primitive characters. In this event possibly hybrid- ization does not occur and the two forms should be treated as specifi- cally distinct, a course which probably also should be followed if limited hybridization occurs. Available material is insufficient to allow of a definite opinion, but I incline to the theory of hybridization which is possibly limited by some factor such as infertility of the resulting offspring. The evidence is against general hybridization, though the two forms occur in many Brasilian states and elsewhere as the following figures show. Where the number of specimens is small those from several areas are grouped together, and all arranged from North to South so that a comparison may be made of the varying incidence of the two forms. altirostris frontalis Bahia?; Minas Gerais; Rio de Janeiro 5 Sao Paulo 2 11 4 BEEVIORA NO. 16 Parana 5 2 Misiones, Argentina; Villeta, Paraguay 6 1 Rio Grande do Sul 8 2 M. lemniscatns multicinctus Amaral (1944, p. 91), with Teixeira, Soares, Parana, Brasil, as a type locality, appears to be a composite of altirostris and frontalis. It is not known which of the two the type might be. MiCRURUS FRONTALIS FRONTALIS (Dumeril, Dmneril and Bibron) Elaps Frontalis Dumeril, Dumeril and Bibron, 1854, Erp. Gen., 7, p. 1223: Brasil (part). The type locality would appear to be Brasil, rather than Corrientes and Misiones, Argentina, as given by Schmidt (1936, p. 199), for Claussen's two Brasilian specimens appear to be the only ones whose description conforms to frontalis {sensu sirido) as currently regarded. The rest of the material listed under this form in the Paris Museum is apparently altirostris. It is interesting to note that Boulenger (1896, p. 428) lists three Brasilian frontalis in the British Museum from Clausen {sic) . (Actually it is not known which way the name should be spelled.) Dr. P. E. Vanzolini tells me that Claussen collected in Lagoa Santa, Minas Gerais; so that is probably the type locality of frontalis. Range. Southern Brasil, where it is known from: southern^ Bahia?, Minas Gerais, Rio de Janeiro (possibly also Espirito Santo and southern Goias, altirostris being apparently unknown and not expected to occur in these five states), Sao Paulo, Parana, Rio Grande do Sul (probably also southern Mato Grosso and Santa Catarina); Paraguay; while in Argentina it probably occurs in the Province of Buenos Aires; and possibly those of Corrientes and Entrerios and the Gobernacion of Misiones. Material examined. Brasil: Bahia?: Santa Cruz (M.C.Z. 3298); Minas Gerais: Belo Horizonte (M.C.Z. 13954, A.M.N.H. 22573), Uberaba (M.C.Z. 12698); State of Parana (C.N.H.M. 37206-7); Rio Grande do Sul: Itaqui (M.C.Z. 17854), Pelotas (M.C.Z. 27652): State of Rio de Janeiro: Rio de Janeiro (M.C.Z. 1375); State of Sao Paulo (M.C.Z. 12699, 16684, 20844-7, 20849-50), eastern part (C.N.H.M. 37738), Butantan (M.C.Z. 17759-60). Paraguay: Villeta: Colonia Nueva Italia (M.C.Z. 47028). > As M.C.Z. 3298 is only labelled Santa Cruz, Brasil, it ia uncertain whether it came from Bahia or one of the other states. 1953 RACES OF MICRURUS FRONTALIS 5 Ventrals 216-236 in c^ c^ , 221-242 in 9 9 ; subcaudals 20-26 in c^ d' , 17-23 in 9 9 ; body triads 9-15; total length: largest d" (M.C.Z. 27652) 1121 (1058 + 63) mm., largest 9 (M.C.Z. 13954) 860 (819 + 41) mm. The possibility of hybridization between frontalis and altirostris has already been dealt with under the latter. Comparison oi frontalis from different parts of its range furnished no grounds for present recognition of further subspecies to those enumerated here. However, some variation was noted. The light bands of the triads, normally 2 or 3 scales in width, are about 4 or 5 scales in width in a Rio de Janeiro cf (M.C.Z. 1375). Sao Paulo snakes tend to have more body triads than specimens from elsewhere, the counts not overlapping with those of more northern snakes but overlapping with more southerly ones ; these counts in turn overlap with those of northern snakes. In the event of a further race being recognizable, the availability of Amaral's multicinctus should not be overlooked. I follow Schmidt (1936, p. 199) in regarding Elaps baUocoryphus Cope (1859, p. 346) as a synonym oi frontalis frontalis. MiCRURUS FRONTALIS PYRRHOCRYPTUS (Cope) Flaps pyrrhocryptus Cope, 1862, Proc. Acad. Nat. Sci. Philadelphia, p. 347: Vermejo River, Chaco, Argentina (ex Schmidt. 1936). Range. Northern Argentina, where it is known from Mendoza and Cordoba provinces, and likely to occur in the intermediate areas between them and the Department of Santa Cruz, Bolivia, from which area it is known. Probably occurs east to the Paraguay River and east to the Parana River below its confluence with the Paraguay. Note also the type locality. Material examined. Argentina: (A.M.N.H. 17592); Province of Mendoza (M.C.Z. 15911-2). Bolivia: Santa Cruz: Buenavista, 450 meters (M.C.Z. 20622). Ventrals 224-232 in c^d", 228-(231) in 9 9 ; subcaudals 24-25 in d'd^, 24-? (tail truncated) in 9 9 ; ventrals about 225 and subcaudals 25 in an unsexed juvenile. Subcaudals appear to average slightly higher than in the other races. The count in parenthesis is that of a 9 (A.M.N.H. 65273) from La Cumbre, Cordoba, Argentina, kindly supplied by Mr. K. P. Schmidt. Body triads 7-9; total length: largest cT (A.M.N.H. 17592) 948 (896 + 52) mm., largest 9 (M.C.Z. 20622) 955 (900 + 55) mm. 6 BREVIORA NO. 16 The median black ring in the BoHvian example is not as wide in relation to the other two black rings of a triad as in the two Mendoza snakes, and it has one triad more than in any others counted. Whether this is an indication of a Bolivian race remains to be seen. It is probable that the ranges of all three races are more extensive than here indicated, at least in Argentina, according to Serie (1936, pp. 52, 53 and 55) whose nomenclature differs, but from whose descriptions it is possible to deduce with some degree of certainty which of the forms he had in hand. BIBLIOGRAPHY Amaral, Apranio do 1944. Notas sobre a Ofiologia Neotropica e Brasilica. Pap. Avul. Dept. Zool., Sao Paulo, 5: 83-94. BOULENGER, G. A. 1896. Catalogue of the Snakes in the British Museum. 3: xiv + 1-727, figs. 1-37, pis. 1-25. Cope, E. D. 1859. Catalogue of the Venomous Serpents in the Museum of the Academy of Natural Sciences of Philadelphia, with Notes on the Families, Genera and Species. Proc. Acad. Nat. Sci. Philadelphia, 1859: 332-347. Schmidt, K. P. 1936. Preliminary Account of Coral Snakes of South America. Field Mus. Nat. Hist. Zool. Ser., 20: 189-203. Serie, Pedro 1936. Nueva Enumeraci6n Sistematica y Distribuci6n Geogrdfica de los Ofidios Argentinos. Inst. Mus. Univ. Nac. La Plata, 1936: 33-61, maps. E V I O R A MiiseiieTi of Coimparative Zoology Cambridge, Mass. May 28, 1953 Number 17 A NEW 0RD()VI(7AX TRILOBITE FROM FLORIDA By H. B. Whittington bitroihirfioii. Tlie trilohite described below came from a core recovered from a deep well in Madison County, Florida. It is the only trilobite found in the numerous wells that have penetrated the Palaeozoic of Florida, and is relatively complete. Of special interest is its relationship to trilobites of central and southern Furope and north- west Africa (not with any so far known in North America), and its being by far the earliest calymenid to appear in North America. I am indebted to Dr. J. Bridg-e. F. S. Geological Survey, for inviting me to study the specimen, and to Drs. Bridge and Jean M. Berdan for allowing me access to the results of their preliminary study of the Palaeozoic rocks from test wells in Florida and adjacent parts of Georgia and Alabama. The advice on the identification of the trilobite offered by Dr. C. J. Stubblefield, Geological Survey of (rreat Britain, is gratefully acknowleflged. Age. The Florida trilobite is here regarded as related to early representatives of the calymenid-homalonotid groups, in particular to species of Colpocorj/phr, but less closely to Placsiaconiia. The former genus occurs in strata of Llanvirn (and probably early Llandeilo) age in Bohemia and Thiiringia (R. and E. Richter, 1927), northwestern France (Rouault, 1S49; Barrois, 1900), Portugal (Delgado, 1908), central Spain (\'erneuil and Barranrle, 1855), Morocco and /Vlgeria (Gigout, 1949; G. and H. Termier, 1950). PJarsiacomia occurs in younger strata, of Llandeilo age, in Bohemia and possibly Normandy (Prantl and Pfibyl, 1948). The Florida strata may therefore be of Llanvirn-Llandeilo age, i. e. in North American terms approximately late Canadian or early Champlainian. 2 BREVIORA NO. 17 Palcogcographical ImpIicatio7is. Evidence brought forward recently by Harrington and Kay (1951) and Wilson (1952) has served to strengthen the impression that the early Ordo\'ician faunas of South America, the Ouachita and Appalachian troughs, and northwestern Europe belong to the same general province. This Hkeness, at least ' between the Appalachians, Ireland, Scotland, and the Baltic, con- tinues into Middle Ordovician times (cf. Stubblefield, 1939, pp. 57-60). Ordovician faunas of central and southern Europe (including West Shrf'psliire and South Wales), and especially those of Llanvirn- Llandeilo time, include a characteristic group of genera, elements of which are rare or unknown farther north and west, i.e. in Scotland, Ireland, and the Appalachians (cf. Stubblefield, 1939, pp. 52-55). The Florida trilobite seems to be a lone representative of this more southerly fauna, and its presence may indicate that the Florida Palaeozoic rocks were laid down in a province faunally separated from that of the Appalachian-Ouachita trough (cf. King, 1950, pp. 657-658). In addition, Professor H. J. Harrington, University of Buenos Aires, informs me (personal communication) that the Florida trilobite is unlike any known in South America. The oldest calymenid-homalonotid trilobites so far known in the Appalachian trough (and elsewhere in North America) are Flcvi- cah/mene and ''Brongniaridhr , from Trentonian rocks of late Cham- plainian time. This is much later than the first appearance of caly- menids in Lower Ordovician deposits of England and Wales, and considerably younger than the presumed age of the Florida strata in question. Thus the Florida trilobite is a representative not only of a different faunal province but of the earliest calymenids in North America. On the other hand, the Ordovician of Florida is a clastic sandstone and shale sequence, dark in color, not unlike the Llanvirn- Llandeilo sediments of central European areas. The occurrence of a "Bohemian" type trilobite in Florida may be because of the suitable environment there, rather than because of the existence of some other type of barrier to faunal migration between Florida and the Appa- lachian province. SYSTEMATIC DESCRIPTION Family CALYMENTDAE IMilne Edwards, 1840. Genus Colpocoryphe Novak in Novak and Perner, 1918. H^53 A NEW ORDOVrCIAN TRILOBITE 3 Tiipc species. Calymene arago Houuult, 1S49, from the ()r(lo\-ician (Llandeilo) of north-western France. Discussion. The outline of the ghil)ena, position of eyes, thorax of 13 bluntly-terminated segments, and triangular pygidiimi combine to suggest that the P^lorida specimen belongs with the calymenid- homalonotid group of trilobites, which appears in Lower Ordovician times. It was first compared with illustrations of the homalonotid Plaesiacomia rara Hawle and Corda, 1847 (p. 55, PI. 3, fig. 30; see also Harrande, 1852, pp. 581-582, PI. 29, figs. 21, 22; 1872, p. 40, PI. 5, fig. 27; Prantl and Pfibyl, 1948, pp. 19-21, PI. 2, fig. 6). The outline of the glabella, presence of two faintly defined pairs of lateral lobes, posi- tion of the eyes (far outwards and forwards), form of the thoracic segments and pygidium, are strikingly similar. P. ram is distinguished by the angulate outline of the pseudo-frontal glabellar lobe, the lack of ring furroM'S in the pygidial axis, lack of broad border and border furrow on the pleural lobes of the pygidium, and presence of the first pair of pygidial pleural furrows. The "small circular elevation" on the inner posterior part of the fixed cheek of P. rara, mentioned by Prantl and Pfibyl (1948. p. 19), is apparently not present in the Florida specimen. Dr. C. J. Stubblefield (personal communication) suggested that the Florida trilobite should also be compared with species of the calymenid Colpocori/phe Novak, 1918 {in Novak and Perner, 1918), and especially with C. inopinaia Novak (in Pompeckj, 1898, p. 249; see also Novak and Perner, 1918, p. 37, PI. 4, figs. 1-16; R. and E. Richter, 1927, pp. 73-75, PI. 5, figs. 10, 1 1 ). The pygidium of Colpocoryphr is characteris- tic, the axis showing several ring furrows, the pleural lobes with few, Faint pleural furrows, but with a broad, gently convex border. The border furrow runs from about the mid-point of the anterior margin of the pleural lobe, sub-parallel to the margin, to abut against the tip of the axis (e.g. Barrande, 1872, PI. 2, figs. 35, 39). The Florida trilol)ite has this type of pygidium. The cephalon, however, differs from that of C. arago (Barrande, 1872, PI. 2, figs. 34, 35; PI. 8, figs. 11, 12) in that the eye is farther forwards and outwards, and the glal)ella lacks the deep glabellar furrows and well-defined lateral lobes. The cephalon of C. inopinaia has the eyes as far out and forwards as the Florida specimen, and the glabellar furrows are faint, the lobes lacking independent convexity. Thus the Florida trilobite, while it bears a considerable resemblance to Plaesiacomia rata, has the distinctive 4 BREYIORA NO. 17 pygidium of Colpocori/phr, and the cephalon is like that of one species of the latter. It is therefore placed in Colpocoryphe, and regarded as a distinct species. Colpocoryphe exsul sp. nov. PI. , figs. 1, 2. Holotype. External mould in dark-grey, micaceous shale from core taken between 5154 and 5162 feet depth in Hunt Oil Company's J. W. Gibson No. 2 well, sec. 6, T IS, R lOE., Madison County, Florida. This well was drilled in May, 1944, surface elevation 107 ft., and reached the top of the Palaeozoic rocks, beneath probable Lower Cretaceous rocks, at 4628 ft. depth. The total depth drilled to was 5381 feet, so that 753 feet of Palaeozoic rocks were passed through, mainly dark shale, some sandstones. Geological Horizon. Late Lower or early Middle Ordovician (see above) . Description. Glabella of length (sagittal) 3.7 mm., maximum width immediately in front of occipital ring 3.8 mm., narrowing to 1.6 mm. across pseudo-frontal lobe, gently convex, antero-lateral margins straight, anterior margin bluntly rounded. Occipital ring .17 mm. long (sagittal), defined by a shallow furrow curving slightly forwards to the mid-line, deeper in the outer part. In front of the outer one- third of the occipital furrow is a low, narrow, transversely-directed inflation, strongest at the inner end, dying out distally. In front of this inflation the basal glabellar lobe is exceedingly faintly outlined, the first furrows commencing at the mid-length and running diagonally inwards to about one-third the width. The posterior edge of the lobe is defined by the transverse inflation. Second glabellar lobes are indicated by the second furrows, commencing at one-third the glabellar length, running parallel to the first and extending in to one- third the width at that point. Axial furrows deep, continuous with deep, preglabellar furrow. Fixed cheeks gently convex, broad, outer parts and borders not preserved. x\nterior to the glabella a sagittally short preglabellar field is continuous with the cheeks, and the anterior margin of this field (probably the rostral suture) is a curve concave forwards, so that the length (longitudinal) of the preglabellar field increases outwards from the mid-line. Right palpebral lobe repre- sented by a small, low swelling at the margin of the fixed cheek, situated 1953 A NEW ORDOVICIAN TRILOBITE 5 in line with the preghil)ellar furrow and 1 .(> mm. out from the antero- lateral corner of the glabella. Thorax of 13 segments. Axial ring of second segment 2.7 mm. wide, of last segment 1.6 mm. wide. Axis gently convex, articulating furrows deep. Pleurae of left side only preserved, width (transverse) of pleura of second segment 2.7 mm., of last segment 2.1 mm. Pleural furrow deep, straight, slightly diagonal, ending against broad facet. Tips of pleurae bluntly rounded. Pygidium of length (sagittal) 2.7 mm., width measured over surface approximately 4.8 mm. Axis more than one-third total width, gently convex, the lilunt tip not reaching the posterior margin. First three ring furrows distinct, fourth (situated just beyond half length) faint. Pleural lobe gently convex, no pleural furrows, divided at about half width by prominent longitudinal furrow which runs from the anterior margin sub-parallel to the lateral margin to abut against the tip of the axis. Narrow border behind tip of axis. External surface of exoskeleton with reticulate pattern of fine, raised lines. REFERENCES Barrande, J. 1852. Systeme Silurien du Centre de la Boheme. vol. 1. Prague and Parii^. 1872. Systeme Silurien du Centre de la Boheme. Supplement au vol. 1. Prague and Paris. Barrois, C. 1900. Bretagne, Livret-Guide, VHP Congres geolog. internat. 36 pp. Paris. Delgado, J. F. N. 1908. Systeme Silurien du Portugal. Etude de Stratigraphie Paleonto- logique. Comm. Serv. Gool. Portugal, pp. 1-245, pis. GiGOUT, M. 1949. Le synclinal silurien et devonien de Foucauld (Maroc occidental). C. R. Acad. Sci., Paris, vol. 228, no. 15, pp. 1303-1304. Harrington, H. J., and M. Kay 1951. Cambrian and Ordovician Faunas of Eastern Colombia. J. Paleont., vol. 25, no. 5, pp. 655-668, pis. 96, 97. Hawle, I., and A. J. C. Corda 1847. Prodrom einer Monographie der ixihmischen Trilohiten. Prague, pp. 1-176, pis. 1-7. 6 BREVIORA NO. 17 King, P. B. 1950. Tectonic Framework of Southeastern United States. Hull. Am. Assoc. Petrol. Geol., vol. -ii, no. 4, pp. 635-671, figs. Nov.iK, O., and J. Per,\er 1918. Die Trilobiten der Zone Ddi7, von Prague und Umgebung. Paleontogr. Bohemiae, Ceska Ak. C. Frant. Jos. pro v6dy etc., no. 9, pp. 1-51 (Czech and German text), pis. 1-4. Prantl, F., and A. Pribyl 1948. Classification of the Bohemian Homalonotidae (Trilobitae). Bull. Int. de TAcad. tcheque des Sci., 46th yr., no. 9, 24 pp., pis. 1-2. Pom PECK J, J. F. 1898. Uber Calymmene Brongniart. Neues Jahrl)., Min. Geol. Pal., 1898 (vol.' I), pp. 187-250. RiCHTER, R. and E. 1927. tJber zwei fiir das deutsche Ordovicium bedeutsame Triloliiten. Senckenbergiana, vol. 9, ))t. 2, pp. 64-82, ])1. 5. ROUAULT, M. 1849. Memoire 1) sur la composition du test des Triloliites; 2) sur les changements de formes. . . . l-5ull. Soc. Geol. France, ser. 2, vol. 6, pp. 67-89, pis. 1, 2. Stubblefield, C. J. 1939. Some Aspects of the Distribution and Migration of Trilol)ites in the British Lower Palaeozoic Faunas. Geol. Mag., vol. 76, pp. 49-72. Termier, G. and H. 1950. Palcontologie Marocaine, vol. 2. Actual. Sci. Industr. no. 1095, Palis. Verneuil, E. de, and J. Barrande 1855. Description des fossiles trouves dans les terrains silurien et dcvonien d'Almaden, d'une partie de la Sierra Morena et des montagnes de Tolede. Bull. Soc. Geol. France, ser. 2, vol. 12, pp. 964-1025, pis. 23-29. Wilson, J. L. 1952. Cambrian and Ordovician Faunas from the Marathon Region. Guidebook, 1952 Spring Field Trip. W. Texas Geol. Soc, pp. 62-64. y *a.*--i >' >^- #-« EXPI AXATIOX OF PLATE 1 Citlpoccnjphr t.rsu' sp. nov. x 6. IT. S. Nat. Miis. Collections. Fig. 1. Huhher mould of holoty])('. The black spots on the left side of the cephalon are caused by minute holes in the mould. Fig. 2. Holotype, ex- ternal mould. E V I O R A Museum of Comparative Zoology Cambridge, Mass. September 23, 1953 Number 18 CHARACTERS AND SYNONYMIES AMONG THE GENERA OF ANTS Part II By William L. Brown, Jr. Museum of Comparative Zoology Harvard University What follows is a continuation^ of the inquiry into the characters and interrelationships among the genera of ants, having as its goal the eventual reclassification of the Formicidae along more natural and useful lines than those presently holding. Perhaps some readers will be surprised to find that several of the genera mentioned below were originally described in the wrong tribe or even in the wrong subfamily, and that many monotypic genera are actually synonymous with more familiar genera with which they have never been associated. Nevertheless, such cases are commonplace among the largely very artificial tribes of the larger subfamilies, and more of them will duly be exposed, and must be exposed, before the really basic work on ant classification can begin. A careful perusal of these cases, as well as those treated in Part I and in some of my other papers, will make clear the impossibility of constructing a workable key to the ant genera, and should also open the eyes of those who now regard the existing generic keys as basically sound. I think that I need not labor the fact that a key bringing what now prove to be synonymous genera out to different tribes is a key of very limited usefulness and is scarcely reliable. Only after the synonymy is fully uncovered, justified, and formally recorded can we begin to think of large-scale generic keys. It is suggested, therefore, that useful generic keys will be greatly hastened if more mjTmecologists will spend a little of their time investigating and formally publishing new synonymy. 'Part I: Breviora, Mus. Comp. Zcjol. 11: 1-13 (1953). 2 BREVIORA NO. 18 Anochetus Mayr Anochetus Mayr, 1861, Europ. Formicid., p. 53. Genotype: Odontomachus ghilianii Spinola, 1851, by designation of Bingliam, 1903. Myrrnapatetes Wheeler, 1929, Amer. Mus. Novit., 349: 6. Genotype: Myrrna- patetes filicornis Wheeler, 1929, loc. cit., fig. 3, male, monobasic. New synonymy. Wheeler described Mynnapatetes filicornis, "obviously one of the Dolichoderinae," from a unique male taken on Larat Island, near New Guinea. His original figure, and even more clearly the type specimen in the Museum of Comparative Zoology, show that he had before him not a dolichoderine, but a ponerine male apparently representing one of the smaller Indo-Australian species of Anochetus, and typical of males of the latter genus in every respect save for its larger-than-usual compound eyes. When males and workers are associated in one nest series, further synonymy may result at the species level. Meanwhile, Anochetus filicornis (Wheeler) becomes the necessary new com- bination. AsPHiNCTOPONE Santschi Asphinctopone Santschi, 1914, Boll. Lab. Zool. Portici, 8: 318. Genotype: Asphinctopone silvestrii Santschi, 1914, loc. cit., fig. 6, worker, monobasic. Asphinctopone (lucida) Weber, 1949, Amer. Mus. Novit., 1398: 7, figs. 5-7, worker. Lepidopone Bernard, 1953 (1952), Mem. Inst. Fran(,-aise Afr. Noire, 19 (1): 207. Genotype: Lepidopone lamottei Bernard, 1953, op. cit., p. 208, fig. 4, worker, monobasic. New synonymy. Professor Bernard has raised the genus Lepidopone for a species {lamottei) having the general characters of Asphinctopone, but sup- posedly differing in details of the coxae of the posterior legs, post- peduncle of the petiole and basiventral process of the gaster. How- ever, it is difficult to understand how these features may be cited as differences marking a separate genus when none of them has as yet been reported upon for either species of Asphinctopone. From the descriptions and figures of A. silvestrii and A. lucida, it seems probable, not that the characters are absent, but rather that they are obscured by glue in the types or that they otherwise escaped observation. Bernard offers no further evidence concerning these structures of either older species, so it is premature to insist that their presence in the new species constitutes a valuable difference. The basiventral 1953 CHARACTERS AND SYNONYMIES AMONG ANT GENERA 3 process particularly should not be singled out as a diagnostic feature unless it is found to be developed in some unique direction; actually, this process in some form is found widely throughout the Ponerinae in many genera and species. However, even should the two Asphincto- pone species prove, in the long run, to lack the unusual development of the characters in question as seen in lamottei, it would seem that the strong relationship of all tlu-ee species named above overrides con- siderations based on minor characters. It seems preferable by far to avoid setting up what is at best a very weak monotypic genus when the generic unity of the three species is so clear. Asphinctopone lamottei (Bernard) becomes the necessary new Combination. Hylomyrma Forel Pogonomyrmex {Hylomyrma) Forel, 1912, Mem. Soc. Ent. Belg., 20: 16. Genotype: Pogonomyrmex. (Hylomyrma) columbicus Forel, 1912, loc. cit., by original designation. Lundella Emery, 1915, Bull. Soc. Ent. France, p. 191. Genotype: Tetramorium reitteri Mayr, 1887, by original designation. New synonymy. When Forel described Hylomyrma, he neglected to mention the spurs of the middle and posterior tibiae. By his association of the new group with Pogonomyrmex, the original author at least created the impression that the spm-s were present and pectinate as in other members of the tribe Myrmicini. We still do not know whether pectinate spurs are present in Forel's two Hylomyrma species, but this question hardly matters now in the face of several exceptions to the same character in members of the Myrmicini. Forel also mentions as a character that "the labrum, with two lateral teeth, surpasses the anterior clypeal border." This is certainly incorrect; what Forel refers to here is the depressed anterior clypeal border which, in an indeterminate Hylomyrma specimen from Central x\merica, has a low, blunt clypeal tooth on each side in such a position that a superficial examination might lead to the belief that they were on the labrum. This same indeterminate specimen has the spurs present and very minutely and sparingly barbulate as seen at great magnification. It seems evident from Forel's description of the mandibles and petiolar node, as well as other features, that Hylomyrma agrees well with Emery's Lundella, even though Emei-y assigned his genus to tribe Tetramoriini. Lundella speciosa Borgmeier, 1937, [described in Arch. Inst. Biol. Veg. 2: 241, figs. 25-29 (worker)] appears to me indis- 4 BREVIORA NO. 18 tinguishable from Hylomyrma gocldii Forel, if the descriptions con- cerned are accurate. Surely, L. speciosa is a typical Hylomyrma, and should be compared with authentic specimens of H. goeldii. Another specimen from Nova Teutonia, Santa Catharina (F. Plaumann), con- sidered to be a Lundella near reifferi (Mayr) by Father Borgmeier, is also a Hylomyrma, and may be the same as goeldii, with the description of which it agrees. The recognition of the generic synonymy thus throws five species together and creates problems of specific synonymy that will only be certainly settled when more is known about the tjqjes concerned than is given in their mostly inadequate descriptions. I believe that Irogera, considered by Emery to be a subgenus of Rogeria, is either very close to or synonymous with Hyloimirma. The two Irogera species from the New World, /. procera Emery and 7. tonduzi (Forel), are known to me only from the original characteriza- tions, which are very inadequate, and I therefore suspend judgment until the types can be examined. Irogera should, however, be con- sidered as a genus distinct from Rogeria for the time being. Further it is clear that the Oriental-Pacific species placed by Mann in Irogera, and by Santschi in Rogeria, do not belong there, but instead should be shifted to Lordomyrma Emery. The further study of species belonging to the genera mentioned in this paragraph requires the review of cer- tain crucial types before a solid rearrangement can be made. It is apparent now, however, that Rogeria has never been clearly defined, and that it has served as a dumping ground for ambiguous myrmicine species thi'oughout its existence. While the species concerned are mostly rare and economically unimportant ones, little-known taxo- nomically or biologically, their systematic placement will affect myrmicine classification most profoundly. Blepharidatta Wheeler (tribal transfer) Blepharidatta Wheeler, 1915, Bull. Mus. Comp. Zool., 59: 484. Genotype: Blepharidatta brasiliensis Wheeler, 1915, monobasic. Wheeler and other authors have considered Blepharidatta to belong to tribe Attini or tribe Dacetini, but a recent study of the types of B. brasiliensis convinces me that the genus is really a member of the Ochetomyrraicini. The single species is very closely related to the species of Wasmannia Forel, differing chiefly in its more elongate head with produced posterior angles and in having a long, low petiolar node. Wasmannia seems scarcely to be separable from its sister-genus 1953 CHARACTERS AND SYNONYMIES AMONG ANT GENER-A. 5 Ochdomyrmcx Mayi- by means of the presently-employed character based on the form of the clypeus. Wasmannia wiUiamsoni Kusnezov is one ambiguous species, and there may be still others in which the clypeal form is intermediate. I have already shown that Hercynia J. Enzmann is a junior synonym of Wasmannia (Brown, 1948, Ent. News, 59: 102). Technomyrmex Mayr Technomyrmex Maj-r, 1872, Ann. Mus. Civ. Stor. Nat. Genova, 2: 147. Genotype: Technomyrmex sh-enuus Mayr, 1872, by designation of Bingham, 1903. Aphantolepis Wheeler, 1930, Psj'che, 37: 44. Genotype: Aphantolepis quadri- color Wheeler, 1930, loc. ciL, fig. 2, worker, monobasic. New synonymy. In his figure of A. quadricolor, Wheeler shows an ant agreeing in every respect with a Technomyrmex of group sophiae Forel, but with a small, clearly-drawn poison-ejecting conule at the tip of the gaster. This conule is characteristic of the subfamily Formicinae, and Wheeler assigned his new genus to this subfamily. The type worker of A. quadricolor has long since been dislodged from its point and lost, but two whole legs and other fragments remain stuck to the glue on the point. Placed beside the type pin in Wheeler's collection are other specimens, determined as this species by Clark and evidently con- firmed by Wheeler, collected by T. Greaves at Cairns, northern Queensland (the type locality is "Cairns dist., rotting leaves; A. M. Lea"). These specimens agree in detail with Wheeler's characteriza- tion and figm-es, and their legs correspond with the parts on the type point, but they do not have an ejaculatory conule. I myself took a specimen of this ant among dead leaves on the rain-forest floor at Kuranda, near Cairns (probably the precise locality of the original type collection), and this agrees perfectly with the other material. It is obvious that the ant in question is a dolichoderine, not a formicine, and that it is most closely related to Technomyrmex sophiae Forel of southern Queensland; Wheeler's depiction of the gastric apex is due to an illusion or an artifact that led him to think the conule was present in the type; he was probably further misled by certain large paired setae on the alitruncal dorsum that resemble those of Para- trechina, etc. The necessary new combination is Technomyrmex quadricolor (Wheeler); the ant is distinguished from other Tech- nomyrmex species by its rounded propodeum and its rather smooth. 6 BREVIORA NO. 18 shining integument; it is ferrugineous yellow in ' are deposited in the British Museum of Natural History, London. All figures X L Figs. 1, 2. Gri/poceras (Gri/poccras) t>idorsatoidcs n. s])., B.M.N.H. — C10428, holotype. Figs. 3, 4. Menuthionautilus kieslingeri Collignon. Immature specimen; hgure 3 shows clearly the ventral position of the siphuncle. B.M.N.H. — CI 0438a, i)aratype. i^^ -■ '^h platp: 1 EXPLANATION OF PLATE 2 The specimens illustrated on this plate are from the Lower Ceratite Marl, zone of Prionolohus rotundatus, Chideru, Salt Range, Pakistan. They are deposited in the British Museum of Natural History, London. All figures X 1. Figs. 1-3. MenuthionautUus kieslingeri Collignon. I. Mature phragmo- ■cone (the cross section is illustrated in text figure 3), B.M.N.H. C10433, holo- type. 2. Ventral view of small paratype, B.M.N.H. Cl043Sc. 3. Side view of small paratype, B.M.N.H. C10438h. Fig. 4. Grypoceras {(iri/poceras) aenndans n. sp. B.M.N.H. C47068, holo- type. PLATE 2 E V I O R A Mesemm of Comparative Zoology Cambridge, Mass. September 23, 1953 Number 21 THE ANCESTRY OF THE FAMILY NAUTILIDAE By Bernhard Kummel The evolutionary history of the ammonoids and nautiloids is ((uite parallel. The nautiloids had their greatest radiation in the early stages of their history — in the Ordovician and Silurian — and declined rapidly in numbers of genera per period until the Recent. The ammonoids, however, displayed a slow start in the middle Paleozoic hut reached a peak of development in the Upper Triassic and again in the Jurassic and Cretaceous. Both groups experienced periods of mass extinctions with few surviving lines which formed the starting point for new anfl generally even greater radiations. The transition from the Permian to the Triassic witnessed wholesale extinctions among the ammonoids with only the Xenodiscidae and Pronoritidae giving rise to the early Triassic stocks (Spath, 1934). The nautiloids at this critical boundary did not undergo such radical changes. A number of late Paleozoic families did become extinct but three major stocks continued uninterrupted into the Triassic. The Triassic-Jurassic boundary was, however, a time of crisis for both the ammonoids and the nautiloids. Only a single stock of ammonoids survived the Rhaetic to give rise to the vast numbers of Jurassic and Cretaceous forms (Spath, 1934). The Triassic families of nautiloids became extinct by the Norian. No Rhaetic nautiloids have as yet been described. The general character of Triassic nautiloids is quite distinct from those of the Jurassic. Spath (1927) was the first to present a comprehensive summary of post-Triassic nautiloids. He placed all the post-Triassic nautiloids in ."i families and 2S genera, many of which were established by him. On the problem of origin of the.se post-Triassic nautiloids Spath (1927. p. 23) discussed ''G'ri/pocrras cf. mi:so(licum'" (Hauer) recorded by Trechmann (191.S, p. 181) from the Upper Triassic (Carnian, bed C) of the Hokonui Hills, New Zealand. Spath considered this species a 2 BREVIORA NO. 21 Bisiphytcs and stated that "... there is more prol)ability of BIsiphytes being an invokite, glol)o.«e, development of the family Syringo- nautilidae Mojsisovics, with annular lobe." Neither Trechmann (1918) nor Spath (1927) illustrated this species and Trechmann's description is inadequate. The purpose of this paper is to illustrate this species and discuss its affinities, and review the problem of ancestry of post-Triassic nautiloids. The author is indebted to Dr. L. F. Spath and Dr. W. N. Edwards for many courte- sies during his visit to the British Museum (Xatin-al History) and for permission to study Trechmann's specimen. The photographs were taken in the British Museum l)y its staff photographer. The uniqueness of this species (here named Bisipluitis trcchmanni n. sp.) lies in the fact that it is the only tangible clue we have to deciphering the evolutionary patterns of the nautiloids across the Triassic-Jurassic boundary. Spath (1927, p. 23) recognized that Trechmann's specimen was not a Gri/poceras and goes on to mention it as ". . . indistinguishable from typical BIsiphi/tcs, with their strong strigations, especially on larger whorls, annular lobe and only slightly sinuous septa." With this statement I am in complete agreement. The evolutionary' patterns of Triassic nautiloids can be summarized as those of culminating trenfls which began in the Carboniferous. Three distinct lines of development are represented. One includes the Paranautilidae, C'lydonautilidae, and Gonionautilidae characterized by globular, occluded conchs and tendency towards sinuosity of the suture. The second major group consists of the Tainoceratidae in- cluding most of the "ornamented" Triassic nautiloids. This stock is generally evolute and with nodes and ribs. The Tainoceratidae in- cludes Pennsylvanian, Permian, and Triassic genera. The third major group comprises the Grypoceratidae and Syringonautilidae. The Grypoceratidae include evolute to involute, generally smooth, compressed nautiloids with tendency for modification of the venter. This family also includes Pennsylvanian, Permian, and Triassic genera. The Upper-Triassic developments of this family, Grypoceras and Gryponaufihis, are completely unlike Bisiphytcs trcchmanni. Within the Syringonautilidae there are forms which are very similar in conch patterns to Bisiphytcs trcchmanni. Syringoccras and Syringo- nautilus are the most similar to the species described here. These two genera include evolute nautiloids with rounded whorl sections in the early volutions that become more quadrate in outline adorally. The outer whorls have distinct umbilical and ventral shoulders and steep 1953 ANCESTRY OF THE FAMILY NAUTILIDAE CO en < d &i y ID < 5 z c 5 c .9 c c .2 External Anatomy of Hypognatha elaborata sp. nov. Figures 1-5 Fig. 1. Eyes, clypeus, and cephalic horns of male, from in front. Fig. 2. Median cephalic horn of male in profile. Fig. 3. Lateral cephalic horn and LE of male. Fig. 4. Left male palp, ventro-lateral view. Fig. 5. Left male palp, distal end of tarsus. 1953 A NEW SPECIES OF HYPOGNATHA 3 by 9/10 of their diameter. Laterals separated from one another by one half the diameter of PLE. Height of clypeus equal to sHghtly more than three diameters of AME. Clypeus with a robust curved median horn arising from just above the middle of this area; with a lateral horn on each side bearing at its distal end the two lateral eyes (Figs. 1-3). Chelicerae. Basal segment .65 mm. long; other features essentially as recorded for the female allotype. Maxillae, Lip, and Sternum. Essentially as described for the female allotype. Legs. 1243. Width of first patella at "knee" .184 mm., tibial index of first leg 14. Width of fourth patella at "knee" .13 mm., tibial index of fourth leg 15. Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in millimeters) 1. 2. 3. 4. Palp Spines. Even more reduced than recorded in the female. Probably here it would be well to regard them as being replaced by bristles. Palp. Patella and tibia short, without apophyses. Tarsus compli- cated; with embolus coiled broadly at distal end of bulb and near the tip it passes into a deeply grooved conductor and completes a second circle at a right angle to the first. Near the middleof the ventral surface there is a long, slender, undulating, lash-like apophysis; near the base there is a stout, two-pronged apophysis (Figs. 4, 5). Altogether different from the palp of H. nasuta 0. P. Cambridge. If what I have called the embolus at the distal end of the bulb were lacking, then what I have termed the median apophysis probably would be termed the embolus. This raises certain questions about the identification of parts in H. nasuta O. P. Cambridge as well as in this species. Abdomen. 2.34 mm. long; 2.470 mm. wide about two fifths from base where it is widest; the chitinous ring nearly surrounding spinnerets and anal tubercle less well developed than in female; covers somewhat less of the carapace than in the female; dorsum much less regularly subdivided into angular areas than in female; four dorsal sigilla very clear. .975 .487 .769 .975 .430 3.636 .968 .462 .682 .858 .412 3.382 .682 .252 .396 .418 .308 2.056 .902 .330 .550 .770 .363 2.915 .352 .176 .240 .572 1.340 4 BREVIORA NO. 23 Color in alcohol. Essentially as recorded for the female except as recorded below: carapace and chelicerae somewhat lighter in color; testudinate dorsum without such distinct angular areas; with many whitish sub-chitinous granules so that nearly the whole dorsum is whitish; venter much less clearly marked with brown oblique and transverse bands. Female allotype. Total length 3.185 mm. Carapace so nearly covered by abdomen that ordinary description of this part is nearly impossible. Observations on a paratype show that median longi- tudinal thoracic groove is lacking; considerably constricted opposite interval between first coxae and palpi; posterior two thirds smooth; anterior third granular on surface; appears to be devoid of spinules and with a scant supply of hair. Eyes. Eight in two rows; all dark except PLE which appear light perhaps because of degeneration; both rows would ordinarily be considered strongly procurved. Viewed from above, posterior row nearly straight; viewed from in front, both rows strongly procurved; central ocular quadrangle equally wide in front and behind, slightly longer than wide. Ratio of eyes AME : ALE : PME : PLE = 11 : 8 : 11 : 7. AME separated from one another by a little less than their diameter, from ALE by four times their diameter. PME separated from one another by a little less than their diameter, from PLE by five times their diameter. Laterals separated from one another by one eighth of the diameter of ALE. PLE irregularly oval; all others circular (Fig. 6). Height of clypeus equal to 37/11 of the diameter of AME. Chelicerae. Basal segment .75 mm. long; somewhat receding; quite robust; basal boss well developed; fang well developed and finely dentate along inner margin particularly in basal half; fang groove with five teeth along promargin and four along retromargin (observed clearly on dissected paratype but confirmed on allotype) . Maxillae. Slightly convergent; broadly rounded along outer margin where serrula is very extensive. Lip. Wider than long in ratio of about 2:1; gradually narrowed to a rounded point distally; deeply grooved transversely at base. Sternal suture distinctly procurved. Sternum. Of a modified cordiform shape; wider than long in ratio of about 4:3; deeply notched at posterior end; the notch is filled by a sclerite in the form of a peg; with posterior coxae separated by about 1.3 the width of one of them; with a scant covering of short stiflf bristles (Fig. 7). 1953 A NEW SPECIES OF HYPOGNATHA External Anatomy of Hypognatha elahorata sp. nov. Figures 6-10 Fig. 6. Eyes and clypeus of female allotype, from in front. Fig. 7. Sternum of female allotype. Fig. 8. Female allotype, dorsal view. Fig. 9. Epigynum of allotype, posterior view. Fig. 10. Epigynum of disvsected paratype to show complete posterior exposur Tarsi Totals Legs. 1243. Width of first patella at "knee" .195 mm., tibial index of first leg 14. Width of fourth patella at "knee" .1624 mm., tibial index of fourth leg 15. Femora Patellae Tibiae Metatarsi (All measurements in millimeters) .520 .715 .921 . .520 .650 .814 .380 .418 .418 .440 .616 .780 1. .975 2 .975 3. .812 4. 1.100 .440 3.571 .418 3.377 .352 2.380 .400 3.336 6 BREVIORA NO. 23 Spines. True spines almost completely lacking. A few weak spines may be recorded as follows: First leg: femm- dorsal one near middle, one prolateral near distal end; tibia prolateral one near proximal end. Third leg: patella apparently with a dorsal distal spinule. Three tarsal claws throughout. Palpal claw: pectinate in a single row. Ahdomcyi. 2.73 mm. long; 3.25 mm. wide near middle; very much flattened; with a soft membranous ring nearly surrounding the six spinnerets and anal tubercle; tracheal spiracle with a well chitinized lip just anterior to base of anterior spinnerets. Covers nearly two thirds of the carapace. Dorsal angulate areas clearly separated by narrow lines (Fig. 8). Epigynum. Epigynal plate a broad, rounded ridge with openings and other characteristic parts directed posteriorly and partly hidden by genital groove (Fig. 9). Posterior margin more clearly shown in dissected paratype (Fig. 10). Color in alcohol. Carapace a light rich reddish brown with nearly black indefinite flecks and streaks. Sternum much the same but with more dark flecks. Chelicerae a rich amber color. Legs yellowish with many dark flecks varying much on the difl^erent segments. Abdomen: the strongly chitinized dorsum is generally light brown but is finely pitted and dotted with many brownish specks; the angular areas are separated by yellowish lines; the venter is pro|jided with several brown oblique and transverse bands alternating with narrow light lines. Type locality. Male holotype from the Barro Colorado Island, C. Z., July, 1934; no mature male paratypes; female allotype from the same locality July, 1936. Paratype females from the same locality July- August, 1936; July, 1939; July- August, 1950. Several immature males and females July, 1936, 1939. Hypognatha nasuta O. p. Cambridge, 1896 (Figures 11-15) H. nasutn F. P. Cambridge, 1904 H. nasuta Petrunkevitch, 1911 The following notes are based upon a study of specimens of this species loaned by the British Museum (Natural History); collected April, 1905 at Teapa, Tabasco, Mexico; from the Godman and Salvin collection. Male. At the base of each lateral cephalic horn there is a groove ending in a shallow pit beneath each AME. The bases of the lateral 1953 A NEW SPECIKS OF HYPOGNATHA horns are relatively much broader than in H. clahorata sp. nov. and the clypeus is extended into a broad shelf from near the ventral border of which the medial horn extends (Figs. 11, 12). The cephalic part of the cephalothorax is roughened by shallow pits. The palp appears to follow the same general pattern as in H. clahorata sp. nov. with the long slender median apophysis extending much as in the new species, but what is here regarded as the long coiled terminal embolus in the new species is a relatively short tube following closely the conductor and very slightly coiled (Fig. 13). The dorsum is essentially like that in the new species. External Anatomy of H ypognatha nasuta O. P. Cambi-idj. Figures 11-15 Fig. 1 1 . Eyes, clypeus, and cephalic horns of male. Fig. 12. Median cephalic horn of male in profile. Fig. 13. Left male palp, ventro-lateral view. Fig. 14. Dorsal view of female abdomen. Fig. 15. Epigynum of female; a somewhat posterior view. 8 BREVIORA NO. 23 Female. Angular areas on the chitinized dorsum are bordered by much broader margins than in the new species (Fig. 14) and the whole abdomen is much deeper and more robust. Total length 3.51 mm.; abdomen 3.055 mm. long; greatest width 3.445 mm. Ratio of eyes AME : ALE : PME : PLE = 9:6:9:6. AME separated from one another by slightly less than their diameter, from ALE by slightly less than four times their diameter. PME separated from one another by slightly less than their diameter, from PLE by slightly less than five times their diameter. The epigynum has a central depressed region posterior to a strongly chitinized ridge (Fig. 15). BIBLIOGRAPHY Cambridge, O. P. and F. P. Cambridge. 1889-1905. Arachnida-Araneida. Vols. I-II. In: Biologia Centrali-Americana. Dulau & Co , London. Simon, Eugene. 1892-1903. Histoire Naturelle des Araignees. Deuxieme Edition. 2 Vols. Libraire Encyclopedique de Roret, Paris. Petrunkevitch, Alexander. 1911. A Synonymic Index-Catalogue of Spiders of North, Central, and South America, etc. Bull. Amer. Mus. Nat. Hist., 29: 1-809. E V I Museiiiiin of Compsirative Zoology Cambridge, Mass. January 27, 1954 Number 24 A NEW FAMILY, A NEW GENUS, AND TWO NEW SPECIES OF BATOID FISHES FROM THE GULF OF MEXICO By Henry B. Bigelow AND William C. Schroeder^ The West Indian-Gulf of Mexican region is proverbially poor in skates as compared with the more northerly coastal waters of the Atlantic; so much so, indeed, that two specimens only, of Raja akleyi Garman 1881, from the Yucatan Bank and nearby, were the only skates that had been reported in scientific literature from any part of the Gulf prior to 1921 . This is fewer than one is likely to find, stranded, along the beaches of Cape Cod during an hour's stroll on any summer day. And while a second species, R. texana Chandler 1921, is now known to occur in some numbers around the northern shores of the Gulf from Florida to Texas, it was not until the winter and spring of 1938-1939 that the trawling campaigns of ATLANTIS brought to light the presence of a varied skate fauna at depths greater than 200 fathoms around the coasts of Cuba. Examination of these, of collections of skates in the U. S. National Museum that had been taken by the x\LBATROSS many years ago, and of others taken recently by OREGON of the U. S. Fish and Wild- life Service, had brought to light ten new skates of the genera Raja, Breviraja and Cruriraja,^ up to 1950, from Cuban waters and from the northern part of the Gulf; also of a new genus, Springeria Bigelow and 1 Contribution No. 657 from the Woods Hole Oceanographic Institution. 2 These are Raja lentiginosa Bigelow and Schroeder 1951; R. olseni Bigelow and Schroeder 1951; R. teevani Bigelow and Schroeder 1951; Breviraja atripinna Bigelow and Schroeder 1950; B. colesi Bigelow and Schroeder 1948; B. cubensis Bigelow and Schroeder 1950; B. sinua- mexicanus Bigelow and Schroeder 1950; B. yucatanensis Bigelow and Schroeder 1950; Cruriraja atlantis Bigelow and Schroeder 1948 and C. poeyi Bigelow and Schroeder 1948. 2 BREVIORA NO. 24 Schroeder 1951, representing the little known family Anaeantho- batidae. We can now report the capture by OREGON of another new Raja from the northern part of the Gulf, and of a skate-like fish, from the southern part, for which a new family seems needed. Family PSEUDORAJIDAE, Fam. Nov. Family characters. Rajoidea without dorsal fins; with well developed caudal fin extending around tip of tail and forward along lower side of tail about as far as along upper side, supported by a great number of very slender ray-like strands, apparently cartilaginous;' with outer- posterior margins of pelvic fins nearly straight, or even slightly convex if spread widely; anterior wall of spiracle with a transverse row of low vertical ridges, representing the vestiges of the embryonic gill fila- ments ; pelvis with transverse element nearly straight, each of its outer corners with a short projection directed forward. The batoid fish, described below as Pseudoraja fischeri, falls clearly among the Rajoidea because of the shape of its pelvis and of the persistence of vestiges of the embryonic gill filaments on the anterior wall of its spiracles (mentioned above). It is rajoid, too, in general appearance; in the nature of its dermal armature; and in the fact that its snout terminates in a fleshy process, for a corresponding structure tips the snout in the rajoid family Anacanthobatidae, represented in the Gulf of Mexico by the genus S-pringcria Bigelow and Schroeder 1951, and perhaps in the skate Psammobatis m?Va(Garman)1877^ but is not known to occur in any of the Myliobatoidea. The presence of a rostral projection (albeit a short one) from the front of its cranium, and its lack of a tail spine, point toward a rajoid rather than a mylio- batoid relationship, though neither of these two features is strictly alternative, between the two suborders, for a few rajoids lack the rostral cartilage, while it is a matter of common knowledge that some myliobatoids lack the tail spine. On the other hand, a myliobatoid relationship, rather than a rajoid, is suggested for Pseudoraja, superficially, by the shape of its pelvic fins with straight or slightly convex outer contour, and by the large size and the shape of its caudal fin. But, again, we are dealing with 1 The caudal fin-folds of various species of Raja, that we have examined under the microscope, are supported similarly. 2 Gannan (Proc. Boston Soc. Nat. Hist.. Vol. 19, 1S77, p. 207) described the snout of mira as terminating in an "acute point", which is pictured as pricklv in his classic monograph (Mem. Mus. Comp. Zool., Vol. 36, 1913, PI. 27, Figs. 3-.^)). 1954 NEW FAMILY, GENUS AND SPECIES OF BATOID FISHES 3 characters that are not as strictly alternative as seemed once to be the case. Thus the pelvics of Pseudoraja are approached closely, in shape, by those of Sympterygia (undoubtedly a rajid) in which they are so weakly concave in outline, when spread, that they can hardly be characterized as "bilobed".^ And while the caudal fin is a close counter- part, in size and shape, of the caudals of the myliobatoid family Urolophidae, it is similar in shape in the rajid genus Springeria Bigelow and Schroeder 1951, though smaller. With Sympterygia and Pseudoraja wholly bridging the gap, in shape of pelvic fins, between the typical Rajoidea in which these are definitely bilobed, and the Myliobatoidea, in all of which their outer contour is continuously convex, it is evident that the precise shape of these fins must be abandoned, as alternative between these two suborders. But the shape of the pelvis itself is sharply diagnostic in this regard, for while it is nearly straight transversely, or bowed forward only very slightly, in all the Rajoidea for which its shape is known^ and has a projection (longer or shorter) directed forward from each of its outer corners, its transverse element is bowed strongly forward, and it has no prepelvic projections at the outer corners in all the myliobatoid families where it has been studied, but has a prepelvic process in the mid-line in some of them.^ The presence or absence of vestiges of the embryonic gill filaments on the anterior wall of the spiracle, after birth, seems equally diag- nostic, as between the suborders Rajoidea and Myliobatoidea, for these are present in various species of Raja, in Breviraja, in Cruriraja, in Springeria and in Pseudoraja, whereas no trace of them is to be seen in any of the myliobatoids where we have sought them, which include representatives of Dasyatis, Taeniura, Gymnura, Urolophus, Mylio- hatis, Actobatus, Rhinopiera and Mobula. The taxonomic significance of this character was forecast, in fact, more than a century ago, by Johannes Mueller's (Arch. Anat. Physiol. Jahrg. 1841, p. 274) discovery that these vestigial gill folds are present after birth in skates and in torpedos, as they are in many sharks, but not in Myliobafis, in Dasyatis (referred to by him as Trygon), or in Taejiiura. But while the contrast in this respect, between skates and sting rays, has been mentioned repeatedly since Mueller's day, 1 For illustration of the pelvics in Sympterygia, see Garinan, 1913, PI. 27, Fig. 1. - This includes representatives of the genera Raja, Breviraja and Cruriraja (Family Rajidae), and of Springeria representing the Anacanthobatidae. The shape of the pelvis is not known for the somewhat problematical family Arhynchobatidae. ' See Garman (1913, Pis. 53, 54) for illustrations of the pelvis in Dasyatidae, Potamotry- gonidae, Gymnuridae, Urolophidae, Myliobatidae and Mobulidae. 4 BREVIORA NO. 24 we can not find that any of our predecessors have taken account of it in defining the two subfamilies in question. Pseudoraja, in short, falls clearly among the Rajoidea. But its lack of dorsal fins with its well developed caudal fin forbid its reference either to the Rajidae or to the somewhat problematical family Arhynchobatidae, while the nature of its pelvic fins, the large size of its caudal, and its well developed dermal armature set it apart, equally sharply, from all known members of the Anacanthobatidae. Hence the necessity for a new family, lest it be left a taxonomic orphan. Genus PsEUDORAJA Gen. Nov. Generic characters. Pseudorajidae with shape of disc, and of tail relative to disc, as in skates of the family Rajidae; pelvic fins very large, wing-like, with anterior outline directed outward, about trans- verse to main axis of disc; front of cranium with a rostral projection, longer or shorter; a deep pit on ventral surface of head on either side, close posterior to nostril but entirely separate from latter. Other characters those of family Pseudorajidae. Type species Pseudoraja fischeri Bigelow and Schroeder. Pseudoraja fischeri,^ Sp. Nov. Figures 1, 2 Type. Female, 479 mm. long, southern part of Gulf of Mexico, near Campeche Bank, Lat. 22°42'N, Long. 86°41'W, 225 fathoms, OREGON Sta. 726; bottom temp. 47.7°F; U. S. Nat. Mus. No. 163368. Study material. Also female 428 mm.; female 262 mm., and juven- ile male 262 mm. from same station. Description. Proportional dimensions, in per cent of total length, of female 479 mm. long (type) and female, 428 mm. Disc: Extreme breadth, 48.6, 50.3; length 39.5, 40.0. Snout length in front ofr orbits 9.2, 8.4; in front of mouth 12.3, 12.1. Orbits: Horizontal diameter 4.4, 4.7; distance between 3.0, 3.0. Spiracles: length 2.3, 2.2; distance between 6.0, 6.1. 1 Named in recognition of E. N. Fischer's skillful portrayals of elasinobranchs. s From base of rostral filament. 1954 NEW FAMILY, GENUS AND SPECIES OF BATOID FISHES 5 Fig. 1. Pseudoraja fischeri, type specimen 479 mm. long. Rostral filament and side view of caudal fin to larger scale. U. S. Nat. Mus. No. 163368. 6 BREVIORA NO. 24 Mouth: breadth 4.7, 4.6. Nostrils: distance between inner edges 6.4, 6.3. Gill openings: lengths, 1st 1.0, 1.1; 3rd 1.2, 1.2; 5th 0.7, 0.7; distance between inner ends, 1st 11.2, 11.8; 5th 6.9, 7.0. Caudal fin: length of base, upper 7.7, 8.2; lower 6.9, 7.3. Distance: from tip of snout to center of cloaca 39.5, 38.5; from center of cloaca to tip of tail 60.5, 61.5. Pelvics: anterior margin 17.7, 17.7. 28 . 30 . „ , „ „ 28 . , ^ m male of 262 mm. Teeth: ^ in type; o^ in female of 428 mm. 28 Disc, from base of rostral filament, about 1 .2 times as wide as long, so broadly rounded in front that its anterior contour is not susceptible to angular measurement; the rostral filament about ^ as long as orbit, soft, narrow triangular, tapering to very slender but blunted tip. Fig. 2. Fseuduraja Jischcri. Left, ventral view of oronasal region with right hand nasal flap rolled back to show nostril and oronasal pit (0). x about 1. Right, skin from central part of pectoral, to show arrangement of prickles. X about 3 1/2. Anterior margins of pectorals weakly concave abreast of eyes and spiracles, convex thence rearward; outer corners broadly rounded; posterior margins moderately convex; posterior corners rather ab- ruptly rounded; inner margins weakly convex, to axils. Tail slender, its dorsal surface nearly flat, but its sides and lower surface rather broadly arcuate in cross section; its breadth at axils of pelvics about 80 per cent as great as length of orbit, its length, from center of cloaca to tip, about 1.5 times as gi-eat as distance from center of cloaca to base of rostral filament; the lateral caudal folds confined to about posterior 3^ of tail and extending a little beyond origin of caudal fin, very narrow along their anterior part, but widening rearward until 1954 NEW FAMILY, GENUS AND SPECIES OF BATOID FISHES 7 Hearing the level of the caudal fin. The breadth of the folds is about 3^2 S'S great as the height of the caudal above its axis, or about % as great as the length of the first gill openings. Entire upper surface of disc including skin above eyes, and of tail close set with minute, sharp pointed prickles curving rearward, except close along extreme posterior edge of pectorals ; also two large and one small thorn around anterior contour of orbit, the small one the outer- most; a group of two large and one smaller thorn close behind each orbit, with the largest member of the group the outermost; two thorns on each shoulder, with a median line of three from nuchal region to pectoral girdle; these are followed, after a short gap, by a row of about 32 along mid-line of disc and of anterior % of length of tail, the first five to pelvic girdle small, the next 17 or 18 large and conspicuous, with strongly striate bases, those farther rearward, along tail, pro- gressively smaller; and the posterior }/i of tail without large thorns, recalling conditions in Raja senta. Either side of mid-dorsal belt of disc and upper side of tail with 3-4 irregular rows of closely crowded thornlets ; no thorns large or small (apart from the prickles) anywhere on pectorals. Caudal fin generally prickly above caudal axis, but with only a few scattered prickles below axis. Upper sm-face of pel vies wholly naked. Lower surface of disc and of pelvics naked, also of anterior part of tail rearward to abreast of rear corners of pelvics, but densely prickly thence rearward to caudal fin. The dermal armature of small specimens essentially similar, except the mid-dorsal thorns fewer (about 26 in 2 specimens of 262 mm.). Snout (from base of terminal filament) to front of orbits about 2.9 times as long as distance between orbits, its length in front of mouth about 1.9 times as great as distance between inner edges of exposed nostrils. Orbit about 1.5 times as long as distance between orbits and about 2.1 times as long as spiracle. Upper eyelid bowed downward, and eye with a rounded black velum with crenate margin above pupil, as in skates of the genera Raja and Breviraja. Outer lip of spiracle smooth; vestigial gill-ridges on anterior surface of spiracle about 12 in number, rounded, their edges completely fringed with minute lobelets. Nasal cm-tains smooth edged; those of the two sides of head actually separated by a space about }/2 as wide as distance between exposed nostrils, but seemingly interconnected there by a weakly outlined fold of skin. Outer (posterior) margin of nostril smooth, slightly expanded in scoop-like form; the exposed nasal aperture noticeably small. 8 BREVIORA NO. 24 An interesting feature of this new skate is that if the nasal curtain, on either side, be rolled inward, and the skin between nostril and corner of mouth be spread outward, a deep pit is exposed, close behind the nostril, but its entrance separated from the latter by a bar of stiff tissue. The pit is directed forward-upward, and it extends so far that a probe inserted into it can be felt clearly from the dorsal side of the head, through the overlying skin. And it is so voluminous on the two larger specimens as to allow the entrance of an ordinary match stick, or of a slender lead pencil. The presence of this pit was wholly unexpected, for nothing comparable to it is to be seen, either in Raja, in Dasyatis, or for that matter in any of the other batoids that we have examined in this respect, and these include representatives of all known families, both of Rajoidea and of Myliobatoidea, excepting only the Arhynchobatidae. Mouth nearly straight transversely in females and in juvenile males, its contom- not known for adult males. Teeth ofi7t~npW9S' those of females and of juvenile males low, rounded, in quincunx mosaic; those of mature males not seen; the tooth bands attached rather loosely to the jaws. First gill openings about ^q as long as distance between exposed nostrils; third gills a little longer than first, fifth gills about 7/10. Distance between inner ends of first gills about 1.8 times as long as between inner edges of exposed nostrils and about 1.7 times between inner ends of fifth gills. No dorsal fins. Caudal fin a little longer than distance between exposed nostrils, its height above caudal axis about }4 as great as its length, its depth below axis about j^ as great as its height above latter; upper and lower caudal origins about even. Upper margin continuously rounded; lower margin less strongly so; tip broadly rounded, notched abreast of tip of axis in type specimen, but merely ragged there on slightly smaller specimen (428 mm.) and continuously rounded terminally in small specimens (262 mm.), evidence that its terminal contour on the type has resulted from mutilation. Pelvic fins more widely spreading than in most other rajoids, and of very diagnostic shape, the anterior outline directed nearly transversely to the main axis of disc and tail, and so long that the tips of the pelvics reach outward considerably beyond the margins of the overlying portion of the pectorals. Anterior margin nearly straight; outer corners broadly rounded; outer posterior (distal) margin straight, or 1954 NEW FAMILY, GENUS AND SPECIES OF BATOID FISHES 9 very weakly convex if spread widely; only very weakly scalloped to conform to positions of tips of radial cartilages; the rear corner abruptly rounded. The inner margin is so short that the pel vies of the two sides appear to be separated only by a shallow notch. But the condition of the clasper described below shows that this is not evidence of a partial fusion of the inner edges of the pelvics with the sides of the tail, but only that these edges are relatively shorter than they are in the skates of the genera Raja and Breviraja. The anterior margin of the fin soft and fleshy; the first radial cartilage noticeably stout, especially toward its base; the other radialia slender and flexible. Point of separation of clasper from inner margin of pectoral, in juvenile male, about even with the axil of the fin; tip of clasper already reaching a little beyond rear corner of pelvic. Claspers of mature male not seen. It is interesting, in this connection, that the claspers of juveniles of Springeria originate considerably in advance of the apparent axils of the fins, evidence of partial fusion of the inner margins of the latter with the sides of the tail (Bigelow and Schroeder, 1951, Fig. 1). Front of cranium with a low, blunt tipped rostral projection, reaching forward a little beyond level of fronts of nasal capsules; tips of anterior rays of pectorals reaching nearly to tip of snout. Pelvis with transverse element nearly straight, each outer corner with a short prepelvic projection. Color. Upper surface ashy gray, deepening rearward to sooty gray along posterior % of tail; disc rearward from orbits with many small vaguely outlined pale spots, most conspicuous on head, and along mid- belt thence rearward, less so outward over pectorals; outer parts of pectorals unspotted. The large thorns white basally, their tips gray. Anterior part of disc also with a rather conspicuous pattern of black dots marking the mucous pores as follows: a) an irregular row along outer margin of each pectoral from a little behind tip of snout to about abreast of axis of greatest breadth; b) a row of 7-8 on each side, diverging from tip of snout to a little in front of level of front of orbits; c) a row of four, extending in anterior-posterior direction close in front of each orbit; d) a larger black spot, marking a cluster of three or four pores a little outward and rearward from the posterior edge of each orbit; e) a similar black spot (3 or 4 pores) either side of the mid-line in nuchal region, close in front of the first large mid-dorsal thorn; f) one black dot (1 pore) on inner part of each pectoral, in shoulder region; a second in line of orbit a little rearward from shoulder 10 BREVIORA NO. 24 region; and a third about midway thence toward axil of pectoral. Tail irregularly pale spotted along anterior Yi, with a few scattered pale spots thence rearward to caudal fin; also with a vaguely outlined dark cross-bar a little posterior to level of tips of pelvics, followed, after a short gap, by a second such bar. Upper surface of pelvics ashy gray, without pale spots. Lower surface of disc ash gray with irregular sooty cloudings, these most conspicuous on abdominal region; the pectorals of a more brownish, the mid-belt of a more bluish cast. Lower surface of pelvics ashy gray, with bluish cast; lower surface of tail pale grayish white with irregular ashy-gray cloudings, and with the dark cross-bars of the upper surface encroaching downward across the sides rather conspicuously. The pale spots on the disc are much less conspicuous, and less definitely outlined on small specimens than on large, but the dark cross-bars on the tail are more conspicuous. &ize. The larger specimens being female, and our only male a juvenile, their sizes give no clue to the dimensions to which this skate may grow. Habits. Apparently confined to rather deep water. Range. So far known only from the southern side of the Gulf of Mexico, at the locality listed above, page 4. Family RAJIDAE Raja fuliginea, Sp. Nov. Figures 3, 4 Study material. Type specimen. A juvenile male, 306 mm. long, OREGON Sta. 534, northwestern part of Gulf of Mexico, Lat. 27°32'N, Long. 93°02'W; trawl haul at 400-450 fathoms, April 11, 1952; U. S. Nat. Mus., No. 163367. Distinctive characters. This skate resembles R. hathyphila Holt and Byrne 1908 so closely in the general arrangement of thorns and prickles, in proportional dimensions in general, and especially in the very dark coloration of the lower surface of disc and tail that we took it for a specimen of hathyphila on our first cursory glance. But a closer examination showed that it differs from hathyphila in a considerably more obtuse anterior contour of its disc (Fig. 3A) and especially in the fact that the entire lower surface of the tail, apart from a very narrow median stripe, is densely prickly from base to tip (naked in 1954 NEW FAMILY, GENUS AND SPECIES OF BATOID FISHES 11 bathyphila). The prickles, also, on the upper surface of disc and tail are coarser than on bathyphila, and there are no naked areas, while its pelvics are largely prickly on the upper surface (naked in bathyphila) The only skate yet known from the northwestern Atlantic, or from the Gulf of Mexico, other than R. bathyphila, with which fuliginea shares a uniformly dark colored lower surface is R. olscni. But it differs from olseni in blunter snout and tail prickly below. It shares a tail prickly below with R. mollis. However, it is marked off from mollis, not only by its dark lower surface {mollis is pale yellowish or whitish below) but also by a much thornier tail and disc, by its considerably more convex anterior contour, also by its harder rostral cartilage. Since these divergences concern characters that are not generally subject to much individual variation among the members of its genus, a new specific name seems requisite for it. We suggest fuliginea because of the sooty chocolate hue of its lower surface. Fig. 3. A. Outlines of front of disc; outer, Raja fuliginea, type specimen, and inner, Raja bathyphila from southern slope of Georges Bank to show difference in shape. B. Upper teeth of Raja fuliginea from near center of mouth. X about 15. R. fuliginea, like bathyphila and olseni, parallels R. badia Garman 1899, R. trachura Gilbert 1892, and R. abyssicola Gilbert 1895 of the Pacific coast of Central and Nortli America in its uniformly dark lower surface. But it differs sharply from all three of these in various respects. Description. Proportional dimensions, in per cent of total length. Juvenile male, 306 mm. long; ORP^GON Sta. 534, northwestern part of Gulf of Mexico, Lat. 27°32'N; Long. 93°02'\V; 400-450 fathoms. Disc: Extreme breadth 46.3; length 42.8. Snout length in front of: orbits 9.8; mouth 12.1. Orbits: horizontal diameter 3.9; distance between 3.6. 12 NO. 24 Fig. 4. Raja fuliginea, type specimen 306 mm. long, U. S. Nat. Mus. No. 163367, with right hand nostril and nasal curtain x about 3.4, and under side of tail in advance of first dorsal fin, x about 1.1. 1954 NEW FAMILY, GENUS AND SPECIES OF BATOID FISHES 13 Spiracles: length 2.3; distance between 6.7. Mouth: breadth 5.9. Nostrils: distance between inner ends 6.2. Gill openings: lengths 1st 1.3; 3rd 1.3; 5th 1.0; distance between inner ends, 1st 12.9; 5th 7.5. Distance: from tip of snout to center of cloaca 39.3; from center of cloaca to 1st dorsal 45.4; to tip of tail 60.7. First dorsal fin: vertical height 2.9; length of base 5.9. Second dorsal fin: vertical height 2.6; length of base 5.6. Interspace between: 1st dorsal and 2nd dorsal 0.0. Pelvics. anterior margin 14.4. Teeth: ^. Disc about 1.1 times as broad as long, so broadly and continuously rounded in front that the anterior contour is not susceptible to angular measurement; the tip of the snout projecting slightly, and blunted. The pectoral margins only very slightly concave abreast of eyes and spiracles, and broadly and continuously rounded thence rearward, around to posterior corners, without definitely marked outer corners; posterior corners rather abrupt; inner margins nearly straight. Axis of greatest breadth about % (66^) of distance rearward from tip of snout toward level of axils of pectorals. Tail with very narrow lateral folds along posterior % (38-39 /^) of its free length posterior to axils of pelvics; its length from center of cloaca about 1 .2 times as great to first dorsal and about 1 .5 times as great to tip as from center of cloaca to tip of snout. Upper surface of disc, also upper surface and sides of tail, densely set everywhere with rather coarse prickles curving rearward, or minute thornlets, a very narrow band close along posterior edges of pectorals being the only naked area. Also a group of small recurved thorns along anterior part of rostral ridge; four larger thorns in a line around inner side of each orbit; one thorn in mid-line in nuchal region; six large thorns on scapular region, two of these on either side with two in mid-line in pattern shown in Figure 4; a line of six smaller thorns along mid-line of disc from scapular region nearly to level of axils of pectorals, followed, thence rearward, by 3-4 irregular rows along anterior % of tail, succeeded by two thorns in mid-line, to first dorsal fin. First and second dorsals, and caudal membrane sparsely prickly; anterior lobes of pelvics 14 BREVIORA NO. 24 naked; posterior lobes rather densely prickly over inner and posterior portions. Lower surface of disc naked throughout. Lower surface of tail densely set with prickles or thornlets similar to those of upper surface, except that a very narrow median stripe is naked, both along anterior % of tail, and, again, rearward from level of origin of first dorsal fin. Snout in front of orbits about 23^ times as long as orbit; its length in front of mouth about twice as great as distance between exposed nostrils. Orbit about as long as distance between orbits and about L7 times as long as spiracle. Nasal curtain deeply fringed, with about 18 lobelets; expanded posterior (outer) margin of nostril fringed, also. Mouth nearly straight, the lower jaw arched forward only a very little centrally; its breadth about 1/5 (21%) as great as breadth of disc at level of mouth, and only about 6 per cent as great as distance from tip 41 . of snout to tip of tail. Teeth j^, with low triangular cusp, blunted at tip, arranged in quincunx in juvenile male, probably also in female. P'irst pair of gill openings about 22 per cent as long as breadth of mouth; distance between inner ends of first gills about 2.1 times as long as between inner edges of exposed nostrils, and about 1.2 times between inner ends of fifth gills. Dorsal fins about alike in shape and equal in size, their bases con- fluent, without intervening thorn or prickles. Caudal membrane, posterior to second dorsal, about % as long as base of second dorsal. Pelvic fins deeply concave outwardly; outer margin strongly scalloped around the concavity with three marginal lobes on the one fin, four on the other; but with the positions of the radial cartilages only faintly indicated thence rearward; anterior lobe narrow, fleshy with rounded tip; posterior lobe strongly convex; rear corner abrupt; anterior margin of anterior lobe about 90 per cent as long as distance from its point of origin to rear corner of posterior lobe; rear corners extending back from about 14 of distance from level of axils of pectorals toward first dorsal fin. Firm rostral cartilage detectable by touch as reaching very nearly to extreme tip of snout; tips of anterior rays of pectorals falling a little short of level of tip of rostral cartilage. Color. Upper surface of disc, tail and pectorals uniform dark ashy gray, darkest on anterior lobes of pelvics, but without definite dark markings anywhere. Lower surface of disc sooty chocolate to nearly black, and noticeably darker than upper surface on head and around 1954 NEW FAMILY. GENUS AND SPECIES OF BATOID FISHES 15 outer belts of pectorals. A sub-triangular area in region of cloacareach- ing forward about to pelvic girdle, and vaguely outlined, irregularly interrupted areas on the inner parts of the pectorals rearward from the gill region are of a somewhat paler sooty chocolate hue, perhaps partly as a result of rough treatment in the trawl. Lower surfaces of pelvics dark sooty, except pale on tips of anterior lobe; sides and lower surface of tail of a very dark ashy-gray, except somewhat paler along a narrow median stripe. Size: The type (and only known) specimen being a juvenile male, it gives no clue to the size to which this skate may grow. Habits and range. The depth of capture (400-450 fathoms), added to the dark color of its lower surface, marks fuliginea as a deep water species. Present indications are that the upper boundary to its usual range lies not far from 400 fathoms, else specimens of a skate made so conspicuous by its dark lower surface would almost certainly have been noticed among the catches of the many trawl hauls that have been made in the Gulf at lesser depths. Nothing more than this is known of its habits. The locality of capture lies in the northwestern part of the Gulf, some 100 miles off Galveston. It is an interesting question for the future whether fuliginea is restricted to the Gulf, or whether it has simply been overlooked in the open Atlantic. BIBLIOGRAPHY BiGELOw, Henry B., and William C. Schroeder 1948. New genera and species of batoid fishes. Sears Found., .Jour. Mar. Res., Vol. 7, No. 3, pp. 543-566. 1950. New and little known cartilaginous fishes from the Atlantic. Bull. Mus. Comp. Zool., Vol. 103, pp. 385-408, 7 pis. 1951. A new genus and species of anacanthobatid skate from the Gulf of Mexico. Jour. Washington Acad. Sci., Vol. 41, No. 3, pp. 110- 113. 1951a. Three new skates and a new chimaerid fish from the Gulf of Mexico. Jour. Washington Acad. Sci., Vol. 41, No. 12, pp. 383-392. Chandler, Asa C. 1921. A new species of ray from the Texas coast . . . Proc. U. S. Nat. Mus., Vol. 59, pp. 657-658. 16 BREVIORA NO. 24 Garman, Samuel 1877. On the pelvis and external sexual organs of selachians . . . Proc. Boston Soc. Nat. Hist., Vol. 19, pp. 197-215. 1881. Report on the selachians. Bull. Mus. Comp. Zool., Vol. 8, No. 11, pp. 231-237. 1899. The fishes . . . Mem. Mus. Comp. Zool., Vol. 24, 431 pp. Atlas, pis. 1-85, A-N. 1913. The plagiostomia. Mem. Mus. Comp. Zool., Vol. 36, xiii, 515 pp. Atlas, 75 pis. Gilbert, C. H. 1892. Descriptions of thirty-four new species of fishes . . . Proc. U. S. Nat. Mus., Vol. 14, pp. 539-566. 1895. The ichthyological collections of the steamer "Albatross" . . . ' Kept. U. S. Comm. Fish. (1893), pp. 393-476, 16 pis. Holt, E. W. L., and L. W. Byrne 1908. Second report on the fishes of the Irish Atlantic Slope. Sci. Invest. Fisheries Ireland (1906), No. 5, pp. 1-26, 4 pis. Mueller, Johannes 1841. Fortgesetzte Untersuchungen iiber die Pseudobranchien. Arch. Anat. Physiol., Jahrg. 1841, pp. 263-277. E V I O R A mseimmi of Cooiiparative Zoology Cambridge, Mass. JAxrARY 2S, 1954 Number 25 A NEW MIOCENE SPECIES OF PELU8I0S AND THE EVOLUTION OF THAT GENUS By Ernest Williams Amoiij,^ the reptilian remains from the island of Rusinga in Lake \'ietoria, Kenya ( olony, sent for determination to tlie British Museum (Xatural History) are the greater part of the carapace and a smaller part of the plastron of an apparently new species of Pcluiiios. Dr. ^^'. E. Swinton, who suggested that I examine the unidentified chelonian remains from British East Africa, has kindly consented that I describe the new form. Accordingly I name it: Pehsios ursLXGAE, new species Ti/pc: Coryndon Museum Ru FodlT a partial carapace and plastron. Horizon: Miocene of iiusinga Island, Lake \'ict()ria, Kenya ( olony.' Diaf/no.s-i.s: A Pciusios belonging to the a(laii.soiiii-(/aboiif7i.si.s section of the genus, distinguished by the following combination of characters: a very depressed shell (height included in length about four times); the carapace expanded posteriorly; the x'ertebral region \ery shallowly excavated, quite without keel; first vertebral scvite much larger than \ertebral 2 and wider than long; vertebrals 2, 3, and 4 slightly longer than wide; mesoplastra extremely narrowed medially, barely meeting. The living species of Pclusios fall into two sections: One, which is northern and western in distribution, comprises two species, I', (iddiisoiili and /-". gahoticu.sis. This group is characterized by having the anterior lobe of the plastron relatively long and the abdominal scutes relatively short, so that the sulcus between the abdominal scutes is included more than twice in the length of the anterior lobe. Also the mesoplastra are more or less tapered medially, so that the hyoplastra anteriorly and the hypoplastra posteriorly (or the hypoplastra only) are longer medially than laterally, projecting ' Por a suniiijarv of the geologv and the Mioeene fauna of Rusinga and adjacent areas .see Kent (1944). 2 BREVIORA NO. 25 into and filling up the inter\'a,l left l)y the tapered margins of the mesoplastra. The other group within the li\ing membei's of the genus is less restricted in distribution. One of its species — P. suhnigcr — overlaps most of the range of the first group and in fact extends bej'ond that range on the west to the (ape Verde Islands. On the east this same species extends to Zanzibar, the Seychelles, Mauritius and ^ladagascar. On the north, however, this species does not extend beyond British East Africa into the Sudan range of P. adansonii. The group typified by P. subnigcr is distinguished l)y having the anterior lobe shorter and the abdominal scutes longer so that the sulcus between the abdominal scutes is included less than twice in the length of the anterior lobe, and by having the mesoplastra not tapered and presenting straight transverse contacts with both hyo- and hypoplastra. The relationships of P. rnsingac are clearly with the first of these two living groups: the tapered mesoplastra clearly indicate this position. From P. adansonii, however, P. nisingac difters (1) in the more depressed shell, (2) in the absence of any vertebral keel, (.3) the first vertebral wider than long, (4) greater size. From P. gahoncnsis it difl'ers in (1) the posterior expansion of the shell, (2) the absence of any trace of vertebral keel, (3) the second to fourth vertebrals longer than wide. From both species it differs in the more extreme medial narrowing of the mesoplastra. The table below summarizes the shell characters of the two Recent and the fo.ssil species (I utilize the data of Loveridge, 1941, which I have, however, verified on other material). P. adansonii P. gabonensis P. nisinyae Sulcus between humerals Sulcus between humerals Unknown 3-4 times as long as 1 1 2-2 times as long as that between pectorals. that between pectorals. Mesoplastra tapered Mesoplastra tapered Mesoplastra strongly medially only poste- medially both anteriorly tapered medially riorly, thus a and posteriorly, thus an anteriorly and transverse hinge with ol)lique suture with both posteriorly, the hj'oplastra but an hyo- and hypoplastra. hardly m(>eting. oblique suture with the hypoplastra. A keel on the anterior A nodose keel in the No keel, the four vertebrals young, lost in the vertebral region throughout life. old. somewhat depressed. | 1954 EVOHTIOX OF PEI.I SIOS 3 Vortebrals alwut as lonj; At least vcrt('l)ials 1 V(Mtcl)ial 1 wider as broad in adults. to H hroadcr tlian lonjj than long, in adults. vcrtebrals 2-4 longer than wide. Height in longtli Height in length Height in length about 2.(1 times. 2.;^ to 3.8 times. about 4 times. Shell (iistineti.\- Shell not broadened Shell distinctly broadened posteriorly. ])osteriorl\-. broadened posteriorly. Known maxinnun size: Known maximum size: Estimated size: 185 mm. 2r)!) mm. 24.T mm. /'. rufiinqar thus contrives to combine some of the characters of both the two living memliers of its group. It occurs, also, outside — south and east — though not far outside, the present limits of its group. Only P. srihuujcr of the alternative group is known from Lake Victoria today. Three fossil species of Pcliisios have been previously descril)ed: P. rudolphi Aramhourg from the Lower Pleistocene of Omo, founded on a partial plastron and carapace (type in Paris Museum); P. df'icitzianus \. Reinach represented by fragments from the Middle Pliocene of Wadi Xatrun (type formerly in Munich, now destroyed); and P. blanckcnhorni Dacque, a skull from the Lower Miocene of Moghara (type in Berlin?). In addition and not previously recorded there are ahimdant fragments (Nairobi Museum) and a complete shell (British Museum No. R oTOl) of P. sinuatus (a still living species of the P. subnigcr group) from Bed I, Pleistocene of Oldinai. The fossil P. sinuafiis need not be compared with P. ru.simior. The Olduvai material is clearly referrable to the Recent species which still occurs in this area. P. rudolphi needs as little attention. The type (examined at the Paris Museum) resembles closely ohl specimens of Recent P. niuuatus. It may provisionally be accepted as ancestral to the P. sinuafiis of the later Pleistocene and of the Recent. P. drwitzianiis was originally described on quite inadequate material which, however, was still sufficient to place it as a member of the P. subnigcr group. It was redescribed from much better material by Dacque (1912), who at the same time discovered that a supposed Pliocene species of Pclomcdu.s-d {P. pUoccnica v. Reinach) was a synonym of this species. P. dcnihiaiuis as a member of the alternative 4 brf:viora no. 25 group requires no comparison with P. nisiiu/(tc. There remains, however, P. hlnnckriihorni, which is from a deposit apparently equivalent in age and Aery similar in fauna to that of the Lower Miocene of Rusinga Island, but 2000 miles distant. P. hlauclcm- horni and P. rusitigac cannot be compared, since one is l)ased on a skull, the other on a shell. The skull of P. hJanckcnhorni, inadequately described and figured only in dorsal view by Dacqur, seems similar to that of P. gahoncnsis. It may, therefore, belong to the same group within the genus as P. rusingac, and it is not impossible that the latter is a synonym. But to hazard the identity of forms 2000 miles distant from one another and represented by incomparable parts would be without substantial basis. Furthermore, a special element of doubt attaches to species be- longing to this section of the Pelomedusidae which, as with P. hkmckrnhorni, are founded solely upon the skull. It is a remarkable fact that Pchisios and Pclomcdusa, though quite distinct in shell characters, have extremely similar skulls. The .skulls of the two Recent genera can be told apart only by characters which in many other groups would be counted of specific value only. Reference of a fossil skull, therefore, to either genus is a doubtful procedure unless there is the confirmation of an associated shell. In the present case this leaves us with the possibilities that P. blanckcuhorni may be either specifically identical with P. nisingae, or specifically dift'erent, or it may belong to a dift'erent genus. This conclusion may appear as absurd as it is unsatisfactory, but this is a dilemma not uncommon in paleontology, and it is decidedly worthwhile to recognize and emphasize the difficulty of evaluation of fossil species l)a'-ed on parts not comparable. All that can be suggested as a method of decision, which, while arbitrary, is still not devoid of reasonableness, is that material from deposits of the same or equivalent ages and geographi- cally close may be provisionally associated if any apparently valid grounds for such association exist; but geographic distance or dift'erence in geologic age carry with them a presumption of distinctness which must be countered by stronger arguments than those that — for the moment — suffice in the other case. P. rusingac and P. hlanckcnhorni (if this is really a Priusiu.s), occurring in the Lower IVIiocene, aie the oldest members of this genus. It will be useful to consider them against the background of the early history of the family of which they are a part. The family Prioiiicdiisidae is certainly very old; it probably stems 1954 EVOLUTION OF PELT'SIOS 5 ultimately from the pleurosternuls of the Upper Jurassic and the ( "retaceous. Unfortunately the form which has l)een suggested as the oldest representative of the family, PJatiirhclnidcs ni/nssdc Haughton, is incompletely known and doubtful as to agi'. It li;is small laterally placed mesoplastra and is therefore not an obviously primitive form. Mesoplastra meeting in the midline are certainly primitive for turtles and Pilu.sio.s would therefore be more primitive in this respect, Kulrss the larger mesoplastra of P. lusios are a secondary development (see below). The single known character in which Platychdoides differs from Pelonu'dusa as ordinarily conceived — the absence of the median plastral fontanelle — does not in fact separate it from that genus, since, as I have been al)le to determine on British Museum specimens from Uganda and the Sudan, the median fontanelle is sometimes lacking in even small specimens of PcJomrdusa. The beds from which PJatiichcJoidcs ni/a.s-.sai' derives are Cretaceous in age, but to what part of the ( 'i-etaceous they belong is not known. This African form is therefore not certainly older than the better known pelomedusids of the Upper Cretaceous of North and South America and Europe, though it is probably as old. Widespread already in the Cretaceous, the pelomedusids continued so in the early Tertiary with represent- atives in Xorth and South America, England, Italy (del Zigno, 1887), Egypt, Congo and India. It is a curious fact that every one of these older members of the family that arc sufficiently known is pelomedusine in type rather than pelusiine, that is: the mesoplastra are small and lateral elements, as in Pcluiitcdusd ami Pl(tti/clirli>ld(;s, not large elements meeting in the center of the plastron as in Pciu^ios and the pleurosternids. Nor is it at all likely that this o!)servation is an artifact resulting from a failure to recognize as pelomedusids those with complete mesoplastra. A pelomedusid with large complete mesoplastra is immediately dis- tinguishable from a pleurosternid by the total absence of inframarginal scutes. The uniformity in the ccmdition of the mesoplastra in the oldest members of the family is an intimation that the pelomedusine type of mesoplastra (small and lateral) may be primitive for the family and that the pelusiine type (large and centrally meeting) may be second- arily derived from the pelomedusine. With this suggestion the known facts about Pdusios are fully congruent. The members of the genus Pclusios form a structural series in regard to the size of the mesoplastra, P. rus'nKjar having the most f) BREVIORA NO. 25 reduced mesoplastra, P. gahouru.'n'.s the next, P. adnnsonii next, and the P. siihnlgrr group the most fully developed mesoplastra. A structural series is always ambiguous unless the time dimension can be added, but P. rusitigac as the oldest shell belonging to the genus seems to provide this time element. The similarity of the skulls of Prloiiitdu.sa and Prlu.s-ioff further suggests relationship, and the existence of a species of PclomeduMi (P. progahata v. Reinach) anterior in time (Lower Oligocene) to the earliest (Lower ]VIiocene) Pchi.sios further supports the view that Pehmos is a relatively late and specializefl genus directly derived from Pelomcdusa. P. ru.s-ingac is thus a fortunate discovery, offering a much needed term in an evolutionary series — a series apparently affording an example of the reversal of an evolutionary trend. Acknowledgements: I am indebted to Dr. W. E. Swinton for the opportunity of examining and describing this fossil, to Prof. (". Arambourg for the pri^•ilege of examining the type of P. rudolphi at the Paris Museiuu, to Dr. H. W. Parker for permitting me to examine comparative material of the Recent genus in the Reptile Section, British Museum (Natural History), and to AL Jean Guibo for similar permission in Paris. The photographs in Plates 1 to 4 were made by Peter Green and are reproduced by permission of the British Museum (Natural History). This study is part of a series of researches made possible by the grant of a (niggenheim Fellowship during the year 1952-53. RKFEKEXCES Arambourg, C. 1948. Contrihutioii a I'otudc geologic [uc ft j)iileoiitologique du bassin du Laf Hodolphe et de la basse vakV tie i'Omo. 2me Partie. Paleontologio in Mis-sion Scientific [ue de I'Omo. 1932-1933. Vol. 1, fasc. 3, pp. 231-559. Dacqce, E. 1912. Die fo.'^silen Schildkioten Aegypteiis. Geol. Paleont. Ahhandl., vol. 14, i)p. 273-333. Hatghton, S. H. 192S. On some reptilian remains from the dinosaur bed.** of Xvassalaiid. Trans. Roy. Soe. S. Africa, vol. 16, pp. 67-75. Kknt, p. H. 1944. The Micjcene beds of Kaviroiido, Keii\a. (^uart. .Jour. Geol. Soe. London, vol. 100. ])p. S5-ll(i. 1954 EVOLUTION OF PELUSIOS 7 LoVERIDCK, A. 1941. Revision of the African tcnapin of the family Pclomcdusidac. liull. Mils. Comp. Zoo!., vol. SS, pp. 4()7-524. Rkinach. a. von 1908. Sc-hiklkroteiHcstc aus dem uegyptischen Tcrtiar. Al)han(li. SenckenlxMg. Xaturf. (les., vol. 29, pp. l-()4. ZlGNO, A. DEL 1887a. Chelonii scopcrti nei terreni cenozoici drlle i)i('al|M' Veiieti. Mem. R. Inst. Veneto, vol. 23, pp. 119-129. lS87b. Chelonio scoperto nel calcare nummulitico de avesa pressa Verona. Ibid., vol. 23, pp. 135-145. --^■m^. PLATr: 1 Pehisios rusliH/(U', dorsal view of type shell PLATi; -2 I'liufiioa rumngue, ventral view of iy\)v sli PLATE S Feliisios sinuatus {B.M.y^.H. H. o7(il). dorsal vit-w of shell Irom Bed 1, Pleistocene of Olduvai. PJ.ATl'] 4 Pelasios sinuutiis (B.M.X.H. H. oTOl), ventral view of shell Pleistocene of Olduvai. E V I O R A Meseem of Comparative Zoology Cambridge, Mass. February 3, 1954 Number 26 A PRELIMINARY LIST OF THE EARTHWORMS OF NORTHERN NEW JERSEY WITH NOTES By H. Davies Dover, New Jersey The literature of North American earthworms contains few references to New Jersey and no paper has appeared describing worms collected in that state. Moore (1895, p. 473) refers to three species found within 30 miles of Philadelphia {Allurus tetraedrus, Bimastos palustris and Sparganophilus tamesis), and Eaton (1942) mentions four lumbricids found at Alpine, N. J. and Edgewater, N. J. This paper is therefore presented as a contribution to the knowledge of the fauna of New Jersey. (Collections of earthworms were made above a line drawn east and west through Princeton, N. J. with a concentration in the area of Morristown and Dover. Notes are given on the living conditions of the various species where such are thought to add to our knowledge. The term 'clitellate' is used in this note to describe worms with recognizable featiu-es of clitellar development regardless of the stage. Where the tubercula pubertatis only are present, 'aclitellate' is used, while worms having no indications of clitellum or tubercula are con- sidered 'juveniles'. Family LUMBRICIDAE Genus AllOLOBOPHORA Eisen 1874 Allolobophora arnoldi Gates 1952 Morristown, Mt. Kemble Road, in soil by stream; April 15, 1950, 2 clitellate specimens. Andover, in rich loam in woods, April 12, 1953, 4 clitellate speci- mens, many juveniles. 2 BREVIORA NO. 26 Dover, in garden soil, April and May 1953, 3 clitellate specimens, many juveniles. Morristown, James St., garden, May 19, 1953, 13 clitellate and I aclitellate specimens. Mt. Freedom, in sandy ditch, May 19, 1953, 5 clitellate specimens. Beatystown, in meadow. May 1953, 7 clitellate specimens. Mt. Tabor, under stones in garden, June 8, 1953, 3 clitellate and II aclitellate (post sexual?) specimens. This is the third record of a species which has hitherto been found in Massachusetts and New Hampshire. The specimens correspond to the diagnosis given by Gates except that the number of segments covers a slightly wider range as the data from 21 specimens show: 139(2), 140(1), 141(2), 143(1), 147(1), 153(1), 154(1), 160(1), 164(1), 174(1), 177(2), 181(1), 182(2), 183(1), 184(2), 187(1). In each of these the anal segments were normal in appearance; however amputation is prevalent in this species and some might prove to be amputees. Allolobophora caliginosa (Savigny) 1826 Dover, in garden soil (clay), June 8, 1953, 2 clitellate specimens. Mt. Tabor, under stones in garden, June 10, 1953, 2 clitellate specimens. Pompton Plains, in marshy meadow, June 1953, 3 clitellate specimens. Paterson, in garden top soil, June 1953, 2 clitellate specimens. These specimens are tentatively assigned to caliginosa, but do not conform to the description given by Cernosvitov and Evans (1947, p. 13), as Table I illustrates. In every case a grey-brown pigment is present on the dorsal surface, the clitellum being a dull yellow. Description of tubercula pubertatis : Specimen No. 150a band-like, indentation at xxxii on upper side. 150b band-Hke, indentation at xxxii on upper side. 157a band-like, tripartite origin, on xxxi-xxxiii. Intersegmental lines distinct. 157b L. As 157a. R. Band-like, slightly indented at 31/32 and 32/33. 161a \ ) Band-like, tripartite origin, on xxxi-xxxiii. 162a I 162b ) 1954 EARTHWORMS OF NEW JERSEY 3 In each case the lateral portion of the tuberculum is translucent while the median portion is opaque and appears as a hand. Table I External characteristics of Allolobophora caliginosa Number Specimen of Tubercula Genital No. Locality segments Clitellum pubertatis tumescences 150a Dover 133 27-135 31-33 30 32 33 34 150b Dover 179 27-i35 ^30-33 30 32 33 34 157a Mt. Tabor 167 27- 34 31-33 30 32 33 34 157b Mt. Tabor 156 27- 34 31-33 30 31 32 33 34 161a Pompton Plains 161 27- 34 31-33 30 32 33 34 161b Pompton Plains 1.56 27- 34 31-33 33 34(R) 162a Paterson 159 27- 34 31-33 30 32 33 162b Paterson 130 (amputee) 27- 34 31-33 30 32 33 34 Allolobophora limicola Michaelsen 1890 Morristown, Glen Alpine Rd., in thick mud (pH 5.5) near branch of Primrose Brook, May 27, 1952, 25 clitellate specimens, April 26, 1953, 8 clitellate specimens, June 6, 1953, 5 clitellate speci- mens, many juveniles at each visit. Dover, 2nd Street, marshy ground, June 8, 1953, 6 clitellate specimens. Beatystown, wet meadow, June 10, 1953, 1 clitellate specimen. The Morristown location is undisturbed marsh and remains wet throughout the year except during the winter freeze. Castings were noted under logs and debris but were not apparent on the surface. Both the Dover and the Beatystown locations are unimproved and are saturated throughout the year. On the June 6 visit to the Norris- town location, worms were observed ?'?? copula approximately 4 inches underground. This is the second record of this species in North America, Gates (1953, p. 518) having found it in the x\rnold Arboretum in Boston. The specimens conform to the description given by Gates, the number of segments corresponding very closely and illustrating the narrow range of this species. Number of segments in 17 specimens: 104(2), 111(1), 112(1), 113(1), 114(3), 117(2), 121(2), 122(1), 123(2), 124(1), 129(1). NO. 26 Allolobophora longa Ude 1885 Dover, sandy soil in garden, Nov. 1951, 3 elitellate specimens. Morristown, in garden soil (clay), May 10, 1953, 1 elitellate specimen. Dover, in garden on 2nd St., June 8, 1953, 3 elitellate specimens. Beatystown, damp soil in meadow, June 9, 1953, 1 elitellate specimen, 1 aclitellate (post sexual?) specimen. Genus BiMASTOS Moore 1893 BiMASTOS PALUSTRis Moorc 1895 Flanders, in wet moss in rocky stream near Route 31, Nov. 1951, 4 elitellate specimens. Hardwick, in ditch in forest (running water), April 1952, 3 elitellate specimens. Ironia, in wet moss by stream, Succasunna-Ironia Road, May 3, 1952, 1 elitellate specimen. Swartswood, under moss on log by stream. May 29, 1952, 15 elitellate specimens in association with E. tetraedra. Mt. Freedom, in wet moss on marshy ground by stream, April 27, 1952, 7 elitellate and 1 juvenile specimens. Shongum, Raynor Road, in very wet sandy loam by stream, March 29, 1953, 4 elitellate specimens. This species seems to have an affinity for running water and is apparently more restricted in this respect than Eiseniella tetraedra with which it is sometimes found. The above confirms Smith (1917, p. 169) who states that B. -palustris is found in the wet earth of rivers and ponds and has been collected from New Jersey. Sperm atophores were noted on many specimens, the majority having one pair and two specimens having two pairs. One specimen from Swartswood has 53 segments, and one from Mt. Freedom 52 segments, and in each case there was no indication of amputation, the anal segment being normal in appearance. These numbers are smaller than the 80 to 100 given by Smith (1917). Genus DendrOBAENA Eisen 1874 Dendrobaena mammalis (Savigny) 1826 Morristown, Mt. Kemble Road, in moist soil by edge of stream, May 1, 1951, 4 elitellate specimens. 1954 EARTHWORMS OF NEW JERSEY 5 This is the first record for this continent of this species. It is con- sidered to be endemic in the British Isles and has also been collected from a few localities in France. Its occurrence in North America is not unexpected since it has been intercepted on plant materials im- ported into this country (Gates, 1953, p. 530). The specimens con- formed externally to the description given by Cernosvitov and Evans (1947, p. 20). Dendrobaena octaedra (Savigny) 1826 Morristown, Glen Alpine Road, in rotting wood, June 30, 1952, 17 clitellate specimens in association with D. ruhida. Little is known of the distribution of this species in North America although it has been collected in Massachusetts and Michigan in addition to Newfoundland and Greenland. Dendrobaena rubida (Savigny) 1826 Dover, under rotting grass cuttings in garden, Nov. 10, 1951, 2 clitellate specimens. Ironia, in damp moss by Succasunna-Ironia Road, May 3, 1952, 2 clitellate specimens. Morristown, Glen Alpine Rd., in rotting wood, June 30, 1952, 8 clitellate specimens and 1 juvenile with D. octaedra. The records of this worm in North America refer only to collections from New Hampshire, Massachusetts and Maine. However as the writer has also collected it in Michigan, it may be widely distributed. Dendrobaena subrubicunda (Eisen) 1874 This species is recorded by Eaton (1942) as occurring at Alpine, N. J. It was not found in recent collecting. Genus ElSENIA Malm 1877 EisENiA FOETIDA (Savigny) 1826 This species is found plentifully in farm manure heaps in many North Jersey localities and is sold for bait in the area. It was found in every manure heap where a search was made and only one specimen was collected in any other habitat. However, the current farming practice of spreading manure daily instead of accumulating it, is probably reducing the numbers of E. foetida which cannot be said to be as "abundant" as Smith (1917, p. 165) implied. 6 NO. 26 EiSENiA LONNBERGi Michaelsen 1894 Morristown, Mt. Kemble Road, in mud at bottom of stream, April 1950, 2 clitellate specimens. Stokes State Forest, Naponock Brook, in wet moss, April 12, 1952, 1 juvenile. Ironia, under moss and in wet earth liy stream. May 3, 1952, 2 clitellate and 4 juvenile specimens. Shongum Lake, in wet soil by stream. May 1952, 1 clitellate and 4 juvenile specimens. Dover, Millbrook Valley, in swampy ground, Aug. 1952, 5 clitel- late and 4 juvenile specimens. Great Swamp, Madison, in wet moss and debris by water (pH 6.0), April 26, 1953, 2 juveniles. In addition to the above, 4 clitellate and 9 juvenile specimens were obtained from the Pennsylvania bank of the Delaware above Montague, N. J. Some of these were completely submerged under water in gravelly mud. Table II External characteristics of Eisenia Innnbergi Number of Tubercula Regener- Condition segments Clitellum pubertatis ation Remarks aclitellate 77 — 26-28 56/57 clitellate 62 24-30 26-i29 — Amputee clitellate 110 P3-30* 26-28 — Amputee aclitellate 129 — 26, 27, 28 — juvenile 76 — — — juvenile 116 — — — immature 48 — ■ — 7/8 Apparent head regenerate clitellate 44 §24-30 26-28 — Amputee clitellate 79 ?24-30* 26, 27, 28 — Amputee clitellate 51 24-30* 26-^29 — Amputee clitellate 142 ?24-30* 26-i29 — juvenile 133 — — 101/102 juvenile 132 — — 47/48 juvenile 74 — — — Amputee juvenile 118 — — — juvenile 132 — — — juvenile 135 — — — juvenile 104 — — — clitellate 125 24-30* 26-i29 — *Clitellum feebly developed 1954 ElARTHWORMS OF NEW JERSEY There are few references to this worm and Httle has been published regarding its habitat. It has been found in Georgia, Virginia, North Carolina, Connecticut and Massachusetts. Evidently it is widely distributed in northern New Jersey and is probably endemic. Its habitat appears to be restricted to very wet locations in or along the banks of rivers or streams. The range in segment number of normal specimens from 125 to 142, extends the previous record of 138 segments (Smith 1917, p. 164). The incidence of amputation is very high in the specimens collected, whether this is due to predatism or to autotomy is not known, but none of the specimens autotomized during handling in collection or preservation. EisENiA ROSEA (Savigny) 1826 Stokes State Forest, in soil. May 30, 1950, 2 clitellate specimens. Dover, in clay soil in garden, Nov. 17, 1951, 1 clitellate specimen. Morristown, in marshy soil, (pH approximately 5.5), June 30, 1952, 1 clitellate specimen. Swartswood, in mud at edge of stream, May 29, 1952, 1 clitellate specimen. Layton, in muddy ditch by road, April 3, 1953, 1 clitellate specimen. Andover, woods near Lackawanna R.R. (Cutoff), in rich black loam, April 12, 1953, 2 clitellate specimens. Table III External characteristics of Eisenia rosea Number Tumescences Tumescences Specimen of Tubercula including including No. segments Clitellum pubertatis ab cd 44a 112 25-32 29-^31 26-31 12(L), 13(R) 44b 114 25-32 29-131 26-31 11, 12(L), 12(R) 61 109 25-31 29- 30 26-31 11, 12 55 128 26-32 29- 31 26-32 12(R) 122 131 26-32 29- 31 25-32 None 120 130 i26-32 30- 31 12(L), 11,12(R) 26-32 None Further description of tubercula pubertatis : No. 44a Tubercula bluntly elliptical from 28/29 to |31 without marginal incisions. Intersected by furrow 29/30 only. 8 BREVIORA NO. 26 44b Shape as 44a but extending to |31. Slight indication of intersegmental furrow 29/30 on L. side. 61 Tubercula elliptical, from 29/30 to 30/31 with incisions at 29/30 on both margins and distinct furrow across tubercula at 29/30. 55 Tubercula somewhat ill defined ellipse from 28/29 to 31/32 intersected by distinct furrows at 29/30 and 30/31. 122 Single tuberculimi on 29 and elliptical mass from 29/30 to 31/32. Furrow 29/30 very distinct. Slight incision on median margin at 30/31 with trace of furrow. 120 Tubercula elliptical from 29/30 to 31/32 with incisions in both margins at 30/31. Indistinct furrow at 30/31. The presence of tumescences in the clitellar region would appear to identify these specimens as var. maccdonica Rosa 1893, which is said to be characterized "by the presence of small, mostly lightly coloured glandular papillae along the border of the clitellum" (Cernosvitov and Evans 1947, p. 23). However, this variety is ill defined in comparison with the typical form and positive identification cannot be given. Genus ElSENIELLA Michaelsen 1900 EisENiELLA TETRAEDRA (Savigny) 1826 forma typica Morristown, at roots of grass in stream, April 15, 1950, 1 clitellate specimen. Stokes State Forest, in mud under stone in stream, June 10, 1951, 1 clitellate specimen. In wet moss, Naponock Brook, April 12, 1952, 3 clitellate specimens. Swartswood, under moss in stream, May 29, 1952, 2 clitellate specimens in association with B. palustris. Dover, in marsh near Orchard St. Cemetery, June 20, 1952, 18 clitellate specimens, many juveniles. Great Swamp, Madison, under wet leaves. May 1952, 8 clitellate specimens. Buttzville, under wet leaves at edge of Pequest River, April 16, 1952, 8 clitellate and 3 juvenile specimens. Morristown, Glen Alpine Road, in wet gravel at brook, March and April 1953, 10 clitellate specimens. One specimen collected at Morristown has the male pores on segment ix with the clitellum correspondingly forward, apparently due to hypomeric regeneration. 1954 EARTHWORMS OF NEW JERSEY 9 On many specimens, setae ab of segment xxii are genital, occasionally on prominent lightly colored tumescences. Genus LuMBRICUS Linnaeus 1758 LuMBRicus RUBELLus Hoffmeistcr 1843 Stokes State Forest, Big Flat Brook, under leaves, May 30, 1950, 1 clitellate specimen. LuMBRicus TERRESTRis Linuacus 1 758 Throughout that part of New Jersey with which we are concerned, this worm is very common in lawns and gardens. It is also found in meadows, particularly where the ground is marshy, and in ditches. It is much in demand by local fishermen and is sold for bait in many places. Genus OCTOLASIUM Oerley 1885 OcTOLASiuM LACTEUM (Oerley) 1881 Buttzville, under rotting leaves at edge of Pequest River, Aug. 16, 1952, 1 clitellate specimen. One clitellate specimen was also collected in May 1950 from gravelly mud under water level from the Pennsylvania bank of the Delaware above Montague, N. J. Both specimens have tumescences on ab of xxii (one specimen also on xxi, right). Family GLOSSOSCOLECIDAE Subfamily SPARGANOPHILINAE Genus SpaRGANOPHILUS Benham 1892 Sparganophilus eiseni Smith 1895 Morristown, Glen Alpine Rd., in mud under water (pH 5.5) in bay of Primrose Brook, May 27, 1952, 28 clitellate specimens, many juveniles. June 6, 1953, 2 clitellate specimens, many juveniles. Dover, Second St., in mud at water level of pond (pH of water approximately 8.0), July 16, 1952, 10 clitellate specimens, many juveniles. In stream feeding pond, June 8, 1953, 2 clitellate specimens, 3 juveniles. Castings were produced by this species at both localities. Early in 10 BREVIORA NO. 26 April, castings were formed on the surface of the mud, each casting being approximately | in. in diameter and ^ in. high and generally conical in shape. As time passed, the number of such castings became greatly increased until in four weeks the whole area was covered. In early May at the Morristown location, castings were observed under water. These increased in size progressively and became supported by grasses growing out of them, becoming eventually a beehive shape with a height of 6 to 8 inches and projecting above water level. As warmer weather arrived, the water level receded, leaving the casting exposed. Examination of these castings revealed numbers of juveniles, adult specimens being obtained only by digging in 4 to 6 inches of mud. In many cases, both of clitellate and juvenile worms and particu- larly at the Morristown location, the last 10 to 30 segments were brownish in color, probably indicative of parasitic bodies accumulating in the coelom. Specimens collected at Dover were noticeably shorter than those at Morristown where one specimen possessed 258 segments (165 to 225 according to Olson 1940, p. 9). Posterior regeneration was noted in six specimens. Family MEGASCOLECIDAE Genus PherETIMA Kinberg 1867 Pheretima agrestis (Goto & Hatai) 1899 Maplewood, rotting leaf pile in private garden, June 1953, 2 clitellate specimens. Pheretima levis (Goto & Hatai) 1899 Maplewood, rotting leaf pile in private garden, June 1953, 3 clitellate specimens. Pheretima sp. Maplewood, rotting leaf pile in private garden, June 1953, 2 aclitellate and 1 clitellate specimens. The last three specimens were athecal as well as anarsenosomphic and could not be referred to either of the two above-mentioned species. All the pheretimas were found in a garden which contains many oriental shrubs. The species are known to have been present for at least three years and may possibly have been imported directly from the orient. The leaf pile is under a large conifer and the protection thus afforded, together with the heat generated by the pile may preclude severe winter freezing. The writer is indebted to Dr. G. E. Gates for the identification of the specimens of Pheretima sp. 1954 EARTHWORMS OF NEW JERSEY 11 DISCUSSION Earthworms of twenty species have now been recorded from northern New Jersey and of these fifteen are recorded from the state for the first time. As might have been anticipated, the majority of species are widely distributed peregrine lumbricids of Eurasian origin. Dendrobaena mammalis was found for the first time outside of Europe where it is endemic in the British Isles. This species has been inter- cepted on plant materials imported into North America (Gates 1953, p. 530), and its appearance, therefore, might have been anticipated. Bimastos palustris and Eisenia Idnnbergi are evidently endemic, and relatively common in their restricted habitat along the banks of streams. Neither of these is listed by Olson (1940) as occurring in New York State and this may imply that northern New Jersey, corre- sponding roughly to the extent of the ice cap in glacial times, is the northern limit of their general distribution. The glossoscolecid Sparganophilus eiseni is presumed to be endemic, although its original source may well be south of New Jersey. That each of the endemic species has a limnic habitat may be of significance as the lack of competition and the ease of colonization thus afforded may have enabled these species to quickly repopulate the barren areas left by the retreating ice cap. Allolobophora Uinicola, found in three locations, has just recently been recorded from Boston (Gates, 1953), where it was thought to have been introduced from Europe. As the three New Jersey localities are respectively 10 and 16 miles apart, possibility of a single intro- duction seems unlikely. Allolobophora arnoldl Gates, recently de- scribed and only known heretofore from two New England states, was found to be relatively common. It is tentatively considered to be of European origin. The species of Phcrdima were not expected as no attempt was made to search in ornamental gardens or greenhouses. They are probably of East Asian origin and may have been imported directly with exotic shrubs. Of the anticipated species, Bimastos tenuis, Allolobophora chlorotica, and Dendrobaena subriibicunda were not found in recent collections. A. chlorotica was probably overlooked as it has been found in Rockland County, New York, within a few miles of the New Jersey border. Altogether, collections were made in 55 locations, some of which were found to be heavily populated. Five species were taken from a Dover garden and six species from a pond and adjacent swamp near Morris- 12 BREVIORA NO. 26 town. Allolohophora arnoldi, A. caliginosa, A. limicola and A. longa were found together in one piece of marshy ground in a garden. In northwestern New Jersey, comprising Warren and Sussex counties, the terrain is characterized by dense woodland rising to 1500-1800 feet. The soil of these woods is largely glacial drift with numerous rocky ledges and outcrops. Searching for earthworms in such localities proved to be most unproductive, possibly due to the fact that the soil drains and dries out very rapidly. However, Bimastos ■palustris and Eiseniella tctraedra were found in woodland ditches and streams at isolated locations where the top soil was entirely without worms. Similarly, the top soil of wooded areas in Morris County was found to be largely without worms except where the ground was damp as often denoted by the presence of skunk cabbage {Symplocarpus foetidus). SUMMARY Twenty species of earthworms have now been recorded from northern New Jersey, including sixteen lumbricids, one glossoscolecid and three megascolecids. Thirteen of these are recorded for the first time from this state, Dendrobaena viammalis being reported for the first time outside of Europe. Bimastos palustris and Eisenia lonnhergi are considered to be endemic and are widely distributed in the northern part of the state. Allolohophora limicola was found for the second time outside Europe and in three localities. Three species of Pheretima were present in a garden. REFERENCES Cernosvitov, L. and A. C. Evans 1947. Lumbricidae. No. 6 in Synopses of British Fauna. London, Linnean Society, 1-36. Eaton, T. H., Jr. 1942. Earthworms from the North Eastern United States. Jour. Washington Acad. Sci., 32, No. 8: 242-249. Gates, G. E. 1952. New species of earthworms from the Arnold Arboretum, Boston. Breviora, 9: 1-3. 1953. On the earthworms of the Arnold Arboretum, Boston. Bull. Mus. Comp. Zool. Harvard, 107, No. 10: 497-534. 1954 EARTHWORMS OF NEW JERSEY 13 Moore, H. F. 1895. On the structure of Bimastos pahistrls, a new oligochaete. Jour. Morph., 10: 473-496. Olson, H. W. 1940. FJarthworms of New York State. Amer. Mu.s. Nov., No. 1090: 1-9. S.MITH, F. 1917. North American earthworms of the family Lumbricidae. Proc. U. S. Nat. Mus., 52: 157-182. E V I MeseiLim of Comparative Zoology Cambridge, Mass. February 5, 1954 Number 27 ANTERIOR REGENERATION IN A SEXTHECAL SPECIES OF LUMBRICID EARTHWORM By G. E. Gates Records of regeneration, in an anterior direction, at known levels, by posterior pieces of lumbricid earthworms have been brought to- gether in two recent contributions, the first (Gates, 1949) containing all that had been found for FAsenia foctida (Savigny) 1826, the second (Gates, 1953) those for the other species that had been studied. Records also have been found of anterior regeneration by a number of specimens, unwittingly used along with E. foctida, that must have belonged to yet another species. Identification of this species is the problem with which this note is concerned. The records involved are available only because it had been found to be "a tedious operation cutting off a definite number of segments". Accordingly, "the number of segments cut off was not counted at the time" . . . but was calculated "after regeneration by the position of the vasa deferential or segments containing the seminal receptacles" (Morgan, 1895, p. 452). The latter organs, now usually called sperma- thecae, were said to be normally in segments "9-10-11" [idem, p. 455). Such a characterization with reference to paired organs of earthworms certainly can be understood to indicate presence of three pairs of the spermathecae, located one pair each in segments ix, x and xi. But, E. foctida has only four spermathecae, usually present, according to various authorities, in ix-x though occasionally in x-xi. The variation does not involve location of spermathecal pores which are on inter- segmental furrows 9/10 and 10/11. In view however of the difference in internal location might it be possible that two pairs of spermathecae were so located as to require reference, in most cases, to three segments? ' This means: segmental loration of the external openings of the deferent ducts, i. e.. the male pores. GEO 2 BREVIORA NO. 27 In this connection location of spermathecae was determined in each older individual (30 clitellate and 4 late juvenile) of the first entire colony of E. foetida that was accessible. Some variation as to which of two consecutive coelomic cavities any spermatheca got into as it grew through the parietes was indeed found (Table below). Never- Table 1 Segmental location of spermathecae in a colony of E. foetida Segments Number of specimens 9-10 10 10-11 9, 11 L9, LIO, RIO. R9, RIO, LIO. RIO, LIO, Lll 17 L9, LIO, R9, Lll. L9, LIO, RIO, Rll R9, RIO, L9, Lll. R9, Rll, LIO Spiral abnormality in metamerism of the spermathecal region 29 * On the left ide spermathecae in ix and x, on the right side both sperma thecae in x. Each worm had four spermathecae theless, only in four of the thirty-three metamerically normal worms is it necessary to refer to three segments to indicate spermathecal location, and in two cases all spermathecae are actually in one segment. These results, together with previous findings, show it is unlikely that twenty-two out of twenty-seven specimens of E. foetida would have had spermathecae in three segments. Furthermore, the method by which location of spermathecae of the other five specimens was indi- cated, e. g., "-7-8", as if the first of tliree pairs^ had not been found, also supports the interpretation of "normal" that was first suggested above, i. e., presence of three pairs in three consecutive segments''. - In one case, "-6-7?", in which the question mark seems to indicate uncertainty but as to what, was not explained. 3 Seminal receptacles cannot be regarded as a lapcftis calami for seminal vesicles. The latter are four pairs, in ix-xii, in E. foetida. Other species may have only three pairs but they are not in three consecutive segments, ix, xi-xii. 1954 ANTERIOR REGENERATION IN AN EARTHWORM 3 Accordingly, it is concluded that Morgan's specimens with sperraa- thecae in two segments were quadrithecal and, as he supposed, E. foetida, but that the others were of some sexthecal form with sperma- thecae in ix, x, xi. Three lumbricid species of this country are so characterized: Allolobophora chlorotica (Savigny) 1826 and Eisenia lonnbergi (Michaelsen) 1894 with spermathecal pores on 8/9-10/11, Dendrobaena odaedra (Savigny) 1826 with the pores on 9/10-11/12. E. lonnbergi, though at present within the same genus as foetida, has a quite different habitus, is a native of the southern states, has never been reported from the vicinity of the region where the worms in question apparently were secured, has not been found in manure heaps — the som-ce of the material — and is unlikely to have been involved. A. chlorotica normally has an obvious yellow or green color- ation, as well as other characteristics that would immediately dis- tinguish it at a glance from foetida, and has only once been reported from the vicinity of a manure pile. D. octacdra does have the same red pigmentation that characterizes E. foetida, uniformly distributed (instead of in transverse bands) as in many individuals of E. foetida, has been found occasionally in manure piles, and does have much less special glandularity in the region around the male pores (so that site of the male pores would be less easily recognizable). Of the three species, D. octaedra is the one most likely to have been inadvertently accepted as E. foetida. Nothing, however, has been known of re- generation at levels in front of 15/16 in any species of Dendrobaena and D. octaedra has not, apparently, been available to others who have studied regeneration in earthworms. An unusually large octaedra proportion of the population (in the manure heap that provided the experimental material under consideration) seemingly is indicated by such figures as are available : one of eleven specimens (Table 3, Morgan, 1895), five of sixteen (Table 10), six of eleven (Table 12, first half), four of nine (Table 13), twenty-two of the twenty-seven cases in which spermathecae were mentioned. The number of records (22) of anterior regeneration that can now be recognized as of octaedra is larger than for any other lumbricid (Gates, 1953) except E. foetida and two species which have not yet been studied in this country, Allolobophora terrestris (Savigny) 1826 and Lumhricus rubellus Hoffmeister 1843. These records (Table below) show that head regeneration is possible at each level from 1/2 to 8/9 inclusive. The five segment regenerate shows that equimeric regener- ation can be expected at each level from 1/2 to 5/6 inclusive at least. 4 BREVIORA NO. 27 Failure to secure equimery at levels behind 3/4 indicates that con- ditions, either of the external or internal environment or both, were less than optimal for regeneration by octaedra, as well as for E. foetida'^. The results obtained from four worms that were deliberately cut diagonally instead of transversely (note under table) may indicate that head regeneration is also possible at levels back at least to 12/13 Table 2 Number of segments in head regenerates of Morgan's sexthecal earthworms Number of segments Record quoted from Level of regenerated Morgan, 1895, amputation* 2 3 4 5 pages EL 2/3 1 - - - 447 EL 3/4 3 1 - - 447, 455 EL 4/5 - 2 - - 448, 456 EL 5/G - 1 1 - 455 EL 6/7 1 1 1 - 455, 456 EL 7/8 - 1 1 - 456 EL 8/9 1 1 1 1 455, 456 * A symmetrical homoeotic would not have been recognizable after operation. EL Estimated level. Estimation made, after regeneration and presumably after preservation, from position of male pores and/or location of spermatheoae. When cuts were made diagonally "anterior segments obliquely amputated" (p. 457), four to twelve segments were said to have been completed. In three of these, which had spermathecae after regeneration in "9-10-11" and in which 4, 8, and 12 segments had been "completed", presumably no segments were completely excised. The other specimen had a hypomeric regenerate presumably (indicated by location of the spermathecae in "8-9-10"). All of i-iii was then removed in addition to parts of iv-x ("7 segments completed"). After excision of a piece estimated to comprise 103^ segments, and subsequent regeneration (p. 455), there were still spermathecae in the first two segments of the substrate. The estimate could have been correct, regardless of species, only if the worm involved had been homoeotic (-|- 1 or more), or if spermathecae had been developed in the substrate during regeneration. Nothing of the latter sort has ever been recorded from the Lumbricidae. < The number of segments in the head regenerates of E. foetida was smaller than has been obtained (Gates, 1949) and in view of this evidence from two different species the less favorable conditions may have been in the external environment. 1954 ANTERIOR REGENERATION IN AN EARTHWORM 5 SUMMARY Individuals of some sexthecal species were frequently used along with E. foctida in Morgan's early studies of earthworm regeneration. From the information available as to distribution, habitat, habitus, etc., of the sexthecal species of this country, it is concluded that only D. octacdra is likely to have been inadvertently taken for E. foetida. Nothing has been known of anterior regeneration in the genus Den- drobaena and the records now attributable to D. octaedra show that it is able to develop a head regenerate, in an anterior direction, at each level back to 8/9, possibly to 12/13, and, in better conditions, complete replacement of excised segments may be expected at least at all levels from 5/6 anteriorly. REFERENCES Gates, G. E. 1949. Regeneration in an earthworm, Eisenia foetida (Savigny) 1826. I. Anterior regeneration. Biol. Bull., 96: 129-139. 1953. On regenerative capacity of earthworms of the family Lumbri- cidae. Am. Midland Nat., 50: 414 419. MORGA.V, T. H. 1S95. A study of metamerism. Quart. Jour. Micros. Sci., 37: 395-476. E V I Museium of Comparative Zoology Cambridge, Mass. Fklriary S. 1954 Number 28 CLEMMYDOPSIS BODA A VALID LINEAGE OF EMYDINE TURTLES FROM THE EUROPEAN TERTIARY By Ernest Williams In 1847 Hermann von Meyer gave the name Kmys turnavrnsis to a small emydine turtle from the Upper (Sarmatian) Miocene of Turnau in Steyermark, Austria. He published no description at that time, but validated the name by a full description and figure in 1858. Kmys fiimauensis was founded on an anterior fragment of carapace (nuchal and both first peripherals, second peripheral of the left side, both first pleurals, and parts of left pleurals 2, 3. and 4). Its most evident peculiarity was the absence on the first pleurals of any indi- cation of grooves for the costal scutes. The first vertel)ral thus ex- tended across the entire front of the shell back of the anterior marginals and the relatively broad nuchal scute. The second and third ^'ertebrals less obviously but unmistakably had a similar great lateral expansion. Von Meyer commented: "Von alien mir bekannten Schildkroten mit Grenzeindriicken zeichnet sich vorliegende durch den Mangel an Seitenschuppen aus. . . . Der Verlauf der Grenzeindriicke sonst ist in dieser Schildkrote so regelmassig dass der Mangel an Seitenschuppen unmoglich fiir eine zufallige Erscheinung oder fiir eine Abnormitat gehalten werden kann: er wird der Species wirklich zugestanden haben, und es wird sich eigentlich nur um Entscheidung der Frage handeln, ob der ganzliche Mangel einer Schuppenart in einer Schildkrote zur Errichtung eines eigenen Genus brechtigt oder nur zu den Kennzeichen gehort, welche bei der Unterscheidung von Species in Anwendung kommen. Mit der Beantwortung dieser Frage mochte ich um so mehr bis zur Kenntniss der fehlenden Theile der Schildkrote, namentlich (les Bauchpanzers zuriickhalten, da die hervorgehobene Abweichung im Hautskelet, so auffallend sie ist, mit einer Abweichung in der Zahl oder Form der knochernen Theile, wenigstens so weit diese verliegen, 2 BREVIORA NO. 28 nicht verbunden sich zeicht; die knochernen Theile sind vielmehr Emys entsprechend gebildet. Zur P^rrichtung jedoch einer neuen Species war wohl hinreiehender Grund vorhanden." Fig. 1. Clemmydopsis turnauensis (von Meyer). A. Type specimen, after V. Meyer. B. Carapace, after Staesche. C. Plastron, after Staesche. Von Meyer mentioned also that he had an anterior fragment from Teitiary formations at Chaux-de-fonds, Switzerland, which might belong to Emys turnauensis, but that he had never seen anything re- sembling this species elsewhere in Switzerland. In 1927 Anton Boda described and figured from the Lower Pan- nonian (Lower Pliocene) beds near Sopron, Hungary, a complete 1954 CLEMMYDOPSIS 3 dorsal shell of a form which he recognized as similar in its vertebral pattern to Eviys turnaucnsis but which he referred to a new genus and species as Clcmmydopsis soproncnsis. The new form differed from Emys turnaucnsis in the shape of the neurals. In Cleminijdopsis so- pronensis neurals one to three were hexagonal, short-sided behind, and neural four quadrilateral, while the first neural was oval, the second Fig. 2. Clenuniidopsis sopronensis Boda. A. Type carapace, after Boda. B. Plastron, after Thenius. to fourth hexagonal, short-sided in front, in Kmi/s turnaucnsis. Boda assigned his new form to the section of the Emydinae which includes Geoemyda {Nicoria of authors), doing so on the basis of the character of neural shape to which great taxonomic weight had been given by Boulenger, Siebenrock and others. In 1931, K. Staesche placed on record from the Sarniatian Miocene of Steinheim in Wiirttemberg material very similar to the unique type of Emys fnr7iavcnsis but which he regarded as new, giving it the name Clcmmys stcinhcimensis. Staesche's material was much more nearly complete than that of either Boda or von Meyer. He had three speci- mens, none individually perfect but together providing satisfactory knowledge of both carapace and plastron. Staesche described his material very fully and discussed the re- lationship of his species to Emys turnaucnsis and Clcmmydopsis so- pronensis. He admitted the possible synonymy of his species with 4 BREVIORA NO. 28 Emys turnauensis but considered that one minor difference might be of specific value: the marginal scutes did not encroach upon the first pleural of turnauensis as they quite clearly did in sieinheimensis. He relegated both forms to the genus Clnnviys, regarding the single point of difference from Clemviys as usually understood — the absence of the first and second costals — as of specific value only. In the case of Clemmydojms soproncnsis, Staesche was not in doubt as to the distinctness of the species, but he was quite doubtful of the validity of the new genus. He argued that the difference in neural shape appeared to preclude inclusion of steinhcimensis in Clemmydopsis and that, therefore, if the character of the horny shields was counted of much value, a new genus would need to be erected for steinheimensis . He concluded: "Vermutlich diirfte daher wohl auch eine nahere .Verwandschaft zwischen CJrmmys und Clemmydopsis bestehen, derart, dasz der Gestalt der Neuralplatten (kurzeste Seiten hinten oder vorn gelegen) keine so iiberragende systematische Bedeutung zukommen, kann, wie dies Boda nach Siebenrock annimmt. Das Fehlen der 1. und 2. Lateralschilder ist nicht als Genus — sondern nur als Artcha- rakter zu bewerten. Dieses Merkmal tritt bei zwei verschiedenen (Neuralplatten !) aber wohl verwandten Gattungen an der Wende von Miocan und Pliocan auf, um alsbald wieder zu verschwinden. Man konnte versucht sein in Dacque'schem Sinne von einer IModestromung zu sprechen, denn ein besonderer Zweck diirfte dieser Einrichtung kaum zugrunde liegen. Mit der Annahme einer naheren Verwand schaft beider Gattungen kann man aber auf diese Deutung verzichten." In 1934 T. Szalai in a list of the fossil turtles of Hungary synony- mized Clemmydopsis Boda with Geoemyda Gray on the basis of the similar neural shapes, rejecting as not significant at the generic level the peculiarities of the horny shields. Erich Thenius (1952) has followed Szalal's generic assignment (pre- ferring, however, the synonymous name Nicoria) in reporting new finds of sopronensis from a new locality Brunn-Vosendorf near ^'ienna. Thenius' material is important in that it completes our knowledge of sopronensis by providing (from dissociated specimens of all the plastral paits) the characters of the plastron, heretofore unknown. Thanks to the kindness of the authorities of the Staatliche IVIuseum fiir Naturkiinde in Stuttgart and especially to the friendly cooperation and diligence of Dr. Karl Dietrich Adam of that museum I have been able to examine the type specimens of Clemmys sieinheimensis Staesche. There is little to add to Staesche's accurate description and excellent 1954 CLEMMYDOPSIS 5 photographs. There are, however, some points of interest in regard to the interpretation and systematic position of the fossils. I wish first to suggest that steinheimemis may best be synonymized with turnauensis. The two named forms are equivalent in age, not far distant in locahty and distinguished by a single quite trivia] character which may well be only an individual peculiarity of the unique type of turnauensis. If this synonymy is correct we are dealing then with four occurrences of only two forms. Yet these two forms have received from the very few authors who have written about them four different generic as- signments: "Emys", "Clemviydopsis, new genus", ''Clemmys", and "Geoemyda ( = Nicoria)". Of these generic names the first may be dismissed at once; it belongs to a period in which almost all fossil emydine species and even some forms not belonging to the family Testudinidae or the suborder (ryptodira were placed in the genus Emys. More serious discussion must be accorded the other generic as- signments, but it appears to me that previous authors have failed to consider one important possibility: that the two forms, which are after all chronologically consecutive, are phyletically related. The absence of the first two costal scutes in the two forms turnau- ensis and soprotiensis is a quite extraordinary phenomenon apparently not closely approached by any recoided aberration of the horny shields. The similarities in detail shown by the two forms are fan- tastically close if they are the result of parallelism only. There is a real difficulty in the plural occurrence of so improbable an event. The difficulty, however, ceases to exist if we assume the event happened only once and if we explain the similarity of the two forms by direct inheritance. The evolutionary and therefore the taxonomic dilemma which the two forms have appealed to present is in all essentials solved by the hypothesis of phyletic relationship. There are, of course, real differences between the two species, and, in fact, one of the differences — neural shape — has been considered of high taxonomic value and is still used in the taxonomy of Recent forms to discriminate genera. Without question in the Recent emydine turtles the character of neural shape tends to have utility in segregating natural groups, although not without some instances of difficulty, but, granting to the character the maximum of utility for Recent forms it must still be used with discretion when dealing with fossil forms. It can never be 6 NO. 28 forgotten that neurals which are hexagonal, short-sided in front, are primitive and that other types have evolved from that condition. Inevitably then some of the ancestors of genera showing the modified types of neurals would be placed — if that character were alone con- sidered — in a more primitive genus. Quite certainly Geoemyda and its relatives have evolved from a form with Clemm.i/s type neurals. There are indeed a number of examples of evolution in neural shape displayed in the fossil record. The fossil tortoise Stylcmys nebrascensis of the Oligocene cf North America regularly has all the neurals after the first hexagonal short-sided in front : the more highly evolved species of the same genus from the John Day beds (lowest jMiocene) regularly have one or more octagonal neurals. Hadrianus with primitive neurals evolved into later subgenera of Testudo with modified neurals. There is also among Recent forms considerable intraspecific variation in neural shape. In the present instance turnaunisis is temporally antecedent to sopronensis and in the matter of neural shape structurally more primi- tive. It is surely a possibility worth attention that the one has a phyletic relationship to the other. Except for neural shape there are no known differences between turnauensis and soproncnsis which are not at the specific level or below it, and, since none of fossils is quite perfect, it is possible in fact that some of the recorded differences are matters of interpretation and reconstruction and not real. I tabulate the differences below: turnauensis 1. gulars narrow, not overlapping humerals 2. femoral scutes longer than pectorals 3. no anal notch 4. first vertebral encroaching on at least the first, second and third peripherals 5. first neural rectangular-oval, neurals 2-8 hexagonal, short- sided in front, all moderately broad 6. fifth vertebral and fourth costal moderate in size soipronensis gulars broad, overlapping hu- merals pectoral scutes longer than femorals a distinct though nan-ow anal notch first vertebral encroaching only on the first and second peri- pherals neurals 1-3 hexagonal, short - sided behind, neural 4 quadri- lateral, neurals .5-8 hexagonal, short-sided in front, all still broader than in turnauensis fifth vertebral and fourth costal quite small relatively to the other scutes 1954 CLEMMYDOPSIS 7 These differences will serve — even if one or two should fail — to distinguish the species. Now, granting that the two species are directly related, to what genus or to which genera should they be referred? Admittedly, if the genus Clcmmiidopsis is recognized, it will be solely on the pattern of the vertebrals and costals. It is, of course, possible that some of the skeletal parts which are not known — skulls, limbs, or vertebrae — might show striking differences from the related forms, Geoemyda or Clcmmys, but this is an untestable hypothesis, which it is useless to consider. The question is then a simple one: Is a single charactei at once striking and superficial — adequate for generic distinction? There is unfortunately no objective way to evaluate such a question. Characters involving quite radical differences are in some cases infra- specific, while other characters which to the non-specialist are utterly trivial separate suprageneric groups. That the rank of supraspecific categories is wholly subjective or nearly so is a point generally agreed upon. In practice it is necessary to steer a middle course between those who would reduce genera arbitrarily on mnemonic grounds — deploring the necessity of remembering so many names — and those who tend to regard as generic any differences which are recognizable immediately and as specific any differences which are determinable with more difficulty. In the present case there are several grounds for accepting — at least for the present — Clemviiidoysis as a valid genus. 1. The peculiar modification of the horny scutes, though a single character, is also one without parallel in the normal characters or the known variants of the chelonian order. Anomalies of the plates and scutes of turtles have been studied intensively by Gadow. Newman. Coker, Grant, Lynn and others using literally many hundreds of specimens, yet nowhere does there appear to be any record of a scute anomaly resembling that seen in these forms from the Miocene and Pliocene of central Europe. The closest parallel is with two cases cited by Grant, one in Tcstudo dcnticulata and one in Chri/seinys pida in which the vertebrals do touch the marginals. But even here the similarity is verbal only, the marginals being enlarged and not the vertebrals while the grossest asymmetries and distortions indicate the abnormality of the condition. 2. There is currently recognized a Recent genus Notochdys with the single species A. plati/nofa which differs from Clcmmys primarily by 8 BREVIORA NO. 28 the intercalation of a small scute between the usual fourth and fifth vertebrals. But this condition, while a population character in pla- tynota, is a rather common variant in numbers of other members of the Testudinidae. This variation is especially common in Homopus in which a very substantial fraction of at least H. areolatus shows a similar condition Thus while there is a weak argument for Clcmmy- dopsis on the grounds of symmetry in generic discrimination, it may in this case be countered by a demand for suppression of the generic separation for N. platynota, which, while certainly a distinct species, does not obviously merit more than subgeneric distinction. 3. However, the best ground for separating Clemmydopsis as a genus is the impossibility of placing the two central European forms, directly related though they seem to be, in any one currently recog- nized genus. On the current definitions of emydine genera the Miocene form belongs in Clcmmy.'i, where Staesche has already placed it, and the Pliocene form in Gcoemyda ( = Nicoria) to which Szalai and Thenius have already referred it. This, however, is an inadmissible solution since it would make the genus Geocmyda polyphyletic. It is entirely possible, indeed probable, that Clemmys and Gcoeiayda as at present defined are mere form genera, but the true relationships within the Clcmmys-Geocmyda section of the Emydinae are still to be analyzed and promise to be difficult of analysis. Thus while it is perhaps possible to look forward to a time when relationships will become known, and larger genera may be recognizable within the Emydinae, it is not now desirable to randomly unite emydine genera, in so doing pretending to a knowledge which is not at hand. Rather it is preferable to split rather finely at the moment, counting the present as that period of analysis which must precede a synthesis. We need first to discover the correct phyletic sequences. We must first recognize the family tree; afterward we may dispute the nomen- clatorial rank of the several twigs and branches. Clemmydopsis Boda should then be retained at present for an ap- parently short-lived but uniquely specialized lineage of emydine turtles from the later Tertiary of middle Europe. Whether or not it should ultimately be retained as a genus in formal nomenclature, it is certainly a "gens" in the sense of Vaughan 1905 (in Cain 1952) (a lineage or phyletic line). Examination of the type of Clemmys steinheimensis in Stuttgart was made possible by the support afPorded by a Guggenheim fellowship 1952-1953. 1954 CLEMMYDOPSIS 9 REFERENCES BODA, A. 1927. Clemmt/dopsis sopronensis, n.g., n.sp. aus der uuteren pannonischen Stufe von Sopron in Ungarn. Zentralbl. Min. etc., Abt. B, 1927, pp. 375-3S3 Cain, A. 1952. Geography, ecology and coexistence in relation to the biological definition of species. Evolution, vol. 7, pp. 7«i-S3. COKER, R. 1910. Diversity in the scutes of Chelonia. ,Iour. Morph., vol. 21, pp. 1-7.5. Gadow, H. 1S99. Orthogenetic variations in the shells of Chelonia. in Arthur Wiley's Zoological Results, part 3, pp. 207-222. Grant, C. 1936. An extraordinary tortoise shell. Copeia, 193'o, p]). 231-232. Lynn, W G. 1937. Variation in scutes and plates in the box turtle, Terrapene rnrolina. .\mer. Nat., vol. 71, pp. 421-426. Meyer H. von 1847. (Letter on various fossils.) Neues .Jahrb. f. Min. etc., 1S47, pp. 181-196. 1858. Schildkroten und Saugetiere aus der Braunkohle von Turnau in Steyermark. Palaeontographica, vol. 6, pp. 50-55. Staesche, K. 1931. Die Schildkroten des Steinheimer Beckens. A. Testudinidae. Palaeontographica Suppl., vol. 8, part 2, pp. 1-17. Szalai, T. 1934. Die fossilen Schildkroten Ungarns. Folia Zool. Hydrobiol., vol. 6, pp. 97-142. Thexius, E. 1952. Die Schildkroten (Testudinata) aus der Unterpliocan von Brunn- Vo.sendorf bei Wien. Neues Jahrb. f. Min. etc., 1952, pp. 318-334. E V I O R A Museeirai of Coeiparative Zoology Cambridge, Mass. February 8, 1954 Number 29 ABSENC^E OF MESOPLASTRA IN A PELOMEDUSA (TESTUDIXES, PELOMEDUSIDAE) Hy Ernest Williams Mesopla.stra (a pair of l)ones intercalated between liyu- and hypo- plastra) are believed to be primitive features in turtles. They were often present in Jurassic and Cretaceous genera, and two pairs of inesoplastra were reported by Fraas (1913) in Profrrochcrsis, one of the Triassic genera. Mesoplastra have, however, been lost se\eral times independently: ( 1 ) in the Jurassic and Cretaceous plesiochelyids and thalassemydids; (2) in Wealden Chiiracrphahis\ (3) in all crypto- dires; (4) in the chelids among pleurodires. In many of the groups in which they are known, they were early reduced. Primitively they extend quite across the plastron and meet centrally. In some Pleuro- sternidae, however — Plafi/chch/s — and in some Baenida.e (within the genus Bai'na, cf. Bai'iut rijxirid. Hay, 19(),Sj they fail to meet in the center. To my knowledge, howe\er, no instance has been recorded previ- ously of the complete absence of mesoplastra in any form in which they were typically present. Such an example is now furnished by a specimen of Pclomcdusa suhrufa from I'ganda, collected by Delme Radcliffe and now 1905-5-19-1 in the collections of the Reptile Section, British Museum (Natural History). In Pclumcdtisa the mesoplastra are typically small and lateral. In HM 1905-5-19-1, however, no mesoplastra are visible, and instead the hyo- and hypo-plastra join for their full transverse extent and do so quite symmetrically on the two sides just as in the C'helidae or the Cryptodira. The carapace again joins the plastron perfectly without aft'ording any place for a trace or rudiment of mesoplastra. 2 BREVIORA NO. 29 Yet this specimen is clearly referable on "habituellen ]\Ierkmalen" to the family Pelomedusidae, the genus FcJoiucdnsa, the species suhrufa (the single species of the genus). The carapace is quite normal except for some minor asynnnetries and the fusion of vertel)rals 3 and 4 to a single scute. The nuchal scute is, as usual, absent. The first neural is rather characteristically tapered in front. In normal fashion there are 7 neurals, and the posterior pleurals meet behind the neurals separating them from the suprapygal. In contrast to chelids the posterior peripherals are all rather narrow and show no suggestion of posterolateral expansion. The plastron is quite typical of the genus in the size and relations of its horny scutes. There is the usual median fontanelle, though it is small in this instance, as seems frec^uently to be the case in northern representatives of the genus and species. The skull is quite characteristically pelomedusid and very unchelid in the temporal emargination from behind and in the marked pro- jection of the opisthotic posterior to the squamosal. All these features and the lack of mesoplastra are well siiown in the excellent photographs (Plates 1 and 2) taken by Peter Green of the photographic staff of the British Museum (Natural History) and reproduced by permission of the Trustees of that institution. The absence of mesoplastra is quite certainly an individual variation. Four specimens from Mt. P^lgon, I'ganda, otherwise \ery similar to the individual without mesoplastra, all show these elements well- developed and only slightly varying in size. All resemble the aberrant specimen in the reduction of the median fontanelle f)f the plastron; one, indeed, has the fontanelle completely closed. The presence of mesoplastra is supposedly a family character of the Pelomedusidae, but except in the genus Prlu,s-ios the mesoplastra are always small and lateral, in effect vestigial, and the disappearance by individual variation of a \estigial feature is not too surprising. It is, howe^•er, of special interest in connection with the suggestion recently made by me that the genus Pelusloft with well-developed mesoplastra (meeting medially) has been derived directly from the genus Priomcdus'a with reduced mesoplastra by a secondary expansion of these elements. If this suggestion be \alid, then, taken in connection with the present case, we see in Pclonialusd an interesting ambivalence of evolutionary potentiality, one type of \ariation reversing a previ- ously well-defined trend to restore (with some differences) an ancestral condition and, on the other hand, the o])posite type of variation 19r)4 ABSF.XCK OF MESOPLASTKA I\ A PP:LO!MKDrSA iMjiurc 1. Plastni of viirious lurtlcs to show |)rcsciicc or ;il)sciicc of nicsoi,l;isti;i. M(>so|)l;is1 rn shaded. A. I'rlu»iv,l u^ii witliout nicsoplastra. H. \oYm-A\l',l(,m,'ditsa. ('. a v\n'\yu\ {II ijilmwrihim). D. f>,h,sios. E. Po- dornemis. {W to [• alter Houlcnii'T.^ 4 BRKVIORA , NO. 29 carrying to completion the former trend and realizing a more modern- ized type of shell. It does not seem advisable on the basis of a single, probably rare, variation to revise the definition of the family Pelomedusidae to include forms without mesoplastra. Such a variation would appear — on the face of the present evidence — to have as little taxonomic weight as the occasional men ])orn without legs have for the definition of the family Hominidae. Yet there is this difference: a man without legs is not likely to be a selecti\'ely valuable variant, but, as the majority of the living forms testify, a turtle without mesoplastra is not only perfectly A'iable but highly successful. It cannot be assumed, therefore — given a forward- looking evolutionary point of view — that the pelomedusids will always and in all cases be forms possessing mesoplastra. There is indeed a special caution pertinent for paleontologists. If the present specimen hafl been found as a fossil, it would almost certainly have been misallocated as to family. There is in fact no assurance — the pelomedusids having once been world-wide — that all past species have always had mesoplastra. The use of key charac- ters in identification without regard to total habitus and neglecting the ever surprising power of animals to vary and to undergo evo- lutionary change will only lead to error. The obser\'ations recorded here were made during a \isit to the European collections, made possible by a Fellowship from the Guggen- heim Founflation. Access to the specimens mentioned was generously granted by Dr. H. W. Parker, Keeper of Zoology. British Museum (Natural History), and assistance was given by ]\Ir. J. C. Battersby and Miss A. G. G. Grandison of the Reptile Section. hp:ferexcp:s Fraas, E. 1913. I'luteruchersis, eine pleurodire Schildkrote aut; dein Keuper. Jahreshefte Ver. Vaterlands Wiirtt. Xatuikunde, Stuttgart, 191.3, pp. 13-30. Hay, O. p. 1!)0S. Fo.«sil turtles of Xorth America. Carnegie Inst. Washington Puhl. X'o. 75, pp. 1-.5()S. LoVERIDfiE, A. 1941. Revision of the American terrapins (jf tlie family Pelomedusidae. Bull. Mus. Comp. Zool., vol. 8S, pp. 4(57-524. PLATE 1 Pelomedum s„hn,fa H.M. X.H. Xo. IWa-o-lO-l. IiitiTiial view of jilastmn. PLATE Pelomcilusa snhnifa H.M.X.H. Xo. 19()r)-r)-l<»-l. Dorsal vi<'\v o\' skull and shell. E V I O R A Musceiim of Coimpsirsitive Zoology Cambridgp:, Mass. I"i;iii;rAKV S, 1954 XrMBp:R :>() AKSTIVATIOX I\ A PERAIIAX LUXCJFISH Hy Alfred S. Romeh Musi'uiii of ConipMr.-itivc Zooloj^y, Harvard riiivcr-sitj' and Everett (\ Olson I'liivcrsity of Chicaji-o It is generally recognized that the survixal to modern times of the three living genera of dipnoan fishes — K/jicirdtix/us i Xconrdfodii.s) of Australia. Profa/ifrriis of Africa and Lcpido.sirm of South America, - is due in great measure to adaptations which tide them oxer the seasonal di'oughts characteristic of the tropical regions which they inhabit. .\ primary adaptation lies in their pos.se.s.sion of functional lungs which enable tliem to utilize atmospheric oxygen, supplementing the refluced oxygen content in stagnant waters during drought; indeefl, in Proiopirrus, the hmg-hreathing fimction has attained such domi- nance that the fi.sh is unable to sm-\i\e if air-breathing is not possible. But a.ir-l)reathing in itself is not entirely satisfactory as a means of drought sm-\i\al - particularly if, as may often happen, the stream or pond dries up entirely. The Australian lungfish, under such circum- stances, may wallow into the mud, and have some chance of escaping desiccation or enemies. In the African and .South .Vmerican forms, an additional adaptation is present the de\elopment of a technique of aestivation. Meml)ers of these genera are rather elongate and eel- like in shape, in contrast to I^piccrafndiis and to primitive lungfishes. On the approach of the dry season the fish burrows in coiled form into the mud, which hardens atiout it to form a cocoon; an opening from the surface allows the entrance of air, but the amount neerled is small, due to a great reduction of metabolic processes. The fish passes mto a "summer sleep" compara!)le in great measure to the hibernation of various northern animals — a phase which ends with the i-eturn of the water and tiie resumption of normal a(|uatic life. 2 BREVIORA NO. 30 Nothing has been known until now of the origin of lungfish aesti- vation. The early lungfishes, as far as known, all had a fusiform body shape fairly comparable to that of the modern Kplccratodu.s', and hence were incapable of coiling in a burrow in the fashion of the elongate- bodied Lcpidosiren or Protoptervs; but it is possible that burrowing of some sort accompanied by phsyiological features of aestivation, might have preceded change in body shape. The lungfish (I'nathorhiza, characteristic of the Clear F'ork Permian of Te.xas, has been suspected by both the present writers to be related to the Lepidosirrn-Protopterus group (Romer and Smith 1934, Olson 1951) but there is no strong proof of relationship, and the body structure was unknown. Some years ago Dr. H. J. Sawin and Mr. Adolph Witte, while engaged in work for the Texas Bureau of P>onomic Geology were told by a resident of Willbarger County, of a geologic occurrence which local geologists had been unable to interpret. They visited the locality, made collections of the material for the Bureau and informed others, including the present writers, of the incident. In 1952 Mr. Witte took a Harvard field party, including the senior author, to the site; obser- vations and collections were made at that time and on a return visit the following spring. The locality lies in southern Willbarger ( "ounty, Texas, on the Reed Ranch, in Section 34, Block 3, H. & T. C. RR., Co. Survey. The area has not been mapped in detail geologically but the horizon appears to be a short distance alcove the Lueders limestones and hence in the lower part of the Arroyo Formation of the ( lear F'ork Group, Lower Permian. Gently sloping exposures here run for a hundred yards or so close to the south bank of Minnie's Creek. Between two thin impure limestones lies a bed of red shale. Scattered o\er the surface are numerous circular discs, somewhat like large checkers, with diameters which are for the most part between 5.5 and 10 cm. The material of the discs is essentially the same as that of the shale bed from which they are derived, but is slightly harder and somewhat more calcareous, so that they tend to weather out of their enclosing matrix. On exca- vation, these discs are seen to be segments cleaved from vertical cylinders which are thickly spread through the shale, often only a few inches apart. In no case was a complete cylinder seen; the greatest length preserved in a specimen excavated and collected was 2S cm. The cylinders, in general, weather out from the softer shales with a smooth surface; in some instances portions of the surface show slicken- sides. No upper terminal portions were found; apparently, at the top, 1954 AKSTIVATlOX 1\ A PEKMIA.N UNGFISH 3 the material of the cylinder Mends laterally with the superposed calcareous shale. The lower ends of the cylinders are generally smoothly rounded; in some cases, however, a lower terminal portion has a diameter rather less than the major portion of the "bore". The great majority of the "checkers" fractured off from the cylinders show no organic material on their surfaces. A fair number, however, show fish remains. In some instances these appear to be scales and other disarticulated elements of small palaeoniscoids Most of the remains, however, are clearly dipnoan in nature. The hardness of the shale and the relatively soft nature of the bone makes preparation difficult, and rehance has been had mainly on the structures seen in section on the disc surfaces or, in some cases, weathering of the external surface of the cylinders. Bony scales and plates are frecjuently seen in section and in a few specimens fracture has shown scales in surface view, the largest about 20 mm. in greatest diameter. Numerous discs show hollow ribs, neural or haemal arches and fin supports, sometimes in articulated series. Preparation of the basal segment of one of the cylinders (figured) has revealed a considerable portion of a caudal fin of the sort typical of Paleozoic dipnoans. Vertebral arches and fin elements are present, partly as bone, partly preserved as impressions. The vertebral a.xis, presumably cartilaginous in life, is represented by a vertically oriented band devoid of impressions. The end of the tail, poorly preserved, was curved around the bottom of the cylinder. While much of the lungfish material suggests a considerable degree of decay and maceration, it seems certain that in many instances we are dealing with cross-sections of complete, or nearly complete lung- fish bodies. In a few instances where "articulating" discs composing the greater part of a cylinder have been collected, it is obvious that we are dealing with successive sections of a single fish, larger, frequently rounded, body sections in the upper and middle parts of the series and caudal sections, typically lens shaped, at the lower end. In many cases the internal skeletal materials are rather irregularly scattered across the surface of the section; in others the arrangement of the ribs and fin supports is essentially regular, and toward the disc margins there can l)e made out body outlines, markefl by scale sections with a. surrounding layer of darker matrix. The only lungfishes known from the ( 'lear Fork belong to tlie genus Gnathorhiza, mentioned above; Sagcuodus, the common Pennsylvanian genus, possibly antecedent to (\'rato(lu.s' and Epiccrntodus, is not 4 BREVIOKA NO. 30 recorded in Texas abo\'e the ^^'ichita group. But while it is probable that the Reed Ranch lungfish is Guathorhiza, no tooth plates have been found there to make identification certain. A second fintl, however, lias produced a number of fragmentary teeth. These are unmistakably the blade-like, shearing teeth of (inafhorhiza and are referable to the species G. dikrloda rather than G. .si rrntfi. This locality was discovered by the junior author and Dr. Nicholas Hotton III in 1949 and has been levisited several times since then. The bed in which the cylinders occur is located in north central Knox ( ounty, Texas, on the Waggoner ranch. It crops out along a small tributary to Little Mustang ("reek in locality KI of the junior author (see Olson, 1951 p. 104). The age is middle ^'ale of the Clear P^'ork Group, Lower Permian. The site is approximately 700 to 800 feet higher in the Clear Fork section than the Reed Ranch locality. The physical features of the cylinders from the two sites differ in no important particulars. Those from the ^'ale locality occur in a deep red shale which ^■aries in thickness from two to three feet. Above and below this bed are lighter colored shales that appear to have been deposited under somewhat different conditions. The shale is best exposed along a small arroyo, but crops out sporadically over an area of al)out an acre. The cylinders are distributed in rather irregular groups of ten to twenty in unevenly spaced areas not much more than a yard scjuare. The longest cylinder encountered measured 45 cm., hut as in the case of the Arroyo specimens no upper termination has l)een identified. The vertical orientation shown in the specimen figured is constant throughout the deposit. Organic remains in the cylinders from the two localities dift'er somewhat. The most common remains in the \'ale (ylinders are ribs, skull plates and scales of Giiaihurhizn but, although the ribs in some instances suggest the general body shape of the fish, no case in which the scales show the body outline has been observed. As at the Reed Ranch locality, remains of lungfish are found only in a fraction of the cylinders. Scales of small palaeoniscids and infrequent scraps of skulls of some small captorhinid reptile occur in others. ]\Iore abundant, howevei-, are partial vertebral cohunns aiul single \ertel)rae of the small, worm-like amphibian, Li/.soroj/hiis. In no case have these columns been found in the coiled conflition characteristic of Ijisonipluis specimens throughout the Arroyo formation. There can hardly be any question that the lungfishes from the two localities l)elong to the same genus. Some question does exist, how- 1954 AKSTIVATION IN A PERMIAN LUNGFISH 5 ever, with respect to specific identity. Two species, (1. .scrrata and (>'. dikcludn, are known from the Vale formation, l)nt only the first has been fonnd in the Arroyo (Olson, 1951). Tooth plates of (1. .srrratu are small, lower plates ranging from 10 to 12 nnn. in length, whereas those of G. dikcludd are between 25 and 32 mm. long. The skulls of adult individuals of (1. dikcloda appear to have been between 75 and 100 mm. long; this would suggest that the skulls of (1. terrain, unknown as yet, would have a range of between 25 and 35 nun. The fact that the Arroyo and \'ale cylinders are comparable in size indicates that they were made by fish of about the same dimensions. There is, however, the puzzling fact that no teeth of the larger species have been found in the Arroyo in spite of very extensi\e exploration of its exposures. Except at the Reed Ranch locality lungfish remains ai-e exceedingly rare in the Arroyo; only four or five occurrences have been reported, and these have consisted of single teeth. It may well be that (j. dikcloda was in existence during the Arroyo times, but that it found suitable habitat in few localities in the areas which are now exposed. With increasing aridity during the \'ale, the number of habitats and the chance of discovery probably were greatly increased. From the data given above, the nature of the phenomena described will be as obvious to the reader as to the authors. We are here surely dealing with evidence of aestivation of lungfish at an early. Paleozoic stage in the history of the Dipnoi. The cylinders are aestivating burrows, made by lungfishes in the muddy bottoms of drying lakes or streams. That most of the cylinders are empty is to be expected; most of the fishes which occupied them sur\ived to the return of the waters and swam away. The burrows were later filled by sediments fairly similar to those which had already formed the mud banks into which they had burrowed. P'ortunately, however, the filling material at the localities described was of a somewhat harder, more calcareous, nature than that of the original mud; had they been just the same, differential weathering would not have occurred and the burrows would prol)ably have escaped observation. The lack of definite upper "openings" of the burrows is presumably due to the disturbance of surface materials on the return of the water. The presence of lungfish remains in some of the cylinders indicates that a certain percentage of the aesti\ating population failed to survive until the return of the waters. The finding of occasional actinopterygian remains in the burrows is due to the presence of palaeoniscid minnows in the returning waters before the burrows had been filled l)y sediments. This probably applies to the 6 BREVIORA NO. 30 remains of Li/.s-orophus as well. This amphibian, like (riiatJiorhiza, had both aestivating and free swimming phases. It characteristically was coiled during aestivation. The fact that coiling has not been observed in the burrows indicates that Lysorophus entered the open holes left by the lungfish upon the return of the water, and died during the non-aestivating phase of their existence. The finding of Gnaihorhiza teeth in the material from the \ ale (KI) locality, together with the absence of identifiable remains of any other lungfish type in the Clear P^ork beds makes it seemingly certain that the burrows were made by thi?t lungfish and that alone. This evidence of the adoption of aestivating hal)its by Gnathorhiza strengthens the suggestion, gained from tooth plates and skull pattern, that this genus is possibly ancestral to the modern aestivators, Lcpidosiren and Protoplcrus. One major difi'erence in the mode of aestivation between modern forms and Gnathorhiza may, however, be noted. The Recent genera are elongate, eel-like animals which coil their bodies in a flask- shaped burrow. There has been no previous evidence of body shape in Gnathorhiza. The Permiam liurrows may be somewhat narrower in their upper portions; they are, however, essentially cylinders rather than "flasks", and in no case is there evidence of more than one section through a fish body at a given level of a buri-ow It thus seems certain that the Gnathorhiza body was of a normal, non-elongate fusiform shape; that the burrow was formed by a spiral rotation of the body and tail with the fish's long axis in a vertical position; and that the aestivating fish was posed erect in the burrow, essentially "standing on its tail" with the mouth at the top, close to the surface where air was available. EXPLANATION OF PLATP] Above, a sciich; of cxpos^ed burrows at the Reed Ranch locality; a half-dollar gives ail indication of size. Cciitrr, a terminal section of a burrow developed to show an included lunglish tail, its tip curved upward. The skeleton is partly preserved in bone, partly as impressions which have been painted white. At the left, a lateral view of the specimen; at the right, a view from the under- side. Beloir, left, a burrow in place at the KI locality; length 45 cm. Belniv, right, an incomjilctc !)Uii()w from the Reed Ranch locality; length 25 cm. 8 BREVIORA NO. 30 The genus (inathorhiza is known not only from the Texas beds, l)ut also from the late Pennsylvanian of Illinois. In Texas, however, it is very rare in the Wichita formations of the Permian where Sagrnodus, common in the Pennsylvanian, is fairly abvmdant: on the other hand, there is no trace of Sagcnodus in the ("le.ir Fork, where Gnathorhiza replaces it. This local succession is in agreement with our beliefs as to the climate of early Texas. Various lines of evidence suggest that Clear Fork time, particularly that of deposition of the Vale and Choza of that group, was one of increasing aridity with marked seasonal variations in rainfall (cf. for example Olson, 1948). Accompanying this change is a shift in the local dipnoan fauna from Sagcnodufi, a presumed non-aestivating form, to (Inathorhiza, ))etter adapted for survival under drought conditions. LITEH.\TrRK CITED Oi.sox, K. C. HUS. A pr.'liiniiuu y report on vcrtclirates from t\v' Permian Vale formation of Texas. Jour. Geo!., 56: 1S6-198. 11).") 1. Fauna of Upijer Vale and Choza. 3. Lunj; fish of the Vale 4. The skull of Gmthnrhiza dikeloda Olson. Fieltliana, Geol., 10: 104-124. Ro.MER, A. S. and H. .1. S^^TH 1934. American Carboniferous dipnoans. .Jour. Geol., 42: 700-719. B R E V I Miaseiiei of Coimparative Zoology Cambridge, Mass. Makc h o. 1954 Ximber 31 NEW FRESHWATER ClASTROPOD AIOLLUSKS OF THE AFRICWX GEXUS LAXISTES By T. Pain London, England The two new species and one subspecies flescril)ed herein form jxirT of the collection of the Museum of Comparative Zoology at Har\ard College. I am deeply indebted to Dr. J. C. Bequaert and Dr. A\'. I. Clench for the opportunity to examine and describe them. La.\i.stp;s (LAXISTES) KOBELTi, new species Figures 3-4 l!»lo. Lanistes .«p.? Kobelt, Mon. Amp. in Martini and Chemnitz, Svst. Conch. Cab.. 1, Aht. 20, Xeue Folge, pi. 30, Hgs. 4-4a. Shell sinistral, fairly thin. Whorls strongly shouldered below the suture, with a very slight ridge. Spire produced, apex much corroded. Aperture ovate; parietal lip simple and only thinly calloused; colu- mella thin, not reflected. I'mbilicus very wide and deep, limited behind by a very prominent, angular, thickened ridge, which forms a slight angle at the base of the moderately concave columella. Growth- lines irregular, interspaced with fine axial striae, and cut by very fine, irregular, close-set spiral engraved lines in the holotype (very super- ficial in the paratype). Color pale yellowish-brown, with numerous irregularly spaced reddish-brown bands; interior of aperture pale bluish-white banded with reddish-brown. Operculum corneous, of the usual shape. Measurements (in millimeters) Length Width Aperture Length Width Holotype 23 24 15 11 Fig. 3 Paratype 23 23 1.5 11 Fig. 4 IS 23 15 10 2 BKEVIOKA NO. 31 Localiiy. "Africa" (probably West Africa); holotype M.C.Z. No. 173419; paratype.s M.C.Z. No.'l73420 and in T. Pain ("oil. This unusual little species was apparently figured by Kobelt (1913, pi. 30), but he did not name it or mention it in the te.xt. The umbilical keel is very prominent, more so than in L. (L.) congicus Boettger. The spire is also higher and the suture less flattened than in that species. L. (L.) kohrUi is, as the above table shows, very constant in size and subject to but little variation in this respect. LaiMstes (MELADOMis) coNNOLLYi, new species P'igures 12 Shell sinistral, perforate, fairly solid. Color dark reddish-brown, without external bands. Whorls noticeably flattened and obtusely shouldered above, rounded below, with a blunt angulation around the umbilicus which is deep and moderately wide. Apex entirely corroded, remaining whorls 3. Surface somewhat malleated, with flne and regular growth-lines; spiral sculpture obsolete. Aperture ovate vertical; lip simple, thin, edged with pale blue. Columella thin, evenly curved, not angular at the base. Parietal lip bluish-white and only thinly calloused. Color dark reddish-brown, without external bands; interior of the aperture marked with faint reddish-brown lines. Operculum unknown. Measurements (in millimeters) Length Wi.lt li Ape Length rtu Width Holotype 40 36 31 22 Fig. 2 Paratype 45 40 32.5 26 Fig. 1 " 43 39 30.5 24.5 " 55 50 38.5 30 " 45 44 33 27.5 LocaUtivs. Southern Rhodesia: Mctoria P'alls, holotype ex Connolly Coll., M.C.Z. No. 173421, and paratypes, M.C.Z. Nos. 173422 and 179845; also at Congo Museum, Tervuren, and in T. Pain Coll. Northern Rhodesia: River Kafue at N'kana, near the Congo border, paratype, collected by C. von Hirschberg, M.C.Z. No. 173423. Additional specimens of this species, referred to Lanistes ellipticus von Martens by the late ^lajor Connolly, are from the mouth of the Chobe River, Southern Rhodesia (Holub) and from the Kafue River, Mumbwa District, Northern Rhodesia (Pitman). (See Connolly, M., 1954 NEW FRKSHWATEK (JASTKOPOD MOLLUSKM 3 1938, Ann. Soutfi African Mus., 33, p. 556.) L. connollyi is easily confused with L. ellipticus v. Martens, which occurs in the same general territory. It differs in its less upright elliptical form, decidedly shouldered whorls more or less flattened below the sutures, the deeper and wider umbilicus and the pronounced blunt angulation around the umbilicus. Lanistks (MKi-ADoMisi (j\ I M MWERiEXisis, new subspccies Figures 5-7 Shell small, fairly thick. ( 'olor uniform greenish-brown, without trace of bands. The spire is much corroded; remaining whorls 3, rapidly increasing, the last considerably inflated. Sutures impressed, but little flattened. Umbilicus narrow but deep. Surface with fairly prominent growth lines, interspaced with fine wavy striae, especially on the body-whorl. Aperture ovate, lip simple, interior pale blue. Columella thin, curved, joined by a moderately thick callus of !)luish- ite color. ()j)crcului 1 a s in typical L. (M.) ovum Me isu rements in millimeters) Lrlljrtli Width Apertu! Length Width ^olotype 30 30 19.5 16 Fig. 5 Para type 30 32 21 15 Fig. (i " 'ifi 25 19 13.5 Fig. 7 Localities. Belgian Congo: Lake Mweru (or Moero) at Kilwa, holotype, collected l>y Dr. L. van den Berghe, M.C.Z. No. 109842, and paratypes, M.( '.Z. No. 185683; also at Congo Museum, Tervuren and in T. Pain Coll.; Lake Mweru at Katenge near Kilwa, paratypes, M.C.Z. No. 152483 (received from Dr. ^L Wanson); Lake Mweru at Pweto, paratypes, M.( ".Z. No. 152346 (received from Dr. ^L Wanson) ; Lake Mweru at Moba (Chamfubu River) near Pweto, ^LC.Z. Xo. 185682 (received from Dr. M. Wanson). Only the holotype (Fig. 5) and one of the paratypes (Fig. 6), l)oth from Kilwa, appear to be full-grown. The paratype in Figure 7, also from Kilwa, is immature. Most of the other paratypes are even younger, the smallest being only 9 mm. in length Apparently a dwarf race of L. (M.) ovum Peters, possibly ccMifined to Lake Mweru. When adult it resembles the much larger L. (M.) ovum procerus v. Martens, but has always a much thicker and more solid shell. Immature shells are always of a unifS. coxalis (Roger), S. costata Emery, S. hiroi Emery. B. Second postpetiolar (IV abdominal) segment not costate. 1. Ej'es placed anterior to middle of sides of head (ac- cording to original descriptions) : S. crassicornis Forel, S. spiralis Karawajew. 2. Eyes placed at or slightly behind the middle of the sides of the head: S. bicolor Emery, S. hinghamii Forel, 8. laevior Forel, 8. menadensis Mayr. In addition to the species listed above, with their synonyms, I have also seen two apparently undescribed species of group B,2. Stictoponera laevior Forel Ectatomma (Stioioponera) laevins Forel, 1905, Mitt. Naturli. Mus., Ham- burg, 22: 7, worker. Type locality: Tjibodas, Java. Stictoponera laevior var. avia Forel, 1912, Notes Mus. Leyden, 34: 96, all castes. Type locality: Nongkodjadjar, Java. New Synonymy, A small, slender shining species with sparse foveation. The species was described from a unique, obviously teneral ; var. avia refers to fully pigmented specimens. I have seen one worker from Poentjak, Java, no collector cited. Stictoponera menadensis Mayr Ectatomma (Stictoponera) menadensis Mayr, 1887, Verh. zool.-bot. Ges. Wien, 37: 539, nota, worker. Stictoponera stylaia Menozzi, 1925, Philippine Jour. Sci., 28: 440, pi. 1, figs. a, b, worker. New Synonymy. Stictoponera menaden,^is var. obscura Santschi, 1932, Mem. Mus. Hist. Nat. Belg., (hors serie) 4 (5) : 11, worker. New Synonymy. This species is of medium size, shining, with full adult color dark reddish-brown to piceous, gaster black. Mesonotum with a 1954 ANT GENUS STICTOPONERA MAYR 3 smooth longitudinal median strip free of foveation, extending in some specimens for a short distance onto the posterior portion of the pronotum; the latter otherwise always rather closely foveate. This species is apparently the common lowland and foothill Sticioponera in the East Indies and the Philippines. It occurs on Java and Sumatra together with 8. hicolor, a very closely related species, but no intergrades between the two have yet been reported from these localities. I have seen no specimens of menadensis from mainland Asian localities, and records from the mainland are scarce in the literature. From the present data, S. menadensis appears to be a peripherally distributed form in the process of being replaced by S. hicolor, the latter spreading from southeastern Asia outwards into the archipelagic regions. The types of S. menadensis and the var. obscura come from Celebes, and Santschi's description mentions no characters dif- ferentiating his variety from normal menadensis populations. The type of 8. stylata came from Mt. Makiling, Luzon, and is present in the collection of Dr. J. W. Chapman. Unfortunately, Dr. Chapman's collection was somewhat damaged during the period when it was hidden in the mountains of Negros while the owner was interned in prison camp during the late war, and some of the specimens of Sticioponera became separated from their labels, among them the stylata type. Dr. Chapman and I have, however, segregated a small group of specimens in the collection by elimination of possibilities, and some specimen of this group is the type, of stylata. Since all the specimens in this group are clearcut menadensis, and since menadensis is the com- mon species on Mt. Makiling, we concluded that the above syn- onymy is required. Menozzi's description offers no characters in opposition to this decision, and the description itself could be the basis of the synonymy. Obviously, Menozzi did not have a clear idea of the identity of the typical menadensis, and merely described it under a new name. In addition to the Mt. Makiling record, the collections of Dr. Chapman and the Museum of Comparative Zoology contain series from the following Philippine localities: Sarai, Paete, Laguna (Luzon) (R. C. McGregor). Mt. Manapla, Negros Occidental (F. del Rosario) ; Lake behind Dumaguete; Cuernos Mts., several collections, 1800 to about 3600 feet, Negros Oriental (J. W. Chapman). Momungan and 4 BREVIORA No. 34 Lanao, Mindanao (Chapman and D. Empeso). Los Banos, Luzon (F. X. Williams). Samar I. (McGregor et al). Kolam- bugan, Mindanao (C. S. Banks). There are also series from Borneo: Mt. Tibang, 1400 M., and Pajan (E. Mjoberg). Ku- ching (Hewitt). Sumatra : Wai Lima, Lampongs (Karny). In the Philippines, nests are built in rotten logs and tree fern stems and under moss on rocks. Dr. Chapman informs me that this, the most common Stictoponera in the Philippines, is usually seen running over foliage or resting in the axils of large leaves in wet foothill ravines. Stictoponera bicolor Emery Ectatomma (Stictoponera) hicolor Emery, 1889, Ann. Mus. Civ. Stor. Nat. Geneva, 27: 493-494, worker. Ectatomma (Stictoponera) bicolor var. minor Forel, 1900, Jour. Bombay Nat. Hist. Soc, 13: 316, worker. Stictoponera menadensis subsp. minor, Brown, 1948, Psyche, 54: 264, (teneral) worker. Stictoponera hicolor, Brown, 1950, Wasmann Jour. Biol., 8: 245-246, worker, synonymization of var. minor. Since this form was described, not a few authors, including myself, have been confused by it one way or another. It is very close to /S'. mefiadensis, and has the same head shape, with prom- inently projecting posterior occipital lobes or "ears" and a cor- responding, deepl}' concave, border between them as the head is viewed in full face. The eyes are rather large and situated well posterior to the middle of the sides of the head. The size is as iu menadensis, with some series averaging a trifle smaller, but the color of the alitrunk is lighter and brighter, varying shades of orange-ferruginous. The head varies from about the same color as the alitrunk to pieeous, the darker color being more common. The median smooth strip of the menadensis mesonotum is re- placed in hicolor by fine, indefinite, more or less opaque longi- tudinal rugulation, and the sculpture throughout is usually stronger, closer and less shining. S. bicolor may or may not possess minute propodeal denticles; these appear to be an allo- metric character of the sort that grades through within and be- tween series. The tenerals are straw-colored, and appear very different from fully-pigmented individuals in the same nest series. 1954 ANT GENUS STTCTOPONERA MAYR 5 Like S. menadensis, the present species has often been mis- identified. It is still possible, of course, that menadensis and hicolor are mere geographical color representatives of one species, but from the present material they can still be separated. The record of hicolor from the Philippines is based on a male of un- certain species ; no authentic hicolor records are yet known for these islands. I have reviewed material from the following localities : Indo- china: Coxan and Dong Mo (F. Silvestri). Hainan I.: Dwa Bi; Ta Han ; Loi Molia ; Nodoa (J. L. Gressitt) . Sumatra : Wai Lima, Lampongs (Karny). Java: Pemalang (L. G. E. Kalshoven). Hong Kong: no further locality (Ris). Emery described this species from material taken in various Burmese localities, both upcountry and in Tenasserim. Stictoponera bingiiamii Forel Eotatonvma (Stictoponera) hinghamii Forel, 1900, Jour. Bombay Nat. Hiat. Soc., 13: 137, worker, female. Type locality: Burma [Pegu Yoma?]. Stwtoponera horn,eensis Emery, 1901, Ann. Mus. Civ. Stor. Nat. Genova, 40: 662, nota, worker. Type locality : Sarawak. New Synonymy. This species resembles S. menadensis, but lacks the prominent lateral occipital ears and has a normal full adult coloration of rich ferruginous red. The eyes are at or close behind the middle of the sides of the head, and the posterior occipital border is straight to gently concave in different specimens and according to the view. Small teeth are present on the propodeum of most individuals. The insect is rather strongly shining, the foveation tending to be less dense than in menadensis, and the middle of the pronotum has a variable smooth, shining area free of foveae. The petiolar node seen from above is approximately as broad as long, length being favored in larger specimens. The second post- petiolar segment is very smooth and shining, foveation very in- distinct and shallow and virtually confined to the sides. The antennal funiculi are rather thick, the median segments broader than long even in the largest specimens. At first I had separated hinghamii from horneensis, though the two were obviously closely related, but I now find that specimens referable to both were taken by Dr. Chapman on the same date at the same locality near his camp in the Cuernos 6 BREVIORA No. 34 Mts., Negros Oriental, Philippines; these probably represent a single nest series, and in any case, it is apparent from a study of the full series, including those in Dr. Chapman's personal reserve collection, that the large and small forms are only allometric extremes of one form. I have also seen material referable to the same species from Tutu River, North Borneo (Mjoberg) and a specimen from "Pedada-B., Lampongs, Sumatra," unknown collector, and I feel sure that Forel's Burmese and Emerj^'s Bor- nean types are merely the small and large extremes of one variable species, though I have not examined type material. Dr. Chapman tells me that he found this species nesting in rotten logs in ravines in forest. Stictoponera crassicornis Forel New status Ectatomma (Stictoponera) binghnmi subsp. crassicornis Forel, 1912, Zool. Jahrb. Syst., (suppl.) 15: 51, worker. Type locality: Indrapura, Su- matra. Forel states that the eyes are anterior to the middle of the sides of the head, which if true would separate this form from hmghmnii very decisivel3^ The other characters cited, however, indicate considerable similarity, and Forel may well have been mistaken about the eyes. The description of >S'. spiralis, the next species following, also claims a similar position for the eyes. Among all the series available to me, I have seen no Stictoponera specimens with the eyes in front of the middle of the sides of the head. In any case, it is very unlikely that crassicornis can be a race of S. binghamii, since the known distribution of the latter straddles the Sumatran type locality of crassicornis. Provisional specific rank is indicated for crassicornis until the type can be re-examined. Stictoponera spiralis Karawajew Stictoponfira spiralis Karawajew, 1925, Konowia, 4: 79-81, worker. This species, also described from Sumatra, reads very much like S. crassicornis in what seem to be the significant features, and it is possible that the two names are synonymous. Stictoponera biroi Emery Stictoponera biroi Emeiy, 1902, Term. Fiizetek, 25: 154, worker. 8. biroi, the only species of the genus so far recorded from New Guinea, appears to be quite distinct from the Indomalayan forms on the basis of its original description. 1954 ANT GENUS STICTOPONERA MAYR 7 Stictoponera posteropsis Gregg Stictoponera posteropsis Gregg, 1&52, Psj'che, 58: 77-8§, figs. 1, 3a, 3c, female. This very aberrant species has large eyes situated almost at the extremes of the posteriorly salient occipital corners. The type, a dealate female, came from Sumatra, but I have seen a few additional worker specimens from the Cuernos Mts., Negros Oriental (Dr. Chapman), taken in rotten logs in forest ravines. Some of these workers are larger even than the female type, and their color is dark piceous instead of ferruginous red. Stictoponera costata Emery Ectatomma (Stictoponera) costatum Emery, 1889, Ann. Mus. Civ. Stor. Nat. Geneva, 27: 494, worker. Forel, 1900, Jour. Bombay Nat. Hist. Soc, 13: 316, 317, worker. Bingham, 1903, Fauna Brit. India, Hym., 2: 83, worker. Ponera rugosa Fr. Smith, 1857, Jour. Proe. Linn. Soc. London, Zool., 2: 6^, worker; name preoccupied by Le Guillou, 1840. New Synonymy. Stictoponera costoda Emery, 1901, Ann. Mus. Civ. Stor. Nat. Genova, 40: 662, worker. Stictoponera costata var. unicolor Forel, 1901, Eev. Suisse Zool., 9: 335, worker, male. New Synonymy. Stictoponfira rugosa var. parva Forel, 1913, Zool. Jahrb. Syst., 36: 6, worker. New Synonymy. 'i Stictoponera costata var. simalurensis Forel, 1915, Tijdschr. v. Eut., 58: 23, worker. New Synonymy, with doubt. Stictoponera costata var. pinealis Wheeler, 1929, Boll. Lab. Zool. Portici, 24: 31. New Synonymy. Stictoponera wallacei Donisthorpe, 1932, Ann. Mag. Nat. Hist., (10) 10: 447, noDi. pro F oner a rugosa Fr. Smith. New Synonymy. The key reference to this form is that of Emery for 1901 (loc. cit.) in which he cites variation in size among specimens from Sumatra and Mentawei (the costata type locality is in Burma) and suggests, but does not declare, the sj-nonymy with Ponera rugosa of Smith. He also cites differences between costata and a type of coxalis he received from the Koger Collection in Berlin, the latter differing "from costata in the much less coarse sculp- ture of the whole body, and especially of the abdomen, and in the more slender antennae, with the flagellar segments less thick, the third to the fifth longer than thick (thicker than long in S. costata).'" I am able to confirm and amplify Emery's differentia- 8 BREvroRA No. 34 tion of these two species, as will be seen below under S. coxalis. In the series I refer to S. costata, stemming from ten localities ranging from Malaya to Borneo and Mindanao, there is a large amount of variation in size, color, and degree of development of sculpture; the variation in sculpture of the first gastric (post- petiolar) segment is particularly notable. The prevailing color in North Bornean specimens is rich ferruginous red, but certain specimens from Mindanao are piceous, and the female type of pineaUs Wheeler, from Penang, is deep reddish with the gaster black. The variety imicolor Forel, based largely on a Bornean male, was never really differentiated from the types of costata. The various color conditions are constant within some nests, but not in others; and considering the long teneral period shown by ants of this genus, the relative conspicuousness of tenerals under the collector's eye, and the chances of color changes in preservative and cabinet, I can attach no great importance to color by itself. In the present material, color, size, variation in the shape of the lateral occipital ' ' ears ' ' or lobes, and sculpture are discordant geographically, so that it seems not possible to recognize objective geographical races based on these characters. Emery mentions, as stated above, that certain Sumatran series varied in size, and it seems likely that Forel 's variants parva and simahirensis are merely small costata. The Bornean speci- mens show significant size variation even between relatively close localities. Essentially, the sculpture of costata, except for the very coarsely and regularly cost ate second gastric (second postpetiolar or IV abdominal) segment, is in the form of large, rather deep punctures, foveae or pits, with more or less distinct and smooth, shining spaces between them. On the first gastric segment, the punctures tend to be more elongate, and the spaces between them, particularly toward the posterior edge of the discal surface, begin to form more or less definite longitudinal ridges or costae. In some specimens, such as those from Mt. Penrissen and Mt. Tibang, in Borneo, the spaces are broad and shining, and the punctures definitely prevail over the rather weakly suggested costae, while in others, such as most Bornean specimens and the pinealis type, the costae are closer and sharper on the first gastric segment, at least posteriorly, and the punc- tures are closer together and more elongate. The sculpture of 1954 ANT GENUS STICTOPONERA MAYR 9 the remainder of the body follows that of the gaster more or less closely, but differences are harder to see and describe. In spite of these rather distinct differences between extremes, I find that Borneo alone supplies a fnll set of intergrades w^hich is enough to obscure, any j)ossible taxonomic split based on this character. It therefore seems that Emery was correct in maintaining cosfata as a single taxonomic unit despite the variation he saw. His reasonable suggestion that F. Smith's rngosa, from Sarawak, equalled costata is also accepted here, and since the name rngosa is preoccupied, and Donisthorpe's noineri novum wallacei later than costata, the name costata will remain in use. I have studied series in the Museum of Comparative Zoology from the following localities : North Borneo and Sarawak : Mt. Penrissen; Mt. Tibang; S. Slau ; Brooketon ; Pajan; S. Saranibo ; Baian River (E. Mjoberg). Mindanao, Lanao district: Maria Christina Falls (J. W. Chapman). Momungan (D. Empeso) Malaya: Penang Island (F. Silvestri), type of var. pinealis. Sumatra: Langkat, E. Coast (Jourin). Stictoponera coxalis (Roger) Ponera coxalis Eoger, 1860, Berl. ent. Zeitschr., 4: 308, worker. Ectatomma (Stiotoponera) coxale Forel, 1900, Jour. Bombay Nat. Hist. Soc, 13: 316, worker. Nee Bingham, 1903, Fauna Brit. India, Hym., 2: 84, fig. 44, worker. Stictoponera coxalis Emery, 1901, Ann. Mus. Civ. Stor. Nat. Geneva, 40: 662, worker. Roger described this species very incompletely for modern needs from specimens collected by Nietner in Ceylon. Emery's statement of the differences between coxalis and costata has already been translated above under S. costata. I have been able, through the courtesy of Prof. M. Beier, of the Naturhistorisches Museum, Vienna, to examine a worker (here designated as lecto- type, and so labelled) and a dealate female from the type series of coxalis, that Roger had early sent to Gustav Maj^r. As Emery mentions, the sculpture throughout is considerably finer than in costata; it is also denser and consists more predominantly of longitudinal costulation or coarse striation instead of the large, predominating punctures of costata, although somewhat smaller punctures are still clearly present and often conspicuous. Under 10 BREVIORA No. 34 lower magnifications, coxalis tends to appear more opaque gen- erally than does costata. On the first gastric segment, there are up to two or three irregular, but rather close and fine longitudinal eostae for every one seen in the average costata specimen. Also, as mentioned by Emery, the funiculi are notably more slender in coxalis, with the third through the fifth (I would include the sixth and possibly also the seventh) funicular segments slightly longer than broad. The same segments are broader than long, though somewhat variable in proportions, in the costata samples I have seen. I agree with Emery's separation of the two species on the pres- ent evidence ; obviously, however, the two are very closely related. To my knowledge, S. coxalis remains known only from the type series from Ceylon, though various authors have followed Bing- ham in confusing this species with S. menadensis and S. hicolor. When better collections are available from southern India, it may be necessary to re-examine the relationship of coxalis and costata. Note on " Stictoponera sauteri" The name Stictoponera sauteri (Chapman and Capco, 1952, Check list of the ants of Asia, Monogr. Inst. Sci. Tech., Manila, 1: 30) is a combination proposed in error, and actually refers to Ectomomyrmex sauteri, a species described earlier by Forel. This species has no connection with Stictoponera; the Check List com- bination was purely an unintentional clerical slip. R E V I O R A Mnaseitim of Coimparative Zoology CAMHRIDdE, .AfASS. JuLY 2S, 1 (KM Xl-.MBER 35 NEW OR REI)E8(JRIBED PELOMEDUSID SKULLS FROM THE TERTIARY OF AFRICA AND ASIA (TESTUDINES, PELOMEDUSIDAE) 1. Dacquemys paleonwrphd. new oenus. new species from the Lower Oligocene of the Fayuni, E characters which it sliares with the Pad'tcncmis group of genera, to which it must, indeed, he referred — the enlarged "carotid channeT' and the l)asioccipital-qnadrate contact etc. From Boihrey)iys (Hay, 1908) 1954 DACQUEMYS PALEOMOBI'lIA 5 it differs as radically as from Sfcrctxjciii/s c.o'. in ]a('kiii<|- tiic extraordinary pits in the maxillae so characteristic of Bt^tlncnn/s, in the mnch narrower skull, in the absence of a vomer, etc. Other genera of fossil pelomedusids are known, but in none of these is the skull known, and there is no plausible reason for ascribing this skull to any of them. But while it is thus easy to show that to place the Egyptian skull in any described genus would be erroneous or probably erroneous on present knowledge, it is impossible to be sure that it does not belong to one of previ- ously described Fayum species based on shells. I incline, indeed, to the opinion that it does belong to one of these species. From the Lower Oligocene beds of the Fayum from which Dacquemys derives, three species of pelomedusid turtles have been described: Podocncmis fajuniensis Andrews, i^tercodcnijs libyca Andrews, and Pchjiiudusa progaleata v. Reinach. If Pdomedusa progaleata (material re-examined by Dacque, ]912) is correctly assigned and there appears to be sufficient rea- son to think that it is, D. paleomorpha cannot I)e synonymous with it, since the skulls in the two genera are so very different that they do not belong even in the same subsection of the Pelomedusidae. Unless the Dacquemys skidl belongs to some variety of shell still undiscovered, decision must then lie between Podocnemis fajumensis and Stereogenys lihyca. This task is difficult, requir- ing a rather complex and tenuous argument. The argument may be stated in this fashion : The shell of Podocnentis fajiiiiK iisis differs so little from that of Recent EryninocJielys itiadagasravensis that, considering the shell alone, the distinctness of the fossil species might be cpies- tioned. The skvill of the fossil form might therefore be expected to be very like that of the Recent species, and in fact the Dae- qneniys skull is more similar to the skull of Erymnoehelys or of related Peltocephalus than to any other known pelomedusid skulls. Since it seems evident that skulls in the Podocnemis sidebranch of the Pelomedusidae are subject to much more radical modification than the shells, which tend to be conservative and relatively difficult to distinguish, it is a j^ossiliilty worth con- sidering that the Dacque skull may belong to "Podocnemis" fajurnensis. 6 BREVIORA No. 35 111 the alternative case of Stercofjcin/s lihijcd the sliell whieh is the type of this species is clearly very similar to that referred to S. cromeri. The argument that the shells referred to S. cromeri really belong to that species is plausible even though not con- clusive. Andrews (1906, p. 298) remarks: "Although in no case have the carapace and plastron be(Mi found associated with the skull in such a manner as to leave no doul)t that they belong to the same individual, nevertheless the shell . . . may be regarded with reasonable certainty as belonging to the present species. In the first place, this form of shell, like the skull, is the commonest occurring in the Qasr-el-Sagha beds, and in the second place it differs widely from the shell of Fodooifiiiis. the only other Pleu- rodiran genus found at this horizon." Accepting Andrews' in- terpretation of the situation, if Sfereugcnys libyca is really closely related to N. cromeri, it should have a skull to some degree resembling that of N. cromeri. The Dac(|ue skull is so very dif- ferent from that of *S'. cromeri that it is very unlikely to l)e the skull of »S'. lihyca. Thus the Dacque skull is unlikely t;) lie that of S. lihiica and might be that of P. fajumen.sis. But to assert the truth of the latter suggestion confidently w^ould be a bold venture. In the next paper of this series T shall present additional evidence based on a skull from the Miocene of Moghara that this suggestion may possibly be true, but even with this addition the evidence will still be incomplete. It remains an unfortunate fact that with none of the pelomedusids of Egypt is there an unequivocal association of a skull and a shell. I have therefore described the species paleomorpha as new, preferring to indicate the present state of the evidence : that we have here a very distinct type of skull which may belong to either of two distinct shell types or possibly to a third unknown one. Andrews (1906, p. 295) has described from the same beds witli Dacquemys paleomorpha, P. fajumensis and *Sf. lihyca an anterior portion of mandible which he has referred to P. fajumensis. His comment on this element is as follows : ' ' This mandible differs from that of P. madagascarensis in having a broader symphysis ; the biting surface is also broader and is partially divided in two by a median ridge." Figure 1 is a photogra])h of tlris mandibular fragment made by Peter CTreen and published ])y permission of I 1954 DACQUEMYS PALEOMORPIIA 7 the Trustees of the IJritisli .Museum. It will I)e seen that the ridoiuo' of this niaiiilihh' wouh] fit well with the maxillary ridging of Dacqueuii/s pal(rj)li(i. liut there is no strong hook at the mandibular symphysis such as seems to be implied by the deep premaxillary form of the skull. In this ease, as in the previous one, only discovery of associated material can verify the reference to genus and sjiecies. Fig. 1. Dorsal view of niiiiidil)lt> lefened Ijy Andrews to Prxlooieinis fajiniu'H.si.s. B:M (Nil) R ."wUG, from tlie Lower Oligoceiie of tlie Fayuni, Egypt, xl. Photograiili reinodiiced l)y periiiissioii of tlie Trustees of tlie British Museum. The photographs of the skull of Dacqucmijs paleomorpha were made by the Museum fiir Naturkunde in Stuttgart and 1 am very grateful to the authorities there for the privilege of reproduciiig them. 8 BREVIORA No. 35 PAPERS CITED Andrews, C. W. 1906. A desc-iiptive catalogue of the Tertiary vertebrata of the Fayum, Egypt. London. 324 pp. Dacque, E. 1912. Die fossilen Schildkroten Aegyptens. Geol. Palaeont. Althandl., vol. 14, pp. 275-337. Hay, O. p. 1908. The fossil turtles of Xortli America. Carnegie Inst. Publ. No. 75. 568 pp. Siebenhock, F. 1902. Zur Systematik der Schildkrotengattung Podocnemis Wagl. Sitz. Ber. Akad. Wiss. Wien, math. nat. Kl., Abt. 1, vol. Ill, pp. 157-170. SWINTON, W. E. 1939. A new fossil fresh-water tortoise from Burma. Ree. Geol. Survey India, vol. 74, pp. 548-551. Williams, E. 1953. Fossils and the distribution of chelyid turtles. 1. " Hydraspis leithii" in the Eocene of India is a pelomedusid. Breviora, No. 13, pp. 1-8. J'lnti' 1. 'I'ypc skull (if l>y courtesy of the •Staatliches ^Museum fiir Xaturkunde. xl. E V I O R A MinseiLiitjni of Comparsitive Zoology CA:MBKinGK, ^[\> August (i, 19.")4 Number 36 PRESENT KNOWLEDGE OF THE SNAKE ELACHISTODON WESTFAUIANNI REINHARDT By Carl Gans Carnegie Museum, Pittsl)urgh, Pa. and Erxest E. Williams Museum of Compaiative Zoology, Camlnidge, Mass. INTRODUCTION In 1863 Johannes Reinharclt described a peculiar snake belong- ing to the new genns and species Elachisiodon ivestermanni, the specific name referring to the collector, William Westermann, who obtained the initial specimen in Rnngpore (Rangpur, see Fig. 1), northeast Bengal. Reinhardt's single specimen was characterized by certain specializations of the integument, by an almost complete absence of teeth, and liy certain of the verte- brae bearing hypapophyses which penetrated the esophagus. Comparison with the African snake Dasypeltis, long known to be specialized for egg-eating by a similar series of modifications, was inevitable, and Reinhardt came to the conclusion that Elachi- stodon was also an egg-eater. Differences between Elachistodon and Dasypeltis centered about minuscule grooved teeth and an enlarged row of dorsal scales in the new genus. The presence of the grooved teeth on the posterior aspect of the maxillae sug- gested the placement of the form in the composite section Opisthoglypha, but Reinhardt pointed to the close similarity between this form and the aglyphous Dasypeltis as evidence for the artificialitv of this categorv. No. 36 Since 1863 only four additional specimens have been reported and there has been little additional contribution to the morphol- ogy or natural history of the species. A recent study (Gans, 1952) has demonstrated a number of previously undiscussed modifications in Dasypeltis and has shown that this genus is much more highly adapted to its very special mode of life than had previously appeared to be the case. In view of this new knowledge of Dasypeltis it was felt that a re-examination of the status of Elochisfodon would be especially Fig. 1. The distribution of Elachistodon westermanni. Crosses mark the localities where specimens were found. appropriate at this time. Unfortunately, however, the existing specimens are too few to permit a detailed investigation such as was carried out for Dasypeltis. It has, indeed, l)een possible for one of us (Williams) to make an examination of the specimen in the British Museum and of the osteological preparations from that specimen figured and referred to by Malcolm Smith (1943, p. 403), but the information so derived is suggestive rather than adequate, and it is very evident that further specimens will be needed for thoroughgoing anatomical studies. 1954 ELACIIISTODON 3 This iJciper, tlieivfore, lias two oljjects. Fii-st, it is iiit<^n(l(Ml to provide a recapitulation of the cxistiu-i' data on E'dcltistodon in the light of the findings in the ucinis Ddsiipcltis, and secondly it is desired to reawaken interest in Elacliistodoii in the hope of obtaining the additional specimens upon which the more intensive studies might be based. HISTORICAL RESUME The initial sjiecinien was rather fully described by Reinhardt (1863, p. 198), whose careful investigation of its internal anat- omy was luidertaken l)ecause the enlarged vertebral scales of the new form had left him in doubt as to its taxonomic placement. He prepared the tooth-bearing elements of the right side of the skull and one of the mandibles, and checked the condition of the vertebral hypapophyses and their relation to the esophagus without dissecting out the vertebrae or separating them. He found teeth on both ]-)alatine and pterygoid as well as teeth and minuscule fangs on the maxilla and described the shape and appearance of the dcMitary. in comparing Ehichishxluii with Ddsjiix Ills he pointed out that in both there were two tyjX's of penetrant liypapo])hyscs. though both types in the new form were less well-developed. He figured the palatal and mandibular elements of both genera (see Fig. 1). The stomach and gullet of the type specimen of E. ivestennainii were found to be filled with an amorphous congealed mass which analysis indicated might be either milk or egg fluid. Though he did not find shell fragments mixed in this mass, Reinhardt speculated on the feeding habits of the animal and appears to have leaned toward tlie theory that the species w^as an egg-eater. Reinhardt also described in detail the squamation of Elachisto- don, mentioning the very ])eculiar deep pit on the posterior nasal as well as similarities to Dasypeltis. Further details on these and other structural points will be presented in the dis- cussion of the morphology of E. westermanni. Reinhardt placed the new genus in the family Rachiodontidae {sic). The next to refer to Elachisfodon was Giinther (1861, p. 414) who mentioned it on Reinhardt 's authority only. Pie suggested that the genus should be separated as a distinct group of colubrids. 4 BREVIORA No. 36 Blanford (1875, p. 207) reported the second specimen from Bihar, Purnea (see Fig. 1), and his general description agreed with that of Reinhardt. He suggested that the genus be referred to the Dipsadidae. Cope (1886, p. 494) considered only the familial assignment of the genus, placing it in the subfamily Dasypeltinae with Dasypeltis. Boulenger in the Fauna of British India (1890, p. 362) placed the genus in the Opisthoglypha, Dipsadinae, next to Dipsas. He stated that he had not been able to examine specimens but from the key (pp. 356-357) it appears that the assignment was based on the grooved teeth, the enlarged vertebral shields and the elliptically vertical pupil. He also referred to esophageal teeth capped with enamel. Sclater (1891, p. 48) in listing the snakes in the Indian Mu- seum recorded a specimen from Bengal which seems to have been that earlier recorded by Blanford. He placed it in the Dipsadinae. In the Catalogue of the Snakes of the British Museum, volume 3, Boulenger (1896, p. 263) erected the subfamily Elachistodon- tinae of the Colubridae with the single genus Elachistodon. He regarded this as the opisthoglyphous analogue of the aglyphous Rachiodontinae. His description of the subfamily mentions ptery- goid teeth. The next record is that of Wall (1913, p. 400) who reported a new specimen from the Jalpaiguri district (see Fig. 1) and published a detailed comparison of the three specimens then known. He did not give the sex of the third specimen (it is, how- ever, the female described again by M. Smith, 1943) but did cast doubt on the presence of pterygoid teeth. His later 'Handlist' (1923, p. 878) did not contain any new data. In a serialized discussion of the snakes of northern Bengal and Sikkim, Shaw, Shebbeare and Barker (1941, p. 65) list the specimens known to them. In addition to the original Rangpur specimen in the Museum at Copenhagen they cite a second from Purnea (the Blanford specimen) in the Indian Museum at that time. The specimen obtained by Travers at Jalpaiguri (Wall, 1913) was stated to be in the Bombay Museum (it is now in the British Museum), while the then whereabouts of two addi- 1954 ELACTIISTODON 5 tional specimens obtained hy Travers at Baradiulii ^ ajipeared to be unknown. Malcolm Smith (1943, p. 4m) redesci-ibed the uvnus and removed it from the Elaehistodontinae to tlie Dasypeltinae. lie presented the first published figures of the modified vertebrae and also refigured the palatal bones. Pterygoid teeth are not men- tioned in his text or shown in his figure. Smith and Bellairs (1947, p. 362) mention only that the Plarderian gland is enlarged in this form as in Dux fi pel tin. MORPlIOLOdY (ieneral The following condensed diagnosis is designed to aid identifica- tion of specimens. For this reason it follows the summary pre- sented l)y Smith (1943, p. 404) with only minor modifications. Head fairly distinct from neck ; eye large with vertically ellipti- cal pupil (round fide Reinhardt) ; a large pit in the posterior part of the nasal shield. Body moderately elongate, feelily com- pressed. Tail short, subcaudals paired. Internasals as large as prefrontals ; frontal large, longer than its distance from the end of the snout ; nasal large ; 1 small pre- ocular, the loreal below it entering orbit ; two postoculars ; two long anterior temporals ; H or 7 supralaliials, the third and fourth touching the eye ; 2 pairs of sublinguals. Scales smooth, in 15 rows, 19 on the neck, the vertebral series much enlarged, hexag- onal. Anal single. Ventrals 208-217. Caudals 59-65. Above, dark olive brown to blackish, the vertebral scales yel- lowish-white, except at their outer margins, forming a light vertebral stripe extending the whole length of the body; sides spotted or flecked with the same color ; whitish below, the outer margins of the ventrals and adjacent rows of scales edged with brown ; a yellow stripe along the top of the head from the snout to the angle of the mouth, passing above the eye ; an angular bar or spot on the nape ; lips yellow. 9 784 nun. (670 + 114 mm.). Ventrals 217. Caudals 59. Rung- pore, Bengal. Copenhagen Museum (Reinhai'dt ). 1 We have been unable to find Baradijihi on any map. but the Imlijin Consuhir Office in New York has very kindly informed us that thtTo is a railiuad station by that name about 30 miles north of Calcutta. BREVIORA No. 36 800 mm. (670 + 130 mm.) fide Smith, (762 mm. [635 + 127 mm.] /;V/f Wall). Ventrals213 (210). Caudals 62 (64). Near Mai, Jalpaiguri District. Formerly Bombay jMuseum, now British Museum (Natural History). 222 mm. (186 + 35.7 mm.), Ventrals 208. Caudals 63. Bihar, Purnea. Indian Museum (Blanford). Fig. 2. E. irefulai' scales at all, while in Eluchistodou there is but one row of these. In both of these forms the sublinouals are so rigidly tied in that expansion of the chin region can only take place along- the angle of the mouth during ingestion. In Dasypeltis this expansion is made possible by the presence of extremely distensible skin between the scales of this region. The last labials have also been rearranged to form the first scales of the lowest lateral rows. The same condition is also the ease in Elachistodon, and the resulting lines of expansion are clearly seen in Figure 2. Such a specialization, the value of which is probably solely restricted to egg-eating, would tend to furnish additional evidence for the idea that other described modifica- tions of Elachistodon are adaptations to this habit. A possibly significant point of ditference between Elachistodon and Dasypeltis is the ]n-esence in the form of a large pit on the posterior part of each nasal shield. The function of similar pits in the Crotalidae as well as in cei-tain species of Pythonidae has been determined to be that of a thermosensitive range-finder (Noble and Schmidt, 1987), and it is known that similar struc- tures of as yet uninvestigated function exist in other forms. In Elacliisiodon nothing is known of the function of the pits. Scpiamation of the Posterior liody The enlargement of the vertebral series of dorsal scales char- acteristic of Elarhistodo)! is a feature also found in a large num- ber of dipsadine and other snakes but not in Dasypeltis. Little is known about the function of this enlargement although Peters (MS 1952, p. 27) has speculated on its relation to the arboreal habits of the dipsadines. He believes that in conjunction with the extreme lateral flattening of the body it might impart a stiffening etfect. He offers the analogy of an I-beam, and suggests that the enlargement of the vertebrals might permit the unusual horizontal rigidity and extension observed in climbing members of the Dipsadinae. We cite the suggestion here without passing judgment upon its applicability to Elachistodon, which shows but feeble lateral compression. It is also to be noted that the scales of Elachislodnu ai-e smooth while those of Dasypeltis are strongly keeled. This character is, however, of little value at the generic level or above, in view of 8 BREVIORA No. 36 the number of eases in snakes in whieh it shows intrageneric variation. Elachiatodon e.xhibits no traee of the size reduetion, angling and serration exhibited by the second to sixth rows of dorsal scales in Dasypeltis. Except for the vertebral row all the dorsals are uniform in size and character. Soft Tissues of the Head No pul)lished observations are available on gum structure. Examination of the British Museum specimen indicated that the gum ridges are probably covered with a somewhat less convoluted mass of tissue tlian are those of Dasypeltis. Smith and Bellairs (1947, p. 362) have noted that the Harder- lan gland is notably enlarged in EJachisiodon paralleling the con .ition in Dasypeltis. Fig. 8. E. wesiermanni. A, Palatal liones ; B, anterior hvpapopliysial vertebrae; C, posterior hypapopliysial vertebrae. Ablireviations: e, et-to pterygoid; mx, maxillary; pal, palatine; pi, ptergygoid. After Malcolm Smith (']943). Skull To the best of our knowledge no complete skull has ever been prepared. The palatal elements have been figured by Smith (1943) (see Fig. 3, A). A check of the material by one of 1954 ELACHISTODON 9 US (Williams) found the fifjures accurate in all particulars includ- ing- the fact that pterygoid teeth (described and illustrated by Reinhardt in the Copenhagen specimen) are not to be seen. Comparison with the homologous elements in Dasypelfis in- dicates few differences beyond the presence of the relatively enlarged grooved teeth on the posterior aspect of the maxilla in Elachistodoti. This portion of that bone is also somewhat more strongly developed and there are general though minor differ- ences in the relative proportions of the various structures. The union between palatine and maxillary may be somewhat weaker than the corresponding articulation in Ddsifijeltis. The wing- shaped process of the palatine is much abbreviated though this process is shown as greatly extended in Reinhardt 's drawing. These bones are, however, very fragile, and it may be possible that some of the apparent differences may be traced to the method by which these elements were cleaned. The similarities seen lead us to the conclusion that the palatal elements of Elachistodon are as rigidly joined as those of Dasypeltis. The dentitional formulae are as follows : Maxilla edentulous except for two small teeth followed l)y two larger grooved teeth ; palatine with four minute teeth in its center ; pterygoid — ac- cording to Reinhardt — with three extremely minute teeth. Wall (1913) and Smith (1043) — botli treating of the specimen from Jalpaiguri — do not refer to pterygoid teeth. The dentary has eight to twelve teeth diminishing in size posteriorly. Reinhardt has figured and described the anterior end of the mandible for both Elavil ishxhnt and Dasypeltis. The appearance of the element in Elachisfodoii is very strange, and since the fig- ure of the same element in Dasy pelt's liears veiy little resem- blance to actual specimens seen by us, we do not care to connnent further in this matter (see Fig. 4). \%'rte))ral Cohunn and Eso]>hagus There are certain rough similarities in the appearance of the anterior vertebrae (hypapophysials) in the two forms. In both there are two basic types of hypapophyses — anterior and poste- rior, but in Dasypeltis some of the anterior hypapophyses undergo considerable ontogenetic variation so that it is best to distinguish anterior, intermediate, and posterior types. It is not known 10 BREVIORA No. 36 whether similar ontogenetic variation is found in Elachistodon. For the latter genus, therefore, these comments must of necessity be restricted to the modified vertebrae of the adults. In Dasypeltis the anterior hypapophyses are little modified. They are of the "normal" squarish shape with only their ventral edges enlarged into sled-like runners. The egg glides along these runners during ingestion, but the runners do not serve any cut- ting or sawing function. The egg is pierced by the elongate, spiniform, forward pointed hypapophyses of the posterior series, wliicli penetrate the esophagus in all specimens (Gans, 1952, p. 236). After the shell has been broken, the hea\y rounded boss- like hypapophyses of the intermediate series transmit a crushing force to the egg shell and fold this, while the egg contents are being squeezed into the stomach through the esophageal valve. The forces involved in this action are sizable, and for this reason the articulating surfaces of the neural arches are greatly ex- panded, the pre- and postzygapophyses being laterally displaced and extended by spiniform processes. This offers additional sur- face for the action of the dorsal musculature wliich is kept from lateral slippage by the confining inclination of the surfaces of attachment. The intermediate and some of the anterior hypapophyses are penetrant in adults of Das i/ pelt is, but this character shows con- siderable variation in juveniles. There is also evidence that some seasonal variation of this character may occur in adult specimens as well. In Elachistodon nothing appears to be recorded about the hypapophysial vertebrae immediately posterior to the cervicals, and these have never been figured. Perhaps it may be assumed that they are of ' ' normal ' ' appearance, possil)ly performing the same function as do the homologous structures in Dasypeltis. There is no record as to the number of these "unmodified" verte- brae, though Smith (1943, p. 403) states that the modified hypapophyses start opposite the tenth ventral shield. The amount of skeletal preparation which has been done for this region is very limited. Reinhardt (1863, footnote to p. 202) stated that he had not skeletonized but rather examined the vertebral column in situ. Smith figures two short sections of three vertebrae each (Fig. 3, B and C). One of us (Williams) Fis'. 4. Plate from tli.- (ni^iii; ventral ami side views of the In E. tn.sh iiiKiKiii : K, (leiitary I (les.-iiptioii ]iy Heiiihaidt. J, /.', (', Dorsal, ad of K. irrstrniKiiiin : J), jialatai Ijones of )f E. W(.^trniuunii ; E, palatal bones of Dasypeltis ticabcr; G, mandible of DiusypcUia scab 12 BREVIORA No. 36 has examined these vertel)rae, and the t'ollowinji' statement is based on this examination. The more anterior of the prepared and figured hypapophyses are very strange in appearance. Bosses are present as in the intermediate hypapophyses of the African form, but they are not as closely integrated with the base of the centrum. They are instead mounted on a pedicel beyond which they extend both anteriorly and posteriorly. The vertebrae bearing these hypa- pophyses are relatively undifferentiated. No extension or inclina- tion of the zygapophysial surfaces is apparent. In some ways the anterior hypapophyses of Elachisfodon are reminiscent of stages in the ontogeny of the intermediate hypa- pophyses in Dasypeltis, though th(> mounting on a pedicel sug- gests that they may be specialized in a different direction. Rein- hardt (1863, p. 203) stated that there were nine such anterior hypapophyses, while Smith (1943, p. 403) counted eighteen all of which projected through longitiulinal slits into the esopha- gus. Both authors believed that these are cutting instruments, which hardly seems likely in view of the findings in Dasypeltis. Compared to the posteriormost hypapophyses in Dasypeltis which are such highly effective tools, these same elements in Elachistodon seem even less modified than the anterior hypa- pophyses. In Elachistoelon the hindmost hypapophyses are block- shaped spines whose major specialization lies in the fact that they are directed forward rather than backward. They most nearly resemble the hypapophyses of several of the s])ecies of Elaphe recently mentioned as being specialized for egg-eating (Gans and Oshima, 1952, p. 15), but they also resemble those hypapophysials of very young Dasypeltis that are transitional between the intermediate and the posterior series. Reinhardt speaks of thirteen of the posterior type hypa- pophyses in Elachistodon, with only the first ten penetrating the esophagus, while Smith mentions eight of these structures, none penetrating the esophagus. Smith's observation would of course be in strong contrast to the findings in Dasypeltis, and the con- tradiction in observations here makes it quite clear that the egg- eating habits of this form will have to undergo a separate analysis to determine the extent to which parallelism in function exists, and to explore the dift'erences. 1954 ELACHISTODON 13 111 spite of this, it iiuiy he stated tliat all indicalions exist that eg-os are opened h_v I'JhicliisttxIon in a nmiiner siniilar to that eni]iloy(^d by Dasi/ix His and Elaphc climacophora, i.e. by exertined by Joachim Barrande. Anier. Jour. Sci., vol. 238, pp. 241-259, pis. 1-4. 1941. The Trinucleidae — with special reference to North American genera and species. Jour. Paleont., vol. 15, pp. 21-41, x)ls. 5, 6. 1953. A new Ordovieian trilobite from Florida. Breviora, no. 17, pp. 1-6, pi. 1. Wilson, James L. 1954. Late Cambrian and early Ordovieian trilobites from the Marathon Uplift, Texas. Jour. Paleont., vol. 28, pp. 249-285, pis. 24-27. EXPLANATION OF PLATE Onnin terri/i (Leith, 1938). Caparo-Bellavista Series, Middle Ordovician, near the Caparo River, State of Bariiias, Venezuela. Fig. 1. Rubber east from external mould of cephalon, showing upper lamella of fringe, X 4, MCZ 5183a. Extremities of third external row of pits indicated l)y " E:^. " Fig. 2. Mould of internal surface of glabella and cheek lobes, external surface of lower lamella of fringe, X 3, MCZ 5183b. Position of first (Ei) and second (E2) external row of pits, girder (g), and of first (Ii), second (To), and third (I:^), internal rows shown. Fig. 3. Holotype, original of Leith (1938, fig. 2), internal mould of gla- bella, cheek lobes and inner surface of upper lamella of fringe, X 2, Peabody Museum, Yale University, 15317. Extremity of third external row indicated by "E:^." Fig. -4. Mould of external surface of right cheek lobe, lower lamella of fringe and genal spine, X 3, MCZ 5183c. Fig. 5. Internal mould of cephalon showing upper lamella of fringe, X 21/2, MCZ 5183d. Fig. 6. Rubber cast from external mould of pygidium and four poorly preserved thoracic segments, X 3, MCZ 5183e. Fig. 7. Rubber cast from external mould of pygidium, X3, MCZ 5183f. E V I O R A Miaseuioi of Comiparsitive Zoology Cambridge, Mass. November 24, 1954 Number 39 NEW OR REDESCRIBED PELOMEDUSID SKULLS FROM THE TERTIARY OF AFRICA AND ASIA (TESTUDINES, PELOMEDUSIDAE) 2. A Podoeneniide Skull from the Miocene of Moghara, Egypt. By Ernest Williams In 1952-1953 while visiting the British Museum (Natural History) as a Guggenheim Fellow I was privileged to examine an undescribed skull from the Moghara Miocene of Egypt. The skull evidently belonged to a member of the genus Podocneniis sensu lato or to a close relative of that genus; the critical recogni- tion marks of this section of the Pelomedusidae — the enlarged "carotid" channel, and the contact of basioecipital and quadrate — were readily visible. Further examination left no doubt of the novelty of this fossil as compared with any previously known podocnemide ^ skull, and suggested interesting problems in regard to its proper correlation with a shell name and in regard to its l)hyletic position. The British Museum ]\Ioghara skull is unfortunately im- perfect in many respects. The snout is missing so that important characters and relations here cannot be checked. Major parts of the postorbitals and jugals are broken away, as are both squa- mosals, the quadrate of one side, parts of the parietals and the occipital condyle. Many structural details of taxonomic and phyletic importance are therefore not ascertainable : the presence or absence of a complete temporal roof, the presence or absence of a vomer, the position of the foramina incisiva, the relations of the internal lamina'?' of the premaxillae and maxillae to each other and to 1 I use the term "pddociu'inide" as a convenient and informal designation for a subsection of the family Pelomedusidae typified hy the genus Podocncmix. 2 A PODOCNEMIDE SKULL FROM MOGHARA 1954 the choanal margin, the presence or absence of an anterior process of the palatines dividing the choanal oi)ening, the pres- ence or absence of a median pit in the palatal snrface of the premaxillae and of a hooked process anteriorly on the premaxil- lae. The following points may, however, be made out : The skull must have been rather broad, certainly very little longer than broad. The interpai-ietal scale marked off by grooves on the skull roof is broad also, but tapers posteriorly, so that the parietal scales must have met behind it. There is no hint of a forehead groove but perhaps too little of the interorbital region is pre- served. The orbits are visible in dorsal view. There were two triturating ridges on the posterior portion of the palatal surface of the maxilla — a low, rough, median ridge and, parallel to it, a still lower, rougher, ridge at the margin of the choanae. There was no extreme development of a secondary palate. Grooves on the postorbital bar indicate the presence of a "subocular" scute in the sense of Siebenroek (1902). The "carotid" channel is fully enlarged in podocnemide fashion. The ectopterygoid processes are large, blunt and almost wholly lateral in orientation. The basisphenoid is a conspicuous triangu- lar element, the anterior apex blunted. There is a narrow basioccipital quadrate contact, more longitudinal than trans- verse. The cavum tympani is large and lacks a precolumellar fossa. These determinal)le characters are few indeed on which to hazard an estimate of the relationships of the Moghara form. One method of appraisal, however, is to tabulate the more diagnostic of these characters against the characters of other known podocnemide skulls. Tabulating first against the living podocnemide species (Table 1), w^e find that very little is learned except that the British Museum Moghara skull is not exactly like that of any modern podocnemide. If, now, we compare the British Museum Moghai-a skull with the previously known fossil skulls of podocnemide type (Tal)le 2), the impressive fact is the close agreement, in cited characters, of the British Museum skull and Dacquemys pahomorpha. We need in fact to inquire what characters assure us that we are not dealing with Dacquemys. NO. 39 fci2 O O ij O - & X s =s - A PODOCNEMIDE SKULL FROM MOGHARA 1954 S -— o o o — S g o _2 be 5 o 1 1 M s C3 "oj -•s •n 3d is '5 j3 i ^ cS S 1 a '> 1 P 'B 1 "3 o 2 NO. 39 BREVIORA 5 It is unfortunate that the unique generic character of Dac- qiiemys — the long squamosal-parietal suture — cannot be veri- fied or denied for the Moghara specimen. This region has been broken away. Tliere are, however, differences which may or may not imply generic distinctness. The British Museum skull is broad ; that of Dacqueniys is elongate. The orbits are visible dorsally in the ]\Iiocene specimen ; they are completely concealed in Oligocene Dacqueniys. The two triturating ridges are low and rough in the Moghara specimen, the inner parachoanal ridge especially so ; the same ridges are high, smooth and sharply defined in Dac- quemys. The suggestive point about these differences is the fact that in each case the Moghara skull is closer in these characters to Erymnochelys {=Podocnemis madagascarcnsis of Boulenger) than to Dacqueniys. The skull of Erymnochelys madagascarensis is relatively broad, the orbits are exposed dorsally, there is but one low ridge on the triturating surface of the maxilla. But in regard to the last character there is some roughening of the parachoanal border of the maxilla in Erymnochelys; the condi- tion in that genus could be explained as a further carrying through of a trend initiated in the British Museum skull. In fact the British Museum skull might on all its characters be interpreted as an intermediate between Dacqueniys and Erymno- chelys, perhaps somewhat closer to Erymnochelys. It we assume the reality of this intermediate phyletic station of the British Museum Moghara skull, it is then probable that this skull belongs with the common podocnemide shell of the Moghara deposit, the shell type named by Andrews Podocnemis aegyptiaca, which has every shell character of Erymnochelys and is on shell characters barely, if at all, distinguishable from Erymnochelys madagascarcnsis. I shall hereafter refer to Podoc- nemis aegyptiaca Andrews as Erymnochelys aegyptiaca (An- drews). There is only one other podocnemide shell type known from Moghara — the form named by Fourtau (1920) "Podocnemis" hiamlyi. This form differs from Erymnochelys aegyptiaca and from all Erymnochelys in the larger size of the intergular scale which separates the gulars as in the Recent South American G A PODOCNEMIDE SKULL FROM MOGHARA 1954 podocnemide species (Podocncmis sensu stricto and most speci- mens of Peltocephaliis) . If we could assume that P. hramlyi is a precursor of Peltocephalus then it would not be too anomalous for our Moghara skull to belong to this species, rather than to E. aegyptiaca. In the living species Peltocephalus dumeriliana (in which the gular-intergular pattern is typically very similar to that of P. Iramlyi) the skull has definite, strong similarity to that of ErymnocheJys or of Dacquemys and thus also to the Moghara skull. I know of no grounds for decision between the two alternatives thus presented. The British Museum skull may as plausibly belong to P. hramlyi as to E. aegyptiaca. We have too little of the Moghara skull, and in addition we are probably too close to the branching off point of Peltocephalus and Erymnochelys to expect wide differences in skull structure between these two, then nascent, genera. But, though we must thus remain undecided as to the species allocation of the Moghara skull, I think that one positive statement of some importance may be made. The Mog- liara skull — on whichever fork of the phyletic tree it belongs — is a structural intermediate between the Recent genera, Pelto- cephalus or Erynmochelys, and the Oligocene Dacquemys. Tills, indeed, is the principal suggestion that I wish to make : that there is a phyletic relationship between Peltocephalus, Erymnochelys and Dacquemys of the sort diagrammed below (Fig. 1). Peltocepholus - Recent of dumeriliono S.Americo Erymnochelys — Recent of modaqoscarensis Modaaosca \ "Podocnemis" bromiyi Miocene of Egypt 1 Erynnnocheivs - oegyptioco Miocene of Egypt and E. Africo Dacquemys — paleomorptio Oligocene of Egypt w ? Docquemys fojumensis Fig. 1 Presumed phyletic position of the Egyptian Mioc-ene podoi-neniidc NO. 39 HREVIORA 7 It cannot be denied that this suogested ])hyh)«i('iiy I'ests u])(in a rather elaborate structure of inferences, any or all of which may be wronj;' ; nothinha, new genus, new species, from the Lower Oligoeene of the P'ayum, Lgy]>t. Breviora, no. 35, ])p. IS. Zangerl, R. 1948. The vertel)rate fauna of the 8elma formation of Alabama. Part II. The pleurodiran turtles. Fieldiana: Geol. Mem., vol. 3, pp. 23-o6. PLATE 1. British Museum skull from Mogliara, dorsal view. Posterior end at top of page. f . ^/ 1^ PLATE 2. British Museuiii f-kiill from Moghara, ventral view. Posterior end at top of E V I O R A Miaseium of Comparative Zoology Cambridge, Mass. November 29, 1954 Number 40 SOME MOLLUSKS FROM THE CONTINENTAL SLOPE OF NORTHEASTERN NORTH AMERICA^ By Arthur H. Clarke, Jr. The continental slope area of northeastern North America has always been a relatively unknown and unexplored region. Onl.v meager information has been on record in regard to the kind and abundance of its fauna. More knowledge is desirable, not only for academic interest, but also that proper evaluation may be made with the ultimate aim of supplementing man's ever- increasing need for additional sources of food. With these objects in mind, in 1952 and 1953 the Woods Hole Oceanographic Institution began a survey of the bottom fauna existing on the continental slope area. An eighty-three foot trawler, the Cap'n Bill II was. equipped for the task, and bottom trawlings were made along the slope from a point east of Cape Charles, Virginia, to a point southeast of Nova Scotia in various depths ranging approximately from 100 to 700 fathoms. Otter trawls were used carrying one and one-half inch mesh at the cod end, and measuring thirty-five feet, fifty feet, and sixty- five feet initially across the mouth. The work was under the direction of Mr. W. C. Schroeder of the Woods Hole Oceano- graphic Institution,. to whom we are indebted for the fine array of mollusks collected. Mollusks were not the primary concern of the survey, and the large size of the mesh prevented retention of the smaller specimens, though many of the larger species were obtained. These were submitted to Dr. W. J. Clench, Curator of Mollusks at the Museum of Comparative Zoology to be added to the research collection. 1 Contribution No. 722 from the Woods Hole Oceanographic Institution. 2 BREVIOBA NO. 40 After identification of the mollusks collected, a search of the literature was made so that the data gained from this survey could be compared with previous records. Two papers by A. E. Verrill (1884 and 1885) were found listing the bathymetric ranges of all of the then known western Atlantic species of marine mollusks found below 60 fathoms. Verrill's records have apparently been generally overlooked by many of the more recent compilers of molluscan lists. His data, as obtained primarily from material gathered by the U.S. Fish Commission steamers Fish Hawk and AlMtross from 1880 to 1884, indicate much more extensive bathymetric ranges for a great many species than have subsequently been given. W. H. Dall, in his "Preliminary Catalog ..." (1889) omitted much of Verrill's data, and although Dall frankly stated that his catalog was incomplete, this oversight has been perpetuated by later writers. It is urged that future investigators refer to Verrill's most complete and excellent list as a supplement to modern lists. In the following tabulation of species collected by the Cap'n Bill II, the minimum depth range indicated by the data is given. For instance, if a particular species was found at two stations where the depths of trawling ranged from 200 to 250 fathoms and from 300 to 350 fathoms respectively, the minimum range would be 250 to 300 fathoms. A distinction is also made between living and dead specimens, and bathymetric ranges are given for each, if not identical. For comparative purposes, the broad- est bathymetric and geographic ranges indicated by the litera- ture are also included, and information sources are cited. When geographic ranges are extended, these localities are defined in terms of miles from a position on land. When not extended, approximate location only is indicated. By examination of the following list of 28 species, it will be seen that these new data have extended the previously recorded bathymetric ranges of seven species (marked *) and the geo- graphic ranges of six species (marked f). List of Mollusca Collected YOLDIA THBACIAEFORMIS (Storer)*t One large dead specimen at station 72. Bathymetric range: 400 fms. dead (this survey) ; 29-182 fms. dead (Verrill 1884). 1954 MOLLUSKS FROM NORTHEASTERN NORTH AMERICA 3 Geographic range : 240 mi. east of Sandy Hook, New Jersey (this survey) ; Gulf of St. Lawrence to Massachusetts (Johnson 1934) ; West Greenland (Thorson 1951). Pecten vitreus (Gmelin) Three living and two dead specimens at station 110. Three living specimens at station 208. Bathymetric range : 250-320 fms. (this survey); 50-800 fms. (Dall 1889). Geographic range: Arctic Ocean (Dall 1889) ; off Newfoundland and Nova Scotia (Johnson 1934) ; south of Nova Scotia (this survey) ; off Martha's Vineyard (Johnson 1934) ; east of Sandy Hook, New Jersey (this survey) ; west Florida [Gulf of Mexico] (Dall 1889) ; Europe; Mediterranean Sea; Africa (all Verrill 1884); west Greenland (Thorson 1951). Anomia aculeata Miiller Living- specimens attached to shells and rocks at stations 14, 19, 38, 53, 110, 131, 173, and 195. Bathymetric range : 145-320 fms. (this survey) ; 4-640 fms. (Verrill 1884) ; 0-80 fms. (Dall 1889). Geographic range : Arctic Ocean to Cape Hatteras, North Carolina (Dall 1889) ; south of Nova Scotia to east of Cape Henlopen, Delaware (this survey) ; Europe (Dall 1889). A very variable species. The normal imbricated or scaly sur- face sculpture is entirely lacking in many individuals and in some entire large lots of specimens. Arctica islandica (Linne) Three single valves at station 1. Bathymetric range: 33 fms. dead (this survey) ; 6-90 fms. (Johnson 1934). Geographic range : east of Montauk Point, Long Island, New York (this survey) ; Arctic Ocean to Cape Hatteras, North Carolina (Dall 1889). Astarte subaequilatera Sowerby Three small living specimens at station 86. Bathymetric range : 230 fms. (this survey) ; 22-410 fms. (Johnson 1934). Geographic range: Labrador to Florida (Johnson 1934); south of Nova Scotia (this survey). Panomya arctica (Lamarck) t One very large, apparently freshly dead specimen at station 38. Bathymetric range: 190 fms. dead (this survey) ; 20-506 fms. 4 BREVIORA NO. 40 dead: 300 fms. alive (both, Verrill 1884). Geographic range: Arctic Ocean to Georges Bank; circumpolar (both, Johnson 1934) ; 170 mi. east of Barnegat Bay, New Jersey (this survey). The specimen cited measures as follows : length 109 mm., height 74 mm., width (valves together) 48 mm. A similar specimen, alive and nearly as large, was taken nearby by the Caryn in 1949 (39°57'N., 70°38'W.), thereby confirming the fact that the species exists alive in this area. Calliostoma BAiRDi Verrill and Smith One dead specimen at station 83. One living specimen at sta- tion 146. Bathymetric range : 87 fms. dead ; 65 fms. alive (both this survey) ; 56-640 fms. dead; 64-192 fms. alive (both, Verrill 1884). Geographic range : south southeast of Nantucket (270 mi. east of Barnegat Bay, New Jersey) ; east of Cape Henlopen, New Jersey (both, this survey) ; south of Martha's Vineyard to Flor- ida Keys (Johnson 1934); Caribbean Sea (Verrill 1889). The living specimen is very beautiful. Patches of lavender are seen through the thin iridescent periostracum which appears golden between the spiral rows of tubercules on the upper whorls and golden peach on the lower whorls. The inner nacre of the aperture shines with iridescent pink and green. The circular operculum is multispiral, typical of the family. SOLARIELLA OTTOI (Philippi) One dead specimen at station 72. Three living specimens at station 211. Bathymetric range : 400 fms. dead ; 600 fms. alive (both, this survey) ; 64-1555 fms. (Dall 1889). Geographic range : south of Martha's Vineyard and Georges Bank (Johnson 1934) ; east of Sandy Hook, New Jersey (this survey); New Jersey; Virginia (both, Dall 1889); Florida Keys; West Indies (both, Johnson 1934) ; St. Thomas, Virgin Islands (Dall 1889). Natica clausa Broderip and Sowerby One very large dead specimen (23 mm. long) at station 182. Bathymetric range: 280 fms. dead (this survey) ; 13-1255 fms. dead; 238-843 fms. alive (both, Verrill 1884); 16-1537 fms. (Johnson 1934). Geographic range: Arctic [Ocean] (Verrill 1884) ; Labrador to oif North Carolina (Johnson 1934) ; south- 1954 MOLLUSKS FROM NORTHEASTERN NORTH AMERICA 5 east of Cape Sable, Nova Scotia (this survey) ; Europe (Verrill 1884) ; Arctic and Bering seas to San Diego, California ; Japan (both. La Rocque 1953). PoLiNicES HEROS (Say) One dead specimen at station 144. Bathymetric range : 150 fms. dead (this survey) ; 0-238 fms. (Dall 1889). Geographic range : Gulf of St. Lawrence to North Carolina (Johnson 1934) ; east of Chincoteague Island, Virginia (this survey). POLINICES TRISERIATA (Say)* One very large, dead specimen (33 mm. long) at station 144. One dead specimen at station 200. Bathymetric range : 100-150 fms. dead (this survey); 0-63 fms. (Dall 1889). Geographic range : Labrador to Cape Hatteras (Dall 1889) ; south of Georges Bank to east of Chincoteague Island, Virginia (this survey). Capulus ungaricus (Linne) One small living specimen attached to Pecten vitrevs at station 110. Bathymetric range-. 320 fms. alive (this survey) ; 1-458 fms. (Johnson 1934). Geographic range: south southeast of Cape Sable, Nova Scotia (this survey) ; Greenland to Florida (John- son 1934) ; Europe (La Rocque 1953). Aporrhais occidentalis (Beck)* Living specimens at stations 95 and 177. Dead specimens at stations 38, 83, 86, 95, 164, 176, 182, and 184. Bathymetric range : 340-360 fms. alive, 90-365 fms. dead (both, this survey) ; 115-349 fms. alive; 341/2-1000 fms. dead (both, Verrill 1884). Geographic range : Gulf of St. Lawrence to off North Carolina (Johnson 1934) ; southeast of Nova Scotia to south of Nantucket (this survey); west Greenland (Thorson 1951). BucciNUM ciLiATUM Bruguierc One small living specimen at station 184 (southeast of Cape Sable, Nova Scotia) . Bathymetric range : 265 fms. alive (this survey) ; 26-471 fms. (Johnson 1934). Geographic range: Arctic [Ocean] (Verrill 1884) ; Labrador to Cape Cod, Massachusetts (Johnson 1934) ; southeast of Cape Sable, Nova Scotia (this survey); Europe (Verrill 1884). 6 BREVIORA NO. 40 Neptunea decemcostata (Say)*t Living specimens at stations 88, 95, 110, 165, 173, 182, and 184. Dead specimens at stations 38, 72, 86, 104, 164, 187, and 195. Bathymetric range -. 270-360 fms. alive ; 225-660 f ms. dead (both this survey) ; 41-86 fms. alive; 6-322 fms. dead (both, Ver- rill 1884). Geographic range: Circumboreal (La Roeque 1953) ; Nova Scotia to Massachusetts Bay and Georges Bank (Johnson 1934) ; southeast of Nova Scotia to 240 mi. east of Barnegat Bay, New Jersey (south of Nantucket) (this survey). Great variation of form exists in this species. Some specimens are attenuated to the extent that the length is 2i/^ times the great- est width. Other specimens show a length of only li/^ times the width. Variation is also seen in degree of sculpture, length of aperture in relation to overall length, etc. The possibility of the existence of one or more clines will be investigated. Neptunea despecta tornata (Gould) *t One small, dead specimen at station 159. Bathymetric range : 690 fms. dead (this survey); 10-471 fms. (Johnson 1934). Geographic range : Gulf of St. Lawrence to off Martha 's Vine- yard (Johnson 1934) ; 120 mi. east of Atlantic City, New Jersey (this survey). CoLus iSLANDicus (Gmeliu) Dead specimens at stations 5, 22, 23, 29, 84, 144, 201, and 225. Bathymetric range : 85-415 fms. dead (this survey) ; 20-1650 fms. (Dall 1889). Geographic rayige: Labrador to Norway (Johnson 1934); Arctic Sea to South Carolina (Dall 1889), south of Georges Bank to east of Chincoteague Island, Virginia (this survey ) . CoLus STiMPsoNi STiMPSONi (Morch) Two living specimens at station 152. One dead specimen at station 201. Bathymetric range : 330 fms. alive ; 85-87 fms. dead (both, this survey); 1-471 fms. (Johnson 1934). Geographic range: Arctic Sea to [Cape] Hatteras, [North Carolina] (Dall 1889) ; south of Georges Bank to east of Cape Ilenlopen, New Jersey (this survey). CoLus pitbescens (Verrill)* Living specimens at stations 10, 13, 17, 53, 72, 139, 182, and 1954 MOLLUSKS FROM NORTHEASTERN NORTH AMERICA 7 184. Dead specimens at stations 8, 10, 14, 17, 19, 35, 49, 83, 95, 104, 114, 131, 132, 139, 144, 164, 173, 184, and 225. Bathy metric range: 130-400 fms. alive; 90-415 fms. dead (both, this survey) ; 18-179 fms. dead; 192-640 fms. alive (both, Verrill 1884). Geo- graphic range : Gulf of St. Lawrence to North Carolina (Johnson 1934) ; Nova Scotia to South Carolina (Ball 1889) ; southeast of Nova Scotia to east of Chincoteague Island, Virginia (this sur- vey). Pleurotomella agassizi Verrill and Smithf One dead specimen at station 211. Bathymetric range : 600 fms. dead (this survey); 39-1309 fms. alive; 1608 fms. dead (both, Verrill 1884). Geographic range-. 105 mi. southeast of Nantucket (this survey) ; south of Martha's Vineyard to the West Indies (Johnson 1934). Scaphander piinctostriatus (Mighels and Adams) One living specimen at station 109 and one living specimen at station 189. Bathymetric range: 240-305 fms. alive (this sur- vey); 46-1255 fms. alive; 1362-1467 fms. dead (both, Verrill 1884). Geographic range: Gulf of St. Lawrence to West Indies (Johnson 1934) ; southeast of Nova Scotia (this survey) ; Bar- bados; Norway (both. Ball 1889). Rossia sublaevis Verrill One living specimen at station 69. Bathymetric range : 490 fms. (this survey) ; 45-640 fms. (Johnson 1934). Geographic range: Newfoundland to 32°33'15"N. (Johnson 1934) ; south of Georges Bank (this survey). Illex illecebrosa (Lesueur) Two living specimens at station 68 and one living specimen at station 74. Bathymetric range: 450 fms. (this survey) ; 0-1022 fms.; beaks 109i-1917 fms. (both, Verrill 1884). Geographic range: south of Georges Bank (this survey) ; Greenland to Cape Hatteras (Johnson 1934). Chiroteuthis lacertosa Verrill One living specimen at each of the following stations : 68, 69, and 74. Bathymetric range: 490-500 fms. (this survey); 435- 2369 fms. (Johnson 1934); arms, 2949 fms. (Verrill 1884). 8 BREVIOBA NO. 40 Geographic range: south of Georges Bank (this survey) ; Nova Scotia to West Indies (Johnson 1934). Mastigoteuthis agassizii Verrill* One living specimen at station 102. Bathymetrie range : 600 fms. (this survey); 640-1050 fms. (Johnson 1934). Geographic range: Gulf of Maine to North Carolina (Johnson 1934) ; south- east of Nova Scotia (this survey). Alloposus mollis Verrillf One living specimen at station 69 (south of Georges Bank) and one living specimen at station 90 (southeast of Nova Scotia). Bathymetric range: 330-490 fms. (this survey) ; 238-1346 fms.; fragment, 1735 fms. (both, Verrill 1884). Geographic range: 110 mi. southeast of Cape Sable, Nova Scotia ; south of Georges Bank (both, this survey) ; Nantucket to Chesapeake Bay, eastern Atlantic (both, Johnson 1934). Bathypolypus arcttcus (Prosch) One living specimen at station 69 and one living specimen at station 72. Bathymetric range: 450-490 fms. (this survey) ; 28- 843 fms. (Johnson 1934). Geographic range: south of Georges Bank (this survey) ; Bay of Fmidy to 32°N. (Johnson 1934). Graneledone verrucosa (Verrill) One living specimen at station 69. Bathymetric range : 490 fms. (this survey); 466-1255 fms. (Johnson 1934). Geographic range : south of Georges Bank (this survey) ; Nova Scotia to Delaware Bay (Johnson 1934). Location and Depth of Stations Cited Station Number North Latitude West Longitude Depth in Fathoms 1 40°50' 71°09' 33 5 39°35' 71°57' 240-260 8 39°36' 71°52' 405-410 10 39°46' 71°35' 395-405 13 39°55' 71°27' 180-190 14 39°56' 71°26' 150 17 39°55' 71°17' 290-330 19 39°57' 71°13' 175-180 22 39'53' 70°53' 300-330 1954 MOLLUSKS FROM NORTHEASTERN NORTH AMERICA 90 95 102 104 109 110 114 131 132 139 144 146 152 159 164 165 173 176 177 182 184 187 189 195 200 201 208 211 225 39°52' 70°51' 375-420 39°52' 70°43' 415-440 40°02' 70°24' 105-110 40°04' 70°12' 190-225 39°57' 69°37' 415-560 39° 55' 69°26' 165-180 40°08' 68°19' 400-510 40°10' 68°16' 490 40°12' 68°08' 400-450 40°17' 67°53' 450-500 39°59' 69°32' 87-90 39°59' 69°35' 82-85 42°23' 64°58' 230-245 42°46' 63° 22' 340-350 42"41' 63°33' 300-330 42'45' 63°47' 330-340 42°35' 64°03' 560-600 42°40' 64°08' 350-380 42°20' 65°03' 305-320 42°17' 65°06' 320-360 40°46' 66°40' 290-300 38°13' 73°40' 145-160 38°08' 73° 45' 205-275 37°38' 74°14' 120-130 37°45' 74°09' 150 38°33' 73°18' 63-65 38°39' 73° 05' 330-400 39°26' 72°10' 690-720 42°43' 63° 50' 310-335 42°42' 63°47' 360-370 42°40' 64°10' 240-270 42°33' 64°17' 280-320 42°32' 64°19' 360-420 42°28' 64°31' 280-305 42°23' 64°52' 265-295 42°15' 64° 58' 660-705 42°18' 65°05' 220-240 40°34' 67°02' 290-300 40°46' 66°48' 100-105 40°45' 66°51' 85-87 40°09' 68°24' 250-340 40°00' 68°49' 600-670 39°53' 70°40' 345-355 10 BREVIORA NO. 40 REFERENCES Dall, W. H. 1889. A preliminary catalogue of the shell-bearing marine mollusks and brachiopods of the southeastern coast of the United States, with illustrations of many of the species. Bull. U. S. Nat. Mus. 37: 1-221, 74 pis. Gould, A. A. 1870. Report on the Tnvertebrata of Massachusetts. Edited by W. G. Binney. Boston. Pp. i-v + 1-524, 12 pis. Johnson, C. W. 1934. List of marine Mollusca of the Atlantic Coast from Labrador To Texas. Proc. Boston Soc. Nat. Hist., 40. no. 1 : 1-203. La Rocque, Aurele 1953. Catalogue of the Recent Mollusca of Canada. Nat. Mus. Canada 129: i-i.x + 1-406. NicoL, David 1951. Recent species of the veneroid pelecypod Arctica. Jour. Wash- ington Acad. Scl, 4L no. 3 : 102-106. Thorson, Gunnar 1951. The Godthaab expedition 1928. Meddel. om Greailand, Copen- hagen, 8Lno. 2: 1-117. Tryon, G. W., Jr. 1879. Manual of Conehology (1), 1. Cephalopoda. Acad. Nat. Sci., Philadelphia, pp. 1-31G, 112 pis. Verrill, a. E. 1884. Second catalogue of Mollusca recently added to the fauna of New England and adjacent parts of the Atlantic, consisting mostly of deep sea species, with notes on others previously recorded. Trans. Conn. Acad., 6, no. 1 : 139-294, 5 pis. 1885. Third catalogue of Mollusca recently added to the fauna of New England. . . . Trans. Conn. Acad., 6. no. 2 : 395-452, 3 pis. 1954 MOLLUSKS FROM NORTHEASTERN NORTH AMERICA 11 - 4- I E V I O R A Mmseitim of Comparative Zoology Cambridge, Mass. December 17, 1954 Number 41 COMMENTS ON THE CLASSIFICATION OF RODENTS By Albert E. Wood Biology Department, Amherst College There is no order of mammals where there are so many un- certainties in the current classification as there are in the rodents. The rodents have generally been subdivided into three groups, the Sciuromorpha, Myomorpha and Hystricomorpha, although there have been several attempts (notably Miller and Gidley, 1918, and Winge, 1924) to break away from the tripartite division. The morphologic criteria on which these three groups are based permits living rodents to be marshalled into them with greater or lesser success. But the resulting problems remain very prominent, and well-nigh insuperable when the paleontolo- gie history of the order is taken into account. For example, if the Hystricomorpha are a natural group, how can their present distribution (South America, Africa, southern Europe and south Asia, with scattered recent Latin American immigrants to North America) be explained in view of the complete absence of Tertiary hystricomorphs in North America? Moreover, the morphology of the various genera and families of fossil rodents does not permit these forms to fit into a three-fold classification anywhere near as well as do the living members of the order. As long as we deal only with the living members of the order, the criteria for subordinal classification can be sorted out fairly well. However, studies of fossil rodents have shown that sometimes the structures on which the classification has been based were independently derived by diverse groups. For example, Wood (1937) showed that the sciuromorph type of zygoma and masseter was developed in the Oligocene genus Titanotheriomys, a form that by no stretch of the imagination 2 BREVIORA No. 41 could be considered to be ancestral to anything included among the recent sciuromorphs. As a result of general dissatisfaction with the three-fold classi- fication, there have been a number of attempts to arrive at alternative arrangements. Those of Miller and Gidley (1918) and Winge (1924) developed so many additional problems, while solving a few, that they are not generally followed at present. Wood (1937) basically followed the three-fold arrangement, but proposed a fourth subdivision, the Protrogomorpha, to include the primitive rodents that have not acquired the features of the jaw muscles that characterize the members of the three classic suborders. All members of this suborder are extinct except for Aplodontia, the "Mountain Beaver" of parts of California, Oregon, Washington and British Columbia. Simpson (1945) made a valiant attempt to squeeze all rodents into the three-fold classification, though pointing out in many places that the scheme does not work well. In an extensive review of the early Tertiary rodents of North America, Wilson (1949) followed Simpson's classification but raised many queries and made a number of very significant sug- gestions, particularly as to the probability of a close relationship between the heteromyids and cricetids, heretofore generally con- sidered to be sciuromorphs and myomorphs, respectively. Simpson (1950) and Wood (1950) suggested that the South American "hystricomorphs" should perhaps be separated as a distinct suborder, though without working out the details. Wood (1955), and Wood and Patterson (in press) are formally carry- ing out such a separation, with detailed discussion as to the reasons for the change. Lavocat (1951), in a review of certain Oligocene faunas of France, agreed with this point of view, and showed that the European Eocene and Oligocene Theridomyidae and Pseudosciu- ridae could not be ancestral to either the Old World or New World hystricomorphs of the present day. For many years, Stehlin worked on a monograph of the evolu- tion of rodent dentition, which was finished after his death by Schaub (Stehlin and Schaub, 1951). This deals exclusively with the dentition, and develops the thesis that rodents with five-crested teeth are all descended from animals with the basic tooth pattern represented by the European Eocene-Oligocene 1954 COMMENTS ON THE CLASSIFICATION OF RODENTS 3 theridomyids. Although this monograph contains a brief taxo- nomic section, there is no attempt to present a comprehensive revision of the classification of the order. There are, however, a large number of suggestions and comments which, while not building up any particular classification, must nevertheless be given serious consideration by anyone interested in the super- generic arrangement of the rodents. I cannot agree with all of Schaub's conclusions, and am therefore attempting to explain some of these differences in detail here. The taxonomic results of these differences of opinion are being published elsewhere as a suggested classification of the rodents (Wood, 1955). Schaub has recently (1953a) published a revised classification of the ' ' Hystricomorph " rodents. In this paper and others (1951, 1953b; Stehlin and Schaub, 1951), he also makes com- ments on the "Sciuromorpha" and "Myomorpha". Basically, his idea is that these last two groups are wastebaskets, made up of superfamilies that have had an entirely independent history, and which cannot profitably be grouped in suborders. Although he is not explicit, he gives the impression that he feels that, in one or two cases, there might be some profit in grouping two of the superfamilies into a suborder. With this part of his thesis I feel a great deal of sympathy if not complete agreement, although the classification proposed in Wood, 1955, does not completely follow Schaub in doing away with these suborders. The rodents representing the initial North American Eocene radiation (the Isehyromyoidea), the middle Tertiary to Recent Aplodontoidea, and the Sciuridae certainly seem fairly closely related. At least they all retain primitive features of teeth, of the zygomasseteric region, or of both, and they may Avell be kept in a common suborder, for which the name Sciuromorpha may be retained. The other superfamilies usually called sciuromorphs seem to be quite distinct from these forms, and may belong- in rather diverse parts of the order. The families that Wood places in the Myomorpha, however, are a group about whose relationships there is little or no cer- tainty. There are some suggestions of mutual relationship, and some indications that Schaub is correct in thinking that they have nothing in common with each other. If they are not related, this would be the end of the Myomorpha, and the superfamilies would have to be either left out of all suborders or, since this is 4 BREVIORA No. 41 illegal in formal taxonomy, each placed in its own suborder. It is therefore obvious that Schaub is correct in his contention that the classical concepts of the Sciuromorpha and Myomorpha will have to be abandoned, and either new and perhaps more restricted definitions developed for these suborders, or the sub- ordinal concept will have to be largely abandoned for this order. It is still too early to reach a definite choice between these alternatives. But on one of Schaub 's major points I find myself in disagree- ment. One of the basic tenets of Stehlin and Schaub 's invalu- able monograph is that the Theridomyidae are ancestral to all forms that have a five-crested tooth pattern or that are derived from a five-crested ancestor. Schaub has therefore proposed a suborder, Pentalophodonta, to include all forms with five-crested teeth, or that are derivable from a five-crested ancestry. How- ever, it seems no more reasonable to assume that a given tooth pattern has not arisen independently several times than to as- sume that a given zygomasseteric pattern has not so arisen. Indeed, all the current indications from the study of rodent evolution are that extensive parallelism is the rule rather than the exception in this order. Of Schaub 's Pentalophodonta, the South American forms are surely a natural group, the Suborder Caviomorpha of Wood (1955), and the Infraorder Nototrogo- morpha of Schaub (1953a). But these forms must have been derived from rodents that entered South America from North America, as is now generally agreed (Schaub, 1953a; Simp- son, 1950; Wood, 1950). There is also agreement that this entry must have been no later than late Eocene. Schaub (1953a, p. 393) argues that this ancestral stock could have been theridomy- ids that crossed North America to reach South America in the Eocene, Paleocene, or even earlier. It is impossible to prove that a form did not live in an area or at a time where it is not represented by fossils. But no rodents of any kind are known from South America before the Lower Oligocene Deseadan, at which time they have become quite abundant and are beginning to be diversified, although clearly stemming from a common origin. Because of the rapid diversi- fication of the caviomorph rodents beginning with the Deseadan, it seems almost certain that, in this deposit, we are observing an explosive radiation in the process of exploding. A backward 1954 COMMENTS ON THE CLASSIFICATION OP RODENTS 5 extrapolation of the post-Deseadan rate of evolution would indi- cate that this explosion must have begun not very long prior to the deposition of the Deseado beds. Although such an extra- polation is dangerous, and although we obviously cannot be certain of the reasons for such an evolutionary explosion, by far the most reasonable interpretation, in view of what is known of the general Tertiary history of South America, is that the explosion began with the arrival of rodents in South America. Here was a virgin field, with no highly developed competition, and, once the rodents were established, they rapidly took over all gnawing niches. If this explanation is not correct, it is exceed- ingly difficult to understand the rapid post-Deseadan develop- ment of the caviomorphs. Hence it seems very questionable that rodents could have been present in South America for any appreciable time before the Deseado. If this postulate is ac- cepted, and if their North American origin be granted, most authors consider it more probable that the caviomorphs were derived from known North American types (such as the Par- amyidae or Sciuravidae) than from the European Theridomyi- dae, still unreported from any of the richly fossiliferous North American deposits. An additional objection to Schaub's suggestion of a Paleocene or even earlier passage of the theridomyids to South America is the complete absence of any rodents from the record prior to the latest Paleocene, when only the single genus Paramys is known. Presumably there must have been rodents prior to their earliest appearance in the record, and they may have been in existence for some time, although there is no valid criterion for fixing the length of time necessary to differentiate the rodents from their ancestors. This time may therefore be fixed as long (extending back into the Cretaceous) or short (reaching only the Middle Paleocene), depending on the individual student's opinion as to the probable rates of evolution involved and the significance of the morphologic differences between Paramys and whatever group is postulated as the ancestor. Moreover, all the earliest rodents, both in North America and in Europe, are referable to the Paramyidae, and to the most primitive stages of that family. The only non-paramyid lowermost Eocene genus known is the pseudosciurid Adelomys, which Wood, in his 6 BREVIORA No. 41 monograph of the Paramyidae, will indicate was derived from the paramyids. The absence of theridomyids from North America is, of course, negative evidence. It is possible that this family may have lived in eastern North America where there are no continental Eocene deposits. But there is no evidence that the family was so repre- sented, and it seems highly probable that the rodents that reached South America went by way of what is now Mexico and Central America, and hence were probably derived from the western part of the continent. Therefore, while admitting that all this is negative evidence, it seems most probable that a representative of one of the known North American Eocene groups reached South America late in the Eocene, to give rise to the South American radiation. A further point of disagreement with Schaub is in the phylo- genetic position of the Paramyidae. Wood (1946) referred to the specialization of the Paramyidae, which indicated that they could not have the fundamental ancestral position generally given them. With this point of view, Schaub is in complete agreement (1953b, p. 35), considering that the Sciuridae have the most primitive tooth pattern in the order. Wood's 1946 opinion was based on preliminary investigation of the paramyids, which includes many rather specialized genera. Further study of Lower Eocene forms, however, has caused a revision of this opinion, and I now feel that the Upper Paleocene and Lower Eocene paramyids are the most primitive known rodents, al- though it is certain that the earliest sciurids have made little or no advance over them in the structure of the cheekteeth. The data for this conclusion have not been published, but are in- cluded in a monograph on the Paramyidae, now nearing comple- tion. Therefore, the North American Paramyidae could have been ancestral to the South American radiation, although the intermediate steps have not been traced in detail. If the para- myids were the ancestors, Schaub 's Suborder Pentalophodonta immediately becomes a composite group, the two infraorders having nothing in common. Schaub places the Theridomyoidea, Castoroidea and Hystri- coidea in the Infraorder Palaeotrogomorpha. But Lavocat (1951) has presented evidence, derived from structures other 1954 COMMENTS ON THE CLASSIFICATION OF RODENTS 7 than the cheekteeth, that the theridomyoids cannot be related to the hystricoids, so that this infraorder would also seem to be composite. Schaub places a number of other groups in the Palaeotrogomorpha, particularly the Spalacidae and Rhizo- myidae. This seems perhaps somewhat less likely than that they belong in the Myomorpha, if there is any such group. Certainly there is not enough known of the history of either family to justify placing them with confidence in any particular category. Schaub 's new proposals for classification of rodents are very useful and thought-provoking. The fact that there are disagree- ments with them is not surprising. The current status of rodent phylogeny and classification is such that anyone can point out inconsistencies in anybody else's classification. The rodents are a sufficiently large and complex group so that no one person can be familiar with the entire order. Progress will be made only if various students attack the problem, even though it will result in presenting solutions that others can show obviously to be incorrect. Wood's current (1955) classification has been pro- posed with this object in view, in the hope that there will be as many criticisms, both constructive and destructive, as possible, since this is the only way to progress. Schaub has made one further suggestion that very probably will be incorporated in the ultimate classification of the order. This is the idea that many of the superfamilies are entirely separate from all others, presumably having an independent history since the Eocene. There very clearly was a sudden diversification of the rodents, at or near the end of the Eocene, which apparently resulted from a rather sudden change from gnawing based on the use of the temporal muscle to that based largely on the use of the masseter, with consequent enlargement and modification of the latter. The present evidence suggests, but does not prove, that this occurred a great many times inde- pendently. It is very probable that these lines all developed from different groups of Eocene rodents, and that, therefore, each of these lines will deserve separate taxonomic treatment. Generally speaking, such superfamilial lines could be grouped into sub- orders only artificially. Although the present evidence suggests the probability of such a bushy phylogeny (rather than a tree with half a dozen main branches), it is perhaps more conserva- 8 BREVIORA No. 41 tive to struggle with an attempt to group rodents into suborders for a while longer, while keeping in mind the possibility (or probability) that a classification with entirely independent superfamilies will ultimately prove to be the only one that will fit in with the facts of rodent paleontology, as these gradually come to light. Certainly it will be some time before an adequate classification of the order is compiled. EEFERENCES Lavocat, E. 1951. Revision de la faune des mammiferes oligocenes d'Auvergne et du Velay. Editions "Sciences et Avenir," Paris, 153 pp. Miller, G. S., and J. W. Gidley 1918. SjTiopsis of the supergeneric groups of rodents. Jour. Wash. Acad. Sci., 8: 431-448. SCHAUB, S. 1951. Review of: H. G. Stehlin et S. Schaub, "Die Trigonodontie der simplicidentaten Nager. " Mammalia, 15: 204-210. 1953a. Remarks on the distribution and classification of the "Hystri- comorpha." Verh. Naturf. Ges. Basel, 64: 389-400. 1953b. La Trigonodontie des Rongeurs simplicidentes. Ann. Paleont., 39: 29-57. Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85: XVI + 350 pp. 1950. History of the fauna of Latin America. Amer. Scientist, 38: 361-389. Stehlin-, H. G., and S. Schaub 1951. Die Trigonodontie der simplicidentaten Nager. Schweiz. Pal. Abh., 87: 1-385. Wilson, R. W. 1949. Early Tertiary rodents of North America. Carnegie Inst. Wash. Publ., 584: 67-164. WiNGE, H. 1924. Pattedyr-Slaegter, vol. 2. Rodentia, Carnivora, Primates. Copen- hagen, H. Hagerups Forlag, 321 pp. 1954 COMMENTS ON THE CLASSIFICATION OF RODENTS 9 Wood, A. E. 1937. The mammalian fauna of the White Eiver Oligocene by W. B. Scott and G. L. Jepsen. Part II. Rodentia. Trans. Amer. Phil. Soc, 28: 157-269. 1946. Early Tertiary rodents of the Family Paramyidae. Bull. Geol. Soc. Amer., 57: 1245 (Abstract). 1950. Porcupines, palaeogeography, and parallelism. Evolution, 4: 87-98. (In Press.) A revised classification of the rodents. Jour. Mammalogy, Wood, A. E., and B. Patterson (In Press.) The rodents of the Deseadan Oligocene of Patagonia and the beginnings of the South American rodent radiation. Chicago Nat. Hist. Mus. E V I O R A MeseiiMM of CoaMparative Zoology Cambridge, Mass. February 28, 1955 Number 42 A NEW SALAMANDER OF THE GENUS PARVIMOLGE FROM MEXICO By GrEOEGE B. Rabb Museum of Zoology, University of Michigan 111 the Museum of Comparative Zoology there is a specimen of a Mexican plethodontid salamander superficially similar to in- dividuals of Pseiidoeurycea cephalica. Examination indicates that it actually represents an undescribed form of the genus Parvimolge Taylor, which may be known as Parvimolge praecellens, new species (Figure 1) Type. An apparently immature female, Museum of Compara- tive Zoology No. 24701, collected at Hacienda el Potrero, near the city of Cordoba, Vera Cruz, Mexico, on December 24, 1940, by Archie F. Carr. Diagnosis. A small plethodontid salamander related to Parvi- molge townsendi by virtue of syndactylous feet, with the par- tially free digits having pointed tips, and by the presence of conspicuous though small glands about the middorsal line. It differs from P. townsendi in larger body size, greater number of teeth, and in having small nostrils. It differs from the only other species assigned to this genus, richardi, not only in these charac- ters but also in having conspicuous dorsal glands. Descr-ipiion. General aspect of body robust. Length of tail approximately six-sevenths of distance from snout to posterior end of vent. Head length (snout to center of gular fold) con- tained about four times in snout- vent distance ; maximum head BREVIORA NO. 42 ^) t! 1955 A NEW SALAMANDER OF THE GENUS PARVniOLOE 3 width contained in same about five times. Head width three- fourths the snout-gular fold distance and nearly equalling snout- posterior angle of jaw length. Width of an eyelid less than minimum distance between eyelids. Horizontal diameter of eye greater than one and a half times snout-anterior angle of eye distance. Internarial distance four-sevenths the interorbital measurement. Viewed dorsally, the snout is gently rounded, though short in relation to anterior angles of eyes. Canthus rostralis not sharply defined, the area between anterior mid- dorsal line of head and edge of upper jaw being a gradual slope. Nostrils small with nasolabial grooves forking on weakly de- veloped nasolabial protuberances. Snout not projecting beyond tip of lower jaw. Lower jaw semi-oval in outline, ventrally viewed. Lateral extensions of gular fold meet grooves extending from posterior angle of eye and continue dorsally to join the posterior ends of a V-shaped groove originating parietally. Epi- branchial ridge extending beyond axillary groove. A deep furrow, perhaps partially due to preservation, follows the mid- dorsal line from scapular region onto tail. The fairly thick skin of dorsum has about fifty small but conspicuous glands (pre- sumably of the "poison" type). These glands appear to be arranged in four irregular longitudinal rows, two of them very close to midline. On the neck the middorsal rows fork and curve laterally. There are a few conspicuous glands on the tail. Maxillary-premaxillary teeth, counting both sides, total 75. They are practically all of uniform size, those in the premaxil- lary position not piercing the lip. Eighty-seven mandibular teeth. Prevomerine teeth 20 on right and 23 on left side mostly in double rows which arch very slightly posteriad in approaching the midline from initial points lateral to the small choanae. Distance between prevomerine series at midline is equal to three times the diameter of a choana. Parasphenoid teeth in two ob- long patches, each of which has about 70 teeth arranged in diag- onal rows. Anteriorly the groups are separated by slightly more than one choanal diameter, posteriorly by about three diameters. Tongue large and free, the fleshy sublingual fold anterior to it well developed. The moderately chunky body has twelve costal spaces. It is difficult to ascertain whether there are twelve or thirteen costal 4 . BREVIORA NO. 42 grooves because of the contorted condition of the specimen. There are about twenty-two caudal grooves indicated posterior to the vent. Tail quadrangular in cross-section, slightly broader dorsally than ventrally. Throughout most of its length, it is deeper than broad. Limbs slender and moderately long, hind liml>- slightly larger and longer than forelimb. Limbs fail to meet by about one costal space when pressed to sides of the body. Hands and feet well developed but not expansive, with slight pads ventrally. The someM'hat slender, terminally free digits have pads which are discrete from the palmar pads. Innermost and outermost digits of hand and foot completely enclosed in web. Terminal phalanges of second and fourth toes and second finger free of web, while third toe and finger have almost two phalanges free. Third and fourth toes and second and third fingers noticeably pointed distally, while the second toe appears more rounded. Toes in order of decreasing length : 3, 4, 2, 5, 1 ; fingers in same order: 3, 2, 4, 1. In alcohol, the ground color of dorsal surfaces of head, body, tail, and limbs is brownish-black. Ventral surfaces show a faded version of this color with some areas being almost completely unpigmented (gular fold, anterior part of chin, and palms and soles). Underside of tail light brown. Jaw margins appear mottled due to lack of pigment in spots. Bluish- white spots and blotches on lower sides of body and on tail were possibly white in life. A distinct but broken bar of this color on the head between the eyes. Measurements (in millimeters). Snout-posterior end of vent, c. 35; tail, c. 30; snout-gular fold, 9.4; snout-angle of jaw, 7.5; snout-anterior angle of eye, 1.6; horizontal diameter of eye, 2.6; head width, 7.0; internarial distance, 2.0; interorbital distance, 3.5; arm, 8.9; leg, 9.4. Remarks. Before commenting on relationships I wish to place on record some information on P. townsendi not obtainable from previous accounts (Dunn, 1922 and 1926; Taylor, 1944). The major source of the data is a series of ten specimens collected at the same time and place as the type of P. praecellens. Varia- tional data for the eight alcoholic adults (MCZ 24712-19) are given in Table I. The males have two or three enlarged premax- 1955 A NEW SALAMANDER OF THE GENUS PARVIMOLGE 5 illary teeth, prominent mental glands, and well developed naso- labial protuberances, with more truncate snouts and with larger external narial openings than the females (average 0.3 mm. versus 0.2 mm.). The parasphenoid teeth groups are separated in half of the eight specimens. Counts of prevomerine tooth rows averaged 5.8 on a side. One male, not included in the aver- age, has these teeth in an irregular patch of ten on each side. UMMZ 111305, an immature specimen (snout-vent, 15 mm.; tail, 13 mm.) from this series, has fewer teeth (prevomerine, 3-5; maxillary-premaxillary, 34; mandibular, 47) and fewer glands (31) than any of the adults of Table I. A female, UMMZ 63944, taken at Cordoba, Vera Cruz, on June 19, 1927, is larger than any recorded specimen, measuring 24 mm. from snout to vent, with a tail of the same length. The ovarian eggs are large, about 0.7 mm. in maximum diameter. Many minor variations in color were found. The ground color was diverse shades of brown, usually lighter ventrally, and sometimes dorsally, than on the sides. The dark bars described by Dunn are most pronounced on the tail, with the most con- spicuous bar or ring occurring at the weakly marked basal con- striction of the tail. A herringbone or V pattern can be made out dimly in some specimens held under water; it is apparently the result of concentration of dark pigment in the costal grooves. Most of the specimens have an incomplete light bar between the eyes well indicated, but two do not. Three have some evidence of an additional incomplete anterior interocular bar. In two the rostral region appears as a brilliant white square, due to absence of dark pigment, Avhich is slightly reduced in this area in the other specimens. The limbs, throat, and venter have scattered light, probably pigmentless, spots. Light streaks on the sides are well developed in some, practically absent in others. In one specimen the middorsal area is light brown, contrasted with dark brown dorsolaterally, which in turn is set off by a double row of light streaks or spots on the lower sides. In the same specimen there are striking dorsolateral white patches (apparently pigment- less areas) in the nuchal region. The tenth specimen from the Potrero series, an adult male measuring about 23 mm. from snout to vent, was stained and cleared (UMMZ S-1556). The testes of this specimen were 6 BREVIORA NO. 42 relatively enormous and were in the bilobate condition which I have found to characterize the larger, and presumably older, classes of males in Chiropterotriton and Pseudoeurycea. The more interesting skeletal features are described here. Skull broadly elliptical, maximum width at the otic capsules three- fourths of the premaxillary-oecipital condyle distance ; f rontals and parietals ossified completely to midline ; facial lobe of maxilla large, forming with nasal bone a posterior border of nasal cap- sule ; nasal fairly prominent, angling forward medially, about two and a half times as long as broad ; prefrontal two-thirds the length and one-half the width of nasal, not participating in the margin of the nasal capsule ; basally fused frontal processes of premaxilla diverge considerably forming a wide fronto-premaxil- lary fontanelle (a tenth of otic capsular width) ; these processes laterally compressed anteriorly ; more posteriorly and dorsally they gradually twist until they appear horizontally flattened where the ends articulate with grooved projections of the frontal bones ; two large premaxillary teeth, twenty-three teeth on each maxilla, twenty-eight teeth on each of the well ossified mandibles, eight prevomerine teeth on a side; no median suture between prevomers ; parasphenoid teeth in two patches. Skull measure- ments (mm.) made with ocular micrometer: otic width, 2.9; premaxillary-condylar length, 4.4 ; prefrontal length, 0.5, width, 0.15 ; nasal length, 0.75, width, 0.3 ; fronto-premaxillary fonta- nelle width, 0.3. Hyoid structures as described by Tanner (1952) , but second ceratobranchials definitely osseous or calcified, which is also the condition in the only cleared individual avail- able of Lineatriton lineola (UMMZ S-1594). Phalangeal for- mula of foot, 1, 2, 3, 3, 2 ; of hand, 1, 2, 3, 2 ; eight carpals, ulnare and centrale partially fused on l)oth hands, on left hand all ossified to some extent but on right hand radiale and carpal im- mediately distal to it are cartilaginous. At least seven, probably nine tarsals, four ossified on each side. Atlas weak, not ossified middorsally ; only the second vertebra with a noticeable neural ridge. Long, proximally forked ribs present from second to fourteenth vertebra, fifteenth having both transverse processes but no rib. Sixteenth vertebra is the sacral, followed by two transitional postsacral or precaudal vertebrae. Twenty-three caudal vertebrae, the anterior ones with a fin-like process ex- 195.': A NEW SALAMANDER OF THE GENUS PARVIMOLGE tending forward from the ventroposteriorly directed haemal process to the anterior edge of the vertebra. Short transverse processes, more laterally angled in the first few caudal vertebrae, more anteriorly directed farther caudad. This account of the skeletal features is at variance with Tay- lor's description (1944: 223) on several points, notably the pres- ence of the prefrontal and the ossification of the carpals and tarsals. The latter character may be variable, but the condition in this specimen does not support contrasting the cartilaginous carpals and tarsals of Parvimolge with the osseous ones of Thorius. This is especially true in view of the fact that Hilton (1946, 1948) has recorded them as cartilaginous in T. pennatu- h(s, and since the single cleared specimen referable to T. dubitus that I have examined likewise has entirely cartilaginous ones. However, I believe that most of the discrepancies between the two descriptions of P. fownsendi may be ascribed to different types of material used, and it should also be borne in mind that the foregoing account is based on a single specimen. Table I Quantitative Characters in Mexican Parvimolge snout- dorsal premax- mandibu- vent tail costal body maxillary lar length length spaces glands teeth teeth townsendi E 20-23 20-26 32-42 45-58 54-69 males (4) M S£.0 £2.5 3.7 39.0 50.0 61.5 townsendi R 19-22 16-21 35-41 48-58 55-73 females (4) M 21.0 19.0 4.2 ss.o 53.0 64.0 praecellens (female) 35 30 1 50 75 87 E, range; M, mean; tooth counts are totals of both sides; measurements in mm. ; costal spaces refer to number between adpressed limbs. 8 BREVIORA NO. 42 Relationships. Parvimolge praecellens is closely related to P. townsendi, as indicated by the foot structure, coloration, con- spicuous dorsal glands, and shape of tail. A third form, richardi, described with reservations as a Parvimolge (Taylor, 1949, 1952), lacks the conspicuous glands. This fact, apparent differ- ences in details of the foot, and the enormous disjunction in range — Vera Cruz to Costa Rica — imply that richardi is preferably not to be associated with townseiidi and praecellens. Determining the nature of the affinity between P. praecellens and townsendi necessitates much more material. However, the more numerous teeth, longer legs, and small nostrils of praecel- lens are probably correlated with its greater size. These charac- ters and the smaller size of the dorsal glands appear to indicate that praecellens is less specialized. The morphological and eco- logical relationship may be of the same general type existing between sympatric large and small species in Pseudoeurycea and Chiropterotriton. Indeed, in the latter genus, C. dimidiata shows seemingly paedomorphic features very like those found in town- sendi. That Parvimolge is allied to Lineatriton as inferred by Tanner {op. cit.) is supported by the distinctive osseous character of the second ceratobranchials found to be common to P. townsendi and L. lineola. Nevertheless, the derivation of townsendi and prae- cellens was probably from a less specialized form, perhaps in the Pseudoeurycea cephalica group or its ancestral stock. The foot shape in very young specimens of cephalica, the presence of ir- regularly, protruding poison glands on the dorsum in some forms of the group, and the guanophore-spotted, brown to black ground color in the group suggest this. The syndactylous foot structure could have become established by a paedogenetic process in Parvimolge stock as it differentiated from some such ancestor. Acknowledgments. I wish to thank Arthur Loveridge for the privilege of examining most of the specimens described in this paper. Charles F. Walker and James E. Mosimann have read the manuscript and given useful suggestions. The drawing was executed by Harold J. Walter. The abbreviations MCZ and UMMZ respectively designate the Museum of Comparative Zool- ogy and the University of Michigan Museum of Zoology. 1955 A NEW SALAMANDER OF THE GENUS PARVIMOLOE 9 LITEEATUEE CITED Dui«sr, E. E. 1922. A new salamander from Mexico. Proc. Biol. Soe. Washington, 35: 5-6. 1926. The salamanders of the family Plethodontidae. Smith College, Northampton, Mass. viii + 441 pp. Hilton, W. A. 1946. Skeletons of Mexican and Central American salamanders of the family Plethodontidae. Jour. Entomol. and Zool., 38, No. 1 : 1-8. 1948. The carpus and tarsus of salamanders. Jour. Entomol. and Zool., 40. No. 1 : 1-13. Tanner, W. W. 1952. A comparative study of the throat musculature of the Plethodon- tidae of Mexico and Central America. Univ. Kansas Sci. Bull., 34. Pt. II, No. 10: 583-677. Taylor, E. H. 1944. The genera of plethodont salamanders in Mexico, Pt. I. Univ. Kansas Sci. Bull., 30, Pt. I, No. 12: 189-232. 1949. New salamanders from Costa Eica. Univ. Kansas Sci. Bull., 33, Pt. I, No. 6: 279-88. 1952. The salamanders and cjiccilians of Costa Rica. Univ. Kansas Sci. Bull., 34, Pt. II, No. 12 : 695-791. R E V I O R A imim of Comparative Zoology Cambridge, Mass. x\pril 7, 1955 Number 43 SPEED-INDUCED SKIN FOLDS IN THE BOTTLE-NOSED PORPOISE, TVliSIOPS TRUNCATUS ' By Fkank S. Essapian ^rarineljuid Kesearcli T.alioratoiy While engaged in a study of the Ix'havior of bottle-nosed por- poises, Tursiops irmicdliis (Montagu), at Marine Studios, Marineland, Florida, over a period of three to four years, the author has oliserv(nl on hundreds of occasions a remarkable modification of the skin contoui- of these animals whenever they resorted to I'apid movement about the tank (Essapian 1953, p. 3!)!;)). Tlie circular tank, where the i)orpoises are on exhibit, is seventy-five feet in diameter and about fourteen feet deep at the center. It liouses an average of ten animals. Generally lost to view due to swiftness of action, this trans- fornmtion of the skin contour takes place when an animal accel- erates its s])eed in excess of its normal rate, or, when swimming rapidly, comes to a sudden stop. It is then that transverse skin folds may stand out in I'elief on ])art of an animal, or may extend over its entire lengtli, from head to tail. The pattern of these formations depends on the speed and sex of the swimming animal, but is not restricted to any age group. The folds may be s(^en even in an animal only a day or two old. The skin folds generated b}' a sudden burst of speed are by far the more common of the two categories, and are usually oc- casioned when the animals engage in pursuit of one another, are frightened and trying to escape, or when racing for a food fish tossed into the tank. The duration of folds is then conditioned ^ Coiitriliutiou No. 7.")! from the Woods Hole Occaiiogrjiiihie Institution. 2 BREVIORA No. 43 by the factors which motivate tlie animals' movements. In the case of a brief dash, for example, when an animal exerts itself to snatch a fish, the skin folds may last only a second, or even a fraction of a second (Pis. 1 and 2). However, in the case of prolonged rapid movement around the tank, the animals may intermittently exhibit skin folds for a jieriod of one or two or more seconds. Often an animal at pla}' or during the mating season may suddenly leap out of water, exhibiting folds at the peak of acceleration, just ])rior to its emergence from water, and again after it has reentered the water (Pis. 3 and 4). At this point it may be of interest to note that the fat adult females tend to exhibit fewer folds, but greater in size and at less regular intervals, than are produced in the adult male or in the younger slimmer animals when swimming at comparable speeds. The same female in Plate 4 is also shown in Plate 5 ; here this animal 's folds are shown in a transitory stage, just prior to full accelera- tion. The skin folds of the second category occur at less frequent intervals and are of very brief duration. These take place when an animal swimming rapidl.y without exhibiting any folds comes to an abrupt halt. Then the folds may suddenly be dis- played (PI. 6). The duration of folds of this sort is indicated by the fact that this entire sequence occupies only twelve frames of 16 mm. film at 16 frames per second. This w^ould indicate a total elapsed time of three quarters of a second for the formation and disappearance of folds. Plate 6 is the fourth frame, showing that these folds were fully formed in one quarter of a second. One feature in the production of skin folds is of particular interest. In an animal maintaining a high and sustained rate of speed the folds are stationary and do not progress wave- fashion. At times, when extremely high speeds are achieved by adult animals, the skin folds tend to slope posteriorly. Observa- tions also indicate that the folds are more likely to occur on ventral and lower lateral surfaces. While the superficial layer of the porpoise's skin is extremely thin, the blubber is of considerable thickness and is tough and fibrous. Tn a Tursiops truncatus fifteen days old and weighing thirty pounds, the skin, including blubber, is approximately half 1955 BOTTLE-NOSED I'ORPOISE TURSIOl'ti TRUNCATUS 3 an inch thick. In an adult it is nearly one and a half inches thick on the ventral surface. Althouoii normally not a])j)arent to the view or touch, the jjoi-poise's skin is, nevertheless, pliant and loose on the body. This slackness of the skin in a live animal held out of water in a slini Tschudi, 1838, Diemiciylus Rafinesque, "1820, Eu proof us Gene, 1838, Hypselotriton Wolterstorff, 1934, Paehytriton Boulenger, 1955 PALAEOTARICIIA OLIGOCENICA 3 1878, Taricha Gray, 1845, and several species of Trittirus^ Rafin- esque, 1815. The fossil genera Archaeotriton von Meyer, 1860, Heliarchon von Meyer, 1863, Oligosemia Navas, 1922, and Poly- semia von Meyer, 1860, likewise all seem to have frontosquamosal arches and paroccipital processes, but information on the pre- maxillae is lacking. The phalangeal formula of the Oregon specimen is 1-2-3-2 (manus) and 1-2-3-3-2 (pes), as also in Cynops, Hypselotriton, Pachytrito7i, and Taricha. Cynops, Hyp- selotriton, and Diemictylus (the only Recent American sala- mandrid genus other than Taricha) can be excluded from identi- fication with the fossil because of their relatively long, narrow- skull, inconspicuous dorsal opening of the cavum internasale, and failure of the ascending processes of the premaxilla to extend beyond the anterior border of the cavum. Diemictylus can be further ruled out by its paired longitudinal ridges extending from the nasals to the posterior border of the parietals, and an abrupt angle in the arches posterior to the frontal-squamosal suture. Pachytriton is excluded by its long, narrow skull ; a ridge (joining the maxillae across the premaxilla) situated between, and concentric with, the outer tooth row and the anterior border of the vomer;- and the very small dorsal cavum internasale opening. Euproctus is ruled out by its phalangeal formula of 2-2-3-2, 2-2-3-3-2, an elongated skull with a spatulate anterior end, concave lateral maxillary borders, and the paroccipital processes extending farther back than the occipital condyles. The five or six species of Triturus that have complete fronto- squamosal arches are excluded by their phalangeal formula, 2-2-3-2, 2-2-3-3-2, and the elongated skull. Of all the known salamandrid genera there is no doubt that Taricha is the closest to the fossil. Skull, body and limb proportions are similar, and other important characters are : the frontosquamosal arches ; paroccipital processes; large openings of the cavum internasale, the dorsal one bordered anteriorly and laterally by the premax- illa. All these similarities to the Recent Taricha species {iorosa, granulosa, rivularis, and sierrae), and the fact that the fossil comes from the present geographic range of one of them (granu- losa), may justify assigning the fossil to that genus. But some liis iisfd liy \\(ill.Tst(irfl- ,\nd IleiTe (li»;-(5). - voiiHTopalatiiif (ir prevomer. 4 BREVIORA NO. 45 very important characters distinguishing; the fossil from the Recent species lie in the nature of the vertebrae : characters as fundamental as some upon which several genera, known from vertebrae only, have been based. The pelvis is also peculiar, and the tarsal basale count is different, as described below. With these differences, and the belief that if the animal were alive, neoherpetologists would distinguish it generically because of the spinal column, I feel that erecting a new genus for the fossil is justified. Class AMPHIBIA Order CAUDATA Family SALAMANDRIDAE PALAEOTARICHA, new genus Genotype: ralaeotaricha oligocenica, new species. Diagnosis: Precaudal neural spines high, with laterally ex- panded, broad, rectangular, sculptured, flat tops, which are in contact with each other fore and aft. Fourteen presacral verte- brae. Pelvis with relatively large, knob-like, laterally directed prepubic processes. Tarsal basale formula 1, 2, 3, 4-(-5. Maxillae extending caudad, united by suture Avith quadrates. Palaeotaricha oligocenica, new species Holotype: University of Oregon Museum of Natural History no. F-5405, a nearly complete specimen. Collector: Mr. George R. K. Moorhead, August 1, 1951. Horizon: Upper Oligocene plant-bearing tuffs, not older than Eugene and Fisher formations, the Willamette formation of Dr. Ralph W. Chaney (in lift.). Locality: About 3 miles (5 km.) southeast of Eugene, Lane County, west central Oregon; southeast Y^, sec. 14, T. 18 S, R. 3 W ; " about 550 feet east of the Southern Pacific railroad tracks on the south side of a small hill in a cut made for the newly con- structed highway near the east end of the overpass over the railroad. This is about 400 ft. north of lat. 44 deg. and 3000 ft. west of long. 123 deg. ... in the Eugene quadrangle" (Dr. Herman Clark, in litt.). Diagnosis: As for the genus. 1955 PALAEOTARICHA OLIGOCENICA 5 Description: Total length (as preserved), 120 mm.; original length, ca. 135 mm.; length of skull (anterior end to foramen magnum, dorsal surface), 15 mm.; width of skull (across paroc- cipital processes), 13 mm.; length and also width of ischiopubis, 6 mm.; length of limb bones: humerus, 9.5 mm.; radius, 7 mm.; ulna, 7.5 mm.; digit 3, manus, 6.5? mm.; femur, 9 mm.; tibia, 5.5 mm.; fibula, 6 mm.; digit 3, pes, 6 mm. The paroccipital processes are large and are produced laterad, abutting the squamosals from behind. The frontosquamosal arches are complete and robust. The dorsal aperture of the cavum internasale is long and wide, and is bordered anteriorly and laterally by the ascending processes of the premaxilla. The vomerine opening of the cavum internasale is large (probably in part artificially) ; it lies wholly within the vomer. The anterior border of the vomer is an arc of a circle (roughly), with no sharp l)reaks in its periphery. No teeth are preserved, but the vomerine tooth rows are evident though crushed and incomplete. The vertebrae are relatively shorter than in Taricha. There are 14 presacral vertebrae (13 in Taricha). Even the atlas has an expanded top, which, however, narrows anteriad. Instead of the dorsal surface of the atlas sloping downward anteroventrally as in Taricha, in Fahieotaricha oUgocenica it is horizontal for most of its length, then there is a short vent rally directed slope toward the anterior end, and then a sudden vertical drop to the condylar facets (PI. 1, figs. 2, 3). The neural spines of the caudal verte- brae lack the expanded tops. Five^ thoracic ribs of the left side are exposed in good condi- tion (PI. 1, %. 1; PI. 3, fig. 1). All but the cauclalmost have conspicuous uncinate processes, which, in successively posterior ribs, become shorter and more distal. The great length of the scapula suggests that the suprascapula was ossified (or calcified). There are four digits in the manus. The phy'angeal count is 1-2-3-2, which, together with the carpal basale ff mula (1+2, 3, 4), is the same as in Taricha. The pchic girdle has floated free and, with the limbs, is fairly well preserved as a unit. The pelvic girdle is apparently a highly ossified and coossified structure. Zones of weakness may perhaps exist anteriorly between left and right halves, and between the 1 Some A tiie things that look like rihs in Plntf 1, figure 1 are phalanges of the left niaaus. 6 BREVIORA NO. 45 pubes and ypsiloid cartilage. The latter is heavily ossified (or calcified ; there is no obvious difference in preservation between it and any other part of the skeleton). About 5 mm. of it are preserved, of which 2 mm. are proximal to the bifurcation. The prepubic processes are rounded and robust. They project later- ally and apparentlj^ slightly posteriad in P. olujocenica, but anterolaterad in Taricha, where they are also less demarcated from the pubes. There are five digits in the pes; the phalangeal count is 1-2-3-3-2. The tarsal basale formula is 1, 2, 3, 4-(-5, whereas the Taricha specimens have 1+2, 3, 4, 5. On the whole the skeleton gives the appearance of having great strength, with the girdles and limb elements all well ossified, and the top-heavy neural spines. Discussion: The elongation of the maxillae seems to be a primitive character of Salamanc'ridae. In all the fossil genera of which the .skulls are known the maxillae articulate with the quadrates: Archaeotriton, Brachycormus von Meyer, 1860, Heli- archon, Oligosemia, Palaeopleurodeles Herre, 1941, and Poly- semia (all Miocene; Oligosemia also Oligocene). In three of the more primitive Kecent salamandrids, Pleurodcles Michahelles, 1830, SaJamandrina Fitzinger, 1826, and Tylofoirifon Anderson. 1871, the maxillae may very nearly or actually touch the quad- rates. In the Palaeotaricha oligoccnica specimen, although the maxillae clearly articulate (fuse?) with the suspensoria, the in- distinct quadrate-squamosal sutures make it difficult to see exactly which suspensorial element is involved with the maxillae. I have assumed that the quadrate is involved, because of the condition in the forms just cited. Other than the maxillae, the P. oligo- cenica skull seems to fall within the variation of the skulls of the several Taricha species. In T. granulosa the vomerine tooth rows generally diverge slightly, from the anterior to the posterior ends; in the other species they are parallel or very slightly divergent for most of their length, then suddenly they are strongly divergent for the remaining 14 or Vs- There is, how- ever, a good deal of individual variation and intergradation be- tween these tooth-row patterns. The available skulls of T. sierrae and T. rivularis (three of the former, four of the latter) seem to be rather similar, and to stand apart from the skulls of the other two species. For one thing, the skull of rivularis and sierrae tends to be somewhat deep and domed, in contrast to the 1955 PALAEOTARICHA OLIGOCENICA 7 more-or-less flattened and widened roof of the others. T. rivularis and sierrae tend to have the dorsal opening of the cavum inter- nasale especially long and wide. In anterior view it is seen as a notch whose bottom lies ventral to the dorsal border of the nares, which is not the case in torosa or granulosa; nor does it extend so far forward in the latter two species. The above characters are more pronounced in rivularis than in sierrae, but the fol- lowing character is equally developed in the two. The ventral (vomerine) opening of the cavum internasale is greatly elongated — roughly the size and shape of the dorsal opening — and ex- tends beyond the anterior border of the vomer into the pre- maxilla. In the other species this aperture is usually circular or slightly elliptical, and is wholly situated within the vomer. In about ten alcoholic specimens each of rivularis and sierrae, and many more of torosa and granulosa, this difference is corrobo- rated. In one skull of granulosa, however, there is a very small, extra, opening into the cavum internasale anterior to the vomer- premaxilla border — a feature that would be very difficult to note in an alcoholic specimen. It would be very interesting to know whether the elongated aperture is an evolutionary advance (cf. Palaeotaricha) or is primitive (cf. primitive extant genera of salamandrids, which have the premaxilla paired, and some have the aperture extremely large). A full description of the skull of T. torosa is given by Bolkay (1928) (under the name Diemyctylus [sic] torosus; his specimen could be a granulosa, judging by the vomerine tooth rows), as well as descriptions of the skulls of most of the Recent sala- mandrid species. The condition in the vertebrae — the expanded tops — of Palaeotaricha oligocenica is not only different from that in Taricha, where the neural spines are high but narrow, but quite different from any other urodele vertebrae that I can find. Among Recent salamandrids Tylototriton comes closest in this respect, jide Herre (1949) and corroborated in the specimens available to me. But in Tylototriton the expansion is much less prominent, and the width decreases anteriad ; the tops of the spines are, however, sculptured, and they do contact in series, as in P. oligocenica. The closest approach of all seems to be in the Miocene Ckelotriton Pomel, 1853, w^hich has a broad top on the neural spine also. But (at least in the specimen figured by Herre, 8 BREVIORA NO. 45 1949, p. 225, fig. 8) the anterior half of the spine comes to a rounded point, and the posterior end is bifurcated by a wedge- shaped notch. Thus it seems that the Chelotriton vertebra has merely a greatly expanded neural spine, and is not fundamentally different from that of most salamandrids (e.g. Taricha, PI. 1, fig. 5). Now arises the question (considering the absence of visible sutures between the vertical and horizontal portions of the P. oligocenica neural spines) of whether we have a structural series of increasingly expanded neural spines — Taricha, Tyloto- triton, Chelotriton, Palaeotaricha — or whether the latter is not in this series, the spinal expansions being dermal plates plastered on top (as suggested by the sculpturing) . I favor the latter view : the expansions look like dermal plates ; and, of course, the series just given has only descriptive validity. The uncinate processes of the P. oUgocen'ca ribs are longer than in the Recent Taricha specimens at hand, but these struc- tures are quite variable in individual .salamanders. In three specimens of Taricha g. granulosa I have seen, one has uncinate processes on ribs 2 and 3 on the right side ; on rib 2 only, on the left. Another has uncinates on ribs 2 to 4 on both sides ; the third specimen has them on 2 to 7 on both sides. One specimen of Cynops pyrrhogaster has uncinates on ribs 2 to 6 on the left side, on 2 to 7 on the right ; another has no uncinates at all. And so it goes. In all these cases the larger and more fully ossified the individual, the greater the number of uncinate processes. In nearly every specimen I have seen there are incipient uncinates on the first one or two ribs distal to the last one that has a distinct uncinate process. The tarsal basale formula in P. oligocenica — 1, 2, 3. 4+5 — • contrasts with 1+2, 3, 4, 5 in the Recent Taricha, Chioglossa Barbo.-'a du Bocage, 1864, Salamandra Laurenti, 1768, Mertensi- ella Wolterstorif, 1934, and Pleurodeles Michahelles, 1830, while Salaniandrina with four toes on all feet apparently has 1, 2, 3, 4, and all other genera have 1+2, 3, 4+5, fide Bolkay (1927), and in agreement with additional material seen by me. It is interesting to note that the carpus, as compared with the tarsus, is apparently much more nearly constant throughout the Salamandridae. As figured by Bolkay (1927) and corrol)orated and extended b}' the specimens available to me, the carpus has 1955 PALAEOTARICHA OLIGOCENICA 9 basales 1+2, 3, 4, prepollex basale,' eentrale, radiale, inter- medium, and iilnare, in all genera. The latter two may or mav not be fused, but usually are: in one specimen of Triturus vul- garis the intermedium of the right manus is separate ; the left is fused with the ulnare. While the observations of Hilton ( 19-48) agree in general with those of Bolkay and me, Hilton illustrates various eases of fusion of carpal and tarsal elements, and extra elements (such as the non-fusion of tarsal basales 4 and 5 in a specimen of Cynops pyrrhogaster) . In the light of Hilton's data, the difference be- tween the PalaeotaricJia and the Taricha tarsus may not be sig- nificant. Yet it must be pointed out that all of the 16 specimens of Taricha at hand have the same tarsal and carpal structures. The large size and knob-like shape of the prepubic process of P. oligocenica also seem to be unique. When these processes exist in the Recent specimens at hand, they are always smaller, more pointed, and directed at least partly anteriad. Nor do any of these Recent pelves have the fusion of the left and right halves to the extent seen in P. oligocenica, where scarcely even a raphe remains of a former suture. From all appearances PalaeotaricJia is not ancestral to Taricha, but is its nearest relative. I do not feel that Palaeotaricha helps determine the further affinities of Taricha, for the one character that might seem to indicate relationship with other genera — the elongated maxillae in Palaeotaricha — is merely primitive and not unexpected. Material examined : I have been fortunate in being able to examine skulls and skeletons of every genus of Recent 8ala- mandridae, with the single exception of Pleurodeles, and to see alcoholic specimens of nearly every species of all genera. It must be emphasized, however, that except for 16 specimens of the three species of Taricha, in no case have I seen more than two skeletons. Hence I do not feel that my knowledge of intraspecific variation is sufficient for more than tentative authority on the above com- parative skeletal studies. Note on the stereophotography The ventral view of the P. oligocenica skull (PI. 2. A) was 1 Bolkay follows another usa{,'e in calliiit? this element laipal hasale 1 ; and what he calls 2 is here called 1+2. Likewise he desi^ates the prehallu.x basale as tarsal basale 1. and 1+2 as 2. 10 BREVIORA NO. 45 made by Dr. Donald Baird, of this Museum, and myself, with a binocular (stereoscopic) microscope, and a Leica 3C camera fitted with a Leitz Focaslide and Leitz Micro-Ibso. The ocular lens of the latter was inserted alternately into the two tubes of the microscope, and exposures were made in the two positions of the apparatus, at an interocular distance of 21/0 inches. The latter refers to the positions of the Focaslide, which was used because it was found that the side telescope of the Micro-Ibso was not sufficiently accurate for the depth of focus required. This method may be theoretically superior to the tilt method usually used, since it takes care of matters of interocular distance and angles more easily. But for subjects as large as the present ones considerable trial-and-error choice of microscope lenses was neces- sary to get the entire image into view. All the other figures were made with a Kodak 2-D 5" x 7" view camera fitted with a 135 mm. lens, and mounted on a pivoted arm. One picture of each stereoscopic pair was taken vertically, the other at 7° from the vertical. Exposures were 20 seconds at f. 22, using panatomic film, and illuminated by two 100- and one 200-watt bulb with reflectors, each about 2 feet from the subject. ACKNOWLEDGMENTS r am very grateful to Dr. Donald Baird for his help and the use of his equipment in the photography ; Mr. Benjamin H. Banta, Stanford University, for the gift of a number of speci- mens ; Mr. Charles M. Bogert, American Museum of Natural History, for the loan of material ; Dr. Ralph W. Chaney, Uni- versity of California (Berkeley), for stratigraphic information on the fossil locality, and for permission to use his unpublished term, "Willamette formation"; Dr. Herman Clark, Willamette University (now retired), for sending the Palaeotaricha speci- men, and for providing and permitting me to quote his informa- tion on the fossil locality ; Dr. Max K. Hecht, Queens College, New York, for much valuable help ; Mr. Arthur Loveridge, Museum of Comparative Zoology, for permission to examine and skeletonize a number of specimens ; Dr. Richard M. Ritland, College of Medical Evangelists, for the use of material from his personal collection ; Dr. Jay M. Savage, Pomona College, for valuable help; and above all to Dr. Ernest E. Williams, Museum of Comparative Zoology, for first suggesting that I undertake this sturly, and for much help and continued advice. 1955 PALAEOTARICHA OLIGOCENICA 11 BIBLIOGRAPHY BOLKAY, StEPHAN JoSEP 1927. Materialen zu einer vergleiehenden Morphologie der Carpal- und Tarsalelemente der Unterfaniilie Salamandrinae. Glasnik Zeiiial .iskog Miizeja u Bosiii i Hercegovini, vol. 39, pp. 59-67. 1928. Die Schiidel der Salamandrinen, niit besonderer Riicksicht auf ihre systeniatische Bedeutiing. Zeitsehr. Aiiat. Entwicklungs gesch., vol. 86, pp. 259-319. GoLUFUs.s, Georg August 1831. Beitriige ziir Kenntniss verschiedener Reptilien der Vorwelt. Nova Acta Acad. Leop. Carol., vol. 15, pp. 61-128. llKKUii:, Wolf J 934. Die systeinatisclie Stellung von TaricJia torosa Esehliolz. Blatter t'iir Aquarieii- und Terrarienkunde, vol. 45, pp. 250-252. 1935a. Die Sehwanzlurche der initteleociinen (oberlutetisehen) Braun kolile des Geiseltales und die Phylogenie der Urodelen unter Einschluss der fossilen Formen. Zoologica, vol. 33, art. 4, pp. 1 85. 19351). I'her 0/i(;o.v< »u'« spinosa Navas, einen fossilen Schwanzlurch aus deni spanischen Tertiar. Palaeont. Zeitsehr., vol. 17, pp. 91-105. J 941. Falaeopleurodeles haitffi nov. gen. nov. spec, ein fossiler Schwanz lurch aus dem Mioziin Siiddeutschlands. Zool. Anz., vol. 134, pp. 1-17. 1949. Neue Tatsachen zur Stammesgeschichte der Sohwanzlun-he. Zool. Jahrb., Abt. Syst., vol. 78, pp. 217-236. 1950a. Schwanzlurehe aus dem PaleocJin von Walbeck. In Neue Ergeb nisse und Probleme der Zoologie (Klatt-Festsehrift). Leipzig, Geest und Portig, pp. 286-301. J 9501). Der dorzeitige Stand unseres Wissens iibcr die fossilen Urodelen, zugleieh einige kritisehe Bemerkungen iiber Boomgardia sala- viaiidriformis v. Huene. Neues Jahrb. Geol. Paliiont., Monatsh., li»50, pp. 19-25. IIU.TON, WlLLIAJI A. 194S. The ('ari)us and tarsus of salamanders. Jour. Entomol. Zool., vol. 40, pp. 1-13. l.AlliK, Gl'STAV C. 1901. Synopsis iler Wirbelthierfauna der Ijciliin. Braunkohlforniation und Besehreibung neuer, oder bisher unvollstiindig bekannter Arten. Abhandl. deutsch. naturwiss.-med. Ver. Bohmen "Lotos," vol. 2, art. 4, i)p. 1-80. 12 BREVIORA NO. 45 MjiYER, Hermann von 1860. Salamandrinen aus der Braunkohle am Ehein und in Bolimen. Palaeontographica, vol. 7, pp. 47-73. 1803. Heliarchon fiirciUatus, ein Batrachier aus der Biaimkohle von Eott ini feiebengebirge. Ibid., vol. 10, pp. 292-298. VoKES, H. E., Parke D. Snavely, Jr., and Donald A. Myers 1951. Geology of the southern and southwestern border areas of the Willamette valley, Oregon. U. S. Geol. Surv. Oil and Gas Invest. Map. OM 110. WOLTERSTORF, WiLLY and WOLF HeRRE 1935. Die Gattungen der Wassermolche der Familie Salaniandridae. Arch. Xaturgesch., new ser., vol. 4, pp. 217-229. EXPLANATION OF PLATES PLATE 1 Figs. ], 2, 4. Palaeotarictlia oligocenica, holotype, University of Oregon Museum of Natural History, no. F-5405. Fig. 1. Entire specimen (except end of tail). Skull seen ventrally, spinal column from right side. Fig. 2, A, A'. Stereograph of vertebrae 1 to 3, left side. Fig. 4. Vertebrae 10 to 12, left side. The neural spines of all precaudal vertebrae are bent sinistrad. Figs. 3, 5. Vertebrae of a Taricha torosa. Fig. 3. Vertebrae 1 and 2. Fig. 5. Vertebrae 9 and 10. od, Odontoid process. Fig. 1x1; figs. 2-5 x 6. PLATE 2 Palaeotaricha oligocenica, skull of holotype. A, A', stereograph of dorsal view. B, B', ventral view, a, Atlas; del, dorsal opening of cavum internasale ; en, external naris ; /, frontal ; fsa, f rontosquamosal arch ; fv, foramen vesti- buli; hy, parts of hyoid apparatus; in, internal naris; Ij, lower jaw; m, maxilla; n, nasal; nld, nasolacrimal duct?; oc, occipital condyle; op, occi- pito-petrosal; os, orbitosphenoid (the white dot may be the optic foramen) ; p, parietal; pam, palatine part of maxilla; pm, premaxilla; poc, paroccipital process; prf, prefrontal; ps, parasphenoid ; pt, pterygoid; q, quadrate; s, squamosal; rci, vomerine opening of cavum internasale; v, vomer. All figures x2y4. PLATE 3 Palaeotariclui oligocenica, holotype. Fig. 1. Eight pectoral girdle and limb, and ribs 3 to 7. Forearm and manus in dorsal view, other parts in right side view, ho l-\-2, Carpal basales 1+2; be /-/-^, carpal basalejf i^/4; /(, humerus; mc 1, metacarpal 1; pph, prepollex basale; r, radius; ra, radiale; rG, rib (i ; .s, scapula; u, ulna; ul-\-i, ulnare + intermedium. Fig. 2. Pelvic girdle and limbs. Pelvis and left limb in ventral view, right limb (left side of photograph) in dorsal view, ht 4-\-5, Tarsal basales 4+5; c, centrale; /, femur; /e, fibulare; fi, fibula; i, intermedium; phh, prehallux basale; prp, prepubic process; f, tibia; ti, tibiale ; y, ypsiloid process. Both figures X 2^4. A <9 PLATE 1 PLATE PLATE 3 BREVIORA Meseiuiei of Comparative Zoology Cambridge, Mass. June 24, 1 955 Number 46 CAVE-FOSSIL VERTEBRATES FROM CAMAGUEY, CUBA By Karl F. Koopman Department of Biology, Queens College, Flushing, New York and RoDOLFO Ruibal Division of Dife Sciences, University of California, Riverside Of all the Cuban provinces, Camaguey is the least known in regard to its former and present vertebrate fauna. With this fact in mind the junior author, in January of 1952, visited the Sierra de Cubitas in Camaguey with the purpose of locating cave-fossil material. The Sierra de Cubitas is a range of low (maximum altitude of 750 feet) Tertiary and Cretaceous limestone hills running north- west to southeast along the northern half of Camaguey. The hills rise abruptly from the flat serpentine Camaguey savanna, and are covered by one of the few remaining tracts of tropical forest in the province. Two caves were visited by the junior author. Both are located on the south slope of the hills, at the Trinchera de Martin Castillo (the site of a battle during the war against Spain), which is about seven kilometers west of the village of Banao. The first cave was the Cueva del Indio (re- ferred to as Cave 1 henceforth in the text), an extensive cave that is on occasion visited by excursionists from the city of Camaguey. The second cave lies a short distance east of Cave 1, and will be designated as Cave 2, since it bears no known name. The present entrance to Cave 1 is a broad, high-domed opening which is continuous with the large chamber from which the deeper passages of the cave extend. Due to the ample entrance the entire chamber is well within the twilight zone. The floor of the chamber has been excavated for bat guano. The entrance to the cave is higher than the chamber floor, and consists of rocks fallen from the eroded cave front. The red soil of the chamber 2 BREVIORA NO. 46 was damp, but at higher levels near the entrance the soil was dry. However, during rains, water enters the cave and material has been deposited at different levels near the entrance. All of the fossil material collected was found as surface litter deposited at different points near the cave entrance. Material was ex- tremely abundant, any handful from the surface yielding many bones and snail shells. Collections were made at random from the many pockets where the litter consisted almost exclusively of bones and snail shells. In the case of Cave 2 erosion has cut back so that it has exposed to direct sunlight a former high dome, about seventy feet high) with massive stalactites and stalagmites along the sides. In this cave, digging for bat guano had only ceased two years previous to the time of the visit and consequently the entire floor of this shallow cave is covered with the mounds of red litter that were screened out by the guano collectors and the material washed out from these mounds by the rain. The abundance of bones in this cave can only be described by saying that there are millions. To walk in any part of this cave is to Avalk upon a floor of bones and snail shells. Collecting in this cave was essentially the same as in Cave 1, but of a greater volume. The material was scooped up at random from the mounds and from the floor. Due to the little time available no adequate investigation was done in any of the more protected parts of the cave where material was still un- disturbed by the guano collectors. The snail shells collected were kindly identified by Dr. W. J. Clench of Harvard University. Eight species were found and according to Dr. Clench all the species are at present common in Camaguey. MATERIAL COLLECTED Class MAMMALIA Order INSECTIVORA Nesophontes micrus. Represented by one maxilla, eight man- dibles, three femora, and one humerus from Cave 1, and by two partial skulls, eight mandibles, and three femora from Cave 2. This species was first described from the provinces of Matanzas (Allen, 1917b), and later from Oriente (Anthony, 1919) and Isla de Pinos (Allen, 1918). In addition there is material in the Museum of Comparative Zoology from Las Villas and Pinar 1955 CAVE-FOSSIL VERTEBRATES FROM CUBA 3 del Rio, and in the United States National Museum from Habana province. The present material from Camaguey completes the record and demonstrates that the species existed throughout Cuba. Solenodon cuhanus. Represented by a portion of one mandible and six isolated teeth from Cave 1. At least two individuals are represented in the material collected. This species is known today as a living animal only from Oriente {8. poeyi is con- sidered a subspecies, following Aguayo, 1950). It has previously been found fossil in both Oriente (Allen, 1918) and Pinar del Rio (Aguayo, 1950). It apparently also occurred in the Sierra de Trinidad in the province of Las Villas up until the middle of the nineteenth century (Gundlach, 1895). The present material agrees well with fresh specimens from Oriente, but differs by somewhat larger size and robustness. Since the species was dif- ferentiated into two subspecies in Oriente, and the animals from the Sierra de Trinidad appear to have been distinct in color from the Oriente forms, it is possible that the Camaguey material may represent a different subspecies. However, the material available is not adequate to determine this. Order CHIROPTERA Macrotus waierhousei. Represented by two mandibles from Cave 1, and a portion of one skull and three mandibles from Cave 2. This species was previously known from Isla de Pinos and all the provinces except Camaguey. Apparently a common and widely distributed form. Monophyllus ciihanus. A single mandible from Cave 2. A single periotic bone from Cave 2 is also allocated to this species. AH recent records of this bat appear to be from Oriente, but Gundlach (1868, pp. 48, 49) records it from Rangel in Pinar del Rio. Brachyphylla nana. Represented by two mandibles from Cave 1, and nine partial skulls and 18 mandibles from Cave 2. This species was previously known from all the provinces of Cuba except Camaguey and Isla de Pinos. A common and widespread species. Artibeus jamaicensis. Represented by five partial .skulls and two mandibles from Cave 1, and eight partial skulls and twenty- 4 BREVIORA NO. 46 six mandibles from Cave 2. Presumably known from all the provinces of Cuba and from Isla de Pinos, though we have been unable to find any definite record for Matanzas. Several speci- mens of this species were shot in Cave 1. A very common and widespread species. Phyllops falcatus. Represented by two partial skulls and three mandibles from Cave 1, and ten partial skulls and twelve man- dibles from Cave 2. Previously recorded from Oriente, Las Vi- llas, Matanzas, and Pinar del Rio, but probably occurs through- out Cuba and Isla de Pinos. None of the fossil material from Camaguey represents P. vetus, still known only as fossil from Oriente (Anthony, 1919). Erophylla sezekorni. Represented by two mandibles from Cave 1, and two rostra and seven mandibles from Cave 2. Pre- viously known from Habana, Las Villas, and Oriente, it is prob- ably rather widespread in Cuba. Phyllonycteris poeyi. Represented by one mandible from Cave 1, and by ten rostra and fifteen mandibles from Cave 2. Previ- ously known from Pinar del Rio, Habana, Matanzas, Las Villas, and Oriente, and is thus a widespread species. Natalus primus. Represented by a single mandible and a fragmentary rostrum from Cave 2. This species was first de- scribed as a fossil from Oriente by Anthony (1919), but it is also known from Las Villas (Koopman, unpublished). Our record is the first from anywhere else and the first rostral material, the other finds consisting entirely of mandibles. Known only as fossil, this species was apparently rare but probably fairly widespread. Tadarida 'brasiliensis. (We follow G. M. Allen, 1911, in con- sidering T. muscula a subspecies of T. hrasiliensis.) Repre- sented by two rostra and two mandibles from Cave 2. Previously known from all the provinces of the Cuban mainland. Evidently a common and widespread form. Tadarida molossa. Represented by one rostrum and one man- dible from Cave 2. These appear to be the first Cuban specimens of this species with definite locality data, T. macrotis, a synonym, having been described only from "the interior of Cuba" (Allen, 1911). Apparently a very rare or local form. The geographic data for the bats were obtained from Miller (1904), Allen (1911), Peterson (1917), Anthony (1919) and from the specimens in the collections of the Museum of Compara- 1955 CAVE-FOSSIL VERTEBRATES FROM CUBA 5 tive Zoology at Harvard, the American Museum of Natural His- tory in New York City, and the United States National Museum in "Washington, D. C. Order RODENTIA Boromys offella. Represented by one mandible and two iso- lated teeth from Cave 2. Originally described as fossil from Oriente (Miller, 1916) it has since been recorded from Las Villas (Allen, 1918), Pinar del Rio (Museum of Comparative Zoology specimens), and Isla de Pinos (Allen, 1918). It apparently was formerly a widespread species. Boromys torrei. Represented by one mandible from Cave 1 and by two rostra, fifteen mandibles, and three isolated teeth from Cave 2. Ten auditory bullae from the latter deposit are also tentatively referred here. This species has previously been found in Oriente (Anthony, 1919), Las Villas (Museum of Comparative Zoology), Matanzas (Allen, 1918), and Pinar del Rio (Museum of Comparative Zoology), and also on Isla de Pinos (Allen, 1918). Evidently this was formerly a widespread and common species. Capromys (Mystateles) nana. Represented by a single tooth from Cave 1, and by one rostrum, four mandibles, and six isolated teeth from Cave 2. This species is known as a living animal only from the Cienega de Zapata (Allen, 1918 and 1942), and as a fossil was only knowm from Matanzas (Allen, 1917a), Las Villas (Allen, 1918), and Oriente (Anthony, 1919). Apparently orig- inally rather widespread. Capromys {Mystateles) prehensilis. A single rostral fragment from Cave 2 is best allocated to this form. The species occurs in Pinar del Rio, Habana, Matanzas, and Las Villas, and is repre- sented by a subspecies on Isla de Pinos (Mohr, 1939). A wide- spread form in western and central Cuba. Capromys (Geocapromys) columhianus. Represented by one mandible and five isolated teeth from Cave 1, and by three man- dibles and five isolated teeth from Cave 2. This species was originally described from cave-fossil material from Las Villas (Chapman, 1892), and it has since been recorded from Oriente (Anthony, 1919), Matanzas, Pinar del Rio, and from Isla de Pinos. Apparently once a common and widespread form. Battus. Represented by many partial skulls, mandibles, and 6 BREVIORA NO. 46 postcranial skeletal parts from Cave 1 and Cave 2. Mus. Skeletal remains from both Cave 1 and Cave 2. Not as abundant as the Rattus remains. A great deal of postcranial and some cranial material, particu- larly of rodents, remains unidentified. Class AVES The bird remains collected at Cave 1 and Cave 2 were not identified. Class REPTILIA Order SQUAMATA Anolis lucius. A few dentaries and maxillary elements from Cave 1 are referable to this species. In Cave 2 only a single parietal of this species was found. This is, at present, a common lizard inhabiting- the limestone walls at the entrance of the caves. Anolis equestris. Represented by the proximal portion of a lower jaw and portion of one maxilla from Cave 2. This species was observed inhabiting the forest outside of the caves. Leiocephalus sp. A large dentary that seems to represent a species of Leiocephalus was collected at Cave 1. Part of a maxilla from Cave 1, and a dentary and a maxilla from Cave 2 are also referable to this genus. Tarentola americana. This rare gekkonid species is represented by a single dentary from Cave 2. Alsophis angulifer. A few vertebrae from Cave 2 are referable to this species of snake. Some cranial and postcranial reptilian material from Caves 1 and 2 remains unidentified. Class AMPHIBIA Order ANURA Hyla septentrionalis. Cranial and postcranial elements of this common tree frog were the most abundant non-mammalian re- mains to be found in Cave 1 and Cave 2. Some of the bones found appeared to be more rugged and larger than those of the two skeletons prepared for comparison. 1955 CAVE-FOSSIL VERTEBRATES FROM CUBA 7 DISCUSSION Most of this collection probably represents the result of owl predation. However, it is probable that some of the material does not represent owl pellet remains but rather bones washed into the cave, the remains of animals that took refuge in the cave, or remains of native kitchen middens. None of the material was stratified, all of it having been inter- mixed by the diggings for bat guano. It is therefore impossible to establish the relative age of the fossils. The presence of two introduced European species — Battus and Mus — proves that at least part of the material collected is post-Columbian in age. However, the presence of species now extinct in Cuba — Neso- phontes, Natalus primus, Boromys, and Capromys (Geocapro- mys) columhianus — or extinct in Camaguey — Solenodon and Capromys (Mystateles) nana — suggests an early pre-Columbian age for some of the material. There is no evidence of difference in age or type of deposition between Cave 1 and Cave 2. The collections from these caves also establishes the presence of a number of species previously not recorded from Camaguey. BIBLIOGRAPHY Aguayo, C. G. 1950. Observaciones sobre algunos mamiferos cubanos extinguidos. Bol. Hist. Nat. Soc. Felipe Poey, 1: 121-134. Allen, G. M. 1911. Mammals of the West Indies. Bull. Mus. Comp. Zool., 40: 173- 263. 1917a. An extinct Cuban Capromys. Proc. New England Zool. Club, 6: 53-56. 1917b. New fossil mammals from Cuba. Bull. Mus. Comp. Zool., 61: 1-12. 1918. Fossil mammals from Cuba. Bull. Mus. Comp. Zool., 62: 131-148. 1942. Extinct and vanishing mammals of the Western Hemisphere. Lancaster, Pennsylvania. Anthony, H. E. 1919. Mammals collected in eastern Cuba in 1917, with descriptions of two new species. Bull. Amer. Mus. Nat. Hist., 61: 625-643. Chapman, F. M. 1892. Notes on birds and mammals observed near Trinidad, Cuba, with remarks on the origin of West Indian bird life. Bull. Amer. Mus. Nat. Hist., 4: 279-330. BREVIORA NO. 46 GUNDLAOH, J. 1868. Eevista y catalogo de los mamiferos cubanos. Repertorio Fis. Nat. de la Isla de Cuba, Habana, 2: 40-56. 1895. Notes on Cuban mammals. Proc. Linn. Soc. N. Y., 7: 13-20. Miller, G. S. 1904. Notes on the bats collected by William Palmer in Cuba. Proc. U. S. Nat. Mus., 27: 337-348. 1916. Bones of mammals from Indian sites in Cuba and Santo Do- mingo. Smithsonian Misc. Coll. 66. (12) : 1-10. MOHR, E. 1939. Die Baum- und Ferkelratten-Gattungen Capromys Desmarest (sens, ampl.) und Plagiodontia Cuvier. Mitt. Hamburgischen Zool. Mus. Inst., 48: 48-118. Peterson, O. A. 1917. Report upon the fossil material collected in 1913 by the Messrs. Link in a cave in the Isle of Pines. Ann. Carnegie Mus., 11: 359-361. E V I O R A Mmseiamri of Compsirative Zoology Cambridge, Mass. July 12, 1955 Number 47 A NEW SPECIES OF WHIPTAILED LIZARD (GENUS Cnemidophorus) FROM THE COLORADO PLATEAU OF ARIZONA, NEW MEXICO, COLORADO, AND UTAH By Charles H. Lowe, Jr. Department of Zoolog.v, University of Arizona In a previous paper it has been shown that the holotype of the form of Cnemidophorus described by Burger (1950) as C. sacki stictogramnius represents a distinct species and is not a member of the sympatric species C. sacki to which it was assigned. In the present paper C. sacki innotatus also described by Burger (op. cit.) in the same study is similarly shown to be a different species from C. sacki; it is Cnemidophorus velox described by Springer (1928) as C. gularis velox. Both of these forms described by Burger {stictogrammus, and innotatus = velox) are sympatric with C. sacki in Arizona. Neither Burt (1931) nor Burger (1950) properly assigned Springer's (1928) material. Springer {op. cit.) designated no specimen as type and his four original specimens were cotypes. Burger {op. cit.) accepted M.C.Z. No. 37208 as lectotype of velox by Burt's (1931:124, 127) designation; this specimen is also accepted here as the type specimen. However, Burger continued the error of the identity of this specimen with the statement (1950 :3) that "It is an example of C. inornatus," and proceeded to redescribe velox as C. sacki innotatus. M.C.Z. No. 37208 is not an example of C. inornatus (= perplexus auct.) as is shown by analysis of characters of scalation ; it is an example of C. velox Springer. Burger (1950:4), in redescribing C. velox as C. sacki innotatus, states that "Springer (1928) may have had specimens of this lizard [innotatus] on hand when he described velox; in BREVIOBA No. 47 some respects his description fits it better than inornatus." This is quite true. During the past few years I have collected topotypes of C. velox from Oraibi, Arizona, and Pueblo Bonito (in Chaco Can- yon National Monument), New Mexico; both of these localities were spoken of by Springer (1928) as localities from whence came his type series of four specimens of C. gularis velox (Oraibi, 3; Pueblo Bonito, 1). The topotypic material of C. velox is species-identical with a series of topotypes of C. sacki innotatus INORNATUS N'25 TOTAL, 54 140 150 170 DORSAL SCALES (OCCIPUT TO RUMP) Comparison of C. inornatus Baird and Girard and C. velox Springer. The white rectangles enclose two standard errors of the mean. One white plus black rectangle encloses one standard deviation. The differences between the two species are highly significant. The differences between the two populations of C. velox are not significant. Data from Tables 1 and 2. Burger that have also been collected for this study in the vicinity of Kanab, Utah, the type locality. Moreover, this species (C velox) has been collected together with C. sacki at several localities in Arizona, as further discussed below. In Table 1 are presented results of statistical analysis of char- acters of scalation of C. sacki, C. inornatus, and C. velox. Table 2 presents a comparison of C. inornatus with samples of C. velox from the two type localities involved: (a) the type locality of 1955 NEW SPECIES OF WHIPTAILED LIZARD 3 Springer's (1928) C. gularis velox (Oraibi, Navajo County, Arizona), and (b) the type locality of Burger's (1950) C. sacM innotatus (vicinity of Kanab, Kane County, Utah). Both of these type localities are in Pinyon-Juniper Woodland. Only one other species of lizard in the genus Cnemidophorus is known to occur in the vicinity of Oraibi or Kanab ; this is C. tigris which is represented by the northern subspecies C. tigris septentrionalis Burger which is a valid and especially distinctive race. Prom analysis of scalation and of color pattern it is clearly evident that C. gularis velox Springer (1928) = C. sacki innotatus Bur- ger (1950). The question now arises as to whether velox is or is not con- specific with sacki (= gularis) as now understood. This question has been unequivocally answered by having found velox and sacki to be sympatric over a wide area where the ranges of the two overlap in Arizona. At such localities each maintains its own respective genetic differences without evidence of gene exchange. One is able to shoot adults of both species {velox and sacki) while standing in the same spot at localities where they are sympatric (e.g., Pinal Mountain, Gila County; Prescott Mountains, Yavapai County; Oak Creek, Coconino County). See Table 1 for certain similarities and differences in morpho- logical characters of these two species. Cnemidophorus velox is a whiptail of Woodland and Coniferous Forest habitats. It is the common striped whiptail of the so-called Colorado Plateau of northern Arizona and New Mexico, southern Utah and Colorado. It exceeds in both altitude and north lati- tude the ecologic distribution of C. sacki in Arizona and New Mexico. C. sacki does occur (marginally) in Coniferous Forest and occurs in Pinyon-Juniper Woodland. Thus on outlying mountain masses south of the main body of the Colorado Plateau, the two species are found to be sympatric in Woodland and Coniferous Forest habitats in central Arizona, near the northern limit of the geographic distribution of C. sacki and the southern limit of C. velox. The much larger C. velox is also specifically distinct from the small C. inornatus (= perplexus auct.). These two species re- semble one another superficially in color and pattern, particularly in the total absence of light spots on the body of both juvenile and adult; small individuals of the tAvo species have been most 4 BREVIORA No. 47 commonly confused. The highly significant difference in the dorsal scalation of these two species is shown in Table 2 and Figure 1. The species rather than subspecies relationship of the two is also clearly evinced by the fact that the much larger bodied form (C velox) is the one with the northward distribution (on the Colorado Plateau) while the smaller sized form (C in- ornatus) is a predominantly Mexican population ranging only into the southern United States. This relationship of surface- volume ratio to environmental temperature is opposite to that which is to be predicted on the basis of the Bergmann Principle applied to ectotherms (see Fitch, 1940:123; Cowles, 1945). Also C. inornatns is a species primarily of grasslands and plains and does not reach the higher elevations attained by C. velox in Conif- erous Forest habitats (7000-8000 ft.) in the United States. The following diagnosis of C. velox Springer (1928:102) is based upon the type specimen, topotypes recently collected, and additional series recently collected in Colorado, Utah, Arizona, and New Mexico. Cnemidophorus velox Springer Plateau Whiptail Type specimen. No. 37208, Museum of Comparative Zoology. Collected in August, 1928 by Stewart H. Springer. The type locality is here restricted to Oraibi, Navajo County, Arizona. See discussion below. Diagnosis. A moderately-sized species of Cnemidophorus char- acterized by : (1) a maximum snout-vent length of approximately 85 mm.; (2) 6 or 7 longitudinal light body stripes, with the seventh (vertebral) stripe, when occurring, less distinct than the remaining stripes; (3) total absence of light spots in the dark fields on the body; (4) ground color of the upper surfaces of the body black to blackish-brown; (5) ventrum immaculate whitish very faintly tinged with bluish; (6) distal portion of tail light bluish in adults and bright blue in juveniles; (7) 171.1 ±: 1.3 (160-184) scales on the midline from occiput to rump; (8) 73.1 ± 0.63 (63-85) scales around midbody, excepting the en- larged ventrals; (9) 7.7 ± 0.16 (5-10) scales between paraverte- bral stripes at midbody; (10) mesoptychial scales conspicuously enlarged and abruptly differentiated from the adjacent granu- 1955 NEW SPECIES OF WHIPTAILED LIZARD 5 lar scales of the gular fold; (11) postantebrachial scales not greatly enlarged (not enlarged as in C. sacki) to 4 or more times the size of adjacent scales. Distribution. On the so-called Colorado Plateau of northern Arizona and New Mexico, southern Utah and southwestern Colorado. The range extends southward on outlying mountain ranges to at least central Arizona and to central or southern New Mexico. The relationships to C. velox of a group of ap- parently non-spotted, small, mountain-dwelling populations of northern Mexico is currently being investigated. Comparisons. C. velox is distinguished from C. sacki, with which it is sympatric, by characters of color pattern and body size as well as scalation (see Table 1). C. velox is smaller, non- spotted, has a black to blackish-brown ground color, and a light bluish tail (adults) that is bright blue in juveniles; C. sacki is larger, spotted, has a light browTi to dark brown ground color, and a brownish tail. C. velox, while superficially resembling C. inornatus in color pattern, is clearly distinguished from inornatus in body size and scalation (see Tables 1 and 2). C. velox is considerably longer, larger, and with a greater number of dorsal scales. Discussion. Smith and Taylor (1950:184) give the type speci- men of C. gularis velox Springer as Butler University No. 848, thereby following Burt (1931:124, 127) while at the same time accepting and using Burger's (1950:3) nomenclature (C. i7i- ornatus with C. gularis velox a synonym) based on the type specimen in question as Harvard M.C.Z. No. 37208. The problem is not so simple as stating or assuming that Butler Univ. No. 848 is now M.C.Z. No. 37208. Mr. Arthur Loveridge of the Museum of Comparative Zoology has kindly checked details of the record concerning M.C.Z. No. 37208. He has recently informed me that the M.C.Z. has had no recorded dealings with Butler University, that the museum register records this specimen as having been received from one J, Piatt on January 11, 1934, and that it is the holotype of Cnemidophorus gularis velox Springer, collected at Lee's Ferry, Arizona, August, 1928, by Stewart H. Springer. The data on the specimen label gives the same information, including ' ' Gift from J. Piatt, Indiapolis[i-/cj." The specimen in question (No. 37208) is recorded as the holotype of Cnemidophorus gularis 6 BREVIORA No. 47 veJox Springer in the second list of type material in the M.C.Z. (Barbour and Loveridge, 1946 :92) ; this action, with type locality given as Lee 's Ferry, Arizona, antedates the action of Smith and Taylor (1950) who restrict the type locality to Pueblo Bonito, New Mexico. While it is possible that M.C.Z. No. 37208 could be Butler Univ. No. 848, there is nothing in the record to so indicate. As far as can be now determined, there is no record at Butler University for the final disposition of this specimen or any of the others in the type series of C. gidaris velox Springer. Dr. N. E. Pearson, Head of the Department of Zoology of Butler Univer- sity, informed me on December 1, 1954, that he was unable to find either a record or specimen used by Springer in describing Cnemidophorus gularis velox. In a paper entitled "An annotated list of the lizards of Lee's Ferry, Arizona," Springer (1928) described Cnemidophorus gularis velox and recorded the collection of his four specimens (cotypes) as follows: Oraibi, Arizona, 3; Pueblo Bonito, New Mexico, 1. My study of M.C.Z. No. 37208 shows it to be a speci- men of C. velox described by Springer, and not a specimen of C inornatus {^= perplexus auct.) as is stated by Burger (1950 :3) . I consider the holotype of C. velox (== C. gidaris velox) Springer to be M.C.Z. No. 37208, the type locality of which is here changed to Oraibi, Navajo County, Arizona, rather than Lee's Ferry, Coconino County, Arizona {fide Springer, 1928). Ample topotypes are now available from Oraibi and vicinity. Burger (1950) concurs in the citation of M.C.Z. No. 37208 as the type of C. g. velox Springer. The later action by Smith and Taylor (1950:184) in their restriction of the type locality to Pueblo Bonito, San Juan County, New Mexico, is not acceptable. Dr. Hobart Smith has recently concurred in this opinion (iri litt.). ACKNOWLEDGEMENTS I Avish to express my appreciation of Dr. Doris M. Cochran of the United States National Museum, Dr. Raymond B. Cowles of the University of California, Los Angeles, Dr. Howard K. Gloyd of the Chicago Academy of Sciences, Mr. Arthur Lover- idge of the Museum of Comparative Zoology, and Dr. T. Paul Maslin of the University of Colorado, for the loan of specimens 1955 NEW SPECIES OF WHIPTAILED LIZARD 7 under their care, and to Dr. J. T. Marshall of the University of Arizona for the collection of specimens important to this study. LITERATURE CITED Barbour, T. and A. Loveridge 1946. First supplement to typical reptiles and amphibians. Bull. Mus. Comp. Zool., 96:(2): 57-214. BUUGER, W. L. 1950. New, revived, and reallocated names for North American whip- tailed lizards, genus Cnemidoplioru^. Chicago Academy of Sci- ences, Nat. Hist. Misc., 65: 1-9. Burt, C. E. 1931. A study of the teiid lizards of the genus Cnemidophorus with special reference to their phylogenetic relationships. Bull. U. S. Nat. Mus., 154: 1-286. COWLES, R. B. 194.3. Surfat-e-niass ratio, paleoelimate and heat sterility. Amer. Nat., 79: 561-567. Fitch, H. 1940. A biogeographical study of the ordinoides Artenkreis of garter snakes (genus Thamnophis). Univ. Calif. Publ. Zool., 44(1): 1-150. Smith, H. M. and E. H. Taylor 1950. An annotated checklist and key to the reptiles of Mexico exclu- sive of the snakes. Bull. U. S. Nat. Mus., 199: 1-253. Springer, S. 1928. An annotated list of the lizards of Lee's Ferry, Arizona. Copeia, 169: 100-104, «« 6 * s S *> ■5 -S ft H 5 b §> i-i 1 C^ p, aj 53 5: s ^ &H ° =« .9 •S ^ ^ 41 O rrl S3 I c3 =» g 6 o ° ^6 00 CI M o +1 +1 ici in lo (M cq CI +1 +1 +1 Cl CI C] +1 +1 +1 e fin > ,^ be ^ ^ _g § .a 1 II 1 be B rt M o iS ft o q-i -^ J rt ^ 1 O a d bfi -a Fr ti ^ 1^ o S S3 g M -2 fS '^ ^ o ^ -o .2 "^ c3 "S 'til cj lO ,-1 rH O . 05 O ! J 1 co- CO ! 5 i CO CO s 2 r-i in Ci r-l !0 1 03 x .a" < o el "3 5 o 1 H O 1 CO BREVIORA MmseiLinii of Comparative Zoology Cambridge, Mass. October 10, 1955 Number 48 NOTES ON AMERICAN EARTHWOBMS OF THE FAMILY LUMBRICIDAE. I-II. By G. E. Gates Information as to variation of taxonomically important char- acters in our little known endemic lumbricids hitherto has been lacking. Some now can be recorded through the kindness of Dr. G. E. Pickford, Prof. Walter Harman, Ottys Sanders, and others. The author's thanks are extended to all those who have supplied or loaned material. EiSENiA LONNBERGi (Michaelscn) 1894 Mt. Carmel, New Haven, Conn. In wet soil at bank of river, Oct. 28, 1928, 1-0-3. Dr. G. E. Pickford. Cockaponsett, Conn. Near Indian caves, Nov. 9, 1941, 0-0-1. Dr. G. E. Pickford. Cockaponsett State Forest, Conn. In wet soil (pH ca. 5.5) under water-logged logs at edge of lake (water level low leaving muddy shore with stumiDs), Oct. 6, 1940, 0-1-2. Dr. G. E. Pickford. (Clitellate specimens were copulating when taken.) Meshoraasic State Forest, Conn. By stream (pH ca. 5.0), April 28, 1935, 0-0-3. In soil (pH ca. 5.0) in bank of stream below Hypochilus colony, Nov. 4, 1935, 0-0-1. Dr. G. E. Pickford. Woodbridge, Conn. In marshy soil of alder copse near Bald- win's Road, Wehawaug River valley, Oct. 21, 1928, 1-0-2. Same locality, in marshy soil (pH ca. 6.2), Oct. 16, 1928, 2 BREVIORA NO. 48 1-0-1. At edge of small pond, June 14, 1931, 0-0-1. Dr. G. E. Pickford. Tallulah Falls, Georgia. In ravine, in sandy loam in seep of stream, July 27, 1931, 1-0-13. J. M. Valentine per Dr. G. E. Pickford. ("Abundant" according to collector.) Mt. Vernon, Virginia. From nearby swamp, April 22, 1917, 0-0-1 (macerated). Walter D. Webb Jr., per Prof. E. Carpenter. (Two other Virginia specimens, kindly loaned by Prof. Harman, have been examined.) External characteristics. Length, 77-104 mm. (Conn.), -130 mm. (Georgia). Diameter in elitellar region which is flat on ventral side, 6-8 mm. Body becomes transversely rectangular in cross section posteriorly. Segments, 114-134 (cf. Table). Pig- mentation in dorsum, dark, rather nondescript, at least in present condition, possibly lighter in viii-xi near mD. Prostomium epilo- bous; tongue short, open or closed (3 specimens). Setae begin on ii (33) on which all are usually present, located at four comers of the body posteriorly where AB < or ca. = CD, BC < AA, AA ca. = DD. Nephropores recognizable only on a few of the segments, first seen on iv, usually in or close to CD on iv-xiii, between levels of female and male pores on xiv-xvi, elsewhere in median part of BC or dorsal to D but without regular alterna- tion or segmental symmetry, occasionally in DD on one side even in xiv-xvi. First dorsal pore on 4/5 (32), 6/7 (1). Spermathecal pores on 8/9-10/11 (30 specimens), 8/9-11/12 (1), just lateral to mD, each on a small tumescence and in a longitudinal groove so that four small tubercles seem to be present on each intersegmental furrow. Female pores at eq/xiv just lateral to B. Male pores each at bottom of a slight trans- verse depression, at eq/xv and much nearer B than C. Clitellum red (formalin preservation), saddle-shaped, reach- ing ventrally nearly to B or to tubercula pubertatis, 23/24- eq/xxx (1), xxiii-xxx (5), xxiv-xxx (22). Tubercula pubertatis longitudinally placed, rather broad bands just lateral to B, on xxvi-xxviii (31), occasionally continued onto anterior half of xxix and/or posterior half of xxv but there gradually narrowed, markedly protuberant (several) and then sometimes with b setae on median face, the b setae on several other specimens slightly 1955 AMERICAN EARTHWORMS OF TETE FAMILY LUMBRICIDAE I-II 3 displaced mesially. Epidermis in AA of clitellar segments about as thick as in genital tumescences. Genital tumescences present (or recognizable?) only on three worms, around c-d of viii (1), of xi-xii (1), around a-b of x (1), of xx-xxx (1, ex copula), of xxiii-xxx (1, ex copula), of xxiii (1). The a and h setae of xx-xxx (1), xxi-xxx (1), xxii-xxx (8), xxiii- xxix (10), xxiii-xxx (11), xxiv-xxix (1), xxiv-xxx (5), xxvi- xxviii (1) are deeply retracted into the parietes and the margin of each follicle aperture is protuberant as a slight but obvious annular tumescence. The a and h setae of x (3) and/or the c and d setae of x (7) likewise are retracted and follicle apertures have an annular tumescence. Male pore tumescences slight, median to the pore, usually not reaching either intersegmental furrow, occasionally quite unrecognizable. Internal a^iatomy. Septa 12/13-14/15 muscular and increas- ingly thickened posteriorly. Oesophageal wall slightly thickened posteriorly in x but with- out any trace of sacs. Calciferous glands large, in xi-xii, not quite reaching insertion of 12/13. Intestinal origin in xv (33). Gizzard in xvii-xviii, some special muscularity in xix in at least one of the worms. The typhlosole begins abruptly in region of xxiv-xxvii or may be recognizable from ca. xx as a flat horizontal band gradually thickened and heightened posteriorly. The shape in cross section behind xxviii is for a time like that of an inverted T but the ventral margin gradually becomes more rounded. The typhlosole is high, nearly reaching floor of gut and when the ventral portion is cut off readily separates into two lamellae, the space between with only a few delicate fibres and a little granular matter. The end, rather abrupt, normally is in region of c-cxv (c/. Table I). The subneural trunk, adherent to nerve cord, usually is fairly large and filled with blood but when empty may be unrecogniz- able. A large vessel passes up from each extra-oesophageal trunk in vicinity of 9/10. Last hearts in xi (33), usually small and empty. Hearts of x often smaller than those of ix. Hearts of vi or vi-vii sometimes also are small. Nephridia have been found only from iii posteriorly (lacking in ii?). Nephridial ducts apparently pass into parietes in region of B. Seminal vesicles three pairs, in ix, xi-xii, those of xii the larg- 4 BREVIORA NO. 48 est, those of ix the smallest though still of fair size. Male deferent ducts appear to be slightly thickened in the entalmost portion and rather sinuous but without definite loops on posterior faces of septa just behind the funnels. Spermathecal ducts slender, usually longer than the ampullae which are in ix-xi or ix-xii. Ovisacs .small, slightly lobed in each adult. Glandular tissue is present on the parietes median to A (and also lateral to B) in a, band that extends from xxx to or well towards xxii and mesially nearly to the nerve cord. Similar tis- sue is present on the parietes laterally in xi-xii of one worm but none is present over the male pore tumescences in any of these specimens. Follicles of the retracted genital setae are markedly protu- berant into the coelom, thick and opaque, only the ental end of a shaft visible at inner end of a follicle. Such follicles (c and d) were found in x of several specimens on which annular tumes- cences of the apertures had not been noticed prior to dissection. Juveniles. Glandular tissues on the parietes were unrecogniz- able but the a and h follicles of xx-xxx were conspicuously pro- tuberant into the coelom, especially in the middle of the series. Abnormality . Male funnels of x though plicate and iridescent in one sexual worm are only about half the size of those in xi. Each spermatheca of ix of one worm has two ampullae both of which contain sperm. The ampullae of the spermathecae in xii of the octothecal worm are smaller than those in the other segments and are transparent though other spermathecae are distended and with spermatozoal iridescence. Brown hodies. Ovoidal, spheroidal and discoidal bodies of various sizes, in a posterior amputee, fill coelomic cavities of several segments just in front of the anus. These bodies some- times contain setae and apparently differ from the usual "brown" bodies only in being white. Habitats. E. lonnbergi, in Connecticut, according to Dr. Pick- ford {in lit.) inhabits wet places, commonly along with Eiseniella tetraedra (Savigny) 1826 and "is not found in cultivated soils, fields, forest soils, nor in stagnant boggy ground but rather in banks of streams and rivers or in marshy backwaters." Presence in soil with a pll of 4.5 also was mentioned by Dr. Pickford. 1955 AMERICAN EARTHWORMS OF THE FAMILY LUMBRICIDAE I-II 5 Life history. Reproduction presumably is biparental ordi- narily as evidence of sperm maturation and/or reception was recognized in all but one of the adults. The exceptional worm (April 28) obviously had not yet copulated nor matured sperm though the clitellum was fairly well developed (presexual clitel- late). Marked iridescence in distended spermathecal ampullae and on male funnels, like the large soft seminal vesicles, in the other two April and the June worms, just as clearly indicated sexual maturity and copulation. Iridescence on male funnels and in spermathecae of July specimens (Georgia) was much less marked. Spermathecal ampullae of two of those worms were filled with a watery fluid or pink jelly in which there was no iridescence, and ovisacs were occupied by a brown granular debris. The slight iridescence on male funnels and in sperma- thecae of some of the fall worms (Oct. 10, 21, 28) and especially the brown debris filling the seminal vesicles again indicated a postsexual state. Two ex copula individuals (Oct. 6) and another (Nov. 4) that had recently copulated show that reproduction may occur in the fall also. Slight iridescence on the male funnels showed that some sperm already had been matured in the aclitel- late (Oct. 6) worm which had not yet copulated. Whether a summer period of sexual inactivity separates spring and fall breeding seasons remains to be determined. Parasites. Nematodes were present in the coelomic cavities of x-xi of one worm. Fairly large cysts, presumably of gregarines, are present in seminal vesicles of several worms where they are quite obvious as soon as the specimen is opened. Other cysts are present in the parietes of several specimens, some even recognizable from the exterior with which they seem to be in communication by a minute aperture. Large, transparent cysts in coelomic cavities of last few segments (several worms) have two hemispheroidal opaque bodies each with a single nucleus. Smaller opaque cysts are numerous in the same region. Ovoidal bodies with an opaque center and a transparent pe- riphery are attached each by a slender and relatively very long stalk to the parietes of segments behind xix ( one worm ) . Remarks. The coloration in vivo, according to Dr. Pickford {in lit.) is a peculiar, dark grey, irregularly interrupted by 6 BREVIORA NO. 48 lighter areas and so characteristic as to enable easy recognition of the species in the field. Pigment may have been lysed at maturity in the dorsum of viii-xii in one worm, just as happens in Eisenia foetida (Savigny) 1826, and may have been leached in others by the formalin. The coloration in some of the Georgia specimens was about the same in the ventrum as in the dorsum and was not externally distinguishable from the darkening of unpigmented worms that takes place, in certain conditions, in formalin. The clitellum of one ex copula worm is markedly tumescent, perhaps more so than on any other specimen. The tumescence of the clitellar region of the copulatory partner is however only very slight. Whether glandular tissue on parietes lateral to B in clitellar region is associated with GS follicles and tumescences or with tubercula pubertatis is unknown. An anal segment had been reconstituted after amputation (or autotomy) in 16 specimens. Regeneration of a tail in this species, as in Lum'bricus terrestris L., may not take place. In the evolution of E. lonnhergi a pair of spermathecae opening on 7/8 has been lost. Two pairs opening on 9/10-10/11 have been acquired. The genetic factors responsible for addition of a pair of spermathecae in xii of the octothecal mutant presumably have been involved in the evolution of various lumbricid species. II Eisenia carolinensis Miehaelsen 1910 Helodrilus (Eisenia) carolinensis Miehaelsen 1910, Ann. Mus. Zool. St. Petersburg, 15, p. 15. (Type locality, presumably Fayetteville, North Carolina. Type, in Hamburg Mus., intercepted in a plant shipment at Hamburg.) Eisenia pearsei Stephenson 1933, Proc. Zool. Soc. London, 1932, p. 935, fig. 28. (Type locality, Durham, North Carolina. Types, 13, in Brit. Mus.) Eisenia caroli^iensis, Cernosvitov, 1942, Proc. Zool. Soc. London, 111, (B), p. 222. (14 specimens, 5 immature, from Durham, N. C, in Brit. Mus.) 1955 AMERICAN EARTHWORMS OF THE FAMILY LUMBRICIDAE MI 7 Chapel Hill, North Carolina, April 19, 1932, 0-0-1. J. M. Valentine per Dr. G. E. Pickford. Highlands, North Carolina. In peaty leaf mould (pH 4.5) among tree roots under leaves near path on way to Pri- meval Forest, July 26, 1931, 0-0-1. In pocket of sandy black soil (pH 5.0) under dead leaves by stream in woods on path to Primeval Forest, July 25, 1931, 0-0-2. In leaf mould and sandy loam (pH ca. 5.0) near path leading up to Whiteside Mountain at altitude of ca. 4500 ft., August 21, 1932, 0-0-1. Dr. G. E. Pickford. Linville, North Carolina. In earth and leaf mould (pH be- tween 5.0 and 6.0) by stream in ravine, near road at 4000 ft. and 14 miles from Blowing Rock, July 31, 1931, 1-0-3. Dr. G. B. Pickford. Indian Gap (now Newfound Gap, in Great Smoky Moun- tains), Tennessee. In leaf mould and loamy soil (pH be- tween 5.0 and 6.0) among tree roots and rotting wood, at ca. 5000 ft., July 30, 1931, 0-0-2. Dr. G. E. Pickford. Chilhowie, Virginia. In pocket of loamy soil among grass roots at bank of stream by bridge at ca. 3500 ft., about 101/2 miles on road to Whitetop Mountain, July 21, 1931, 0-1-4. Dr. G. E. Pickford. (Several other Virginia worms, loaned by Prof. Harman, also have been examined.) Tickfaw River Valley, Louisiana, 0-7-0-0. (Mus. Comp. Zool. No. 2061.) External characteristics. Length, 80-99 mm. (normal worms only). Diameter, 5-7 mm. Segments (122?) 135-142 (c/. Table). Pigment unrecognizable in alcoholic and most formalin material. A light red characterizes dorsum of region in front of xv in freshly preserved material ( Harman 's), and at parietal incisions circular muscle layer looks red. Body four-sided posteriorly, the dorsal surface slightly wider than the ventral, the sides slightly concave. Prostomium epilobous (28 specimens), tongue open and usually quite short. Setae begin on ii (28) on which all usually are present, one pair at each angle of the body posteri- orly where CD < or ca. = AB, BC < AA < DD < iC. Nephropores first recognizable on iii, usually close to or dorsal to D on iii-vii, viii or ix, slightly lateral to B on xv-xvi (16) and thence posteriorly often at that level but occasionally in 8 BREVIORA NO. 48 DD, the location often not the same on opposite sides of the same segment. First dorsal pore on ?5/6 (2), 5/6 (25). Spermathecal pores on 9/10-10/11 (27), close to mD, each in a quite small tumescence. Female pores equatorial, just lateral to B, on xiv. Male pores equatorial, on xv, each in a transverse slit-like depression in median portion of BC. Clitellum red (formalin preservation) or white (one alcoholic), saddle-shaped, reaching ventrally nearly to B, on (xxiii)xxiv- xxxi (3), (xxiii)xxiv-xxxi(xxxii) (1), xxiii-xxxi (3), xxiii-xxxii /2 (1), xxiii-xxxii (1), xxiv-xxxi (6), xxiv-xxxii/2 (1), xxiv/2- xxxii/2 (2),xxv-xxxi (1). Epidermis in AA of clitellar segments thickened but not as much as in genital tumescences. Tubercula pubertatis lateral to B, band-like but not as wide as in E. Id7in- hergi, on xxvii-xxix (25), perhaps occasionally extending slightly onto XXX and/or xxvi. Genital tumescences conspicuous and transversely elliptical, always including both a and h setae, on ix (1), xi (3), xii (3), xvi (2), xvii (21), xviii (5), xx (7), xxi (16), xxii (1), xxiii (1), xxiv (3), xxv-xxx (27), xxxi (14), xxxii (2), xxxiii (1), c-d of xxiv (1). Each may have a greyish translucent central area. Posterior tumescences are not delimited laterally from the clitel- lum. Male pore tumescences restricted to xv (7), extending onto xvi but not onto xiv or only very slightly (5), extending well into xiv and xvi (7). Internal anatomy. Septa 12/13-14/15 muscular and increas- ingly thickened posteriorly. Longitudinal muscle band at mD and peritoneum apparently without pigment. Oesophagus with no sacs in x. Calciferous glands as in E. lonnhergi Michaelsen 1894, large, moniliform, in xi-xii, sepa- rated by a deep constriction, the posterior gland often not reach- ing insertion of 12/13. Gizzard in xvii-xviii, muscularity occa- sionally extending slightly bej^ond insertion of 16/17 or 18/19. Intestinal origin in xv (18). Typhlosole begins in region of xxii- xxvi, abruptly or rather gradually as a flat band that shortly becomes vertical, reaching a height of 11/2-2 mm. and to or nearly to floor of gut, thickly lamelliform, the interior with only a few delicate fibres and some granular debris. The highest anterior portion, as in lonnhergi; may have a deep groove on the ventral face. The end normally is in region of xcvii-cv (cf. Table II). Subneural trunk quite small, adherent to nerve cord, some- 1955 AMERICAN EARTHWORMS OF THE FAMILY LUMBRICIDAE I-II 9 times quite unrecognizable. Last hearts in xi (18), the hearts of x-xi usually slenderer than the ones in ix, none found in v. Nephridial ducts pass into parietes in the b gap. Seminal vesicles three pairs, in ix, xi-xii (18, including the juvenile), those of xi sometimes (5) smaller to much smaller than the others. Male deferent ducts apparently slightly thick- ened in the entalmost portion which is in 2-4 very short loops crowded close together on the posterior face of the septum be- hind the funnel, recognizable (as in lonnhergi) on parietes into XV. Spermathecae with short ducts not confined to parietes, ampullae in ix-x (18, including the juvenile). Follicles of the genital setae (those in copulatory tumescences) enlarged, obviously protuberant into coelom but not as long (nor as muscular?) as in lonnhergi. Glands, atrial and such as are present on the parietes in lonnhergi and species of Bimastos are lacking. Ahnormality and Variation. Spermathecal ducts of the Chapel Hill worm are confined to the parietes and each ampulla is deeply constricted by the septum, the portion anterior to the septum of quite different appearance from that behind. Seminal vesicles of ix are lacking. Genital tumescences are present on xxxi-xxxiii as well as on ix and in addition were developed around c-d of xii. Genital tumescences in front of xv and behind xxx otherwise were only on worms from Linville and Chilhowie. Regeneration. Many of the worms clearly are posterior ampu- tees (cf. Table II) and two others may be. No regeneration, except reconstitution of an anal segment, had taken place al- though the amputation in some cases probably had not been recent. Life history. Keproduction presumably is biparental as sperm had been matured and/or received in each adult. Sperma- tophores (at 25/26 in BC, on xxviii at mV) and brilliant iri- descence in spermathecal ampullae presumably indicated recent copulation by a July 31 worm (Linville) in which no iridescence was recognizable on male funnels. Funnel iridescence was bril- liant on the April (Chapel Hill) and four July 21 (Chilhowie) specimens. Male deferent ducts were iridescent in the April worm which is the only one with maximal clitellar tumescence. The single aclitellate individual (July 21, Chilhowie) with dis- 10 BREVIORA NO. 48 coloration in dorsum of clitellar segments, slight iridescence only at periphery of male funnels and empty spermathecal ampullae clearly is postsexual. Spermathecal ampullae were empty in two worms (Linville) and iridescence was lacking in opaque matter in ampullae of another as well as in transparent watery content of two, in all of which funnel iridescence was not brilliant though still unrestricted. Ovisacs of one of the worms were filled with a brown granular debris. The peak of reproductive activity in these worms presumably was passed at time of preservation. The breeding season presumably extends from April well into July. Parasites. Nematodes were present in the coelomic cavities of x-xi of one worm. Large and medium-sized cysts, opaque or translucent, probably of gregarines, were numerous in coelomic cavities of the last ten to fifteen segments in several of the pos- terior amputees. Remarks. The red coloration of the elitellum appears to be a formalin induced artifact as it was lacking in the single alcoholic specimen. Nephropores of one worm were lateral to B on the right side of xxxiv-xxxix and in DD on the left side of xxxv-xxxix. The spermathecal battery of the ancestral lumbricid presum- ably comprised two pairs opening to the exterior on 7/8-8/9. In the evolution of carolinensis two pairs, opening on 9/10-10/11 have been acquired, and all four of the original ones have been lost. As this is two more than have been eliminated in the an- cestry of lonnhergi, in that respect carolinensis is more advanced. Seminal vesicles of primitive lumbricids presumably were in four segments, ix-xii. Those of x frequently have been lost in the family but in Eisenia they still have not disappeared in some "formae" of a species. Presence of vesicles in x of one (or more?) of the types of pearsei, is about all that remains to dis- tinguish it from carolinensis, and scarcely warrants specific status, if, indeed, any taxonomic recognition. E. carolinensis and E. lonnhergi are the only lumbricids endemic in America that have retained spermathecae. No evi- dence was found in available material of either species that would indicate elimination of those organs is under way. 1955 AMERICAN EARTHWORMS OF THE FAMILY LUMBRICIDAE I-II 11 Table I Typhlosole termination and segment number in Eisenia lonnhergi Typhlosole Atyphlo- ends in solate Number of segment segments segments Locality Eemarlcs 60-61» 1 62 Georgia Amputee 61-64* 64 Georgia Amputee 72 4 76 Virginia Amputee 74 4 78 Georgia Amputee 74 7 81 Amputee 78 10 88 Amputee Juvenile 81 6 87 Georgia Amputee 82 7 89 Amputee 84 3 87 Amputee 84 8 92 Amputee 86 13 99 Amputee? Juvenile 100 3 103 Amputee 100 16 116 101 13 114 103 ? 107+ Broken 104 5 109 Georgia Amputee 105 8 113 Georgia Amputee 100-106* 10 116 Amputee 107 5 112 Georgia Amputee 108 13 121 Ex copula ? f 124 Damaged 110 14 124 ^ 110 17 127 111 17 128 114 18 132 Ex copula 114 3 117 Amputee 115 19 134 Typhlosole ends very gradually rather than abruptly as in other worms. Posterior ends had been broken off from worms not included above (not available), except for one juvenile, presumably of this species, which has 100 segments. 12 BREVIORA NO. 48 Table II Typhlosole termination and segment number in Eisenia carolinensis Typhlosole Atyphlo- ends in solate Number of segment segments segments Locality Bemarhs 65* 6 71 Amputee 73 23 96 Juvenile. Amputee? 79 21 100 Amputee 80 19 101 Amputee 88 26 114 Amputee 82-89t 11 100 Chapel Hill Amputee 90 32 122 Aniputee 93 7 100 Amputee 97 25 122 Indian Gap Amputee? 98 44 142 Linville 100 37 137 Highlands 103 32 135 Chilhowie 105 24 129 Chilhowie Amputee? * Unusually low in seven preceding segments, t Ends gradually rather than abruptly as in other worms. Posterior ends had been broken off from the specimens not listed above at time of collection and anal pieces were unavailable. E V I O R A Mijiseiiiim of Compsirsitive Zoology Cambridge, Mass. November 18, 1955 Number 49 THREE NEW SHARK RECORDS FROM THE aULF OF MEXICO By Henry B. Bigelow, W. C. Schroeder and Stewart Springer' The trawlings recently carried out by ''Oregon" of the U. S. Fish and Wildlife Service in the Gulf of Mexico have yielded two specimens of Eimopterus pusillus (Lowe) 1839, one of Centro- phorus granulosus (Bloch and Schneider) 1801, and one of Dalatias licha (Bonnaterre) 1788, sharks that had long been known in the eastern side of the Atlantic, but which had not been reported previously from the Gulf, though one specimen of Dalatias has been taken on the northern edge of Georges Bank. As the Gulf of Mexico specimens of E. pusillus and of C. granulosus are the first of their kinds to be reported from the western side of the Atlantic or tributary waters, brief accounts follow to emphasize such of their diagnostic characters as may not be immediately apparent from the accompanying illustra- tions (Figs. 1, 2), and to justify our identifications of them. Genus EtmOPTERUS Rafinesque 1810 Type species Squalus spinax, Linnaeus 1758 Whitley (1939, p. 266) has recently revived the genus Aca7i- thidium Lowe 1839, type species Centrina nigra Lowe 1834 which, in 1839, Lowe renamed Acanthidium pusillum, thinking it separable generically from Etmopteriis Rafinesque 1810, type species Squalus spinax Linnaeus 1758. But the differences in fin-characters cited by Whitley do not seem to us sufficient to justify this separation. And the specific name niger is preoccu- 1 Contribution No. 787. Woods Hole Oceanographic Institution. 2 BREVIOBA NO. 49 pied in Etmopterus by nigrum (Cloquet) 1820/ p. 93 and various subsequent authors, proposed as a substitute for spinax Lin- naeus 1758, Etmopterus PUsiLLus (Lowe) 1839 Study material. Male, 445 mm. long, U. S. Nat. Mus. No. 157835, and female of 458 mm., Mus. Comp. Zool. No. 39572, northern part of Gulf of Mexico, "Oregon" Station 1281, Lat. 29° 13' N., Long. 87° 54' W., in 250 fathoms. Also female 278 mm. long, Madeira (Mus. Comp. Zool. No. 1026), and female of 167 mm., off Equatorial West Africa, Lat. 6°08' S., Long. 11°24' B., Mus. Comp. Zool. No. 38002, received through the kindness of Dr. Max Poll. The Gulf of Mexico specimens agree so closely with those with which we have compared them, from Madeira and from tropical West Africa that we have no hesitation in referring them to the same species. The morphological feature the most sharply diagnostic for imsillus, among its genus-mates of the Equatorial and North At- lantic, of the Mediterranean and of the Gulf of Mexico, is that the denticles on the sides of its trunk are low, truncate, the great majority with concave crowns, but an occasional denticle flat, or perhaps even weakly convex, on a conspicuously 4-radiate base.^ In these respects they contrast with the slender, bristle-like den- ticles of E. spinax (Linnaeus) 1758 and of E. schultzi (Bigelow, Schroeder and Springer 1953, fig. 9G), with the stouter, more thorn-like denticles of E. hillianus Poey 1861 (Bigelow and Schroeder 1948, p. 489, fig. 92B), of E. polli (Bigelow, Schroeder and Springer, 1953, fig. 7 C) and of E. princeps (see Bigelow, Schroeder and Springer, 1953, figs. 8 E, 8 F), and with the more conical claw-like denticles of E. virens (Bigelow, Schroeder and Springer 1953, fig. 10 D, E). The color, also, of E. pusillus is distinctive. It recalls E. prin- ceps, Collett 1904, of higher latitudes in both sides of the At- lantic, and E. schultzi Bigelow, Schroeder and Springer 1953, of the Gulf of Mexico, in the uniformly dark slaty to black- tGarman (1913, p. 223) credits the earliest use of niger in this coimectiou to (iiinner 1763, but we have not been in a position to verify this citation. 2 Our earlier characterization of them as a whole (Bigelow. Schroeder and Springer 1953, p. 240) as "flat or weakly convex" was an evident misstatement. 1955 THREE NEW SHARK RECORDS FROM THE GULF OF MEXICO 5§ s O fl d Kl m. long, Zool. N out 40; a p,-^ 00 a M U3 o -** o s- ® ^ * ■-S tn , c3 3 -3 as £ oi* ^ i^.^ lO ^ s ^ -g ^ cc _o o oi ^ 13 .a ^ - .c ts o fl ^ ^ ?. s ., ^ rk Me: nial 1 at c( k ' s o S :S - ^ '^ g ^3*5" i7 ^i ^- '^ ft S3 S . M => g O H. a '^ rt 4 BBEVIORA NO. 49 ish hue of its trunk not only below, but above as well (except for the pale intraocular spot mentioned below), and in the lack of definitely outlined paler and darker areas on its sides. But the outer part of its pectoral fins, with the rear part of its two dorsal fins are white and more translucent, which is not the case either in princeps, in schultzi, or in any other Etmopterus for that matter, that is known yet from the North Atlantic, from the Mediterranean, or from the Gulf of Mexico. This very conspicu- ous feature of pusillus is shown on Lowe's (1843, PI. 6) original illustration of it (as Acanthidium pusillum) ; also on the colored illustration by Braganza (1904, PI. 2, fig. 2, as E. pusillus). Other features that in combination are diagnostic for pusillus are : that the upper margin of its caudal fin is only about as long as from the tip of the snout to the level of the second pair of gill openings ; that the distance from the rear end of the bases of its pelvic fins to the origin of the lower side of its caudal fin is shorter than from the tip of the snout to the level of the first pair of gill openings; that the rear end of the base of its first dorsal fin is much nearer to a perpendicular at the axils of the pectoral fins than to a perpendicular at the origin of the pelvic fins ; that the margins of its pectoral fins are not deeply fringed normally as they are in E. schultzi (Bigelow, Schroeder and Springer 1953, fig. 9 D) ; that the anterior edge of the first pair of gill openings is not concave enough to expose the tips of the gill-folds as it is in E. princeps (see Bigelow, Schroeder and Springer 1953, fig. 8 D) ; and that the upper surface of the head is marked between the eyes with a pale yellowish spot, as it is in E. polli, in E. schultzi, in E. virens, and in E. spinax, (where there is also a white spot "above the hind part of each orbit," Garman 1913, p. 224), but not in E. princeps. Two species of Etmopteriis, schultzi and virens, Bigelow, Schroeder and Springer 1953, had been known previously in the Gulf. The discovery of pusillus there now increases the local list to three. And a fourth member of the genus {hiUianus Poey 1861) is to be expected there for it has been found widespread in Cuban waters and in the West Indian region, and has been recorded from as far to the northward as the offing of Chesa- peake Bay. E. pusillus, originally reported from Madeira, has since been 1955 THREE NEW SHARK RECORDS FROM THE GULF OF MEXICO 5 recorded from various localities in the eastern side of the Atlantic between Equatorial West Africa (Lat. 6°08'S, see above) and the coast of Portugal, the Cape Verde Islands, the Canaries and the Azores. Earlier reports of it from the West Indian region, listed elsewhere (Bigelow and Schroeder 1948, p. 493), seem actually to have been based on E. hillianus. The discovery that pusillus occurs in the Gulf of Mexico shows that its range parallels that of Centrophorus granulosus (p. 9), also that of C. uyato Kafinesque 1810, which had been known only from the Mediterranean and from the eastern side of the Atlantic in low and mid-latitudes until 1951-1952, when "Oregon" trawled 2 specimens of it in the Gulf (Bigelow, Schroeder and Springer, 1953, p. 227). Tanaka (1912, pi. 22; p. 88) also, has referred to pusillus the Japanese shark that was earlier described and pictured by Piet- schmann (1907, p. 395; 1908, p. 654, pi. 1, fig. 2, pi. 2, fig. 2) as E. frontimaculatus. But while the latter agrees with pusillus in the nature of its dermal denticles (Pietschmann, 1908, p. 657, text figs. 1, 2), it differs from pusillus in a more rearward position of the first dorsal fin (relative to the positions of the pectoral and pelvic fins), in a relatively longer interspace between the pelvic and caudal fins, and in color pattern, with the lower surface of its head, its belly, and a definitely outlined area above and anterior to each pelvic fin much darker than the upper part of its sides. A further difference is that the lower surface is as densely clothed with denticles as are the sides and back in pusillus whereas it is largely naked in the Japanese species. Genus CenTROPHORUS Muller and Henle 1837 Type species Squalus granulosus Bloch and Schneider 1801 Centrophorus GRANULOSUS (Bloch and Schneider) 1801 Study material. Female 922 mm. long from northern part of Gulf of Mexico, Lat. 29°15'N., Long. 88°18'W., "Oregon" Sta. 1246, 200-210 fathoms, U. S. Nat. Mus. No. 157833 ; also skin of a male, 855 mm. long, eastern Atlantic, exact locality not known, Mus. Comp. Zool. No. 662. BREVIORA NO. 49 Oj 02 Oj o s f^ n. s ti &t ■-C s Cj to" ^ CD 0} -* ^ ^ TJ w 1 S o be cT a ft --1 c O -si .2 cc rt fe 'O >;; 'S i^ Is a> d re o _o I ,r: g -e - ^ « cc i! S CO - ° ® o ; 5 9 S g S « -^ ^ « i t£ ^o o S o cq ■iliV/ . .„ a - 1 1 'f- fi^ ca ti 'tS -H 12; rt Lo 1955 THREE NEW SHARK RECORDS PROM THE GULF OP MEXICO 7 The most evident diagnostic features that set apart C. granu- losus (with C. machiquensis Maul 1955) from the other two species of its genus {squamosiis Bonnaterre 1788 and uyato Rafinesque 1810) that are known from the North Atlantic prov- ince are: that the inner corner of its pectoral fins is greatly extended and narrowly pointed (Fig. 2) ; that the interspace between its two dorsal fins is longer than the head (to origin of pectoral fins) by a distance nearly as long as the eye; that the denticles on the sides of the body are low, block-like with the ridges on the outer surface converging rearward, closely spaced in quincuncial arrangement, but not overlapping (Fig. 2, D) ; and that the upper teeth are erect and nearly symmetrical all along the central part of the jaw (Fig. 2, B). In granulosus also, as in squamosus, the cutting edge of the lower teeth is par- tially serrate, but so finely so that this feature is visible only on careful examination under a lens (Fig. 2, 0). In uyato, on the contrary, the lower teeth are smooth-edged, like the uppers. Proportional dimensions, in per cent of total length, of female, 922 mm. long, northern part of Gulf of Mexico, U. S. Nat. Mus. No. 157833. Trunk at origin of pectoral. Breadth 9.7, height 10.8. Snout length in front of. Outer nostrils 2.8; mouth 9.1; eye 4.9 Eye. Horizontal diameter 4.9. Mouth. Breadth 7.9. Nostrils. Distance between inner ends 3.6. Spiracles. Distance between inner ends 7.0. Lahial furrows, upper. Length 5.4 ; distance between inner ends 6.4. Gill openings, lengths. 1st 2.6 ; 2nd 2.7 ; 3rd 2.8 ; 4th 3.0 ; 5th 3.3. First dorsal fin. Vertical height 5.4; length of base 11.1; base rearward from anterior beginning of spine 8.5 ; diagonal from emergence of spine to rear end of base 7.1 ; free rear margin of fin 6.9. Second dorsal pi. Vertical height 5.0; length of base 7.0; base rearward from anterior beginning of spine 5.7 ; diagonal from emergence of spine to rear end of base 5.4; free rear margin of fin 4.5. Caudal fin. Upper margin 20.3 ; lower anterior margin 12.3. Pectoral fin. Outer margin 12.9; inner margin 13.2; greatest width 7.6. 8 BREVIORA NO. 49 Distance from snout to. 1st gill opening 16.8 ; to origin of 1st dorsal spine 34.8; to origin of 2nd dorsal spine 69.2; upper caudal 79.7 ; pectoral 22.2 ; pelvics 58.7. Interspace between. Anterior beginning 1st dorsal spine to 2nd dorsal spine 34.4; rear tip of 1st dorsal fin to rear tip of 2nd dorsal 29.5 ; 2nd dorsal and caudal 6.0 ; pelvics and caudal 13.4. Distance from origin to origin of. Pectoral and pelvics 37.6; pelvics and caudal 18.1. Head, to origin of pectoral fins, about 28 per cent of trunk to origin of upper side of caudal fin ; snout moderately rounded, its length in front of snout about ^3 of head to origin of pectorals ; eye about 22 per cent as long as head ; distance from level of front of eyes to tip of snout about as long as eye ; spiracle about 1/4 as long as eye, its anterior edge only slightly posterior to rear corner of eye. Nostrils a little less than i/^ as long as distance between nostrils, approximately transverse, the outer ends pos- terior to level of tip of snout by a distance a little shorter than distance between nostrils. Anterior nasal flap short, narrowly triangular. Distance between nostrils a little less than i/^ as great as from tip of snout to mouth. Mouth very low-arched, its gape when closed occupying about % of breadth of head at level of outer corners of mouth. Distance between inner ends of nasal furrows about 1.8 times as long as between nostrils. Longest gill opening (5th) is almost as long as distance between nostrils. Teeth ii^fTsj of shapes illustrated in Figure 2, B; uppers with the post-functional (outermost) row partly lost, followed by a nearly vertical functional row and then by an oblique backward pointing row that will be next in service ; lowers with two rows visible from outside the mouth, one of which is func- tional, 16 teeth pointing to the right hand side, of which one is a median tooth, and 15 teeth pointing to the left hand side; cutting edge of some of the lowers with microscopically fine and somewhat irregular serrations (Fig. 2, C) but others merely somewhat irregular, a difference probably due to wear. Dermal denticles on sides of body sessile, block-like, sub-quadrate, with the more acute corner directed rearward, closely spaced in quin- cuncial pattern but not overlapping, the crown weakly convex or flat, its anterior 1/2 to % with (usually) 5-7 low ridges, con- verging rearward (Fig. 2, D), the outermost pair the longest. 1955 THREE NEW SHARK RECORDS FROM THE GULF OF MEXICO 9 Denticles at mid-level of side below first dorsal fin avera^fing about 0.9 mm. long. Length of base of first dorsal fin, measured from point of emergence from skin of anterior side of first dorsal spine, about 1.3 times as long as base of second dorsal fin, similarly measured. Interspace between rear end of base of first dorsal and point of emergence of second dorsal spine longer than head (to origin of pectorals) by a distance about equal to length of eye and about lYo times as long as between nostrils. Exposed portion of first dorsal spine about II/3 times as long as that of second spine, the second spine reaching about mid-way along the free anterior margin of the fin. Interspace between rear end of base of second dorsal fin and origin of upper side of caudal fin about as long as base of second dorsal fin measured from point of emergence, from skin, of anterior edge of second dorsal spine. Upper edge of caudal fin about as long as from tip of snout to level of fourth pair of gill openings, its lower edge with obtuse subterminal notch and rounded lower anterior lobe. Interspace between origin of lower edge of caudal fin and rear ends of bases of pelvic fins about % as long as from tip of snout to level of 5th pair of gill openings. Origin of pelvic fins (in female) anterior to a perpendicular at emergence of anterior edge of second dorsal spine, by an amount about equal to interspace between origin of lower edge of caudal fin and rear ends of bases of pelvic fins. Pectorals with inner corner greatly extended and narrowly pointed, reaching rearward (when laid back) nearly to a perpendicular from rear base of first dorsal fin. Sharks referable with certainty to granulosus, not to uyato which Rey (1928, p. 436) considered synonymous with it, had previously been recorded in the Mediterranean, off the coast of Portugal, and at Madeira (specimen in British Museum; Giinther 1870, p. 421). The capture of a typical granulosus in the Gulf now shows that its Atlantic range parallels that of C. uyato Rafinesque 1810, which was recently found to occur in the Gulf (Bigelow, Schroeder and Springer 1953, p. 227). The third species of Centrophorus that is known from the Atlantic {squam- osus Bonnaterre 1788) has so far been reported only off the coast of Portugal and northward to southwestern Iceland and the vicinitv of the Faroes. 10 BREVIOKA NO. 49 Pietschmann (1908, pp. 663-667) credits granulosus to Japan also, from his comparison of two Japanese specimens with gran- ulosus from the Mediterranean. But the eventual decision, whether the northwestern Pacific CeniropJiorus of this general character is indeed indistinguishable from the Atlantic-Mediter- ranean granulosus, is best postponed until a larger number of specimens have been examined, critically, with this question in mind. Genus DaLATIAS Rafinesque 1810 Dal ATI AS Lie HA (Bonnaterre) 1788 Study material. Female 845 mm. long from northern part of Gulf of Mexico, Lat. 28°25'N., Long. 86°02'W., "Oregon" Sta. 1275, 225 fathoms, U. S. Nat. Mus. No. 157834. The specific identification of this specimen as D. licha is so evident, from its close agreement in bodily form, fin characters, teeth, and denticles with the Georges Bank specimen described and pictured previously (Bigelow and Schroeder 1948, p. 502, figs. 96, 97) that no further account seems called for here. As the only previous record for this shark in the western Atlantic was this Georges Bank example, the capture of one in the northern part of the Gulf of Mexico widely expands its known range. In the eastern side of the mid and north Atlantic the known range of D. licha extends from Equatorial West Africa (Rio de Oro) to the Irish Atlantic slope, including the Mediterranean. And it is doubtful whether the representatives of the genus that have been reported from South Africa, from the New Zealand- Australian region and from Japan can be separated specifically from D. licha of the Atlantic. (For discussion, see Bigelow and Schroeder 1948, p. 501). REFEEENCES Bigelow, Henry B., and W. C. Schroeder 1948. Fishes of the Western North Atlantic. No. 1, Mem. Sears Foun- dation for Marine Research, Part 1, pp. 59-576, text figs. 6-105. 1955 THREE NEW SHARK RECORDS FROM THE GULF OF MEXICO 11 BiGELOw, Henry B., W. C. Scheoeder, and Stewart Sprinqek 1953. New and little known sharks from the Atlantic and from the Gulf of Mexico. Bull. Mus. Comp. Zool., vol. 109, no. 3, pp. 213- 276, text figs. 1-10. Bloch, M. E., and I. G. Schneider 1801. Systema Ichthyologiae iconibus ex iliustratum. Ix + 584 pp.; 110 pis.; 2 vols. Bonnaterre, p. J. 1788. Ichthyologie in: Tab. Enc>c. Method. Trois regnes de la Nature. Ivi 4- 215 pp., pis. A, B + 1-100. Paris. Braganza, Carlos de 1904. Esqualos obtidos nas Costas de Portugal. Keaultadoa das Invest. Sci. . . . Yacht "Amelia." Ichthyologia. II; 107 pp., 2 pis. Lisbon. Cloquet, Hippoltte 1820. Le sagre, spinas niger. Diet. Sci. Nat., vol. 1, suppl., p. 93. Collett, R. 1904. Diagnoses of four hitherto undescribed fishes from the depth.*! south of the Faroe Islands. Videns. Selskabs Porhand. Christi ania, 1904, no. 9, 7 pp. Garman, Samuel 1913. The Plagiostomia. Mem. Mus. Comp. Zool., vol. 36, xiii -f 515 pp., 77 pis. GtfNTHER, Albert 1870. Catalogue of fishes of the British Museum, vol. 8, xxv -J- 549 pp. Jordan, D. S., and B. W. Evermann 1896. The fishes of North and Middle America. Bull. U. S. Nat. Mus. No. 47, Part 1, Ix -f 1240 pp. Linnaeus, C. 1758. Systema naturae, 10th Ed., vol. 1, 824 pp., Holmiae. Lowe, E. T. 1834. A collection of fishes made in Madeira. Proc. Zool. Soc. London (1833), Part 1, 1834, pp. 142-144. 1839. A supplement to a sjTiopsis of the fishes of Madeira. Proc. Zool. Soc. London, Part 7, 1839, pp. 76-92. 1843. A history of the fishes of Madeira . . . 196 pp., 27 pis., London. Maul, G. E. 1955. Five species of rare sharks new for Madeira. . . . Notul. Natur. Acad. Nat. Sci. Philadelphia, No. 279, 13 pp., 3 pis. MuLLER, Johannes, aud F. G. J. Henle 1837. Ueber die Gattungen der Haifische und Rochen . . . Arch. Natur- gesch., Jahrg. 3, vol. 2. pp. 394 401, 434. 12 BBEVIORA NO. 49 PlETSCHMANN, ViKTOR 1907. Zwei neue Selachier aus Japan. Anz. Akad. Wiss. Wien. Vol. 44, pp. 394-396. 1908. Japanische Plagiostomen. Sitzber. Akad. Wiss. Wien, math. Nat. Kl., vol. 117, pt. 1, pp. 637-710, pis. 1, 2. PoEY, Felipe 1858-1861. Meinorias sobre la historia natural de la isla de Cuba, vol. 2, 442 pp., 19 pis. Rafinesque, Constantine Samuel 1810. Carratteri di alcuni nuovi generi e nuovi specie di animali c piante deUa Sicilia. pp. iv, 105, 20 pis. Palermo. Rey, Luis Lozano 1928. Fauna Iberica. Feces. Vol. 1, 690 pp., 20 pis. Inst. Nac. Cienciaa, Madrid. Tanaka, Shigeho 1912. Figures and descriptions of the fishes of Japan. Vol. 5, pp. 71- 86 4- [1], pis. 21-25; vol. 6, pp. 87-108, pis. 26 30. Whitley, Gilbert 1939. Studies in ichthyology no. 12. Rec. Australian Mus. Vol. 20, No. 4, pp. 264-277. B R E Mnaseum of Compsirsitive Zoology Cambridge, Mass. November 30, 1955 Number 50 NEW FEOaS OF THE GENERA ASTEBOPHRYS AND OREOPHBYNE FROM NEW GUINEA By Aethue Loveeidge During the past year the Museum of Comparative Zoology has received from Mr. Kenneth R. Slater some interesting microhylids moXudm^ Asterophrys rufescens (Macleay), valvifera (Barbour), turpicola (Miiller) ; Xenohatrachiis rostratus (Mehely) ; Mefo- postira oceUata Mehely and Cophixalus oxyrhinus (Boulenger), taken at Omati, near Port Moresby, Papua. There was also a large and handsome frog which keyed down to Asterophrys hoettgeri (Mehely) in the synopsis to that genus furnished by H. W. Parker in his "Monograph of the Microhy- lidae" (1934:60. Brit. Mus.). However, A. hoettgeri comes from Halmahera Island in the Moluccas lying to the northwest of Dutch New Guinea, while Omati is situated in the southeast of New Guinea. After Dr. E. E. Williams and I had re-examined the Omati specimen and come to the conclusion that it was probably new, it was submitted to Dr. Parker for favor of his confirmation. Meanwhile, in the hope that the collector might be able to secure more material, publication has been deferred for nearly a year. Mr. Kenneth Slater, after whom I take pleasure in naming it, now informs me that there is little hope of his obtaining a series for he will soon be returning to Australia. Asterophrys slateri sp. no v. Holotypc. Museum of Comparative Zoology No. 28205, an adult 9 , from Omati, near Port Moresby, Papua. Collected by Kenneth R. Slater between January and April, 1954. Diagnosis. Apparently closely related to A. hoettgeri (Mehely) from which it differs in having a much longer (not shorter) snout ; 2 BREVIORA NO. 50 an interorbital space that is one and one-third times as broad as (instead of equal to) an upper eyelid; tibio-tarsal articulation of adpressed hind limb reaching eye (instead of nostril) ; some details of coloration; and size, being 52 mm. (instead of 39 mm.; a difference that may be sexual). Possibly it is only racially dis- tinct. Description. Anterior palatal ridge curved, smooth. Snout subacuminate, longer than the horizontal diameter of the eye, which is equal to the distance from nostril to orbit ; canthus ros- tralis rounded ; interorbital space one and a third times as broad as an upper eyelid ; tympanum moderately distinct, its horizontal diameter about half that of the eye ; fingers long, the first much shorter than the second, which is slightly shorter than the fourth, the third being much the longest, all terminating in large sub- triangular disks which are larger than those of the toes; both fingers and toes free of web ; inner metatarsal tubercle indistinct ; tibio-tarsal articulation of the adpressed hind limb reaches the eye. Skin of head, except on occiput, pustular, especially noticeable on upper eyelids ; some glandular folds on the occiput, scapular region and dorsum, the latter also displaying a few scattered tubercles ; forelimb, more particularly the upper arm, studded with tubercles. Below, smooth. Color. Above, purplish brown spotted with black and flecked with white (on body tubercles) ; on the occiput and back are four or five, more or less W-shaped, white lines corresponding to the glandular folds ; posteriorly an obsolete, hair-like, light vertebral line is barely distinguishable ; thighs heavily blotched with black, the elongate blotches separated by irregular white lines; on the tilna these light lines suggest obsolete crossbars. Below, pale brownish spotted w4th darker browai and variegated with indistinct white vermiculations ; these coalesce to form indistinct crossbars on the underside of the forelimbs, larger, broader, and pure white on the hind limbs; lower jaw with two downward-pointing, white, triangular spots on either side of the symphysis, while a less conspicuous pair is situated near the angle of the jaw. Size. Holotype 9 . Length of head (from end of snout to back of tympanum), 18 mm.; length from snout to anus, 52 mm.; length of hind limb (from anus to tip of longest toe), 80 mm. 1955 NEW FROGS FROM NEW GUINEA 3 Diet. The stomach and intestinal contents, examined by my colleague Dr. W. L. Brown, were so finely masticated as to be undeterminable except for some fragments of small shells, the leg of an ant, and the chitinous plates of some arthropod. Parasites. The stomach wall and other organs carried numerous encysted nematodes; one threadworm (preserved) was free in the bodv cavitv. Oreophryne parkeri sp. nov. Cophixalus geislerormn Loveridge (not of Boettger), 1948, Bull. Mus. Com]). Zool., 101: 423. Holotype. Museum of Comparative Zoology No. 13964, an adult 9 , taken from running water in sago-palm forest at Matapan, Australian New Guinea. Collected by E. A. Briggs about Jan- uary, 1923. Paratype. British Museum No. 1955.1.1.17, with same data as type. Diagnosis. In Parker's synopsis to the genus (1934:160) these frogs key down to the a7ithomji-hiroi section but are not very closely related to either species. Oreophryne anthotiyi (Boulen- ger), of which we have a cotype, is a more robust frog than the gravid holotype of parkeri, from which it differs slightly in almost every character, the most conspicuous being the coloring of the underside and : Eye separated from tympanum by a distance equal to IV^ times the horizontal diameter of the latter. Total length of adult 45 mm anthonyi Eye separated from tympanum jjy a distance equal to only Ys the horizontal diameter of the latter. Total length of gravid 5 30 mm parkeri 0. parkeri agrees with hirui, of which we have a series, in size and many other respects, but differs sharply in tympanic char- acters. Tympanum scarcely distinct, very small, % to % the eye diameter; disk of third digit nearly % the eye diameter; ratio of tibia length to head width at commissure of mouth 1.1 to 1.3 (4 ex.) hiroi Tympanum very distinct, large, % the eye diameter; disk of third digit nearly % the eye diameter; ratio of tibia length to head width at commissure of mouth 1.37 to 1.43 (2 ex.) parJceri The head of parkeri, as Dr. R. G. Zweifel has pointed out to me, is relatively small compared with those of other Oreophryne. 4 BBEVIOBA NO. 50 Description. Palate with a row of three transversely elongate tubercles followed by a denticulated dermal ridge in front of pharynx. Snout truncate, subequal to the horizontal diameter of the eye, which is about equal to seven-eighths its distance from the nostril; canthus rostralis rounded; loreal region slightly oblique, distinctly concave ; interorbital space more than one and a half times as broad as an upper eyelid ; tympanum very distinct, almost two-thirds the diameter of the eye to which it is very close ; fingers long, the first much shorter than the second, which is sub- equal to the fourth, the third being much the longest, its disk almost two-thirds the diameter of the eye, all fingers terminating in large subtriangular disks which are somewhat larger than those of the toes; toes webbed at the base, the third slightly shorter than the fifth, inner metatarsal tubercle very indistinct, outer absent ; tibio-tarsal articulation of the adpressed hind limb reaches the shoulder. Skin of head smooth, except for a slightly raised median line and the suggestion of a supratympanic fold ; between the shoul- ders a ) (-shaped glandular fold. Belly smooth ( ? preservation). Color. Above, brow^n, uniform except for some dark crossbars on the limbs and light areas in groin. Below, brown, paling pos- teriorly towards groin; thighs flecked with white, tibiae largely white. Size. Holotype 9 . Length of head (from end of snout to back of tympanum), 7 mm. ; length from snout to anus (which was 28 mm. prior to dissection), now 30 mm. ; length of hind limb (from anus to tip of longest toe), 37 mm. Diet. The stomach of the holotype contained the remains of scores of a shiny black Iridomyrmex ant (smaller than either hatesi or scrutator), a group that is known to forage in trees; also a single head of one of the arboreal Podomyrma ants. Remarks. In 1948, when reporting on the "New Guinean Rep- tiles and Amphibians in the Museum of Comparative Zoology" (Bull. Mus. Comp. Zool., 101: 303-430), I erroneously referred these two frogs to Cophi.mlus geislerorum Boettger. More recent- ly, when Dr. R. G. Zweifel of the American Museum of Natural History was w^orking over our material, he redetermined them as Orcophryne of no known species. I then submitted them to Dr. 11. W. Parker who replied that a clavicle is present though not reaching the scapula, so that they are unquestionably Oreophryne, 1955 NEW FROGS FROM NEW GUINEA 5 also eleutherognathine. I am also indebted to Dr. Parker, after whom this new species is named, for supplying me with the ratios of tibial length to head width at commissure which, in the cotypes of 0. anthonyi is 1.14 to 1.38, with an average of 1.25. Parker also observes that, ac- cording to Boettger, Nieden and Vogt, C. geislerorum is without a tympanum ; the contrary view was stated by van Kampen and so got included in the monograph. E V I O R A Mmseitam of Coimparsitive Zoology Cambridge, Mass. January 27, 1956 Number 51 A SMALL MUSTELID FROM THE THOMAS FARM MIOCENE By Stanley J. Olsen INTRODUCTION For the past ten years the Museum of Comparative Zoology has conducted field work at the Thomas Farm (Florida) for at least one season each year. During the latter part of this decade an intensive search has been made with the view of increasing our knowledge of the microfauna of this unique Miocene deposit. Although minute vertebrate fossils have been collected from every part of the excavation, by far the most productive area has been the boulder bar described by Dr. T. E. White in 1942. The material — mammal, bird and reptile - — occurs here in the matrix between the boulders of Ocala lime- stone comprising the bar, and the latter have served to prevent the crushing and breakage elsewhere characteristic of the quarry. Three large mustelids have been described from this area by White : Mephititaxus ancipidens 1941, Aelurocyon spissidens 1947 and Oligohunis floridanus 1947. The isolated upper molars referred to by White in 1942 (M.C.Z. Nos. 3639, 3640) as mus- telids, have been re-identified as deciduous teeth of a canid. The lower jaws, without teeth, mentioned by him (M.C.Z. Nos. 7029, 7030) are conspecific with the specimen here described. This individual, although incomplete, marks the first occur- rence of associated mandible and tooth in a small mustelid from this locality. BREVIORA NO. 51 Family Mustelidae Subfamily Mustelinae MiOMUSTELA ( ? ) Sp. Bef erred material. M.C.Z. No. 7016 (Fig. 1), incomplete left mandible with Mi present and alveoli for C, Pi, P2, P3, and M2 (incomplete). Horizon and locality. Arikareean Miocene. Thomas Farm, Gil- christ County, Florida. '.^o^C^E^^ Miomustelai'l) sp. Labial, occlusal and lingual vieAvs. M.C.Z. 7016. X 3. Characters. Closest to Miomustela madisonae (Douglass 1903; Hall 1930) among previously known forms but differing as fol- lows : Ml with metaconid smaller, all other crown elements 25 per cent larger. Description. The mandible is short and shallow and the num- ber of premolars is reduced to three. All of the premolars are 1956 A SMALL MUSTELID FROM THE THOMAS FARM MIOCENE double rooted and their arrangement indicates a crowded tooth row. Ml has a large well-defined protoconid with a pronounced cleft between the protoconid and paraconid. The metaconid is about one-half the size of the paraconid. The talonid is basined, with a well-defined rim and the hypoconid retains its individual- ity. The hypoconulid is small but distinct. The paraconid- protoconid shear is nearly parallel with the long axis, making the protoconid-metaconid width narrow as compared with the tooth length. Mo is single rooted. The ascending ramus of the jaw starts just posterior to Mi, causing M2 to be tilted upward and forward. The masseteric fossa extends forward to the an- terior rim of the alveolus of M2. Mental foramina are present below the centers of Pi and P2. MEASUREMENTS (IN MILLIMETERS) OF MANDIBLE a TOOTH LENGTH OF TOOTH ROW AS INDICATED BY ALVEOLI f AMTPBinC lOlAl^mM OC D Tn DnCTCDinO KilADriM rtc vt \ 18.0 5.7 6.0 2.5 M, , HEIGHT OF PROTOCONID FROM LABIAL MARGIN OF ALVEOLUS 5-0 ; M, , " " METACONID " LINGUAL 2.5 M, , " " FWRACONID " " " " " 3.5 Mj , " " HYPOCONULID" " ' L7 DISCUSSION Plesictis (including Miistelavus) agrees with M.C.Z. No. 7016 in the general structure of Mi, but Plesictis has the metaconid and paraconid of equal size, and a jaw that is longer (with four premolars) and lighter in build, the ascending ramus arising well behind Mo (Teilhard de Chardin 1914; Viret 1929; Clark 1937; Simpson 1946; Hall 1951; and original material). Agreement of our specimen with Miomustela is closer than with Plesictis — sufficient at least to justify tentative generic as- signment. Minor differences between the type of Miomustela and M.C.Z. No. 7016 are an apparently lower paraconid and talonid 4 BREVIORA NO. 51 in the type. I believe that these are due to the unworn condition of the described specimen, compared with the worn cusps of the type. This is based on study of the amount of wear under- gone by Ml in a large series of Maries americana, the essen- tially similar cusp arrangement shared by this genus and Mio- mustela permitting a valid comparison. This series shows that such differences may indeed be due to wear. The species represented by this fragment is clearly distinct from M. madisonae but, considering the fragmentary nature of available specimens, I refrain from proposing a new species. Further material may, in fact, show that a new generic assign- ment is necessary. Acknowledgments. I am indebted to Professor A. S. Romer for the use of the material and to Professor Bryan Patterson for his helpful criticism. Also to Dr. J. LeRoy Kay for the loan of the type of Miomiistela. The figure was drawn by Miss Patricia Washer. EEFEKENCES Clark, J. 1937. The stratigraphy and paleontology of the Chadron formation in the Big Badlands of South Dakota. Ann. Carnegie Mus., 25: 261-350. Douglass, E. 1903. New vertebrates from the Montana Tertiary. Ann. Carnegie Mus., 2: 145-199. FiLHOL, H. 1879. Etude des mammiferes fossiles de Saint-Gerand le Puy (Allier). Bib. Ecole Hautes Etudes, Sect. Sci. Nat., 19: 1-252. Hall, E. E. 1930. Three new genera of Mustelidae from the later Tertiary of North America. Jour. Mam., 11: 146-154. 1951. American weasels. Univ. Kansas Publ., Mus. Nat. Hist., 4: 1-466. Simpson, G. G. 1946. Paleogale and allied early mustelids. Amer. Mus. Nov., 1320: 1-14. Teilhard De Chakdin, P. 1914. Les carnassiers des Phosphorites du Quercy. Ann. Paleont., 9: 103-192. 1956 A SMALL MUSTELID FROM THE THOMAS FARM MIOCENE 5 ViRET, J. 1929. Les faunes de mammif^res de I'Oligocene superieur de la Limagne Bourbonnaise. Ann. Univ. Lyon, (n.s. 1), sci. med., 47: 5 305. White, T, E. 1941. Additions to the Miocene fauna of Florida. Proc. New England Zool. Club, 18: 91-98. 1942. The Lower Miocene mammal fauna of Florida. Bull. Mus. Comp. Zool., 92: 1-49. 1947. Additions to the Miocene fauna of North Florida. Bull. Mus. Comp. Zool., 99: 497-515. E V I O R A Mmseiiiiini of Comparative Zoology Cambridge, Mass. April 6, 1956 Number 52 REMARKS ON SOME MIOCENE ANURANS FROM FLORIDA, WITH A DESCRIPTION OF A NEW SPECIES OF HYLA By Walter Auffenberg Up to now, the only reference to Miocene anurans from Flor- ida is by Tihen (1951), who described a new species of Bufo and allocated other remains to Rana sp. All of these fossils were collected at Thomas Farm, Gilchrist County, Florida. While washing a considerable amount of matrix from this same deposit in search of fossil snakes, a rather large number of amphibian remains were found, and these form the basis of this report. The new material permits addition of three more genera to the fauna of the Miocene, and also contributes to a better understanding of the relationships of the previously described Bufo. Bufo praevius Tihen This toad was described on the basis of fragmentary elements of both the axial and appendicular skeletons. The relationships of B. praevius were not discussed at length by Tihen, although he stated that the ilium is somewhat comparable to that in B. terrestris and B. valliceps. Considerable additional material from the same deposit contributes to our knowledge of its prob- able relationships. Of particular interest is the fact that there are now available cranial elements, generally considered to be quite diagnostic between species of toads. This is especially true of the orientation and degree of development of the various cra- nial crests. Information regarding these structures is now avail- able for praevius. The systematic relationships of many of our North American toads have been the subject of considerable controversy during 2 BREVIORA NO. 52 the last few years. The present paper does not deal with the status of the species or subspecies of Recent Bufo, and the latest checklist (Schmidt, 1953) has been followed throughout for tax- onomy. As comparative material, a number of Recent toads have been skeletonized and examined: Bufo t. terrestris (23), B. t. ameri- canus (1), B. w. woodhousei (2), B. w. fowleri (3), B. valliceps (2), B. yunctatus (1), -S. compactilus (1), 5. simus (1), B. gran- ulosus (1), B. quercicus (4), B. perplexus (1) and B. pelticeph- alus (1). The ilium is considered by many workers as being of consider- able value in distinguishing many genera of fossil anurans, and is the element on which the description of B. praevius rests. Sixty-two additional ilia of praevius have been collected at the type locality; these have been deposited in either the collections of the University of Florida (UF), or those of the Museum of Comparative Zoology (MCZ). One of the most important characters of the ilium is the shape and degree of development of the dorsal prominence. In praevius this prominence is fairly low, similar to the same structure in B. terrestris, valliceps and woodhousei. On the other hand, B. com- pactilis and cognatus (latter fide Tihen, op. cit.) have a spine-like prominence, being high, with a narrow base. This fact appar- ently indicates a closer relationship between praevius and the terrestris group than with compactilis and/or cognatus. In B. simus, punctatus, pelticephalus, quercicus, perplexus and granu- losus the prominence is not roughened or knob-like, but forms a smooth, well-defined, rather sharp dorsal edge. Though the ilial prominence appears, in general, to be useful in separating natural species groups in the genus Bufo, changes in shape and proportions with growth are complicating factors that must be recognized. In an examination of the shape of the prominence in a series of ilia of Bufo praevius now available, it is obvious that this character is more variable than indicated in Tihen 's original series. As in modern B. terrestris, this structure may vary from very low and smooth in small specimens, to con- siderably higher and roughened in larger individuals. Tihen stated that the ilial shaft of praevius seemed to be less compressed than in other toads with which he compared it ; but he also indicated that the degree of compression was not beyond 1956 MIOCENE ANURANS FROM FLORIDA 3 the range of variation found in the living forms. The diagnostic importance of this character is considerably lessened by the fact that a comparison of the degree of compression in the larger series of ilia of praevius now available and a series of ilia of mod- ern terrestris shows little, if any, difference between the two species. Another character of apparent diagnostic importance is the curvature of the ilial shaft, which according to Tihen is less than that in other species of Bufo which he had examined, though not beyond the range of variation of modern species. This character, as well as many others, shows considerable ontogenetic change. However, the present series tends to confirm Tihen 's observation, i.e., that the ilial shaft in praevius is, for the most part, less curved than in most other species. In none of the ilia is there a process on the antero-ventral edge of the acetabular expansion such as Tihen observed in one speci- men (MCZ 1933). He states that although the ilium may be aberrant, it might also represent still another species of Bufo. A similar element has not turned up and it is highly probable that this specimen is an aberrant one. Three fra'gmental urostyles were also available to Tihen, who stated that the crest was probably lower in praevius than in most modern species. Although a number of fragmental urostyles have been collected, only two elements have been found complete. In these, as well as in the urostyles of two specimens of Recent valliceps, the height of the crest is equal to, or slightly less than the width of the articular surfaces. This crest is practically al- ways higher than the greatest width of the anterior articular sur- faces in Recent specimens of B. t. terrestris, t. americanus, w. woodhousei and w. fowleri. A considerable number of presacral vertebrae are available for the fossil species. They do not appear to be "heavier" than those of modern species, but are seemingly identical with those of B. terrestris and B. woodhousei. A number of sacral vertebrae are also available. One sacral was found to be fused to the preceding element, forming a very symmetrical unit. This unit is clearly referrable to a bufonid, most likely representing an aberrant specimen of Bufo praevius. Of particular interest in the diagnosis of modern species of toads is the shape of the cranial crests. Fortunately, cranial ele- 4 BRBVIORA NO. 52 ments of the fossil species are now available. These include three frontoparietals (UF 9892), two temporals (UP 9893), three bones of the occipital complex and a number of squamosal stems and mandibles (UF 9894). Thus it is possible to reconstruct the major cranial ridges of the fossil form. In the following discus- sion the terminology used by Sanders (1953) is used throughout, since the system seems reasonable and practical. The available frontoparietals of the fossil species are some- what fragmentary, but indicate a number of important facts. They are cancellous and provided with well-developed crests. Furthermore, the shape and development of these crests clearly show that its relationships lie with the terrestris-woodhousei species complex. From above, the parietal ridges of praevius are slightly diverging posteriorly. The fragmentary elements are not large enough to determine the length and size of the frontal ridges, although such ridges probably existed, as based on an examination of the anterior ends of the fragments. The parietal ridges are low anteriorly, gradually increasing in height pos- teriorly, much in the manner of those of woodhousei. From above, these ridges are relatively narrower than in terrestris or houstonensis (the latter fide Sanders, op. cit., fig. 2), but like those in woodhousei. From the side they are slightly arched, as in valliceps, not straight as in woodhousei, terrestris or houston- ensis. The otoparietal ridge is well-developed and rounded, being somewhat intermediate between that in houstonensis and w. fow- leri. It is not as high and narrow as in valliceps. It is better de- veloped than in terrestris, and of a different shape. The angle formed between the otoparietal and parietal ridges is approxi- mately 90°, not greater, as in valliceps. The occipital groove, if present, is very inconspicuous. The otoparietal plate is not much wider than its ridge, as in houstonensis. However, it should be pointed out that this character, given considerable weight by Sanders (op. cit.), varies with age and possibly with locality in at least B. terrestris. The largest part of the temporal plate is cancellous, provided with a well-developed, but rounded temporal ridge. The tympanic and supratympanic ridges are definitely well developed, but to an unknown degree due to the fragmentary nature of the available elements. The parietal spur is well devel- oped, directed backward and medially. It is not club-like as in t. americanus, but is similar to that in B. woodhousei. The shape 1956 MIOCENE ANURANS FROM FLORIDA and position of the cranial crests in B. praevius and three closely related modern species are shown in Figure 1. From the foregoing description of additional remains of Bufo praevius it is highly possible that its relationships lie with the eastern toads. This has been intimated by Tihen in regard to the ilial prominence, and is y^vy much strengthened by the shape and Fig. 1. Shape of the cranial houstonensis (after Sanders, op. rius; D, Bufo w. fowJeri. crests in four species of Bufo. A, Bufo rit.); B, Biifo i. ferrefitris; C, Bufo prae- development of the cranial ridges. A few of the characters, such as the height of the urostyle crest and the arching of l^e parietal ridge are somewhat suggestive of B. valliceps. The importance of the suggested close relationship of praevius to the terrestris- woodhousei-houstonensis complex is that it indicates that this group was already well f*stablished by early Miocene. 6 BREVIORA NO. 52 Ran A sp. This genus has already been reported from the Thomas Farm by Tihen {op. cii.). Two additional ilia are now available (UF 5919). These remains are too frag^llentary to make a species identification possible, although the shape of the posterior edge of the dorsal crest is highly suggestive of the pipiens "group", including palustris, pipiens, sylvatica, clamitans, etc. They are not as readily referrable to the group including heckscheri, cates- heiana and grylio. SCAPHIOPUS cf. HOLBROOKI In the fossil record, Scaphiopus is known from Pliocene (Tay- lor, 1941) and Pleistocene (Tihen, 1954) deposits of North America. Its occurrence in the early Miocene of Florida is thus of considerable interest. The genus is frequently divided into two subgenera : Scaphi- opus (including h. holbrooki, h. hurteri and couchi) and Spea (including multiplicafns, homhifrons and hammondi). Skele- tons of all species except multiplicatus have been examined. Included in the fossil material from Thomas Farm are four fragmentary ilia (UF 9896), one maxilla (UF 9897), three frag- mentary frontoparietals ( UF 9898) and one presacral vertebra (UF 9899), all of which appear to belong to this genus. The ilium of the subgenus Scaphiopus is quite diagnostic. There is no ilial crest, and the dorsal prominence is absent or but slightly developed. When present it may exist as a small, rounded protuberance, directed dorso-laterally, about half way between the base and the end of the acetabular expansion. In the subgenus Spea the prominence is usually ridge-like, directed more dorsallj^, and contributing to the height and length of the dorsal portion of the acetabular expansion. In the fossil ilia the dorsal prominences are very small and rounded, identical with those found in the subgenus Scaphiopus. The disposition of the maxillary teeth, as well as the shape of the entire element, are also comparable to those of the subgenus Scaphiopus. There is no frontoparietal boss, such as is present in some members of the genus. On the basis of the available fossil remains, the Miocene form is not distinguishable from the modern species, Scaphiopus holbrooki. • 1956 3II0CENE ANURANS FROM FLORIDA Whether or not the Miocene Scaphiopits was identical to the species holhrooM cannot be absolutely determined at this time. However, the fossil form is clearly referrable to the subgenus Scaphiopus. Whether or not the subgenus Spca had differenti- ated by this time, or whether the former is ancestral to the latter is, of course, unknoAvn. Further collecting in other Tertiary de- posits of middle North America may clarify many of our con- cepts regarding the ancestry of this group. Fig. -J,. CJoiupiii isou ot till f u>m1 iiiifi-oliylitl iliuiii fioiu tli locality (A) with lli.it of .1 u-i cat ^pefiiiicii uf Mii-rnlti/ld (■< from Gainesville, Florida. The line represents 1 iinn. Fan. MiCROHYLA sp. Of considerable interest is the fact that one fragmentaj ilium from the Thomas Farm (UF 5144) apparently represents the genus Microhyla. Its very small size and well-developed dorsal prominence (which is triangular from the side and without any protuberances or roughened areas of any kind) are highly sug- gestiAC of this genus. In addition, the acetabular cup is very similar to that in 3Iicro]nila, in that the edges of the cavity are considerably raised above the surface of tbe acetabular expan- 8 BREVIORA NO. 52 sion. Although broken, the ventral portion of the acetabular expansion was apparently narrow and directed downward as in this genus. That this element does not represent a young speci- men of Bufo is indicated in the shape of the dorsal prominence. In small specimens of Bufo this prominence is w^eakly developed, Ijeing low and long, usually with a roughened area on, or very near the dorsal edge. The genus Microhyla has not been reported as a fossil. Additional material, both fossil and modern, is neces- sary before specific identification of the fossil form will be pos- sible (see Fig. 2). In addition to the remains already mentioned, three ilia were found which are referrable to either Hyla, Pseudacris or Acris. They are placed in the genus Hyla mainly for reasons of conven- ience. Neither Acris nor Fseudacris have been found as fossils. Hyla has been reported from the Pleistocene of North America (Brattstrom, 1953). The fossils under consideration appear to represent a new species. I wish to name this form for Dr. Cole- man Goin, who has contributed materially to our knowledge of modern amphibians in life history studies, ecology and taxonomy. Hyla goini nov. sp. Diagnosis. A Miocene Hyla with a relatively high ilial shaft; ilium without a dorsal crest ; ilial prominence rounded, well de- veloped, with the protuberance located rather low on the promi- nence, projected dorsally and laterally, its anterior edge practi- cally even with, or slightly behind the anterior edge of the acet- abulum ; ventral portion of the acetabular expansion broad at its base; acetabulum somewhat sub-triangular. Holotype. MCZ 2277 ; the distal 7 mm. of a right ilium, col- lected by Walter Auffenberg, March, 1954 (Figure 3). Horizon and Type Locality. Hawthorne formation. Lower Miocene, Arikareean; Boulder Bar. Thomas Farm, Gilchrist County, Florida. Referred Material. UF 9900; three fragmental ilia from the same locality and horizon. The ilia of the genera Hyla, Acris and Pseudacris are easily separated from those of Eana and Eleuth erodactylus by the ab- sence of a dorsal crest. From Microhyla, Bufo and Scaphiopvs they are readily distinguished by the well-developed dorsal prom- 1956 MIOCENE ANURANS FROM FLORIDA iiienee, directed dorsolaterally to laterally, and by the slender- ness of the shaft itself. The ilia of Hijla (joiui have been compared with the same ele- ment in the following- uenera and species: Hyla cinerea (12), H. squirello (6), FT. fonorali)^ (2), //. crucifer (2), H. versicolor E F Fig. 3. Interspecific variation in the ilia of Hylidae. A, Hyla femoralis, 4 miles E. Gainesville, Florida; B, Acris gryllus dorsalis, Biven's Arm, Alachua County, Florida; C, Hyla c. cinerea, Gainesville, Alachua County, Florida; D, Pseudacris oimata, 7 mi. northeast of Gainesville, Florida; E, Hyla crucifer hartramiana, Camp Olena, Columbia County, Florida; F, MCZ 2277, type ilium of Hyla goini nov. sp., L. Miocene, Thomas Farm, Gil- christ County, Florida. The lines represent 1 mm. 10 BREVIORA NO. 52 (1), H. septentrionalis (1), Pseudacris nigrita (4), P. ornata (3), P. hrimleyi (1), Acm gryllus (3). The dorsal prominence of Acris gryllus is oval and on a rather long base, located in front of, to very slightly behind, the an- terior edge of the acetabulum. In addition, the ventral portion of the acetabular expansion is shorter at its base than in H. goini. In all the hylids that I have examined, with the exception of crucifer, and to some extent femoralis, the dorsal prominence is oval, usually with a long base, and with the protuberance of the prominence located considerably higher than in goini. In speci- mens of H. crucifer, and some of femoralis, the protuberance of the dorsal prominence is relatively small, rounded, not too high, and on a fairly small base. However, the prominence in crucifer is anterior to the anterior edge of the acetabulum, and the lower portion of the acetabular expansion is considerably longer and narrower than in goini. In femoralis the protuberance is higher on the dorsal prominence and the acetabular expansion is also somewhat higher. In Pseudacris nigrita, P. hrimleyi and P. ornata the dorsal prominence is rounded, never as oval as in most species of Hyla that I have seen. Its base is relatively short and the entire prominence is usually anterior to the anterior edge of the acetabulum. However, goini is considerably larger than any modern Pseudacris. It is apparently the size of an average speci- men of Hyla cinerea. It also differs from modern species of Pseudacris that I have examined, in the shape of the acetabular expansion which is longer at its base. In addition, the somewhat .subtriangular acetabulum apparently separates it from most spe- cies of this genus. In many respects, the ilia of Hyla and Pseudacris are very sim- ilar. This is to be expected if the genera are really as closely related to one another as is generally supposed. The fossil form, goini, is clearly allied to one or both of these genera on the basis of its ilial shape. As a representative of the Hylidae in the Lower Miocene it could, conceivably, be ancestral to both genera. The fact that goini shows characters found in both modern genera may be indicative of an ancestral position, or simply a reflection of the close relationship which evidently exists between Hyla and Pseudacris. In any case, goini can be reasonably placed in the genus Hyla on the basis of its size and the fact that all of its char- acters are duplicated in one form or another in this genus. Ad- 1956 MIOCENE ANURAN8 FROM FLORIDA 11 ditioual specimens representing other parts of the skeleton are needed before its exact status can be clearly fixed. Figure 3 com- pares the ilia in certain hylid frogs, selected to show the typical shape within each species, with the type of H. goini. ACKNOWLEDGMENTS For advice and criticism of this paper I wish to thank Drs. A. B. Grobman, P. Brodkorb, E. E. Williams and A. ,S. Romer. Acknowledgment is due Mr. L. Ogren, University of Florida, who has helped me on innumerable occasions in the field, in iden- tifying fossil remains, and in preparing comparative skeletal material. I also wish to thank Mr. R. Ilighton for the loan of a ai umber of anuran skeletons. LITERATURE CITED Brattstrom, B. ]953. The amphibians and leptiles from Eaneho La Brca. Trans. 8an Diego Soc. Nat. Hist., 11 (14) :365-92. Cooke, C.W. 1945. Geology of Florida. Bull. Florida Geol. Surv., 29:1-339. Sanders, O. 1953. A new species of toad with a discussion of morphology of the bufonid skull. Herpetologiea, 9(l):25-47. Schmidt, K. P. 1953. A checklist of North American amphibians and reptiles. Amcr. Soc. Ichthyologists and Herpetologists. Univ. Chicago Press: Chicago, Illinois. Taylor, E. H. 1941. Extinct toads and salamanders from middle Pliocene beds of Wallace and Sherman Counties, Kansas. Bull. Kansas Geol. Surv., 38(6):177-96. TlHFX, J. 1951. Anuran remains from the Miocene of Florida, with the descrip- tion of a new species of Biifo. Copeia, 1951 (3 ) : 230-35. 1954. A Kansas Pleistocene herpetofauna. Copeia, 1954 (3). -217-21. E V I Mmseiimi of Comparative Zoology Cambridge, Mass. April 6, 1956 Number 53 FOOD-FINDING BY A CAPTIVE PORPOISE (TURSIOPS TRUNCATUSy By William E. Schevill and Baebara Lawrence Introduction 1 Experimental arrangements 2 Passive location 4 Active or echolocation 4 Echolocation to determine presence of fish 6 Echolocation to distinguish between alternate feeding places 7 Passive versus active location 8 Night tests 9 Vision 10 Discussion 12 Acknowledgments 14 References 14 INTRODUCTION It soon appears on acquaintance with porpoises that these ani- mals are well endowed with hearing and sight, and that they use both these senses in their normal environment. A captive Tur- siops truncatus was confronted with the problem of finding un- familiar food under varying and sometimes rather complicated circumstances, all unlike the normal hunting of his previous wild free existence. Experiments to find out which sense was relied on most and in what ways it was most useful have shown a wide and not unexpected variation in a single individual. Like man, the animal used all the clues he could get. Sometimes he listened, sometimes he looked, and sometimes he was so busy doing whatever he had last done that he missed perfectly obvious clues. 1 Contribution No. 832 from the Woods Hole Oceanographic Institution. 2 BKEVIORA NO. 53 Although there was such variation in his responses to similar circumstances that many experiments produced conflicting evi- dence, there was a certain pattern which makes possible a number of conclusions. When he was finding his food he was most eager, swift, and accurate in coming to the sound of a slap on the water. When he had no such clue he would find the fish, evidently by echolocation, if it was in a region where he expected to find food. His final searching for the fish was by eye, though he could not clearly distinguish his preferred butterfish from other offerings. His willingness to use clues, almost we might say his ability to notice them, depended on his memory and other psychological factors as well as hunger. EXPERIMENTAL ARRANGEMENTS Our laboratory on an island at Woods Hole was a pond 34 m. long, 23 m. wide, and 2.5 m. deep, cut off from the sea by a stony beach about 30 m. wide. Our subject was an old bull Tursiops truncatus, 203.6 kg. in weight and 267 cm. long, brought here from Florida especially for this experiment. Some injury had damaged his right eye, which we never observed him using, and towards the end of our work his left eye started to cloud over. Whatever the damage was, it did not affect his hearing, which was sharply directional. Temperamentally, he differed from our earlier subject (Lawrence and Schevill 1954), being bold and aggressive, and for this reason was often seen at the surface. Puzzling situations near at hand often caused him to thrust his head above the water, or "souse ont" in this way to look, and when further off he would "pitchpole" straight up for a third or more of his length. He learned fast, and while his evident pre- ference for doing as he pleased would make any formal analysis of his responses rather meaningless, this independence gave a much more varied and accurate picture of his inclinations and abilities than we could have obtained otherwise. Since he pre- ferred going hungry to being forced into situations he disliked, the tests we devised were as simple as were consistent with get- ting reliable answers. Under these circumstances, it soon bv^came clear that in many cases failure on his part to respond was due not to inadequacy of his senses, but to lack of hunger, temporary unwillingness to approach the feeder, or some other non-sensory reason. 1956 FOOD-FINDING BY PORPOISE 3 The porpoise Tvas ordinarily fed from a punt 3.6 m. long moored against the bank, usually at right angles, but sometimes parallel to it. In some experiments a small dinghy served as a second feeding station. The position of the feeder in the boat varied. The porpoise was fed dead butterfish {Poronotus triacan- thus) about 8 to 23 cm. long, held in the water by hand, and was called by an acoustic signal made by slapping the water, or by hammer strokes on a partially immersed iron pipe, or by tones (ranging usually between 500 and 30,000 cycles per second) from an audio oscillator through an underwater sound projector; the pipe and oscillator signals were always remote from the feeding station. Most of our work was done during daylight, so that we could watch where the animal was and how he responded under differ- ent conditions. It was not difficult to keep track of him because he frequently showed at the surface. Because porpoises see well both above and under water, we had to be sure than an appar- ently acoustic response was not in reality visual. This was the easier because his constant swimming kept the water stirred up and very murky. Transparency by Secchi disk from the surface was rarely as great as 70 cm. and often less than 45 cm. ; the very unusual maximum was 85 cm. on 19 September. At 11 o'clock on a typical sunnj^ day measurements made by diving after dark- adaptation gave a Secchi disk reading of 23 cm. from the diving mask at a depth of 2.4 m., and 61 cm. at the surafee. Under- water visibility will scarcely exceed the Secchi disk reading, and will be appreciably less for less bright objects. This, and observations of the porpoise 's behavior when he was finding fish by eye, led us to believe that we are conservative in saying that through the water he could not possibly have dis- cerned with any clarity objects as much as 1 m. away from him. As a further check we repeated most of our tests at night. We selected the pond because of its great freedom from noise- making animals ; the beach protected us from most of the noises of the sea outside. Our listening gear included an AX58C Rochelle salt crystal hydrophone and a WHOI sound level meter (Suitcase), and was sensitive enough to pick up very plainly the noise of fine beach sand strewed into the water 20 m. from the hydrophone. L 4 BREVIORA NO. 53 PASSIVE LOCATION One of the most conspicuous traits of our porpoise was the accuracy and confidence with which he promptly made his way to the place where the water had been slapped. There was no uncertainty or hesitation in his response to such a signal ; no mat- ter where he was in the pond, he always came with alacrity when he was thus called unless he was not hungry or there were other obvious psychological reasons for his refusing to do so. While failure to reward response to a remote signal often made him ig- nore subsequent calls, this was never the case with a slap on the surface. As often as we repeated the slap he would return quickly, though we never called him more than four or five times in suc- cession without giving him a fish. Not only was he prompt in his response to this slap, but, when conditioned to it, he chose it in preference to other clues. If (as described on p. 8) a fish was slapped in one place and held in the water in another, he would choose to go to the place of the sound instead of to the fish, until he learned this clue was un- reliable. His apparent reliance on this signal alone sometimes led him to miss a nearby fish entirely. The porpoise's hearing was sharply directional and his esti- mates of range (distance) were very close, as is shown by his repeatedly homing directly on a single slap (with no fish in the water) to well within 20 cm. from ranges often as great as 20 to 25 m. through turbid waters and without coming up to look on the way. ACTIVE OR ECHOLOCATION The evidence that our porpoise was echolocating his fish was accumulated over eleven weeks of close observation of his be- havior, when an increasing knowledge of what to expect from him under different circumstances made it possible to understand and to check his various capabilities with a fair degree of accuracy. Early in our work we noticed that our porpoise usually made a characteristic sequence of ''creaks" as he came in for a fish, and that in the last meter or so these sounds were matched to horizontal movements of his head. Bv "creaks" we mean a series 1956 FOOD-FINDING BY PORPOISE 5 of impulsive clicks made at widely varying repetition rates (from less than 10 to more than 400 per second), the slower ones sounding like knocks and the faster ones like snarls or whines. They have also been called "barks", "snores", "rusty hinge", or "rasping and grating sounds," etc., by various authors, such as McBride and Ilebb 1948, Kritzler 1952, and Wood 1954, and have been heard from several odontocetes. Some acoustic details of Tursiops calls have been given by Kellogg, Kohler, and Mor- ris 1953. The other common odontocete sound, the whistle-like squeal, was evidently not employed in echolocation, and appears to be primarily communcative. The acoustic details of these sounds will be reported elsewhere. Before long we noticed that when he was creaking he almost always swam directly to a fish held quietly in the water. When he was not creaking he would not do so. Though this suggested echolocation rather persuasively, we had to be sure that we were not inadvertently giving him other clues and that he was not using sight. His remarkably good hearing and his evident reliance on passive auditory clues made it especially important to make sure he was not coming because he heard us put a fish in water. Very occasionally it seemed highly likely this was just what he was doing, though we ourselves could never detect any sounds, even with our extremely sensitive listening gear. Responses at these few times were discounted. At other times when we were doubtful, we checked his behavior by dipping fish or fingers in and out or dabbling at the surface. This almost never brought him, nor could we get him to come unsignaled to a fish's nose in the water, even though he was coming accurately to a whole fish. Repeated checks also eliminated movements in the boat or the position of the feeder as clues. The possibility that he was using sight to guide him to a fish had also to be investigated, and we did this in two ways. First of all we devised a series of daytime experiments which, because of the murkiness of the water (see Secchi disk readings), elimi- nated any possibility that he was using vision to locate his fish from a distance. These experiments are described in some detail below. After we had learned how he reacted to these different situations we repeated the tests at night, with similar results, which are also described below. 6 BEEVIORA NO. 53 Two rather different sets of circumstances stimulated him to rely on echolocation alone in finding his food. Sometimes he used this means to discover whether or not a fish was waiting for him, and sometimes this was the way he distinguished between alternate feeding places. Echolocation to Determine Presence of Fish His reliance on echolocation to tell him whether or not a fish was waiting showed in a number of ways. Often, in the absence of an expected signal, he would circle the pond, creaking only as he passed the feeding station. This was anywhere from 1 to 5 meters away, and occasionally farther. If a fish were in the water as he passed, he would turn and swim directly to it; if there were no fish he would go on by. If a fish were slipped as silently as possible (and we believe inaudibly) into the water after he had passed when he was starting to circle away, he would usually turn and come back. We tried this many times on seven different days. Of these the first is especially significant. It was early in our work and was our first attempt to bring him to a fish without a signal, he being then conditioned to come only to the slap of a fish on the water. At first, although he passed nearby creaking, he did not come to the fish, but later that morning he was attracted some of the time. Five days later we tried again, and he had apparently learned not to wait for a signal. By now he would come in for a fish held silently in the water if he was creaking as he passed by; his decision to come on in seemed usually to be made at a distance of less than 5 but occasionally as much as 15 m. This was repeated a number of times during the next four days, and again two months later. Our evidence that the porpoise was not seeing the fish before deciding to swim to it was partly his excessive distance when he turned toward it and partly the fact that he often had his bad eye towards the punt as he passed. As well as using echolocation to find fish when he suspected fish should be available, he also used this means to confirm the reliability of a signal. While he almost never refused to come to the sound of a fish slapped on the water, various circumstances on several occasions made him uncertain about other signals. Early in his training he learned that an oscillator note or pipe- banging in various parts of the pond meant that he would get a 1956 FOOD-FINDING BY PORPOISE 7 fish at the feeding station, and, like our earlier porpoise, on being signaled with no fish in the water, he would make his way to the proper place to get it. Sometimes, however, he ignored the remote signal until there was a fish in the water as well. With this to convince him he would come unhesitatingly. This reaction was especially clear on four occasions when we had made changes which he did not like in the feeding station. These were all times when he was in good health and hungry, and the remote signal was one we knew he could hear. Until he was accustomed to the new arrangement, no matter how often we called, he always waited till there was a fish in the water before responding to the signal. The first time we had to signal six times with a fish in the water before he began to come on signal only. The other three occasions, at a much later date, were at successive feedings on two days. The first morning it took twelve fish, that afternoon three, and the following morning seven before he would come with no fish in the water. At these times he was creaking as he swam and for the most part responded to the signal by starting towards the feeding station, but with no fish in the water he usually turned away at 2V^ meters or more, though sometimes he circled as near as a meter and a quarter. When he was thus relying on echo- location to tell him of the presence of a fish, he rarely troubled to look above the surface. Later, when he was coming on the remote signal with no fish held in, he often soused out of the water on the way in, eyeing the situation from a distance of ten or more meters. His ability to arrive at a fish he could not see was further demonstrated one afternoon when his left eye failed. At that time, on a remote signal, he repeatedly swam directly to fish held anywhere over a 51/^ m. radius. echolocation to distinguish between Alternate Feeding Places To test his use of echolocation, two feeders slapped fish simul- taneously on opposite sides of the punt and then one held a fish in the water while the other held a hand out over the water as if feeding. The distance between slaps was about 2 m., and a net projecting 2.5 m. from the end of the punt and hanging to the bottom of the pond prevented the porpoise from circling close 8 BBEVIOBA NO. 53 to investigate by eye. We alternated feeding in irregular fashion and the feeders often exchanged places in order to eliminate any other clues. The first time we tried this he came to the correct side 24 times and the wrong side 11, the next time he made 45 correct responses and 13 wrong. His behavior was similar on both occasions. He never ignored a summons and always came directly and fast, creaking as he approached deep. The Secchi disk reading was 61 cm., and he was at the very least 2.5 m. from the fish when he had to decide which side of the net to go. In contrast to his behavior with a remote signal, he made no optical checks on the way in, though sometimes when feeding was slow he pitchpoled out, looking at us. We cannot say how often his wrong responses were caused by the net interfering with his sound patterns. We suspected this on some occasions, while other errors probably are evidence that his echolocation was not perfect. PASSIVE VERSUS ACTIVE LOCATION Experiments to find out what kinds of clues were most success- ful or preferred produced interesting results. One rather simple but very instructive one, here called the A-B experiment, con- sisted of slapping at A and putting a fish in the water at B or vice versa, with the distance between A and B farther than he could see. This, with minor variations, we tried many times on each of fifteen difi'erent occasions, and though his responses varied, they made a very neat pattern. The first time we tried it, and when w^e returned to it after a period of other work, he would always swim, creaking, directly to the point of slap, search diligently there, and usually swim away unrewarded. If, in leaving, he found the fish, he would thereafter search at the point of slap and then swim directly to the fish no matter how we varied the relation of these two places to each other. Finally, it was possible to destroy his confidence in the slap as a worth- while clue, and then for the most part he would go directly to the fish. When he swam directly to the fish there was no possi- bility that memory could have guided him, because the fish might have been any^vhere over a six meter stretch. Nor could vision have helped, because he was never nearer than 61/2 meters when we signaled, and usually a great deal farther away. When he 1956 FOOD-FINDING BY PORPOISE y looked first at the point of slap and then went on to find the fish, not pausing on the way, the distance between the two places (minimum 1.3 m., usually more, sometimes as much as 5.2 m.) seemed to preclude the possibility of his being guided by sight, the more so because in this secondary finding of a fish it made no apparent difference whether he approached with his bad or his good eye towards it. In one variation of the A-B experiment we fixed the points 2 m. apart. If we slapped at A with the fish in the water at B or vice versa, he went directly to the fish ; if we slapped with no fish in the water, he searched at the point of slap but never in- vestigated the alternate place as he left unrewarded ; if we let him search at the point of slap and then eased a fish quietly into the water at the other place, he would immedately go to it. On a number of other occasions we tried this third modification, vary- ing the place where we put the fish, and he always creaked his way to it. This is not entirely conclusive, as there is a small but unlikely possibility that he might have heard the fish put in, though our efforts to check this led us to believe that this was not passive location (cf. below, p. 11). NIGHT TESTS All of the experiments described above were carried out dur- ing the day. We also tried most of these same experiments on dark nights and found the results closely paralleled our daytime observations. On two nights when we tried to get him to take fish unsignaled, he came, apparently directly, making about ten successful runs each time, though he sometimes swam past, creak- ing, without coming in, and sometimes appeared to search near the fish without taking it. On two other nights we tried the A-B experiment, the first time feeding ten fish which he readily found, though it was too dark for us to see if he made a pre- liminary search at the point of slap. During the second and more prolonged A-B experiment he repeatedly came to the fish, not the point of slap, especially when he came from far down the pond. Occasionally he searched first at point of slap, and twice he missed the fish entirely. The last night feeding was partly from the dinghy in the middle of the pond. Two slaps informed him that fish were to be had in the vicinity ; thereafter 10 BBEVIOEA NO. 53 on remote signal he came directly and accurately seven times in a row to a fish held anywhere in a radius of 6 m. Later, when we fed from the punt, he responded to the correct side 6 times, to the wrong side 2, and did not respond at all 2 other times. VISION While sound was important to our animal, we have good evi- dence that he relied greatly on vision as well, both above and beneath the surface of the water. A bad if not completely blind right eye made him left-sided in his approach to things he wished to see. This was a convenient check when we could not see his eye rolled towards what interested him. When on arrival from Florida he was dumped into the pond, his first cheek of his new surroundings was optical. As soon as he hit the water he swam off fast and silently, sousing high out of the water and blowing frequently, with his good eye towards the shore. It was three minutes before we heard him utter a single sound, and nearly two minutes more before he spoke up as loudly and persistently as he did for most of the rest of his stay. It was hard not to think that he was looking over his new sur- roundings, perhaps searching for a break in the beach. His obvious awareness of things on the shore showed in many ways. During his first two weeks in the pond there were often clusters of people working at different places along the bank. At such times he often blew near them, rolling a little on his right side so that his left eye cleared the surface. Soon he took his sur- roundings more for granted and his inspections of the shore were less frequent, though he quickly noticed changes. Possibly also with the passage of three or four weeks he became more ac- customed to his blind eye ; at all events he rolled more often on an even keel, and without bringing his good eye out of the water. While this kind of check on his surroundings was not impor- tant in helping him find his food, it did have a bearing on his behavior at feeding time. Too many people on the bank near the feeding station made him shy, and he would come in deep and depart hastily. The presence of someone in the boat from which we fed aroused his interest, and he would blow nearby, looking. Sometimes it even seemed as if he reacted differently to different feeders. 1956 FOOD-FINDING BY PORPOISE 11 When we actually called him to eat with the well-understood slap of a fish ou the water, he rarely troubled to make an optical check on the way in as he did with a remote signal, but swam directly from wherever he was, to collect his morsel. At other times when he was less sure of the summons, or when we were slow sending signals, he rolled high, looking towards the feeder, or soused out to see what was going on. On occasions when he was more than ordinarily curious, he would pitchpole out of water as far as his flippers, with his good eye looking ventrad towards us. In addition to keeping track of things above the water, under ordinary circumstances he relied on vision to a great extent in his final accurate taking of a fish from the feeder. Often we could see his eye rolled forward towards the fish. As with our earlier animal, space permitting he would turn over on his side when close, and in the eleven weeks we fed him he only once took a fish with his right eye up. When he lingered at the end of the boat waiting for a fish it was always with the left eye up. When we fed him in a sort of narrow stall 1.2 m. wide so that he did not roll over as he approached, he swung his head from side to side farther to the right than to the left, so that his left eye was in position to scan both sides of the stall as well as the end of the punt. E"'ish put in on his blind side or above his head did not attract his attention unless they were splashed. Fish put in nearby and directly in front of him he also took in more fumbling fashion, and on at least two occasions actually bumped into them before seizing them. On the other hand, anything within his range of vision quickly caught his attention. When a fish was moved 15 cm. or so above the water he would follow it with his eye and when it was held lower would put his snout out to snatch it. Repeatedly, we found a difference in his fish-taking when visi- bility was especially poor or he himself not seeing well. This usually meant that he would begin his search farther from the fish, nodding his head more widely as he approached slowly, and would be more hesitant about taking the fish, sometimes fumbling and dropping it. A few times, when the fish were small, he missed them entirely. This was, true not only when the water was murky, but also when we held the fish deep. These indications that vision is important in his ultimate locat- 12 BREVIORA NO. 53 ing of the fish were borne out by his behavior on two separate occasions when the salinity in the pond was down and the water dirty. At these times it appeared that his left eye also was not seeing well. On the first occasion, for three days his fish-finding was less accurate and his search wider than was usual. Once he even bit at the corner of the punt, though the fish was less than 50 cm. away. The other occasion was one afternoon when his left eye failed; then, lie approached with it shut, and would lie left side up at the end of the punt, not noticing the fish when it was moved above the surface, but coming to grab it clumsily when it was put in the water. Once when he swam past a fish in the water, he worked his way back to it slowly with very exag- gerated head noddings and took the fish awkwardly deep in his mouth. Though his sight recovered after this, he never seemed to see out of his left eye as well as in the beginning. A whitish spot began to form and he would partly close his eye against a low sun, whether because it hurt or dazzled was impossible to tell. At these times his fish-taking was again less accurate, and con- trasted with a greater ease when the sun was not shining directly in his eye. While taking the fish was easier if he could rely on seeing it, he apparently could not discriminate between objects very well. Squid and flattened tin cans don't resemble butterfish much, but even after he had found out he did not like the first two he re- peatedly took them in his mouth when they were offered instead of fish. In the same way he would bite at floating vegetation near the boat, at bits of rope, or even at a rusty pail. Size seemed easier for him to tell, and he not infrequently appeared to in- spect and then leave small butterfish, about which he was not enthusiastic. DISCUSSION In this study we took a gregarious free-ranging animal that in nature is almost always found at least in small groups and often in immense herds, and placed it in solitary confinement. Whereas wild porpoises hunt live food that occurs in schools, ours was obliged to take single dead fish from the hand. Moreover, the single fish was usually close to a boat or a bank, instead of in the more open water to which wild Tiir slops are accustomed. 1956 FOOD-FINDING BY POEPOISE 13 Perhaps the most striking result of our work is the great varia- tion observed in the way our single animal sought his food. His primary reliance on passive auditory clues could have been due to the small size of his target and to the confused echo patterns in the pond. Nevertheless, evidence accumulated that he often echolocated the food we offered him, thus supporting the wide- spread supposition (for example, Kellogg, Kohler, and Morris 1953) that this was how cetaceans hunted. The sounds the porpoise made at these times were faint; indeed, only the very loudest were audible to a submerged man, and, in fact, were picked up by our sensitive listening gear only because we at last had a porpoise in a really quiet place. Thus we learned that the supposed taciturnity of solitary porpoises (Lawrence and Schevill 1954, pp. 229-231) is rather a relative matter; it appears that they merely speak very softly. The noisy listening conditions of our previous experiment had led us into error when we reported (op. cit., p. 229) "the complete absence" of evidence for echoloca- tion although we cited some, unrecognized, at the bottom of page 227.^ The only evidence we had been aware of was in McBride's posthumous note (in press) on net avoidance (what our porpoise taught us about this will be reported in another paper). To demonstrate whether an animal is using echolocation, the most definite way is to show that acoustic intereference affects performance. Thus Griffin and Galambos (1941) and Griffin (1953) by deafening bats and nocturnal birds showed that these animals then collided with obstacles that, undeafened, they had avoided. It is of course necessary to make sure that other senses, such as smell (evidently not available to cetaceans) or sight, have been excluded. Furthermore, it must be shown that sounds suit- able for echolocation are produced. This last point is abundantly proved for porpoises (e.g.. Wood 1952, 1954). The role of sight in our porpoise's food-finding has been discussed under Vision. We did not deafen our animal or interfere with his sound production. Therefore, our evidence for echolocation by the porpoise is essentially that he consistently found fish when we could convince ourselves that no other clue (sight or sound not made by the porpoise himself) was available. 1 On page 414 of an article published while this paper was in press, Griffin (1956. Hearing and acoustic orientation in marine animals. Deep-Sea Re- search, 3, Suppl. (1955), pp. 406-417) suggests that .just such an improved signal-to-noise ratio might reveal evidence of echolocation by porpoises. 14 BREVIORA NO. 53 The porpoise's performance seemed poorer on targets behind him, particularly at the longer ranges. In general, the creaks with the higher repetition rate were heard at the shorter ranges, but this orderly arrangement was usually confused, perhaps because of additional targets. At close range the creaks were timed to a horizontal sweeping of the head (nodding when on his side). These observations maj^ be interpreted as indicating direc- tionality, presumably in his sound production. We have not investigated this arresting possibility further, except to consider that perhaps the pneumatic cephalic sinuses may modify the radiation of sound from the larynx. Echolocation was evidently not a perfect method for our por- poise. Perhaps the fault lay in the special conditions in the pond, where the presence of multiple reflections from the stones in the banks and bottom must have confused the echoes. The primary target was a small fish; behind it was the punt, and behind that was the shore. We noticed that when being fed from the small dinghy (with so much less boat in the water to return an echo), he ordinarily came right to the fish with less hesitation than when feeding at the punt. These are indications that echolo- cation did not give him clear and unequivocal information (human users of this technique will sympathize). ACKNOWLEDGMENTS This work was supported by an Office of Naval Research con- tract with the Woods Hole Oceanographic Institution. The por- poise was generously supplied by Marine Studios, Marineland, Florida. We are most grateful also to the Naushon Trustees for hospitably permitting us to use and modify the pond and its immediate surroundings. And of course we were greatly depend- ent on the ever-ready support of mam' of our colleagues at Woods Hole, without whose help the maintenance of the pond and porpoise would have been impossible. We with to thank Drs. R. H. Backus and C. P. Lyman and Prof. D. R. Griffin for critically reading the manuscript. REFERENCES Griffin, D. E. 1953. Acoustic orientation in the oil bird. Sieatornis. Proc. Nat. Acad. Sci., 39 (8) : 884-893. 1956 FOOD-FINDING BY PORPOISE 15 Griffin, D. E., and Galambos, R. 1941. The sensory basis of obstacle avoidance by flying bats. Jour. Exper. Zool., 86: 481-506. Kellogg, W. N., Kohlee, R., and Morris, H. N. 1953. Porpoise sounds as sonar signals. Science, 117 (3036) : 239-243. Kritzler, Henry 1952. Observations on the pilot whale in captivity. Jour. Mamm. 33 (3) : 321-334. McBride, Arthur F. (In press.) Evidence for echolocation by cetaceans. Deep-Sea Research, 3: (2) McBride, A. F., and Hebb, D. O. 1948. Behavior of the captive bottle-nose dolphin, Tursiops truncatus. Jour. Compar. Physiol. Psychol., 41 (2): 111-123. Lawrence, B., and Schevill, W. E, 1954. Tursiops as an experimental subject. Jour. Mamm. 35 (2) : 225- 232. Wood, F. G., Jr. 1952. Porpoise sounds. A phonograph record of underwater sounds . . . Tursiops truncatus and Stenella plagiodon. Published by the Marineland Research Laboratory. Wood, F. G., Jr. 1954. Underwater sound production and concurrent behavior of captive porpoises, Tursiops truncatus and Stenella plagiodon. Bull. Mar. Sci. Gulf and Caribbean, 3 (2) : 120-133. E V I O R A MmseiMn of Compairative Zoology Cambridge, Mass. June 6, 1956 Number 54 A REVISION OF THE GENUS BBACHYMELES (SCINCIDAE), WITH DESCRIPTIONS OF NEW SPECIES AND SUBSPECIES By Walter C. Brown INTRODITCTION The genus Brachymeles was erected by Dumeril and Bibron (1839) to accommodate a single species, B. bonitae, of this unique group of Philippine skinks. Subsequently, B. bicolor was described by Gray (1845) as the type of the genus Senira; B. gracilis and B. sahadenhergi were described by Fischer (1885) ; and Taylor described B. elerae and B. hurksi (1917), B. suluensis and B. vermis (1918), B. houlcngeri (1922) and B. pathfinderi and B. ivrighti (1925). In 1922 Taylor correctly assigned B. suluensis to the synonymy of B. gracilis and also pointed out that B. houlengeri is closely related to the latter. It now appears that B. houlengeri is probably best regarded as a geographically isolated population which is only subspecificaliy differentiated from B. gracilis. It differs only in minor characteristics from the nominate form and does not overlap in range with any other population of this species as it does with B. schadenhergi. A re- examination of the type of B. bonitae, for which assistance I am deeply indebted to Dr. Jean Guibe, indicates that B. hurksi is Menlo College and Natural History Museum, Stanford University 2 BREVIORA NO. 54 conspecific with it. Two additional species are described in the present paper. The members of this genus are secretive, burrowing lizards that exhibit to a high degree elongation of the body and reduction of limbs, eyes, and ears — specializations that in general charac- terize other strictly burrowing skinks. The most highly special- ized forms have rarely been encountered by field zoologists. As a result, several of the species are represented by only a few individuals and nothing of their inter-island variation is known, assuming for the moment that their range, as in the instance of B. honitae, includes more than one island. Contrarily, at least five of these highly specialized species are known only from type localities, and the possibility also exists that many of these popu- lations have very restricted ranges and are represented by rela- tively small numbers. Substantiating evidence can be pointed out in the instance of B. tridactylus. Four examples of this species were collected in the mountains of southern Negros Oriental, some 20 kilometers west of Bais, during a field period of about two weeks in January and March, 1955, whereas no specimens were uncovered in the Cuernos de Negros area of the same mountain range, about 50 kilometers to the south, although this area was much more thoroughly investigated over a nine month period from July, 1954 to March, 1955. The author is of the opinion that this species does not occur in the Cuernos de Negros area. Other species, as B. gracilis and B. schadenhergi are widespread within the archipelago, and sufficiently large samples are available from several of the islands that well differentiated populations can be recognized. These are regarded as distinct subspecies. No member of the genus has been recorded from outside of the Philippine Archipelago. However, in view of its wide range within this group of islands, including the Sulus and Palawan, both of which are in close proximity to Borneo, it is not im- probable that the genus may occur in North Borneo, at least. Several of the more specialized species can be readily dis- tinguished on the basis of the degree of reduction of the limbs and elongation of the body. Greater difficulty exists in the recog- nition of valid species and subspecies of the pentadactyl forms. P^rt of the difficulty is due to the fact that certain of the differ- entiating characters are not easily quantified, and other, fre- 1956 REVISION OF THE GENUS BRACHYMELES 3 (luently used, key characters may hold for the separation of populations of certain species on a given island, but have been found not to be of value when one is concerned with populations of the same species occurring on a different island. It is hoped that the extent of our present knowledge of the variability of these species may have overcome this difficulty. In the course of the present study 266 specimens have been examined. The name of the institution in whose collections cited specimens are deposited is, in most instances, abbreviated as follows : C.A.S., California Academy of Sciences, San Francisco, California C.N.H.M., Chicago Natural History Museum, Chicago Illinois M.C.Z., Museum of Comparative Zoology, Cambridge, Massa- chusetts N.H.M.S.U., Natural History Museum of Stanford Univer- sity, Stanford, California S.U., Silliman University, Dumaguete City, Negros Oriental SYSTEMATIC DISCUSSION Brachymeles Dumeril and Bibron BmchymtUs Dumeril and Bibron, 1839, Erpt. Gen., 5: 776 (type species: Brachymclcs honitae Dumeril and Bibron, 1839, by monotypy). Senira Gray, 1845, Cat, Lizards Brit. Mus. : 98 (type species : Senira bicoloi Gray, 1945, by monotypy). Brac'hymelvs Agassiz, 1846, Nomeu. Zool. Index Univ.: 51 (emendation). Diagnosis and definition. Pterygoid and palatine bones not in contact mesially, both without teeth; maxillary and mandibular teeth, moderate, conical ; body elongate ; limbs reduced or absent ; eye small; lower eyelid scaly; external ear opening small or absent. The following key will serve to distinguish the known species of the genus Brachymeles. 1. Limbs absent B. vermis Limbs present - •2. Limbs pentadactyl 8 Either fore or hind limbs or both with less than 5 digits 3 3. Digits entirely absent or a single vestigial claw present B. honitae t 4 BREVIORA XO. 54 Two or more digits present (frequently represented only as clawed vestiges) 4 4. Limbs with 2 digits B. samarensis sp. nov. Limbs with more than 2 digits 5 5. Limbs with 3 digits B. tridactylus sp. nov. Limbs with more than 3 digits 6 6. Fore limbs with 5 digits ; hind limbs with 4 B. pathfinderi Fore and hind limbs with 4 digits 7 7. Midbody scale rows -22-; scale rows between the parietals and the base of the tail 83-85 B, elerae Midbody scale rows -28-; scale rows between the parietals and the base of the tail -102- B. wrighti 8. Limbs short, length of the hind limb about 9 to 12 per cent of the snout-vent length for mature individuals; number of scale rows along the vertebral line between the parietals and the base of the tail 92-94 B. licolor Limbs moderate, length of the hind limb about 16 to 25 per cent of the snout-vent length for mature individuals; number of scale rows along the vertebral line between the parietals and the base of the tail very rarely as great as 75 (generally 63 to 73) 9 9. Size at maturity 75-129 mm. snout-vent length for 28 specimens; supranasals large in contact (30 of 45 specimens examined) or narrowly separated; number of midbody scale rows 26-32 (mean ^ 27.9 ±: 0.265) for 49 specimens; number of middorsal scale rows between the parietals and the base of the tail 67-73 (mean = 69.1 ± 0.214) for 49 specimens; venter and lower lateral sur- faces very light wdth no or very few dark spotted scales B. schadenberfji Size at maturity 57.5 to 95 mm. for 72 adult specimens; supra- nasals moderate, generally rather widely separated, in contact in only 2 out of 50 specimens examined ; number of midbody scale rows 24-28 (mean = 26.0 ± 0.104) for 81 specimens; number of middorsal scale rows between the parietals and the base of the tail 63-69 (mean = 65.6 ± 0.132) for 100 specimens; loAvev lateral surfaces and frequently venter with numerous dark Spotted scales (except in B. qracilis houlengeri) B. gracilis Brachymeles vermis Taylor Brachymeles vermis Taylor, 1918, Philip. Journ. Sci., 13:255 — Bubuan Id., Sulu Archipelago. Material examined. Jolo Id. 5 (C.A.S. 60720-22, 60857-58) ; Bubuan Td., Sulu Arch. 1 (C.A.S. 62489). 1956 REVISION OP" THE GENUS BRACHYMELES 5 Diagnosis. Habitus very slender; suout to vent length 64-76 nun. for 3 mature specimens; limbs absent; no ear opening; no postnasal; 22-24 scale rows around the middle of the body (6 specimens) ; 104-109 scale rows along the middorsal line between the parietals and the base of the tail (6 specimens). Range. Known from Bubuan, Jolo and Sulu Islands, Sulu Archipelago. Brachymeles bonitae Dumeril and Bibron Brachymeles bonitae Dumeril and Bibron, 1839, Erpet. Gen., 5:777 — Manila, Luzon Island. Braclnjmeles burTcsi Taylor, ]917, Pliilii). .Tourn. Sci., 12:275 — Sumagui, Mindoro Island. Material examined. Holotype (examined by Dr. Jean Gnibe) ; Mindoro Id. 3 (C.A.S. 62064, C.N.H.M. 22525, S.U. R-20) ; Luzon Id. 3 (C.A.S. 61376-77, 62578) ; Kalotkot Id. 2 (C.A.S. 60556-57); Polillo Id. 3 (C.A.S. 62278-79, 62575). The single specimen (No. 1151, now C.A.S. 62578) referred to B. bonitae by Taylor (1917, 1922), Avith which he compared B. hurksi when describing that species, is somewhat aberrant, whether compared with the type of B. bonitae or B. burksi; it differs primarily in the much greater number of scale rows along the middorsal line betAveen the parietals and the base of the tail, 113 instead of 100 to 106 for 11 specimens of B. bonitae (holo- type 104). The fusion of the first pair of lower labials with the mental, illustrated by Taylor, is characteristic of the type of bonitae and also of 6 of the 9 specimens of B. burksi examined for this character. Whether or not the greater number of mid- dorsal scale roAvs is actually an indiAndual aberration or char- acterizes a more or less isolated population of B. bonitae cannot l)e determined at this time. Diagnosis. Habitus A-ery slender; snout to vent length 57-82.5 mm. for 6 mature specimens ; limbs reduced to vestiges without CA-idence of toes or Avith a single claAv, length of hind limb about 3.5 to 4.3 per cent of the snout-vent length (4 specimens) ; no ear opening; no postnasal; 22-23 scale roAA^s around the middle of the body for 10 specimens ; 100-106 scale rows along the mid- dorsal line betAveen the parietals and the base of the tail. One specimen (C.A.S. 62578) from Los Banos, Luzon, is far out of 6 BREVIORA NO. 54 this range, exhibiting 113 mid dorsal scale rows, but as noted above, is tentatively referred to this species. Range. Known from Luzon, Mindoro, Polillo and Kalotkot Islands in the northern part of the archipelago. Brachymeles samarensis sp. nov. Holotype. C.N.H.M., No. 44472, a juvenile, collected by G. N. Rysgaard, at Guiuan, Samar Island, Philippine Islands, January 10, 1945. Diagnosis. A slender Brachymeles with very short limbs, pos- sessing only two reduced digits on both the fore and hind limbs ; midbody scale rows -22- ; scale rows along the vertebral line be- tween the parietals and the base of the tail -86-. Description. A Brachymeles of very slender habitus; head little wider than the body, tapering anteriorly to the rounded snout; rostral large, in broad contact with the frontonasal; nostril in a small nasal ; no postnasal ; supranasals widely sep- arated in the midline; prefrontals of moderate size, widely sep- arated ; 5 supraoculars, anterior two in contact with the frontal ; 6 superciliaries ; frontoparietals in contact; interparietal large, round-pointed posteriorly; parietals in contact posterior to the interparietal; a pair of nuchals which are narrower than the parietals; two frenals, anterior slightly longer and wider than the posterior ; first upper labial largest, fourth beneath the orbit ; no external evidence of ear; number of scale rows around the middle of the body -22-, number of scale rows along the vertebral line between the parietals and the base of the tail -86-; limbs very short, length of hind limb about one fifteenth the snout - vent length; digits reduced to 2 clawed stumps on each foot. Measurements of holotype. Snout to vent 43.5 mm.; length of hind limb 3 mm. Color (in preservative). Dorsal and lateral surfaces are dark yellowish brown, each scale generally being dark spotted pos- teriorly; the venter is only slightly lighter. Relationship. It would appear to be more closely related to B. elerae than to B. honitae for, although the extent of reduc- tion of the limbs and digits is somewhat intermediate, the elonga- tion of the body as measured by the number of middorsal scale rows between the parietals and the base of the tail is essentially 1956 REVISION OP THE GENUS BRACHYMELES 7 the same as that of B. elerae, 86 as compared to 83-85 for 2 specimens of the latter. The number of scale rows between the parietals and the base of the tail for 11 specimens of B. honitae is 100-106. Range. Known only from the type locality. Bracuymeles tridactylus sp. nov. Holotype. N.H.M.S.U. No. 18354, a probably mature female, collected 1 to 3 kilometers northwest of Mayaposi spring, about 20 kilometers west of Bais, Negros Oriental, at an altitude of about 2000 feet on January 12 to 17, 1955, by Mr. Filomeno Empeso. Fig. 1. Dorsal view of head of Brnchymeles tridactylus. Paratijpes. N.H.M.S.U. Nos. 18355-56 and M.C.Z., No. 54258, collected at the same locality as the holotype. Diagnosis. A slender Brachyineles, with short limbs, possessing three reduced, clawed digits on both the fore and hind limbs; scale row^s 22-24 ; scale rows along the middorsal line between the parietals and the base of the tail 95-103. Description. A Brachymeles of very slender habitus; head not or scarcely wider than the body, tapering anteriorly to a narrow, truncate margin of contact w^ith the frontonasal ; nostril in a minute nasal ; supranasals not in contact in the midline for any of the present series ; prefrontals moderate, rather widely separated ; 4 supraoculars, anterior two in contact with the frontal ; 5 superciliaries ; frontoparietals narrowly or moderately 8 BREVIORA NO. 54 separated in the midline; interparietal large, rather pointed posteriorly ; j^arietals meeting posterior to the interparietal ; a distinct pair of nuchals in only one specimen ; two f renals, nearly equal, or the anterior slightly shorter and wider; first upper labial largest, fourth beneath the orbit; no external evidence of ear; number of scale rows around the middle of the body 22 to 24; number of scale rows along the vertebral line between the parietals and the base of the tail 95 to 103 ; limbs much reduced, the length of the hind limb about one sixteenth to one twentieth the snout-vent length ; digits reduced to three clawed stumps on each foot. Measurements of holofype. Total length 106 mm. ; snout to vent 65 mm. ; axilla to groin 40 mm. ; snout to forelimb 13 mm. ; length of hind limli 4 mm. The largest specimen measures 78 mm. from snout to vent. Color (freshly preserved material) . Dorsal and lateral surfaces are nearly chocolate brown to dark slate brown; the venter is somewhat lighter in shade. Actually the basal part and center of each scale is pigmented, the overlapping lateral and distal edges are not. Relationship. This species is apparently most closely related to B. bo'nitae from which it differs primarily in less reduced limbs and the presence of three stump-like, clawed digits on each limb. B. honitae has lost all the digits or retained a vestige of only one at the tips of the minute limbs. Range. Known o\\\y from the type locality. Brachymeles elerae Taylor Brachymeles clcme Taylor, 1917, Philip. Journ. Sci.. 12:273 — Philippine Islands. Material examined. Luzon Id. 2 (C.A.S. 61499-500). Diagnosis. Habitus very slender; snout to vent length 69-71 mm. for 2 mature specimens ; limbs gi^eatly reduced with vestiges of 4 clawed toes on both fore and hind limbs; length of hind limb 7.2 to 9.3 per cent of the snout-vent length (4 specimens) ; no ear opening ; no postnasal ; -22- scale rows around the middle of the body ; 83-85 scale rows along the middorsal line between the parietals and the base of the tail. Range. Known definitely from Balbalan, Mountain Province, Luzon Island. 1956 REVISION OP THE GENUS BRACHYMELES 9 Brachymeles wrighti Taylor Brachymelcs wrighti Taylor, 1925, Philip. Journ. Sei., 26:106 — Trinidad, northern Luzon Island. Material examined. Holotype (M.C.Z. 26589 : examined by Mr. Arthur Loveridge). Diagnosis. Habitus slender; snout-vent length 130 mm. for one specimen ; limbs greatly reduced with 4 clawed toes on both fore and hind limbs, length of hind limb 7.5 per cent snout-vent length (1 specimen) ; ear opening questionable because of injury; no postnasal; -28- scale rows around the middle of the body; -102- scale rows along the middorsal line between the parietals and the base of the tail. Range. Known only from the mountains of northern Luzon Island. Brachymeles pathfinderi Taylor Brachymeles pathfinderi Taylor, 1925, Philip. Journ. Sci., 26:104 — Glan, Cotabato Province, Mindanao. Material examined. Holotype (M.C.Z. 26581 : examined by Mr. Arthur Loveridge) ; Mindanao 2 (M.C.Z. 26582-83). Diagnosis. Habitus slender; snout to vent length 58-61 mm. for 3 mature specimens; limbs greatly reduced with vestiges of five clawed toes on the fore limbs and four on the hind limbs, length of hind limb 13.6-17.2 per cent of snout- vent length (3 specimens) ; ear opening minute; no postnasal; 22-23 scale rows around the middle of the body; 59-67 scale rows along the ver- tebral line between the parietals and the base of the tail. Ra7igc. Known only from the type locality. Brachymeles bi color (Gray) Senira blcolor (part), Gray, 1845, Cat. Lizards Brit. Mus.: 98 — Philippine Islands. Material examined. Holotype and one additional specimen (examined by Mr. J. C. Battersby). Diagnosis. Habitus slender; snout to vent length 155 mm. for 2 mature specimens ; limbs reduced with 5 clawed toes present on each foot, length of hind limb 9.7 to 11 per cent of snout-vent length (2 specimens) ; -28- rows of scales around the middle of 10 BREVIORA NO. 54 the body; 92-94 rows of scale along the vertebral line between the parietals and the base of the tail. Range. Early records are given as from the Philippine Islands. More definite information as to the distribntion of this species can not he stated at this time. Brack YMELES gracilis (Fischer) Several populations of this species exhibit rather conspicuous and constant differences in color pattern and in certain instances in scalation or other characters. The Negros population does not exhibit the narrow, dorso-lateral light stripe which in gen- eral characterizes other known populations. The venter and lower lateral surfaces of the Polillo population are very much lighter in color than is true of the Mindanao-Sulu population. The dark spotting of the venter is somewhat intermediate for examples from the Negros and Bohol. The postnasal is small and does not contact the second upper labial in the Mindanao-Sulu population, but is larger and in contact with the second upper labial in most examples of the other known populations. Also as pointed out by Taylor (1922), the ear opening is better developed in B. g. hoidengeri than in B. g. gracilis. The population from Bohol Island is in closer agreement with regard to color pattern, distinctness of the dorso-lateral stripe and the size of the postnasal shield with the population of the northern islands (Luzon and Polillo) than it is with that of Mindanao to the south, and is referred to the subspecies bon- lengcri Taylor which was described from Polillo Island. If this interpretation is correct, intervening islands of Leyte and Samar are in all probability occupied by this subspecies. The differ- ences in the number of midbody scale rows and the number of middorsal scale rows between the parietals and the base of the tail exhibited by populations of different islands are not sig- nificant (Table 1). Braciiymeles gracilis gracilis (Fischer^ Eumeces (Kiopa) gracilis Fischer, 1885, Jahrli. Aviss. Aiist. Hamburg, 2: 85 — Mindanao Island. Braciiymeles sulnensis Taylor, 1918, Philip. Journ. Sei., 13:254 — Bubuan Island, Sulu Archipelago. 1956 REVISION OF THE GENUS BRACK YMELES 11 Material examined. Basilan Id. 2 (C.A.S. 60365-66); Minda- nao Id. 32 (N.H.M.S.U. 18596; C.A.S. 15567; C.N.H.M. 52637, 52642-52, 52654-70, 52800). Diagnosis. Habitus rather slender ; snout to vent length 57.5-82 mm. for 10 mature specimens ; length of hind limb 15.3 to 19.2 per cent of the snout to vent length for 10 adults ; 'postnasal short, not in contact with the second upper labial in any of 30 speci- mens examined for this character; 24-26 scale rows around the middle of the body for 36 specimens ; 64-69 scale rows along the middorsal line between the parietals and the base of the tail for 36 specimens ; narrow dorsolateral light stripe generally present, extending anteriorly to the posterior edge of the orbit and posteriorly to the groin or nearly so ; lower lateral surfaces dark spotted ; midventral region distinctly dark spotted ( 24 out of 30 specimens), the dark spots generally occupying the basal and central portion of each scale. Range. Known definitely from the 8ulu Archipelago and Mindanao. Brachymeles gracilis boulengeri Taylor Brnchymeles boulengeri (part) Taylor, 1922, Philip. Bur. Sci. Pub. No. 17:246 — Polillo Island. Material examined. Polillo Id. 6 (C'.A.S. 62272-77) ; Luzon Id. 2 (C.A.S. 61096-97); Bohol Id. 24 (S.U. R-353; N.H.M.S.U. 18271-76, 18707-11, 18714-20; M.C.Z. 54252-3). Diagnosis. Habitus rather slender ; snout to vent length 61-92 mm. for 19 mature specimens; length of the hind limb 17.9-23.8 per cent of snout-vent length for 19 mature specimens ; postnasal long, in contact with the second supralabial in 24 of 26 specimens examined ; 24-28 scale rows around the middle of the body (mean = 26.1 ± 0.175) for 27 specimens; 63-66 scale rows along the middorsal line between the parietals and the base of the tail (mean == 64.2 ± 0.198) for 27 specimens; narrow dorso-lateral stripe (in life near to Monkey skin or Vassar tan, Maerz and Paul, 1930, pis. 6 and 10) generally present (27 out of 28 specimens), extending anteriorly onto the supraorbital region and posteriorly to the groin (this stripe is generally more dis- tinct than in B. g. gracilis) ; lower lateral surfaces dark spotted ; venter with moderate dark spotting or with such spotting absent. 12 BREVIORA NO. 54 Table 1 Comparison of subspecies of Brachymeles gracilis Dorsal scale rows No. of midbody between parietals Dorsolateral scale rows and base of tail stripe present B. g. gracilis R = 24-26 R = 64-69 Yes (Mindanao and M = 25.6 ±0.197 M = 66.4 ± 0.201 Basilan) N=36 N=36 B. g. boulengeri R = 24-26 R = 64-66 Yes (Polillo) M = 25.7 ± 0.305 M = 65.2 ± 0.367 N = 6 X = 6 B. g. boulengeri R = 26,26 R = 63,65 Yes (Luzon) N = 2 N=2 B. g. boulengeri ( '? ) R = 24-28 R = 63-66 Yes (Bohon M = 26.2 ±0.158 M = 64.0 ± 0.199 N = 21 N =21 B. g. taylori R = 24-28 R = 64-68 Xo (Negros) M = 26.5 ± 0.097 M = 65.6 ± 0.150 N=49 X— 49 B. g. taylori R =26 R =66 No (Mindoro) N = l X = l R = range, M = mean, N = number of specimens Color (freshly preserved material). The following notes were based on material which had been in preservative only a day or two. Six to 8 middorsal scale rows are Natal Brown to Chestnut (Maerz and Paul, pi. 7), not uniform but each scale with a darker brown blotch at the base. A dorso-lateral stripe (generally one and a half scale rows in width anteriorly and up to one and two half scale row's posteriorly) is somewhat variable usually near Monkey skin or Army brown (Maerz and Paul, pis. 5 and 6). Five to 7 lateral rows exhibit prominent, dark brown spots; the venter is Rose tan (Maerz and Paul, pi. 12), occasion- ally with small broAvnish spots. Four or 5 rows on the chin, posterior to the mental, are dusky or brownish. 1956 REVISION OF THE GENUS BRACHYMELES 13 Range. Known from Polillo, Luzon and Bohol Islands. The latter population is assigned to this subspecies with some reser- vations, pending the time when collections are available for com- parison from the islands of Samar and Leyte. Brachymeles gracilis taylori subsp. nov. lirachymeles Ixnilengeri (i)art) Taylor, 1922, Philip. Bur. Sci. Pub. No. 17:246. Holotype. N.H.M.S.U., No. 18615, a male, collected by W. C. Brown and D. S. Rabor on low ridge on the north side of the Maite River, about 13 kilometers west of Dumaguete on Cuernos de Negros, Negros Oriental, July 4, 1954. Paratypes. ^Negros Id. 126 (N.H.M.S.U. 17954-55, 18611-14. 18616-24, 18626-28, 18630-31. 18633-37, 18664-80, 18694-700, low Fig. 2. Dorsal view of head of Brachymeles gracilis taylori. ridge north of the Maite River, Cuernos de Negros, about 3-7 km. west of Luzuriaga, Negros Oriental, July, 1954 to February, 1955; 18597-18604, 18659-63, high ridge north of Maite River and on north slope of north peak of Cuernos de Negros, about 6-8 km. west of Luzuriaga, Negros Oriental, February and March, 1955; 18638-58, 18691-93, 18728, 18748-50, ridge, south of Maite River, east slope of Cuernos de Negros, about 4 to 7 km. west of Luzuriaga, Negros Oriental, November, 1954, to March, 1955; 18686-88, east side of Cuernos de Negros, Negros Oriental; 12027, 12224, 12226. Dumaguete environs, Negros Ori- 14 BREVIORA NO. 54 ental, August, 1940 ; 18751, 18729-32, lowlands 2 to 6 km. west of Dumaguete, Negros Oriental, July and August, 1954; 18689, V2 km. S.W. of Palimpinon, Ocoy River Valley, Negros Oriental, August, 1954 ; 18605, 18681-85, 18690, 18752, Mayaposi environs and hills to the north and west of Mayaposi, about 20-30 km. west of Bais, Negros Oriental, January to March, 1955; 18727, llaya sitio, 3-4 km. south of Caliling, Negros Occidental, Decem- ber, 1954 ; S.U. R-117-18, 242-44, 246, 251, 262, 267, 272-73, 275- 78, Cuernos de Negros area, Negros Oriental; M.C.Z. 54254-6, low ridge north of Maite River, Cuernos de Negros, 3 to 7 km. west of Luzuriaga, Negros Oriental, 1954. Diagnosis. Habitus rather slender ; snout to vent length 62-95 mm. for 32 mature specimens ; length of hind limb 16 to 22.5 per cent of snout-vent length for 28 mature specimens; postnasal long, in contact with the second supralabial in 37 of 45 speci- mens examined for this characteristic ; 24-28 scale rows around the middle of the body for 49 specimens; 64-68 scale rows along the middorsal line between the parietals and the base of the tail for 49 specimens ; dorsolateral light stripe not present ; lower lateral surfaces and generally venter dark spotted. Color (freshly preserved material). The dorsum and the upper lateral surfaces vary from light to dark brown, forming a pattern of dark and light, narrow, longitudinal stripes or nearly uniform. The lower lateral surfaces vary from near Rosetan to Blush or darker (Maerz and Paul, pi. 12), frequently with scat- tered to numerous small brown spots. Bange. Known from Negros Island and probably Mindoro. Brachymeles schadenbergi Fischer The Negros population exhibits constant differences in certain characteristics for a sufficiently large series of specimens to jus- tify its recognition as a subspecies which is distinct from the typical one of Mindanao Island. The 3 available examples from •Jolo and 2 specimens from the mountains of northern Luzon are close to the Negros population in the number of middorsal scale rows between the parietals and the base of the tail, and the presence of a more or less distinct, dorsolateral light stripe, and are tentatively referred to this subspecies. The single specimen from Palawan (C.A.S. 15571) lacks the light stripe and would 1956 REVISION OF THE GENUS BRACHYMELES 15 appear to agree most closely with the Mindanao population. if larger samples do indeed prove the Luzon, Negros and Jolo populations to be the same from a taxonomic standpoint, and that of Palawan, Basilan, Mindanao. Ijeyte and Bohol to be taxonomically identical, this distributional pattern, which is not wholly in accordance with logical routes of dispersal in terms of what is known of the geological history of the archipelago, sug- gests three possible interpretations: (1) chance colonization of widely separated islands, across intervening sea barriers, by two distinct subspecies; (2) more or less isolated populations of a polytypic species exhibiting chance convergence with respect to the presence or absence of a dorsolateral stripe; (3) two population groups of sibling species, very similar morphologi- eall}^, which have maintained disjunct distributions within this archipelago. The first interpretation is accepted with some reser- vations in this paper. Brachymeles schadenbekgi schadenbergi (Fischer) Kaineces (Biopa) schadenhergi Fischer, 1885, Jahrli. wiss. Aiist. Hamburg, 2: 87 — • Miudanao Island. Material examined. Basilan Id. 15 (C.A.S. 60305-10, 60312, 60438-40, 60493-98) ; Mindanao Id. 7 (C.N.H.M. 22528-29, 52638- 41, 61968-64) ; Leyte Id. 3 (N.II.M.S.U. 18701; C.N.H.M. 42779, 42792); Bohol Id. 5 (N.II.M.S.V. 18702-06); Palawan Id. 1 (C.A.S. 15571). Diagnosis. A moderately large Brachymeles with rather well developed limbs; number of midbody scale rows 26-28 (mean = 26.5 d= 0.161 for 25 specimens) ; a light dorso-lateral stripe ab- sent ; the second pair of chin shields separated by three scales. Color (freshly preserved material from Bohol Island). Dor- sum and upper surface of limbs are blackish brown; lateral surfaces are near Melon, Baby rose, Crabapple, Burmese gold or Burnt orange (Maerz and Paul, pis. 2 and 3). The venter is whitish, lightly tinted with the above shades. Ratif/f. Known from the Sulu Archipelago, Mindanao, Leyto and Bohol Islands (probably also Palawan Island). 16 NO. 54 Table 2 Comparison of subspecies of Brachynieles schadenbergi Dorsal scale rows No. of midbody between parietals Dorso-lateral scale rows and base of tail stripe present B. s. schadenbergi E = 26-28 E = 67-71 Xo (Mindanao and M = 26.5 ±0.161 M = 68.8 ± 0.250 Basilan) N = 25 X=25 B. s. schadenbergi E =28 E = 68,68,70 No (Leyte) X = 3 X=3 B. s. schadenbergi E =28 H. = 70,71 (4) No (Bohol) N=5 X=5 B. s. schadenbergi ( -') E =28 E = 70 No (Palawan) X = l X = l B. s. talinis E =28-32 E = 67-70 Yes (Negros) M = 30.4 ± 0.276 ^r = 68.2 ± 0.193 X =11 X = 12 B. s. talhnis (?) E = 29,30,32 E 72,72,73 Yes (Jolo) X =3 X=S B. s. talinis ( ='; R = 26,28 E = 69,70 Yes (Luzon) X ^ 2 X = 2 E = range, M = mean, X = number of specimens Brachymeles schadenbergi talinis subsp. nov. Holohjpe. N.H.M.S.U. No. 18358, a female, collected by F. Empeso, December 14, 1954 on the low ridge north side of the Maite River, 5 to 6 km. west of Luzuriaga, Negros Oriental. Paratypes. N.H.M.S.U. Nos. 12225 collected Dumaguete en- virons, Negros Oriental, 1940; 18359, low ridge north of the Maite River, Cuernos de Negros, about 5-6 km. west of Luzuriaga, Negros Oriental, Dec, 1954 ; 18363-64, 18366, high ridge north of the Maite River and on north slope of north peak of Cuernos de Negros, about 6-8 km. west of Luzuriaga, Negros Oriental, Dec. 1953, March, 1955: 18360-62, ridge "south of the Maite 1956 REVISION OF THE GENUS BRACHYMELES 17 Kiver, east slope of Cuernos de Negros, about 6-7 km. west of Luzuriaga, Negros Oriental, Nov., 1954, March, 1955; M.C.-Z. 54257, high ridge north of Maite River, Cuernos de Negros, about 6-8 km. west of Luzuriaga, Negros Oriental; S.U. R-261, ridge south of the Maite River, east slope of Cuernos de Negros, about 6-7 km. west of Luzuriaga, Negros Oriental; C.N.H.M. 22527, Mt. Canlaon, Negros Oriental. Diagnosis. A large Brachynielcs with rather well developed limbs, differing from the typical subspecies in the generally greater number of midbody scale rows 28-32 (mean = 30.4 ± 0.670 for 11 specimens) ; a moderately wide (two and two half scale rows at the shoulder), dorso-lateral light stripe, reddish in life, sometimes fading in older specimens; second pair of chin shields generally wdder, separated by 1 scale in 8 specimens, 2 in 1 specimen and 3 in 2 specimens as compared to 3 in a sample of 23 specimens of the nominate subspecies examined for this character. Fig. 3. Dorsal view of head of BracJujinelc.s scliadnibergi talinlf;. Color (freshly preserved material). Dorsum is dark reddish brown, the lateral margins of the scales tending to be dull red- dish producing an indistinct, lined pattern. This dorsal brown band is rather uniformly six and two half scale rows in width and is bordered on either side by a dull red stripe one and two half rows in width. (These fade to Avhite or whitish tan in preservative.) These stripes begin just posterior to the eye and extend to the region of the hind limlis or the base of the tail. 18 BREVIORA NO. 54 They are most unifonu and conspicuous in juveniles, tending to become somewhat paler and irregular, especially on the posterior part of the body for some adults. The red stripes are bordered ventrally by one to three or four rows of brownish or brown spotted scales. Lower lateral surfaces grayish white to light tan, frequently diffused with orange-yellow and with scattered brown spots especially in the midventral line and the region of the hind limbs. The labials are brownish ; mental, rostral, nasals, supranasals and postnasals grayish. Range. Knowai definitely at the present time from Negros Island (probably also Luzon and Jolo Islands, see p. 14) The name talinis in the native dialect refers to mountains such as Cuernos de Negros which this subspecies inhabits. ACKNOWLEDGMENTS I am greatly indebted to Dr. Jean Guibe, Museum National d'Histoire Naturelle, Paris, and Mr. Arthur Loveridge, Museum of Comparative Zool- ogy, Harvard University, and Mr. J. C. Battersby, British Museum (Natural History) for their kindly assistance in re-examining the types of Brachy- meles honitae, Bracliymeles tvrighti, Brachymeles pathfinderi and Brachymele.s bicolor; to Dr. Eobert F. Inger, Chicago Natural History Museum and Dr. Joseph R. Slevin, California Academy of Sciences for permitting me to e.xamine material of this genus in the collections of their respective institu- tions. I also wish to tliank Dr. Inger for his kindness in suggesting that I describe at this time a new species from Samar Island which is represented Ijy a unique specimen in the collections of the Chicago Natural History Museum. This study avus begun while the author was serving as a Fulbright lec- turer at Silliman University, in the Philippine Islands, under the auspices of the United States Educational Foundation, and it was completed while working as a research associate at Stanford University supported by a grant from the National Science Foundation. Illustrations were prepared by Mr. Walter Zawojski, Stanford Research Institute. LITERATURE CITED BouLEXGER, George Albert 1887. Catalogue of the lizards in the British Museum. Ed. 2. London, vol. 3, i-xii -f 575 pp., pis. 1-40. Duml'ril, Andre Marie Const.\nt, and Gabriel Bibron 1839. Erpetologie general. Paris, vol. 5, i-viii -f 855 pp., pis. -IS-SS. 1956 REVISION OF THE GENUS BEACHYMELES 19 Fischer, Johann Gustav 1885. Ichthyologische uud herpetolog'ische Bemerkungen. Part IV: Uber eine Kollection von Amphibien und Eeptilien von Mindanao. Jalirb. wiss. Anst., Hamburg, vol. 2, pp. 80-81. Gray, John Edward 1845. Catalogue of the specimens of lizards in the collection of the British Museum. London, i-xviii + 289 pp. Maerz, A. and M. Eea Paul 1930. A dictionary of color. New York, McGraw Hill Book Co., i-vii + 207 pp., pis. 1-56. Taylor, Edward Harrison 1917. Brachymeles, a genus of Philippine lizards. Phil. Journ. Sci., vol. 12, pp. 267-277, pi. 1. 1918. Reptiles of the Sulu Archipelago. Phil. Journ. Sci., vol. 13, pp. 233-267, pis. 1-3. 1922. The lizards of the Philippine Islands. Publ. 17, Phil. Bur. Sci. Manila, pp. 1-269, pis. 1-23. 1925. Additions to the herpetological fauna of the Philippines, IV. Phil. Journ. Sci., vol. 26, pp. 97-111. E V I O R A Miuseiim of Cooiparative Zoology Cambridge, Mass. June 6, 1956 Number 55 REDISCOVERY OF HYLA DOESALTS AND LECHBIODUS PAPUANUS IN NEW GUINEA By Arthitr IjOveridge Last year I had occasion to describe a new microhylid discov- ered by Mr. Kenneth R. Slater at Oniati, Papua. More recently we have received part of a collection made by Mr. Slater durino- a trip up the Aramia River. This river lies somewhat to the north of Katow (i.e. Binturi River, Papua), a locality from which Macleay (1877, Proc. Linn. Soc. New South Wales, 2. pp. 135-138) described a number of amphibians, among them Litoria dor ml u Macleay, now known as: Hyla dorsalis (Macleay) This hylid, so far as I know, has not been taken again during the period of almost eight}' years that has since elapsed. Among the Slater material is a 20 mm. gravid 9 Hyla (M.C.Z. 28389) that agrees so closely with Macleay 's brief description of the holotype of dorsalis (whose length is given as "9 lines," i.e. 19 mm., its hind limb as "16 lines," i.e. 33.8 mm. as against 35 mm. in ours) that I have no hesitation in assigning our frog to dorsalis, which must be one of the smallest members of the genus Hyla. Slater's specimen differs only in the absence of a light verte- bral line from snout to anus. On either eyelid are two tiny tubercles that, if constant in the species, may well have been overlooked by Macleay. NO. 55 Lechriodus papuanus (Roux) The same collection contained half a dozen frogs (M.C.Z. 28382-7) bearing- a distinct resemblance to Cornufer, bnt I am indebted to Dr. L. D. Brongersma of Leiden for pointing out that they have arciferal shoulder girdles and more or less dilated sacral diapophyses, so that he would refer them to Lechriodus papua7ius (Roux) which was described in 1927 from a single juvenile ? , 23 mm. in length, taken near Lake Sentani on the northern coast of Netherlands New Guinea. In the thirty years that have intervened since the holotype was taken by Dr. P. Wirz, no second specimen has been recorded so far as I know. No material was available to Parker when he monographed the Leptodactylidae of Australasia (1940, Novit. ZooL, 42, pp. 1-106) and he had but two examples of L. fletcheri (Boulenger) of which Roux made papuanus a subspecies. Parker (pp. 25, 27) treated papuanus as a full species but the dermal skinfolds by which he separates the two forms are so highly variable that it is questionable if they really do so. If Brongersma and I are correct in assigning these frogs to papuanus, the / — \ -shaped interorbital fold figured by Roux is either irregularly transverse, more often \__y -shaped ; the "curved, convergent dorso-lateral folds" are often rather ill-defined and may be represented by a few odd plicae extending no further than the scapula region. In only one frog do they approach the distinctness suggested by Roux's figure; they are quite indistinguishable from our fletcheri ? 9 from New South Wales in this respect. As I find little but color to separate the two, it would seem that Roux was justified in regarding papuanus as only a race of fletcheri. If I were called upon to distinguish the two forms I would do so as follows : Color above pale brown with traces of a blackish brown line from end of snout along the eanthus to the eye and from eye over tympanum to a point above elbow; lower lip dark brown; no interorbital skin fold (on either of our $ 5 from New South Wales). Range: New South Wales; Queensland; and (/id!e Parker) British New Guinea f. fletcheri Color above light gray, a white, finely black-edged, shieldlike patch on snout which may merge with two similar, but smaller, anteorbital patches ; from eye over tympanum to a 1956 HYLA DORSALIS AND LECTIRIODUS PAPUANUS 3 point above elbow a broad or narrow blackish brown line wliich may break up into a series of black specks; lower lip white; an indistinct interorbital skin fold (4 S $, 2 9 9 from Aramia Eiver). Range: British and Dutch New Guinea .... /. pap nanus Our knowledge of /. papuanus, hitherto derived from a single immature 9 , is naturally extended by the Aramia River series of adult $ S and 9 9 ; on the latter the following description is exclusively based. Description. Vomerine teeth in two confluent or semiconfluent series extending in an almost unbroken undulating line right across the roof of the mouth behind the level of, and beyond, the choanae. Head slightly broader than long, i.e. the breadth equals the distance from end of snout to angle of jaw; snout oval to obtusely rounded; nostril slightly nearer the front of the lip than it is to the orbit ; interorbital space narrower than an upper eyelid ; upper eyelid with several small, sometimes scarcely noticeable, tubercles ; a rather indistinct, transverse or v^_^'.shaped interorbital fold ; from the eye a supratympanic fold, which may be broken up, extends posteriorly to a point above the shoulder ; tympanum vertically oval, its horizontal diameter about two- thirds that of the eye ; from behind the eyelids a pair of more or less ) ( -shaped glandular skin folds ( frequently reduced to a pair of short lines) extend to, or beyond, the scapular region of the otherwise largely smooth dorsum. Fingers free, their tips somewhat swollen but not dilated, witliout horizontal grooves, first and second subequal in all six specimens ; subarticular tubercles large, metacarpus with a large supernumerary tubercle ; toes webbed, from first to fifth the following phalanges free (except for a lateral seam) 2, 2, 3, 4, 21/2, their tips slightly dilated and with or without a circum- marginal groove, third toe longer than the fifth ; subarticular tubercles prominent, a conspicuous inner metatarsal two-thirds the length of the first toe, no outer metatarsal tubercle ; tibio- tarsal articulation of an adpressed hindlimb extends almost to the nostril, or to the nostril, or end of snout. Skin of throat and bellv smooth. Color. Above, tip of snout with a conspicuous white, finely black-edged, shieldlike patch that may be disconnected with a similarly light patch in front of each eye, or fork posteriorly and merge with the anteorbital patches; from eye towards flank a more or less conspicuous black line which breaks up into black flecks on the flank immediately posterior to the forearm ; dorsum pale gray, with or without a few fine black lines; supra-anal area pale; circum-anal area black, becoming lighter on the" but- tocks toward the tibio-tarsal joint; fore- and hindlimb more or less faintly crossbarred. Below, white, uniform except that the chin and jaws are grayish in $ $ , while in both sexes the lower posterior aspect of forearms, anterior aspect of tibia, and soles of feet, are black. Size. Length from snout to anus of largest $ (M.C.Z. 28385), 49 mm., of ? (M.C.Z. 28382), 52 mm.; length of head from tip of snout to back of tympanum in 5 , 18 mm. ; in 9 , 20 mm. ; length of hind limb from anus to tip of longest toe in $ , 11 mm. ; in 9 , 86 mm. In size, therefore, papuanus differs little from fletcheri. Stomach contents. That of the S was found by my colleague Dr. W. L. Brown to consist of a ehrysomelid beetle, a termite's head, and remains of ants referable to the genus Odontomachus. E V I Musemiim of Compsirative Zoology Cambridge, Mass. June 15, 1956 Number 56 NOTES ON THE JAMAICAN AND CAYMAN ISLAND LIZARDS OF THE GENUS CELESTUS By Penny Norseen Cousens The collections of the Museum of Comparative Zoology afford a representative sample of the Celesti of Jamaica and the Cay- man Islands ^ and, additionally, include types of 5 of the 8 forms recognized in the last revisions of Chapman Grant (1940b and c). Renewed examination of this material has confirmed the distinct- ness of certain forms but has placed in question the status of certain others. It is hoped to return to the study of this genus in these and other islands in the future but as an aid to students of this difficult group it has been thought desirable to place on record at this time the new conclusions as to the status of the Jamaican and Cayman forms. The characters studied were : hal)itus ; color pattern ; dorsal, ventral and midbody scale counts, numbers of chin shields, and number of subdigital lamellae on the fourth digit ; general body measurements; presence or absence of a keel and number of striae on dorsal scales. Head scalation was checked but not recorded on at least one individual of each species. No internal characters were used. It was impossible to sex the specimens except in the case of females carrying young, but it is doubtful if there is any sexual dimorphism in the species studied. Only four species appear to be valid, as follows : 1 One species, the giant form Celestus occiduiis Shaw, is unrepresented. How- ever, omission of this species from the present discussion appears to be justified in as much as its right to consideration as a still surviving form is very improb- able. It has not been collected in over 100 years. Grant's suggestion that hewardii is a subspecies of this form cannot he seriously regarded. 2 BREVIOKA NO. 56 Key to the Jamaican and Cayman, forms of Celestus: 1. Adpressed limbs meet or fail to meet bj' less thau length of fore- arm. Dorsal color pattern consisting of broken bars or rings 2 Adpressed limbs fail to meet by greater than length of fore- arm. Dorsal color pattern not ringed 3 L'. 19 or fewer subdigital lamellae on fourth toe of hind foot. Throat mottled hewardii Gray 1845 Range : Central and western Jamaica and Portland Point. 21 or more subdigital lamellae on fourth toe of hind foot. Throat not mottled duquesneyi Grant 1940a Range: Portland Point only. 3. 89-116 dorsal scales, occipital to point above vent 4 125-139 doreal scales, occipital to point above vent harhouri Grant 1940li Range: known only from Mandeville in south central Jamaica and Fern Gulley in north central Jamaica. 4. Tail dorsally with the same coror and pattern as back cnisculus crusculus Garman 1187 Range : all of Jamaica, at least in the lowlands. Tail dorsally lighter than back and without pattern cnt^cnlus maculatus Garman 1888 Range: Cayman Brae and Little Cayman. DISCUSSION As the key indicates, and as Grant (1940b) has previously emphasized, the four species fall into two groups — one long- legged, the other short-legged. This may be a quite natural grouping, in these islands at least. In the long-legged group the two species, although similar in general body proportions, differ in several characters. The rela- tively widespread C. hewardii is a stockier, blunt-snouted lizard, its dorsum with dark brown broken bars on a light brown back- ground, its throat mottled, 12-19 lamellae on fourth toe, 4 pairs of chin-shields. C. duquesneyi, confined to Portland Ridge, is a more slender lizard with a pointed snout, its dorsal pattern similar to hewardii but lighter (specimens collected in same year), back- ground tan with light brown bars, the throat immaculate, 21-22 lamellae on fourth toe, 5-7 pairs of chin-shields. Unquestionably C. duquesneyi is a distinct species. M.C.Z. 45181 from Portland 1956 JAMAICAN AND CAYMAN CELESTUS 3 Ridge was misiclentified as hewardii but proved easily separable from a hewardii with the same locality data despite the loss in 45181 of the blue tail so much stressed by Grant as diagnostic for dnquesneyi. In the second, short-legged group the series of C. harbouri studied included the type and 13 specimens collected by Bar- l)our,' all from Mandeville, plus a single specimen from Fern (lulley. One character, dorsal scale count, readily separates C. harhouri from C. crusculiis? This difference is very marked and shows no overlap. Crusculus has the lowest dorsal scale count of any Jamaican Cclcstus (mean 101.52, range 89-116) ; harbouri has the highest (mean 132.27, range 127-139). At Mandeville, where the two species are sympatric, crusculus shows a very low mean of 97.64 (range 96-101) and also the strongest divergence of color pattern. The dorsal pattern of harbouri consists of a series of dark brown chevrons with the Vs pointing caudad. Crusculus shows a variety of patterns: plain tan dorsum with dark brown lateral bands extending the length of the body; a checkerboard dorsal pattern ; checkerboard blending into chev- rons; chevrons with lateral bands; chevrons only, this pattern almost indistinguishable from harbouri in several specimens. At Mandeville, the last type does not occur and the chevrons-with- lateral-stripes pattern occurs only once. These pattern types were established to aid in grouping specimens — gradations and modi- fications are common." r. crusculus is a variable species also in dorsal, ventral, and midbody scale counts. There is, however, no correlation between scale counts and pattern, and the two Jamaican subspecies pro- posed by Grant (1940a) on such supposed correlations are there- fore here regarded as invalid, C.c. cundalli with the type locality Mandeville being untenable on any ground. The Mandeville population cannot ])e distinguished from that at Kingston, the 1 These were labeled by him as impressiis Cope. The status of this name is very much iu doubt and its disposition will depend on a new evaluation of the types in the Philadelphia Academy. 2 Ventral scale counts are. as usual, fully parallel to dorsal scale counts and not really an independent variable. 3 Another color difference has been pointed out to me by Garth Underwood. The side of the neck in harhouri bears rnther larjre pearly spots as compared with smaller punctations in cruHmthm. 4 BREVIORA NO. 56 type locality of crusculus. C.c. molesworthi is more nearly justi- fiable. Specimens from the area ascribed to this latter sub- species, Boston Bay, Ecclesdown and Buff Bay in northeastern Jamaica, show exceptionally high dorsal counts. However a total of five localities shows a count higher than the average for the species as a whole,: Boston Bay (mean 109.4) and Ecclesdown (mean 106.5), on the fringe of the tropical rain forest on the eastern side of the island which may act as an isolate ; Balaclava (mean 110.5), about 15 miles northwest of Mandeville ; Bogwalk and Mona, close to Kingston. The coefficients of difference (CD.) were calculated for Ecclesdown-Boston Bay and Balaclava com- pared to both Mandeville and the species as a whole : Boston Bay-Ecclesdown Balaclava Mandeville CD. = 1.57 CD. = 2.32 entire species CD. = .676 CD. =: 1.01 As is to be expected, as compared to the entire species, the CD. is not significant. The Ecclesdown-Boston Bay and Balaclava populations are sufficiently different from the Mandeville popula- tion as determined by the CD. that in the opinion of some they might well be considered subspecies. Subspecific rank has not been accepted for these populations for two reasons : the samples are so small that there is no probability that the sample mean equals the population mean; furthermore, considered for the Jamaican populations as a whole the distribution of high or low dorsal counts seem to be random. The midbody scale counts are also higher in the northeast populations, but again samples are small, and high and low mean counts are merely the extremes of a cline in which there is overlap between even the southeast and northeast populations. Celestus ynaculaius of Cayman Brae and Little Cayman, re- tained by Chapman Grant as a full species, does not differ from C crusculus in scale counts or measurements but does differ in one detail of color — the coloring of the dorsal side of the tail. In C. maculatus the tail is light tan or ivory in alcohol and so without pattern ; in C. crusculus of Jamaica the dorsal color and pattern are continued on to the tail. In other respects the colora- tion of C. maculatus falls within the range of Jamaican crusculus, and its low scale counts resemble the Mandeville population of the latter. The Cayman population must be regarded as eon- 1956 JAMAICAN AND CAYMAN CELESTUS 5 specific with Jamaican populations but because of the tail color difference it may be recoo:nized as a distinct subspecies, C. crus- culus maculatus. SUMMARY 1. Four species of Celestus occur on Jamaica. 2. They may be arranged in two groups: the hewardii group (long-legged group) made up of C. hewardii and C. duquesneyi. and the crusculus group (short-legged group) comprising C. crusculus and C. harhouri. 3. C. hewardii, C. crusculus and C. harhouri are sympatric at Mandeville. 4. C. hewardii and C duquesneyi (sympatric at Portland Ridge) can be distinguished by habitus, coloration, lamellar connt and number of chin shields. 5. C. harhouri and C. crusculus can be distinguished by dorsal scale count. 6. C. crusculus is extremely variable and composed of two subspecies: C. crusculus crusculus occurring on Jamaica and C. crusculus maculatus occurring on Cayman Brae and Little Cay- man. There may be a third subspecies C. crusculus molesworthi, distinguishable by high dorsal scale counts and found in north- eastern Jamaica. Its validity is, however, not yet firmly demon- strated. 1 wish to express my thanks to Dr. Ernst Mayr and Dr. E. E. Williams for their generous and valuable advice, to Mr. Benja- min Slireve who made the scale counts and measurements on the specimen of Celestus harhouri from Fern Gulley, and to Mr. Garth Underwood for useful criticism of the manuscript. LITEEATURE CITED CopK, E. D. 18G8. An examination of the Eeptilia and Batrachia obtained by the Orton Expedition to Ecuador and the Upper Amazon with notes on other species. Proc. Acad. Nat. Sci. Phila. : 96-140. 6 BREVIORA NO. 56 Gabman, S. 1887. On West Indian Geckonidae and Anguidae. Bull. Essex Inst., 19: 17-24. 1888. Eeptiles and batrachians from the Caymans and from the Bahamas. Bull. Essex Inst., 20: 103-116. Grant, C. 1940a. Notes on the reptiles and amphibians of Jamaica, with diagnoses of new species and subspecies. Jamaica Today: 151-157. 1940b. The Reptiles in Lynn and Grant: The herpetology of Jamaica. Bull. Inst. Jamaica, Sci. Ser., 1: 1-65. 1940c. The herpetology of the Cayman Islands. Bull. Inst. Jamaica, Sci. Ser., 2: 1-65. Gray, J. E. 1845. Catalogue of the specimens of lizards m the collection of the British Museum. London, 289 pp. E V I MmseiLim of Comparative Zoology Cambridge, Mass. JuxVe 29, 1956 Number 57 THREE NEW SPECIES OF MIMETIDAE (ARANEAE) FROM PANAMA By Arthur M. Chickering Allnon College, Albion, Michigan In 1947 I published a brief paper attempting to bring onr knowledge of the Mimetidae of Panama up to date. At that time six species of Gelanor were recognized and two of these were regarded as new. Five species of Mimetus were also recognized at that time and two were considered new to science. As a result of two field trips into various parts of Panama since 1947 (1950 and 1954) a considerable number of specimens belonging to this family have been sorted from my collections. Three new species belonging to the two genera mentioned above have been selected from these later collections. These are described in this paper in accord with my usual procedure. I regret that I cannot yet be more certain regarding the correct association of males and females in several of the known species. The types are being deposited in the Museum of Comparative Zoology. It is again a pleasure to acknowledge my indebtedness to the following persons for their continued encouragement in the pursuit of my studies : Dr. A. S. Romer and Dr. P. J. Darlington, Jr., Director and Curator of Insects, respectively, in the Museum of Comparative Zoology at Harvard College. Without the priv- ileges which have been extended to me for many years in this institution the continued progress of my studies would have been much more difficult. NO. 57 Genus GelAXOE Thorell, 1870 Gelanor depressus sp. nov. (Fig. 1) Female holotype. Total length 4.16 mm. Carapace 2.145 mm. long; 1.69 mm. wide opposite second coxae where it is widest; narrowed to .91 mm. just behind PLE ; about 1.04 mm. tall opposite first coxae where it is tallest ; rises gradually from PE to highest point, then descends fairly steeply to posterior border ; with median thoracic pit about midway of the posterior declivity. Eyes. Eight in two rows as usual ; heterogeneous ; lateral eyes on a moderately prominent tubercle. AME separated from one another by slightly more than their diameter, from ALE by 9/7 of their diameter. PME separated from one another by about 7/10 of their diameter, from PLE by 2.6 times their diameter. Laterals contiguous to one another. Seen from above, anterior External Anatomy of Gelanor and Mimetus Fig. 1. Gelanor depressus sp. nov. ; epigynum, ventral view. Figs. 2-3. Mimetus saetosus sp. nov. ; epigynum, ventral and posterior vievirs, respectively. Fig. 6. Mimetus variegatus sp. nov. ; epigynum, ventral view. 1956 NEW SPECIES OF MIMETIDAE 3 row quite strongly recurved, posterior row gently procurved. Seen from in front, anterior row gently recurved, measured by centers. Central ocular quadrangle wider in front than long in ratio of 13 : 11, wider in front than behind in ratio of 3 : 2. Height of clypeus equal to 9/14 of the diameter of AME. Ratio of eyes AME : ALE : PME : PLE = 14 : 9 : 11 : 10 (long diameters used when differences exist). Chelicerae. Vertical, parallel, somewhat narrowed at their bases ; bases united for about one fifth of their length ; basal seg- ment .975 mm. long; with several long slender spines but none notably enlarged ; basal boss lacking. Fang groove with the usual row of six or seven stiff spines but with no visible teeth; retromargin with a single low tuberculate tooth near middle. Maxillae. Convergent ; moderately slender ; almost contiguoas distal to lip ; palp inserted into basal third. Lip. Clearly united to sternum with no sternal suture visible ; only slightly longer than wide near middle. Sternum. Scutiform in general; moderately convex; widest between first coxae but nearly as wide between second coxae; longer than wide in ratio of about 11 : 8 ; with numerous long slender bristles; posterior end extended between fourth coxae and terminating in a sharp point; with fourth coxae separated by a little less than the width of one of them; first and second coxae robust, third and fourth slender. Legs. 1243. Width of first patella at "knee" .352 mm., tibial index of first leg 7. Width of fourth patella at "knee" .22 mm., tibial index of fourth leg 9. (All measuieineuts iu millimeters) Fiuura Patellae Tibiae Metatarsi Tarsi Totals 1. 3.770 1.235 3.510 3.640 2.015 14.170 2_ 3.510 1.100 2.600 2.510 1.430 11.150 3. 1.950 .660 1.690 1.690 .750 6.740 4. 2.405 .700 1.700 1.755 .750 7.310 Spines. First leg: femur dorsal 0-1-1, prolateral and retro- lateral only one near distal end, ventral ; patella ; tibia only with definite spines on prolateral surface as follows : a series of 4 BREVIORA NO. 57 seven robust spines beginning near proximal end and extending to near distal end and with seven series of shorter spines follow- ing the robust spines and distributed as follows : 2,3,3,3,4,4,5 ; metatarsus with only definite prolateral spines as follows: a series of four robust spines beginning at proximal end and with fourth near middle and four series of shorter spines following the robust spines and distributed as follows : 5,7,8,19. Second leg : femur and patella as in first except femur dorsal 1-1-1 ; tibia only with definite spines as follows on prolateral surface : with five robust spines, the first weaker and near proximal end, the fifth about one fifth from distal end and with shorter spines following the robust spines arranged as follows : 0,1,4,4,5 ; meta- tarsus also only with definite spines as follows : on prolateral sur- face one moderately long and robust proximal spine and near this there is a very long and robust spine with another similar spine about one fourth of the length of the segment from the base ; immediately following the first long robust spine there is a series of six shorter curved spines, and following the second long spine there is a series of 18-20 shorter curved spines diminishing to mere bristles distally. Third leg : femur only dorsal 0-1-0 ; patella only one dorsal distal spine, long and slender ; tibia only dorsal distal 1 and prolateral 0-1-0; metatarsus only dorsal distal 1 (very slender) and prolateral 0-1-0-0-0. Fourth leg: femur 0; patella as in third; tibia only dorsal 0-1-0-1-0; metatarsus ap- parently only prolateral 0-1-0-0-0. Palp : femur apparently only with a pair of dorsal distal spines; patella with a single long slender dorsal distal spine ; tibia with a group of three or four distal spines, lacking ventrally ; tarsus dorsal 0-1-0-0-0, prolateral and retrolateral 0-1-1-0. Palpal claw with one short and one long tooth. Tarsal claws three with spurious claws as usual. The "drum" has been seen on the dorsal side of some tarsi about one fourth of the length of the segment from the distal end. Abdomen. Moderately high, well rounded laterally; slightly broader than long near middle. Anal tubercle and spinnerets as usual in the genus. Colulus well defined. Tracheal spiracle well defined and just anterior to base of colulus. Genital groove well defined just posterior to epigynum. Epigynum. Quite tubereulate; posterior surface with a pair of chitinized areas separated by a shallow slit; ventral surface 1956 ne\v species op mimetidae 5 with a pair of moderately deep depressions separated hy a septum (Fig. 1). Color in alcohol. Le<>s : third and fourth pairs nearly uni- colorous yellowish; first pair darker yellowish with femora red- dish dorsally and patellae also somewhat reddish especially at proximal ends and tibiae reddish at distal ends. Chelicerae a clear yellowish with fangs brownish. Lip and maxillae yellowish with reddish streaks. Carapace yellowish with reddish flecks just behind PE and a pair of irregular spots near apex and two pairs of irregular reddish spots further posterior. Sternum generally yellowish but with a small reddish dot opposite each second coxa and a large irregular spot covering most of the posterior half. Abdomen : anterior third of dorsum yellowish with a few bright red flecks at the base ; about one third from the base are two irregular bright red spots, each with two small white dots nearly enclosed by the red color; behind these spots there is a series of nine or ten very narrow red transverse lines ; the venter is yellowish lateral to the epigynum but reddish centrally behind this organ and with numerous oblique red ventro-lateral lines. Type locality. Female holotype is from Barro Colorado Island, C. Z., June, 1950. There are no paratypes. Genus Mtmetus Hentz, 1832 Mimetus saetosus sp. nov. (Figs. 2-3) Female holotype. Total length 8.71 mm. Carapace 1.95 mm. long; 1.43 mm. wide between third coxae where it is widest; narrowed to .845 mm. just behind PLE ; with no longitudinal thoracic groove but with a shallow depression about midway along the posterior declivity ; rises slightly from PLE to opposite interval betAveen first and second coxae and then after a gentle arch descends steeply to posterior border; with twelve spines in three rows behind PE, the middle row of only two and wdth each lateral row with five, the posterior one on each side being offset laterally. Eyes. Eight in two rows; heterogeneous; LE on prominent tubercles ; AME at base of a tubercle which extends considerably 6 BREVIORA NO. 57 dorsal from the eye level, with a spine at tip of each cone as in M. rusticus Chickering. AME separated from one another by two thirds of their diameter, from ALE by 1.3 times their diameter. PME separated from one another by slightly more than their radius, from PLE by nearly 2.2 times their diameter (long diameters used when eyes are not circular). Laterals separated only by a line. Chelicerae. Vertical, parallel, quite robust, quite convex in front ; basal segments united at bases for about one fourth their length which is .88 mm. ; without basal boss ; with numerous short slender bristles in front and several longer ones ; with one fairly robust spine in medial jiosition near distal end overlapping its mate on opposite side. Fang as usual in the genus. Retro- margin of fang groove with a single tuberculate tooth near medial end ; promargin with a row of ten robust spines. Maxillae. Slender, slightly convergent; palp inserted near base; serrula extends only along outer distal corner. Lip. Wider than long in ratio of about 5:4; reaches well be- yond middle of maxillae ; no definite suture observed. Sternum. Scutiform ; longer than wide in ratio of 13 : 16 ; widest between first coxae ; slightly convex ; with numerous long slender bristles; posterior end rounded, not extended between fourth coxae which are separated by about two thirds of their width. First coxae considerably swollen ; third coxae the smallest. Legs. 1243. Width of first patella at "knee" .286 mm., tibial index of first leg 8. Width of fourth patella at "knee" .198 mm., tibial index of fourth leg 9. (All measurements in millimeters) Femora Patellae Tibiae Metatarsi Tarsi Totals 1. 2.860 .780 2.730 2.275 1.250 9.895 2. 2.340 .650 1.950 1.820 1.105 7.865 3. 1.300 .455 1.040 .855 .840 4.490 4. 1.820 .528 1.690 1.235 .845 6.118 Spines. First leg : femur dorsal 0-1-1-1-1-1-0 with irregularities in placement, prolateral 0-0-1-1-1, retrolateral only one robust spine near distal end but with a somewhat oblique row of about 1956 NEW SPECIES OF MIMETIDAE 7 a dozen short stiff spinules in the proximal fifth, ventral a nearly median row of about fourteen long slender spines together with another row of similar but less regular spines in a prolateral position; patella only dorsal spines 1-1 together with several very slender spines irregularly placed; tibia dorsal 0-1-1-1-0, pro- lateral robust spines 0-1-1-1-1-1-1-1-0 together with short spines between the robust spines and distal to the last in numerical series as follows: 3-2-4-3-4-4-5, retrolateral robust spines 0-1-0 together with numerous long slender spines, ventral with only numerous long slender spines ; metatarsus dorsal only a series of long slender spinules, prolateral robust spines 1-1-1-1-1-1-0 with the expected short curled spines intervening between the robust spines and distal to the last of these in six series distributed as follows : 3-4-5-6-7-9, retrolateral with a row of 9 long slender spines, ventral with another row of long slender spinules, hardly more than bristles. Second leg : femur essentially as in first except foi- the presence of the proximal prolateral row of short stiff spinules presumably used in opposition to the retrolateral row of similar spinules on the first femur ; patella as in first ; tibia dorsal 0-1-1-1-0 plus several long slender spines somewhat irregularly placed, prolateral 0-1-1-1-1-1-0 plus smaller spines between robust spines and distal to the last of these but these are less numerous and less conspicuous than in first, retrolateral and ventral as in first ; metatarsus dorsal as in first, prolateral robust spines 1-1-1-1-0 with shorter spines terminally bent in four series dis- tributed as follows : 3, 4, 6, and 11. Third leg : femur essentially as in first ; patella as in first ; tibia dorsal 0-lp-O-O, prolateral 0-1-0-0, retrolateral 0-1-0, ventral 0-lp-O-lp-O ; metatarsus dorsal 0, prolateral 0-1-0-0, retrolateral 0-1-0, ventral 0-2 ( irregular )-0. Fourth leg : femur and patella essentially as in first ; tibia dorsal 0-lr-O-l-O, prolateral 0-1-0-0, retrolateral 0-1-0 together with sev- eral other slender spines irregularly placed, ventral 0-lp-O-l-O together with several long slender and weak spines irregularly placed ; metatarsus essentially as in third. Palpal claw pectinate with a single row of 5 or 6 slender teeth in proximal half. Ahdomen. 2.08 mm. long; 2.34 mm. wide between "shoulder humps" which are conspicuously developed. With numerous long slender stiff bristles, almost spines. Colulus well developed ; tracheal spiracle a short distance anterior to base of anterior 8 BREVIOBA NO. 57 spinnerets and with a fairly well developed chitinous lip. Spin- nerets as usual in the genus. Epigynum. With a moderately well ehitinized plate; moder- ately protuberant ; somewhat broader than long ; with a narrow notched lip along posterior border (Figs. 2-3). Color in alcohol. Legs and palps yellowish with many reddish brown spots and irregular rings. Chelicerae a rich reddish brown. Carapace in general yellowish with brownish markings; the largest of these begins at the clypeus and extends backward and narrows irregularly to a blunt point near the posterior border; three small irregular brownish spots occur on each lateral side. Sternum : with three pairs of small brownish spots and a single similar median spot at the posterior end. Abdomen : yellowish with many small lirownish spots. Type locality. Female holotype from Barro Colorado Island, C. Z., July 18, 1954. No paratypes. MiMETUS VARIEGATUS Sp. UOV. (Figs. 4-6) Male holotype. Total length 2.795 mm. Carapace 1.3 mm. long; 1.17 mm. wide opposite interval between second and third coxae where it is widest ; narrowed to .58 mm. wide just behind PLE ; with no longitudinal median thoracic groove but with a broad shallow depression about midway along the posterior declivity ; rises slightly from PE to opposite interval between second and third coxae and then descends steeply to posterior border; ap- parently with twelve slender spines in three rows on cephalic area posterior to PE. Eyes. Eight in two rows; heterogeneous; all except PME on prominent tubercles. AME separated from one another by nearly 1.5 times their diameter, from ALE by about one third as much. PME separated from one another by their radius, from PLE by their diameter. Laterals separated from one another by a line. Seen from above, posterior row slightly procurved ; seen from in front, anterior row slightly procurved. Central ocular quad- rangle wider in front than behind in ratio of about 4:3, slightly wider in front than long. Pvatio of eyes AME : ALE : PME : 1956 NEW SPECIES OP MIMETIDAE PLE = 5.5 : 4 : 4.5 : 4.5. Height of clypeus eqnal to about 14/11 of the diameter of AMB. External Anatomy of Mimetus Figs. 4-5. Mimetus variegatus sp. nov. ; tibia and tarsus of male palp; two views. C'helicerae. Vertical and parallel; rather long and slender; basal segment .65 mm. long; united at their bases for about one sixth of their length; without basal boss; with two slender spines on front surface and one more robust spine in medial position near distal end overlapping its mate on opposite side. Fang stout at base, strongly bent at middle, slender in distal half. Fang groove moderately distinct and without ordinary teeth but promargin has a row of 7 or 8 stout spinules (examination made on a paratype to avoid injury to the holotype). 10 BREVIORA NO. 57 Maxillae. Slender, somewhat convergent; reach considerably beyond lip ; with palp inserted into base ; with serrnla only at onter distal corner. Lip. About nine tenths as long as wide ; with a well developed sternal suture; reaches well beyond middle of maxillae. Sternum. Broad, scutif orm ; almost as wide as long ; strongly convex ; widest between second coxae but nearly as wide between first coxae ; with a moderate supply of long slender bristles ; posterior end rounded and not extended between fourth coxae which are separated by about 7/10 of the width of one of them ; first coxae most robust ; third and fourth nearly equal in size. Legs. 1243. Width of first patella at "knee" .198 mm., tibial index of first leg 5. Width of fourth patella at "knee" .132 mm., tibial index of fourth leg 7. (All measuieinents in millimeters) Femora Patellae Tibiae .Metatarsi Tarsi Totals 1. 2.990 .(5S2 3.055 3.510 1.365 11.602 2 2.405 .520 2.112 2.275 1.000 8.312 3. 1.625 .300 .975 1.034 .616 4.550 4. 1.820 .390 1.430 1.430 .616 5.686 Palp .682 .198 .264 .682 1.826 Spines. First leg : femur dorsal 1-1-1-1-1, prolateral 0-1-1-1-0, retrolateral a row of short, somewhat robust spines extending through about three fifths of the length of the segment beginning at proximal end, together with a single long spine near distal end, ventral with several stiff bristles but hardly spines; patella dorsal 1-1 ; tibia dorsal 0-1-1-1-0 plus several weak spines hardly more than bristles between the robust spines, prolateral 0-1-1-1-1- 1-1-1-0, retrolateral 0-0-1-0, ventral 0; metatarsus dorsal nine in the proximal four fifths of varied size and length, prolateral 1-1-1-1-1-0-0 with a few small intervening spines among the robust spines (in contrast to the expected condition), retrolateral with numerous weak spines but none robust of the usual sort, ventral 0. Second leg : femur dorsal as in first, prolateral devoid of long robust spines but with a row of very short stiff spinules extending through proximal third of segment and probably used in opposition to the similar spinules on the retrolateral surface 1956 NEW SPECIES OF MIMETIDAE 11 of first femur, retrolateral apparently with only one near distal end, ventral ; patella essentially as in first ; tibia dorsal 0-1-1-1-0 with several weak spinules and bristles irregularly distributed among the more robust spines, prolateral 0-1-1-1-1-1-0 with some irregularity together with spinules irregularly distributed among the more robust spines, retrolateral 0-0-1-0-1 (weak), ven- tral 0; metatarsus dorsal 0, prolateral 1-1-1-0-0 with short spinules between the robust spines and distal to the last, retro- lateral with a series of 10-12 spines of diminishing lengths toward the distal end w^here there are only bristles. Third leg: femur dorsal apparently 1-1-1-1-1 but with numerous spinules, else- where 0; patella only dorsal 0-1, tibia dorsal 0-1-0-0, prolateral the same, retrolateral 1 (weak) -1-0, ventral lp-lp-0; metatarsus dorsal 0-1-0-0, prolateral essentially the same, elsewhere 0. Fourth leg : femur and patella essentially as in third ; tibia dorsal 0-1-0-1-0, prolateral and retrolateral 0-1-0, ventral 0; metatarsus only dorsal 0-1-0-0 with prolateral essentially the same. A comparison of the spines, especially on legs one and two, shows that they differ from those regarded as typical of the genus especially because of the great reduction of the number, size, and regularity of occurrence of the small spines in between the larger ones of tibiae and metatarsi. Palp : femur with only one dorsal distal spine; patella with two long slender dorsal spines ; tibia with a row of five or six long slender spines or spinules curving along near distal border. Palp. Of moderate length; til)ia and patella short; tarsus moderately complicated ; with structural features best shown in Figures 4-5. Base of cymbium with a lol)ule which suggests that of M. verecundus Chickering but is, nevertheless, quite distinct from that species. Ahdomen. Of moderate height; about 9/11 as wide as long at level of "shoulder humps"; 1.43 mm. long; widest at level of "shoulder humps" w'hich occur slightly anterior to middle; with numerous long slender spinules on dorsal and dorsolateral sur- faces. Anal tubercle and six spinnerets as usual in the genus. The colulus appears to be reduced. Tracheal spiracle near base of spinnerets with a poorly developed chitinous lip. 12 BREVIORA NO. 57 Color in alcohol. Legs generally light yellowish with many grayish and brownish spots together with dark rings, more marked on first two pairs. Each chelicera has a small dark spot in front near base. Carapace : the eyes are ringed with reddish pigment ; there is a median dorsal grayish stripe extending from the PE backward and diminishing to a termination between an irregular V-shaped dark mark which extends forward from the depression at the center of the posterior declivity. The sternum is a clear yellowish. Abdomen : the dorsum from opposite the posterior border of the "shoulder humps'' to base is richly variegated with black, yellow, and red marks of a great variety of shapes ; behind the ' ' shoulder humps ' ' the dorsum is whitish with narrow reddish transverse procurved lines connected lat- erally with black elongated spots ; the venter is provided with several median dark spots and dark lateral bars. Female allotype. Total length 2.925 mm. Carapace 1.49 mm. long; 1.17 mm. wide opposite third coxae where it is widest; narrowed to .65 mm. wide just behind PLE. Otherwise essen- tially as in male. Eyes. AME separated from one another by 7/5 of their diame- ter, from ALE by about 2/5 of their diameter. PME separated from one another by about 5/9 of their diameter, from PLE by slightly more than their diameter. Laterals separated only by a line. Seen from above, posterior row slightly procurved; seen from in front, anterior row the same, measured by centers. Central ocular quadrangle wider in front than behind in ratio of about 4 : 3, wider in front than long in ratio of about 8:7. Ratio of eyes AME : ALE : PME : PLE = 10 : 8 : 8 : 8.5. "Height of clypeus somewhat greater than diameter of AME. Otherwise essentially as in male. Chelicerae. Basal segment .814 mm. long. The spinules along the promargin of the fang groove are more cons^Dicuous than in the male. Otherwise essentially as in that sex. Maxillae, Lip, and Sternum. Essentially as in male. Legs. 1243. Width of first patella at "knee" .231 mm., tibial index of first leg 6. Width of fourth patella at "knee" .154 mm., tibial index of fourth leg 8. 1956 NEW SPECIES OE MIMETIDAE 13 (All measurements in millimeters) Femora Patellae Tibiae Metatarsi Tarsi Totals 2.860 .725 2.925 2.730 1.300 10.540 2.405 .645 2.145 1.885 1.040 8.120 1.495 .390 1.105 1.040 .715 4.745 2.080 .455 1.495 1.300 .780 6.110 Spines. Small spiues between the robust spines on the first and second legs are mueh more conspicuous and more regular in occurrence than in the male. First leg : femur dorsal 1-1-0-0- 1-1, prolateral 0-1-1-1-0, retrolateral with a row of 16-18 short stiff stridulation spinules in proximal third and with one robust spine near distal end, ventral with a double row of long slender spinules; patella only dorsal l(weak)-l; tibia dorsal 0-1-1-1-0, prolateral with 0-1-1-1-1-1-1-1-0 robust spines and short curved .spines between the more robust spines usually four in each series increasing in length distally, retrolateral 0-0-1-0-1 (weak), ven- tral ; metatarsus dorsal a row of 12-15 long slender spinules, hardly spines, prolateral 1-1-1-1-1-0, retrolateral but with numerous bristles, ventral only numerous bristles. Second leg : femur dorsal 0-1-0-0-1-1, prolateral only with a row of short stiff spinules to oppose the corresponding similar spinules on first femur, ventral essentially as in first, retrolateral apparently only a single spine near distal end ; patella as in first ; tibia dorsal 0-1-1-1-1, prolateral 0-1-1-1-1-0 and with a series of shorter weaker curved spinules following each of the larger robust spines, retrolateral only 0-1-0, ventral ; metatarsus dorsal only a row of bristles or slender spinules, prolateral 1-1-1-0-0 with three series of shorter weaker curled spines following the robust spines as follows : 4, 5, and 12 or 13, retrolateral only a row of bristles or weak slender spines, ventral the same. Third leg : femur only dorsal 1-1-0-0 and numerous stiff" bristles; patella as in first ; tibia apparently only dorsal 0-1-0-0 ; metatarsus apparently only dor- sal and prolateral 0-1-0-0. Fourth leg: femur dorsal 1-1-1-0-0, elsewhere except for the frequent weak long slender spines, hardly more than bristles; patella essentially as in first; tibia dorsal 0-1-0-1-0, prolateral and retrolateral 0-1-0, ventral 0-lp- 0-lp-O; metatarsus only prolateral 0-1-0-0 and retrolateral 0-1- 0-0-0. 14 BREVIORA NO. 57 Abdomen. Except for sexual differences essentially as in male. Epigynum. Similar to that of 31. verecundus Chickering but with distinct differences; massively tubercular; with a small shallow depression at each anterior lateral corner (Fig. 6). Color in alcohol. Similar to that of male except that the variegated abdominal area extends over a large part of the dor- sum behind the "shoulder humps." Considerable variation in color has been noted among the paratypes. Type locality. Male holotype and female allotype from the Canal Zone Experiment Garden, August, 1954. Two male para- types from near Pedro Miguel, C. Z., August, 1954 and Canal Zone Experiment Garden, C. Z., August, 1954. Female paratypes from near Chiva, C. Z., July, 1954 ; Summit, C. Z., July, 1950 ; Arraijan, R. P., July, 1950. BIBLIOGEAPHY Chickering, Arthur M. 1947. The Mimetidae (Araneae) of Panama. Trans. Amer. Micros. Soc, 66: 221-248. Cambridge, O. P. and F. P. Cambridge 1889-1905. Arachnida-Araneida. Vols. III. In: Biologia Centrali-Ameri- cana. Dulau & Co., London. E V I O R A Miaseiiim of Cooiparative Zoology Cambridge, Mass. June 29, 1956 Number 58 SEQUENCE OF PASSERINE FAMILIES (AVES) By E. Mayr and J. C. Greenway, Jr. At the XI International Ornithological Congress at Basel, Dr. Stresemann proposed that a committee be appointed which would recommend to the editors of ornithological journals a standardized sequence of the families of Passerine birds. Dr. Stresemann pointed out that some editors follow Hartert, others Sharpe's Handlist, others the sequence of the AOU, and still others the recently proposed sequence of Mayr and Amadon (1951). If a reader wants to find a given family in an article, he may have to look either at the beginning or in the middle or at the encl depending on the particular sequence adopted by the i-espective editor. As a result of Dr. Stresemann 's suggestion a committee was appointed by the president of the Congress with the following members : Dr. G. C. A. Juuge, Chairman Eijksmuseum Nat. Hist., Leiden Prof. J. Berlioz Museum d'Hist. Nat., Paris IMof. Dr. G. Dementiev Zool. Mus. University, Moscow Prof. E. Mayr Mus. Comp. Zool., Cambridge, (Mass.) Mr. R. E. Moreau Edward Grey Institute, Oxford Dr. F. Salomonsen Zool. Museum University, Copenhagen Prof. Dr. E. Stresemann Zool. Museum University, Berlin 2 BREVIORA NO. 58 In the discussion, it was pointed out that more and more editors adopt the sequence of already published volumes of Peters' Checklist, and the editors of Peters' Checklist were re- quested to svibmit to the committee the sequence which they had been planning to use. Dr. Mayr expressed the views of the editors by saying that they would very much like to follow a standard- ized and universally adopted sequence, in view of the fact that there are no decisive arguments available in favor of any of the previously proposed sequences. He stated that the editors of Peters ' Checklist would be willing to adopt whatever sequence the committee would vote for. PvEPORT OF THE EDITOPvS OP PETEPtS' CHECKLIST In order to facilitate the work of the appointed committee, the editors of Peters' Checklist submitted to that body some tabulated historical material on sequences adopted by earlier authors. In the introduction to this report the following sub- ject matter was discussed : The task of the student of avian classification is particularly difficult because too little is as yet known about the paleontology of the songbirds to be of help in devising a sound classification. It is evident for this and more basic reasons that any linear order must be arbitrary. Three considerations usually guide those who attempt to find a suitable sequence : (A) To follow as closely as possible the traditional arrange- ments, except where subsequent work has sho^vn conclusively that a change is advisable ; (B) To place families near each other which are presumably closely related; (C) To place the more primitive families near the beginning and the more advanced families near the end. The following comments may be made on these three principles. (A) Traditional arrangements. In nearly all recently proposed sequences it is acknowledged that the larks and the swallows are aberrant in various ways. These two families are, therefore, usually disposed of at the beginning of the system. The remain- ing bulk of songbirds is usually grouped into three major assem- blages : (1) Old World Insect-eaters and relatives (Campephagi- dae, Pycnonotidae, Sylviidae. Timaliidae, Turdidae, Muscicapi- 1956 SEQUENCE OK PASSERINE FAMILIES (aVES) 3 dae, etc.), (2) New World Insect-eaters and finches, (3) crows, birds of paradise, and associated families. All the more peculiar and isolated families, as well as the Old World Nectar-eaters, are grouped rather irregularly within this broad framework. The sequence of the three major assemblages is by no means standardized, and to achieve a universal (or at least general) acceptance of one should be particularly the object of the com- mittee. The sequence 1, 2, 3 is that of Shufeldt (1!>04), Sharpens Hand- list, Hartert (in reverse here), the official Checklist of Australian Birds of 1926, Chasen's Handlist (1935), Mathews' Systema (1930), and more recently of the Checklist of Japanese Birds (1942), Berlioz (1950), Mayr and Amadon (1951) and Biswas (1952). The sequence 1, 3, 2 was suggested first by Wallace (1874), as far as we can discover. This arrangement was based on the reduction in the number of primaries. As early as 1885 Sharpe remarked that it was difficult to follow it in a linear sequence. However, it was followed in certain sections of the Catalogue of Birds in the British Museum, but in others the arrangements of Sundevall (1872) and Garrod (1876) were adopted. We cannot find that the exact order of the "Catalogue of Birds" has been followed by anyone. This same order (1, 3, 2) was adopted by Stejneger in 1885. Evans (1899) used this order in the Cam- bridge Natural History, Sclater used it in 1930, and Stresemann again in the Handbuch der Zoologie (1934). Finally, the sequence 3, 1, 2 was adopted by Wetmore and Miller (1926), and has been the order used in the American Ornithologists' Union Checklist (1931) and in works following the A.O.U. Checklist. It should be added ])arenthetically that neither Fuerbringer (1888) nor Gadow (1893, 1898) made any attempt to classify the Oscine Passeres into families or subfamilies. (B) Mutual relationship. To determine exact interrelation- ships of these families is in many cases very difficult, if not impossil)le. The system of songbirds abounds in artificial aggre- gates such as the "finches" or the "shrikes" of old, groups which may have little in common, except, for instance, the shape of the bill. In the improvement of the grouping considerable progress has been made in recent years such as the breaking up 4 BREVIORA NO. 58 of the artificial assemblage ''finches" into carduelines, buntings (Emberizidae) and cardinals, or the assembling of the vireos, wood warblers, tanagers, and related families into a single aggregate. There are still many conventional groupings such as Paridae - Sittidae - Certhiidae which are presumably quite arti- ficial. Yet, it would serve no useful purpose to break up these well-known sequences, until additional information sheds new light on the relationship of these families. (C) Phylogenetic sequence. A satisfactory classification into "primitive" or "specialized" families is virtually impossible for the exceedingly similar groups of songbirds. The convenience of placing the rather undifferentiated "Old World Insect- eaters," and particularly the very generalized Campephagidae and Pycnonotidae near the beginning, is now accepted by the great majority of classifiers. But whether to follow the sequence 1, 2, 3, or 1, 3, 2 — that is, whether to consider the finches or the crow and bird-of-paradise group relatively more specialized — depends entirely on the criteria used. If adaptation to a seed diet or loss of the tenth primary are considered characters of great importance, then the sequence 1, 3, 2 is certainly best. However, if the cerebralization of the Corvidae and the extraor- dinary courtship habits of the birds of paradise and bower birds are considered indications of evolutionary significance among songbirds, then the sequence 1, 2, 3 is to be preferred. The recent realization that the cone-billed birds ("finches") form an artificial, polyphyletic assemblage has weakened the sup- port for sequence 1, 3, 2. Not only should the true finches (in- cluding Carduelinae) be removed from the New World Ember- izidae-Cardinalinae assemblage (Tordoff 1954) but likewise the Estrildidae should be removed from the Ploceidae (Steiner 1955). Possibly the Estrildidae are related to the cardueline finches, both groups showing a curious mixture of primitive and advanced characters. The loss of the outermost primary which has happened in many families of birds independently and irreg- ularly is, like all loss-characters, a very unsafe taxonomic cri- terion. All this militates against the sequence 1, 3, 2. However, Aveighty arguments can also be advanced against the sequence 1, 2, 3. To consider cerebralization a most important character (Portmann) is somewhat anthropomorphic, as Wet- more (1951) has rightly observed. Furthermore, high cerebral- 1956 SEQUENCE OF PASSERINE FAMILIES (aVES) 5 ization has also oeeurred auiono^ Non-Passeres (Psittaci), and the relationship Corvidae-Paradisaeidae-Ptilonorhynchidae is by no means unequivocally established. In many purely morphological respects (bill, wing) the families of this group are undoubtedly more generalized than are some of the finches, and particularly the nectar-eating birds. There is, thus, no decisive criterion that would permit coming to an unequivocal conclusion. It should be added that it would be shortsighted to accept a sequence which would satisfy Euro- pean and American ornithologists only, but not those in the remainder of the world. The placement of the birds of paradise and bower birds may be immaterial to the student of European or of North American birds, but no world list of birds can be prepared which does not pay as much attention to the classifica- tion of these famili(»s as to that of the finches or thrushes. It appears to the editors of Peters' Checklist that the 1, 2, 3 sequence is the one most widely used in the ornithological litera- ture of the world. Since they know of no decisive reason for changing it, they have adopted a sequence which agrees roughly with the order 1, 2, ."]. ACTIONS OF THE COMMITTEE The editors of Peters' C-hecklist requested that the members of the Committee express their preference between the sequences ] , 2, 3 and 1, 3, 2 and 3, 1, 2, and they used this opportunity to ask for suggestions concerning the placement of individual families. The result was that the members of the Committee (Berlioz, Dementiev, Junge, Mayr, Moreau, Salomonsen, and Stresemann) voted unanimously in favor of the se(juence 1, 2, 3. They also submitted questions concerning the family rank of certain genera and groups of genera and suggested a number of modifications. Before these are discussed, two matters of policy, which will guide the editors of Peters' Checklist, may be stated. (A) There are a number of natural groups among the Oscines such as the thrush-flycatcher group, the "New World finches," and others delineated in Mayr and Amadou, 1951. In a linear listing unrelated groups of families must often of necessity be placed next to one another. This does not imply that such adjacent families are considered related, but is merely the un- 6 BREVIORA NO. '58 avoidable consequence of having to present a three-dimensional phylogenetic tree cut up into a one-dimensional linear sequence. As a result the Bombycillidae, for instance, are found next to the Cinclidae, the PachycephaJinae next to the Paridae, the Meliphagidae next to the Emberizinae, etc. Family group head- ings will be used in Peters' Checklist in order to minimize the effect of such misleading, but unavoidable groupings. (B) The editors consider it unwise in such an essentially uni- form group as the Oscines to recognize many families containing only a single species. They prefer to place somewhat isolated species in separate subfamilies, near the families with which the genera are traditionally associated. This they intend to do for such genera as Hyposifta, Dulus, and Pityriasis. There are many additional genera of songbirds which could be treated with equal justification as separate families, such as Lamprolia, Ticho- droma, Promerops, and others. Such a multiplication of families would not be constructive in the absence of sound anatomical support for the splitting. The following comments concern the placing of individual families and deal with questions raised by committee members. 1. Alaudidae. Should this family, currently listed near the beginning of the Oscines, be transferred nearer to the Emberizi- dae, a position which it held in some of the older classifications? Answer: This is not advisable. The Alaudidae are a very peculiar family. They differ from all other Acromyiodean Pas- seres by having not only the front but also the back of the tarsus scutellate and in having the pessulus rudimentary. This indi- cates that the larks may not be closely related to &jvy of the other families. Since they are not specialized to any great extent they are probably best placed near the head of the list. Two functional characters, the heavy bill in some of the seed- eating genera, and the reduction in the number of primaries, cannot be considered evidence for relationship to the finches. 2. GralUnidae and Artamidae. Should these families be trans- ferred from a place near the Laniidae to the neighborhood of the Callaeidae ? Answer: Yes. It is ad"^'isable for tAvo reasons to keep together all peculiar Australian families, the relationship of which is obscure. It is probable that these families had their origin in the Australian rocinn and that thev are distantly related. It is also 1956 SEQUENCE OF PASSERINE FAMILIES (aVEs) 7 desirable for purely practical reasons to have all purely Aus- tralian families near each other. In view of a superficial similar- ity, it was once believed that the Artamidae might be related to the Vangidae. However, there is no anatomical evidence avail- able in favor of such an association and much zoogeographical and biological evidence which contradicts it. 3. Bombycillidae. Should this family be transferred from a position near the Sturnidae to one nearer to the Laniidae ? Answer: Yes. The Bombycillidae are presumably more closely related to some of the families in group 1 than to those in group 3. 4. Ptilogonafidae and Dididae. Should not these two groups be retained as families? Answer: They are better placed as subfamilies for the reasons stated above under B. 5. Pityriasis. Does the inclusion of this genus in the Prionopi- dae reflect true relationship? Anstver: Perhaps not, but with the available evidence it is not possible at present to make a satisfactory decision on relation- ship. There is no support for the belief that it might be related either to the Starlings or to the Shrikes (in the restricted sense). Since it is not advisable for the reasons stated above under B to separate the genus in a monospecific family, it will be best to list it as a subfamily in the Prionopidae where the genus has been listed traditionalh^ 6. Esirildidae. Should they be combined with the Ploceidae or be treated as a separate family? Answer: Steiner (1955) has listed much evidence indicating that the Estrildidae deserve family ranking, indeed that they may not even be closely related to the Ploceidae. 7. Turdidae, Sylviidae, Muscicapidae, Fringillidae, Ploceidae, Emherizidae. Should not all these be retained as families? Answer: For the reasons stated by Hartert, the first three should be combined in a single family. Fringillidae and Plocei- dae should be retained as families while the Emberizinae should be retained as a subfamily of the New World finches. The oldest family group name proposed for the New World finches is ap- parently Emberizoidea (Suschkin 1925) . The name of the family then would be Emherizidae. 8 BREVIORA NO. 58 The final sequence of the families of Oscine Passeres which emerged from these discussions was submitted to the committee, which agreed with it, except that Prof. Berlioz stated that he still preferred a placement of the Alaudidae near the Emberizidae. The editors of Peters' Checklist still feel that such an arrange- ment has less to recommend it, for the anatomical reasons stated above. The sequence approved by the committee is as follows: Sequence of Osoine Families Alaudidae Hirundinidae Motacillidae Campephagidae Pycnonotidae Irenidae Laniidae Prionopidae Vangidae Borabycillidae Bombycillinae Ptilogonatinae Duliivae Cinclidae Troglodytidae Mimidae Prunellidae Muscicapidae Turdinae (incl. Zeledauia) Tiinaliinae (incl. Chamaea) Paradoxornithinae Polioptilinae (incl. Rliamphocaenus and Microbates) S.vlviinae (incl. Regnlus, Leptopoeeile, Lophobasileus) Malurinae Muscicapinae Monarchinae Paehycephalinae • Paridae Sittidae Sittinae Hyposittinae Neosittinae Certhiidae Dicaeidae 1956 SEQUENCE OF PASSERINE FAMILIES (AVES) 9 Neetariuiidae Zosteropidae Melipha^dae Emberizidae Eniberizinae C'ardinalinae = Kichniondeninac Tanagrinao = Thraupinac Tersininae Coerebinae Parulidae Drepaniidae Vireonidae (im-l. Viieolaiiiiis + Oyclarhis) Icteridae Fringillidae Fringillinae Carduelinae Estrildidae Ploceidae Sturnidae Oriolidae Dicruridae Callaeidae Grallinidae Artamidae Craeticidae Ptilinorliynchidao Paradisaeidae Corvidae LITERATURE CITED A.O.U. (Amkkicax Ornithologists' Union) 1931. Check-list of North American birds. 4th ed., 5L'(3 pp.. Lancaster, Pa. liERLIOZ, J. 1950. In Grasse (ed.), Traite de Zoologie, vol. lo, Oiseaux, 1164 pp., Paris (Systematics pp. 845-1055). I'.ISWAS, B. 1952. A check-list of genera of Indian birds. Records Indian Mus., 50, pt. 1, pp. 1 62 (in author's copy). Chasen, F. X. 1935. A handlist of Malaysian birds. Bull. Raffles Mus., Singapore, no. 11, 389 pp. 10 BREVIORA NO. 58 Evans, A. H. 1899. Birds. The Cambridge Natural History, 9, 635 pp., London. FUERBRINGER, M. 1888. Untersuchungen zur Morphologie und Systematik der Vogel. 2 vols., 1800 pp., 4°, Jena and Amsterdam. Gadow, H. 1893. In H. G. Bronn's Klassen und Ordnungen des Thierreichs etc. Vol. 6, 4 Abth., II, Systematiseher, 301 pp. (pp. 279, 301). 1898. A classification of Vertebrata, etc. 82 pp., London. Garbod, a. H. 1876-1877. On some anatomical peculiarities which bear upon the major divisions of the passerine birds. Part 1, Proc. Zool. Soc. London, 1876, pp. 506-519 ; Notes on the anatomy of passerine birds, pt. 2, 1877, pp. 447-453, pt. 3, t.c, pp. 523-526; pt. 4, 1878, p. 143. Hartert, E. 1903-1922. Die Vogel der palaarktischen Fauna. 3 vols., Berlin. Mathevs^s, E. M. 1927-1930. Systema avimn Australasianarum. 2 vols., London. Mayr, E. and D. Amadon 1951. A classification of recent birds. Amer. Mus. Novit., no. 1496, pp. 1-42. Ornithological, Society of Japan (A Special Committee ed. ) 1942. A handlist of the Japanese birds. 238 pp., Tokyo. Peters' Checklist (Peters, J. L.) 1931-1951. Check-list of birds of the world. 7 vols, (series not completed), Cambridge, Mass. PORTMANN, A. 1946-1947. Etudes sur la cerebralisation chez les oiseaux. Alauda, 14, pp. 1-20; 15, pp. 1-15 (pt. 2), i.e., pp. 161-171 (pt. 3). li.A.O.U. (Royal Australian Ornithologists' Union, Checklist Committee) 1926. Official checklist of the birds of Australia. 212 pp., Melbourne. SCLATER, W. L. 1930. Systema avium Aethiopicarum. Pt. 1, pp. 1-304 (1924) ; pt. 2, pp. 305-922 (1930), London. Sharpe, E. B. 1885. Catalogue of birds of the British Museum. 10, p. 1. 1956 SEQUENCE OF PASSERINE FAMILIES (AVES) 1.1 Sharpk, R. B. 1899-1909. A luuul-list of the genera niul species of birds. 5 vols., London. Shufeldt, R. W. 1904. An arrangement of families and the higher groups of birds. Amer. Nat., 38. pp. 833-857. Steiner, H. 1955. Das Brutverhalten der Prachtfiiiken, Spermestidae, als Au.s druck ihres selbstandigen Familicncharaeters. Act. XI Cong. Int. Orn. (Basel), pp. 350-3.55. 8TEJNEGER, L. H. 1885. Birds. In Standard Natural History (J. S. Kingsley ed.). Boston. Stresemann, E. 1934. Aves, Ilandlnieh der Zoologie. Kiikenthal and Krumbaeli ed., 7 (2), 898 pp., Berlin and Leipzig. SUNDEVALL, C. J. 1872-1873. Methodi naturalis avium disponendarum tentamen. 2 pts., 4°, Stockholm. SUSCHKIX, P. P. 1925. The Evening Grosbeak (Hesperiphona), the only American genus of a Palaearctic group. Auk, 42. pp. 256-261. TORDOPF, H. 1954. A systematic study of the avian family Fringillidae based on structure of the skull. Misc. Publ. Mus. Zool., Univ. Michigan, no. 81, 42 pp. Wallace, A. E. 1874. On the arrangement of the families constituting the order Passeres. Ibis, pp. 406-416. Wetmore, a. and W. W. Miller 1926. The revised classification for the fourth edition of the A.O.U. check list. Auk, 43. pp. 337-346. 1951. A revised classification for birds of the world. Smithsonian Misc. Coll., 117, no. 4, pp. 1-22. E Museiiim of Comparative Zoology {'amukidge, Mass. September 12, 1956 Number 5!) A NEW SUBGENUS OF VHAMAELIW FROM RHODESIA AND NEW RACE OF MABVYA FROM KENYA COLONY By Arthui^ Loveridge Fift}- years ago when Boiilenger (1906, Ann. Mag. Nat. Hist., (7), 18, p. 346, fig.) described RhampJioleon marshalli from Mashonaland, he remarked on the only occurrence of this genus south of tlie Zambezi as being "of very great interest." In as- signing marshalli to Rliampholeon, Boulenger was undoubtedly guided by his own key (1887, Cat. Lizards Brit. Mus., 3, p. 438) to the genera of chameleons. More recently Parker (1942, Bull. Mus. Comp. Zool., 91, p. 82) examined marshalli for certain osteological characters, in whicli i-espect he found it agreed with other continental species of Ehampholeon, a group I have since suggested should be regarded as only a subgenus of Brookcsia. Dr. V. FitzSimons (1943, The Lizards of South Africa, p. 172, pi. xxi, fig. 5) also treated marshalli as the only South Afri- can representative of Ehampholeon. He personally collected topotypes in Chirinda Forest, Selinda Mountain, as well as an extensive series from Vumba Mountain, also in Southern Rho- desia. Thirty-one of these specimens of marshalli are now in tlie Museum of Comparative Zoology and were examined by me when making a synopsis of the continental African Brookesia (sub- genus Ehampholeon). At that time (1951, Bull. Mus. Comp. Zoo]., 106, p. 182, footnote) I rejected marshalli as a Ehampho- leon, despite its bicuspid claws, referring it to Chamaeleo on account of its prehensile tail. Actually the bicuspid claws (present or absent in Eliam.pho- leon) constitute the only character in which marshalli dift'ers from the forms assigned to Chamaeleo ; thus it bividges the ga]) 2 BRHVIOKA NO. 59 between Chamaelco and the subgenus Bhamplwleon of Brook- esia.^ With a view to inviting attention to the intermediate status of this peculiar little chameleon, I suggest it should be made the type of a subgenus of Chamaeleo, viz. BiCUSPIS subgenus new Type. Rhampholeon marshalli Boulenger; known onl}- from Southern Rhodesia. Diagnosis. General appearance and soft nasal protuberance resembling that of Chamaelco rather than Rhampholeon. Scales on soles smooth ; claws bicuspid ; tail prehensile, half to two- thirds the length of head and body. The tail is included in the length of head and body from 1.6 to 2.1 times, Avith an average of 1.86 times for our entire series of 31 marshalli. The largest of the series, a topotypic gravid 9 (M.C.Z. 44445), measures 103 (68 + 35) mm. The only Brook esia (subgenus Rhampholeon) with a tail any- thing like as long proportionately as that of marshalli, is B. k. kersteni (Peters), the claws of whose forefeet have a secondary cusp — though a secondary cusp is lacking on the claws of the forefeet of its northern representative B. k. rohecchii (Bou- lenger). Mabuya bayonh keniensis subsp. nov. When the Museum of Comparative Zoology received a pair of typical Mahuya hayonii Bocage from Chitau, Bihe District, Angola, in 1936, it was immediately apparent that its East African representatives should be separated, if only on the basis of their strongly tricarinate dorsal scalation. The matter was deferred until such time as a revision of all African Mahuya could be undertaken, or a decision reached as to whether hayonii itself should be treated as a race of gravenhorstii Dumeril and Bibron of Madagascar. As time for any such thorough investiga- tion is lacking, I propose the name Mahuya hayonii keniensis subsp. nov. Holotype. Museum of Comparative Zoology No. 29662, an adult $ , from the northern Uaso (Guaso) Nyiro, Sotik, Kenya 1 As Angt'l (1942, Mem. Acad. Malgache, 36, pp. 154, 178) designated no type for his genus Erohiticauda, I suggest that Brookesia nams Boulenger be so rcgjinlfd. 1956 LIZARDS FROM KENYA AXO RHODESIA 8 Colony. Collected by the Smithsonian-African Expedition, 190!). Paratypes. A specimen from Mount Kenya (U.S.N.M. 40710), another from Wambiigu (U.S.N.M. 40781), a third from Lake Sergoit (Sirgoit: U.S.N.M. 42024), and three others (U.S.N.M. 40947-8 and M.C.Z. 29663) with the same data as the type. Unquestionably the juvenile from a salt marsh on the Loita Plains, recorded by Angel (1922) is referable to this Kenya race, as are also four listings of hayonii by me (1923; 1924; 1929; 1937). Less certain as to which race it should be assigned is Oscar Neumann 's specimen from ' ' Sero " ( ? Ssera, Lake Tanganyika) referred to Mahuia hayoni (sic) by Tornier, 1896 (Die Kreiehthiere Deutsch-Ost-Afrikas, Berlin, p. 42) ; this formed the basis of subsequent listings by Tornier (1897; 1900) and Nieden (1913), who added a series taken between Lake Victoria and Nguruman (i.e. Ngurumani). For the loan of the paratype material in the United States National Museum, as well as an Angolan b. hayonii (U.S.N.M. 26389) for comi)arative purposes. 1 am indebted to Dr. Doris M. Cochran. Diagnosis. Dorsals tricarinate, whereas in h. hayonii they are strongly quinquecarinate. Certain other characters may prove to have an average difference, but this cannot be demonstrated until more material of typical hayonii is available. Description,. Based on the holotype (variations of paratypes are placed in parentheses). Supranasals in contact behind the rostral ; centre of nostril in advance of, even though slightly, the vertical of the suture between rostral and first labial ; post- nasal in contact with first labial only (except on right side of U.S.N.M. 40710, and both sides of U.S.N.M. 40781, where it touches the second labial also) ; anterior loreal in contact with first and second labials (or occasionally second only) ; supra- oculars 3, the first and second being fused into a single shield (3-4 in paratypes); supraciliaries 4-5 (3-5), first largest; sub- ocular narrowed inferiorly, reaching the lip between the fifth and sixth upper labials (in all) ; lower eyelid with a transparent disk that is subequal to, or slightly larger than the ear-opening, which has 2 (2-3) more or less acuminate lobules projecting from its anterior border; frontonasal in contact with the first (as it is fused with the second) supraocular (though usually not in con- 4 HKKVIORA NO. 59 tact with the first) ; frontoparietal single, larger than the inter- parietal, which separates the parietals completely ; nuchals multicarinate. Midbody scale-rows 36 (34-36), dorsals strongly tricarinate; preanals not or but slightly enlarged; scales on soles not or but slightly pointed; subdigital lamellae smooth; toes of the ad- pressed hind limb meet the finger tips (or fail to meet, or extend to wrist) of the backward-pressed forelimb. Color. Above (pale or dark) olive brown; dorsum with several longitudinal series of black (and white) fiecks (or ocelli) ; a (n-eam (or white) dorsolateral line extends backwards from the supraocular region to some distance along the tail ; from the white labials a lateral line extends along the flanks on to base of tail. Below, white, uniform. Size. Total length of type o' (M.C.Z. 21)662), 148 (62 + 86) mm.; of paratype 9 (M.C.Z. 2!)663), 160 i- (82 + 78+) mm., tail-tip missing. E V I O R A Museianti of Comparative Zoology Cambridge, Mass. September 12, 195() Number 60 A NEW SPECIES OF AGRIOGNATIIA FROM JAMAICA, B. W. I. By Arthfr M. Chickerixg III connection with my study of tlic genns TetragnatJia Latreille, 1804 from Panama I have also had the opportunity to study several species believed to belong to the rather poorly- known genus Agriognatha 0. P. Cambridge, 1896. Apparently there is at least one species in Panama not included among those recognized by F. P. Cambridge in 1903 and not described by any later author. It is expected that a description of this new species will be published shortly. For some time prior to her death in 1958, Miss E. B. Bryant of the Museum of Comparative Zoology at Harvard College had l)een engaged in a study of a fairly representative 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, Curator of the Science Museum, Institute of Jamaica, Kingston, Jamaica. The collection is now in my possession and it is hoped that considerable time can be devoted to its study during the next few years. Among the new Tetragnathinae selected by Miss Bryant for description was the species here described in accord with my usual procedure. I have thought it appropriate to name the species in honor of Miss Bryant who for many years gave un- stintingly of her time and energy to the care of the large collec- tion of spiders in the Museum of Comparative Zoology at Har- vard College. So far as I know, this is the only species of Agriognatha reported from the island of Jamaica. ^ BREVIORA XO. 60 The liolotype and allotype Avill be deposited in the collection of the Museum of Comparative Zoology at Harvard Collepre. Genus AgeIOGNATHA 0. V. Cambridge, 1806 The genus Agriognailia 0. P. Cambridge, 1896 was based upon a male from Costa Rica. In 1897 Simon described a species from St. Vincent as Cyrtognatha serrata, apparently following Key- serling who established the genus Cyrtognatha in 1881. F. P. Cambridge had Simon's C. serrata for study and regarded it as congeneric with A. bella (0. P. Cambridge) from Costa Rica. He considered it probable that Cyrtognatha Keyserling was the same as Agriognatha 0. P. Cambridge. He also noted that the name Cyrtognatha was preoccupied and he placed all species regarded by the authors as belonging to either genus in the genus Agrio- gnatha although he had doubts regarding the proper position of A. lepida (0. P. Cambridge). Petrunkevitch (1911) placed all five species then known in the genus Cyrtognatha. Roewer (1942) has retained this practice. I have not seen Cyrtognatha glol)Osa Petrunkevitch, described from a female taken in the San Lorenzo River region of Panama but it seems unlikely that this species belongs in the genus Agriognatha as the latter is now understood. It now seems to the author of this paper that the two genera under consideration here are distinct but that all species recognized by Miss E. B. Bryant (1940, 1945) from Cuba and Hispaniola can safely be placed in the genus Agrio- gnatha. This also seems certain for the species from Jamaica described in this paper. Agriognatha beyantae sp. nov. (Figures 1-5) Male holotype. Total length 3.965 mm. (chelicerae not ex- tended anterior to head region). Carapace 1.745 mm. long, 1.30 mm. wide opposite second coxae where it is widest ; rather sharply narrowed opposite first coxae ; median thoracic pit rather shallow and broad (considerably different from that seen in specimens from Panama). Eyes. Eight in two rows as usual in the genus; LE extend slightly beyond border of head. Viewed from above, anterior 1956 AGRIOONATPIA FROM JAMAICA 3 row strongly i-pcurved, posterior row moderately recurved ; viewed from in front, anterior row moderately recurved, posterioi- row slightly procurved, both measured by centers ; central ocular quadrangle wider in front than behind in ratio of 9 : 8, as wide in front as long. Ratio of eyes AME : ALE : PME : PLE = 9 : 8.5 : 8.5 :7. AME separated from one another by 11/9 of their diameter, from ALE by 5/3 of their diameter. PME separated from one another by nearlj^ three fourths of their diameter, from PLE by about 7/4 of their diameter. Laterals separated from one another by a little less than the radius of PLE. Lateral eyes on a moderately pi'ominent tubercle. Height of clypeus equal to diameter of AME. Chelicerae. Very divergent, nearly horizontal; in i)lace of the basal boss there is a prominent tubercle from which a ridge extends nearly to the base of the fang ; basal segment of each chelicera 1. 235 mm. long and, therefore, nearly two thirds as long as carapace. Fang long, slender, somewhat sinuous, and with a prominent cusp on the inner surface just proximal to the middle. The fang groove has a robust non-bifurcate tooth near the base of the fang on the promargin together with three small teeth; the retromargin has four small teeth as shown in Figure ]. Maxillae. Essentially parallel but with distal third slightly curved outward and considerably widened; slightly more than twice as long as lip. Lip. Considerably wider at base than long; sternal suture gently procurved ; sternal tubercles prominent at ends of sternal suture. Sternum. Cordiform; moderately convex; slightly longer than wide between second coxae where it is widest ; moderately ex- tended between all coxae; terminates in a blunt point betw'een bases of fourth coxae which are separated by slightly more than one third of their width. Legs. 1243. \Yidth of first patella at "knee" .220 mm., tibial index of first leg 5. AYiclth of fourth patella at ''knee" .165 mm., tibial index of fourth leg 7. 4 BREVIORA NO. 60 Femora Patellae (All measu Tibiae lements in Metatarsi millimeters) Tarsi Totals 1. 3.770 .810 3.770 4.192 1.365 13.907 2. 2.990 .745 2.600 2.900 1.105 10.340 3. 1.560 .390 .910 1.170 .640 4.670 4. 2.210 .455 1.820 2.210 .780 7.475 Palp .726 .185 .198 .682 1.791 Many true spines as well as hairs on all legs. It is considered unnecessary to record all spines, hence, only those which seem to be more or less distinctive are emphasized here. First leg: : there is a row of four short but fairly robust ventral spines near the base of the femur together with a row of seven very short but fairly robust retrolateral spines extending through the prox- imal two thirds of the segment and also a long robust retrolateral spine near the distal end of this segment ; the patella has the characteristic retrolateral chitinous ridge ; the metatarsus has a row of about 25 short ventral robust spines extending nearly throughout the length of the segment. The second leg is essen- tially like the first in respect to the characteristics noted with the specialized robust retrolateral spine near the distal end of the femur arising from a raised base. The expected prolateral trichobothria on both third and fourth femora are present but are much less conspicuous than in the Central American species thus far observed. Palp. Complicated; somewhat difficult to describe because of numerous associated apophyses. The paracymbium appears to have a somewhat characteristic terminal enlargement. Near the tip of the cymbium there is a small pit, regarded as a type of sense organ. The same type of presumed sense organ has been noted on related genera and species. See Figures 2 and 3 for details of palpal structure. Abdomen. Unnotched at anterior end which is bluntly rounded; 2.34 mm. long; 1.06 mm. wide near middle; with spir- acle near base of spinnerets ; with a well defined colulus, oval in outline; with genital fold (Fig. 5) only slightly posterior to openings of book lungs. Color in alcohol. Chelicerae brown ; legs yellowish broAvn with many irregular darker areas especially on dorsal surfaces. Carapace : with a narrow lirown marginal stripe from opposite 1956 AGRIOGNATHA FROM JAMAICA 5 PE to opposite third coxae where it becomes a wide stripe which continues to posterior border; also with an irregular broad median dark stripe from PME to median thoracic pit. Abdomen : a broad serrated median dark gray stripe extends from base to spinnerets ; on each side of this there is an irregular light stripe with scattered silvery spangles; the lateral sides are dark gray with oblique extensions dorsally; the venter is yellowish with a row of small silvery spangles on each side. Female allotype. Total length 4.94 mm. ; including the cheli- cerae 5.46 mm. Carapace 1.95 mm. long, 1.43 mm. wide opposite interval between second and third coxae where it is widest ; other features essentially as in male. Eyes. Central ocular quadrangle about as wide in front as long. Ratio of eyes AME : ALE : PME : PLE = 9 : 10 : 9 : 8. AME separated from one another by 10/9 of their diameter, from ALE by slightly less than twice their diameter. PME separated from one another by 10/9 of their diameter, from PLE by twice their diameter. Laterals separated from one another by about 14 of the diameter of ALE. Height of clypeus equal to slightly more than % of the diameter of AME. Other features essentially as in male. Chelicerae. Robust ; only slightly divergent and porrect ; basal segment 1.04 mm. long; fang regularly curved and without special features; promargin of fang groove well marked, with three teeth; retromargin of fang groove with four teeth (Fig. 4) ; some variation in placement and relative sizes of the teeth has been noted and must be expected among paratypes. Maxillae, Lip, and ^iernvm. Essentially as recorded for the male. Legs. 1243. Width of first patella at "knee" .264 mm., tibial index of first leg 7. Width of fourth patella at "knee" .198 mm., tibial index of fourth leg 10. Femora Patellae Tibiae Metatarsi Tarsi Totals (All measurements in millimeters) 3.380 .845 3.140 3.412 1.365 12.142 2.600 .780 2.190 2.632 1.040 9.242 1.495 .435 .845 .975 .715 4.465 2.210 .487 1.495 1.820 .780 6.792 H BREVIORA NO. 60 All legs with stout spines and hairs but the spines seem not to be specialized as in the male. The triehobothria on the third and fourth femora essentially as recorded for the male. Abdomen. Very gibbous dorsally just behind the middle but this swelling seems to be completely lacking in some female para- types; 3.575 mm. long; 2.080 mm. wide opposite the gibbosity where it is widest ; genital groove as shown in Figure 5. Color in alcohol. Chelicerae much lighter than in male with mottled gray markings in front in basal half. Otherwise essen- tially as in male. Type locality. Male holotype and female allotype from Hard- war Gap, Jamaica, B. W. 1., June 27, 1954. Several paratypes of both sexes from the following localities: Blue Mountains, August, 1934 (Darlington) ; St. Andrews, Clydesdale, July, 1950 (Bengry) ; Hardwar Gap, June, 1954'; Hanover, Askenish, trail to Dolphin's ITejid. .luue 24, 1954. BIBLIOGRAPHY Bryant, Elizabeth B. 1940. Cuban spiders in Ihc Miiscuiii of ( 'oiuimiative Z()olo<;y. Bull. Mus. Comp. Zool., vol. 8(), pp. 2-19-531!, pis. 1-22. 1945. The Argiopidae of Hispaiiiola. Bull. Mus. Ooni]). Zool., vol. 9."), no. 4, pp. 359-418, pis. 1-4. Cambridge, O. P. and F. P. Cambridge 1889- Arachnida-Aiaueida. Vols. I and II in: Biologia Centrali- 1905. Americana. Dulau & Co., Loudon. Keyserling, Gr.\f Eugex vox 1881. Neue Spinneu aus .Xmerika. \'eiliaudl. zool. Lot. (ies. Wit^n, vol. 31, p. 270. Petrunkevitch, Alexander 1911. A synonjTiiic index-catalogue of spiders of North, Central, and South America, etc. Bull, .\iner. I\[us. Xat. Tlist., New York, vol. 29, pp. 1-809. 1925. Arachnida from Panama. Conn. Acad. Aits and Sciences, vol. 27, pp. 51-248, 157 figs. iSiuce the completion of this paper two females and one adult male believed to belong to this species have been collected at Hardwar Gap, Jamaica, in July. 1955, by Dr. Allan F. Archer who has kimlly lonned these specimens for examina- tion. 1956 AGRIOGNATHA FROM JAxMAICA IvOEWEli, C. F. 1942. Katalos tier Araneae. Vol. 1. Bremen. Simon, E. 1897. On the spiders of the Iskmd of St. Vincent. Part III. Proc. Zool. Soc. London, pp. 8G0-890. A External Anatomy of Agriognatim bryantae sp. nov. Fig. 1. Male chelieerae, from in front. Figs. 2-3. Tarsus of male palp; lateral and ventral views. Fig. 4. Female cheliceral teeth. Fig. 5. Genital furrow of the female. E V I O R A Mnasemiim of Comparative Zoology Cambridge, Mass. September 14, 1956 Number 61 OX KEGENERATKJN BY EARTHWORMS OF A SPECIES OF THE LUMBRICID GENUS DENDEOBAENA EISEN 1874. By G. E. Gates Regeneration, either cephalic or caudal, after natural or ex- perimental amputation, for any species of Dendrohaena has not been recorded hitherto, though several of the head regenerates attributed by Morgan to Eisenia foctida (cf. Gates, 1953, 1954) probably were D. octaedra. Some data as to regeneration in another species of the genus can now be presented. The worms, presumably all of an athecal morph frequently called Bimastos tenuis in the past, were clitellate and were secured early in the summer in Michigan. Amputation was without anesthesia. Amputees were kept in damp leaves at 22° C for the periods indicated in the tables. The author's thanks are extended to Prof. Murchie for making this material available for study. Anterior regeneration in D. ruhida (Savigny, 1826) All regenerates (cf. Table I) are hypomeric. Metamerism in most regenerates deviates more or less from normal. Regenera- tion of a head clearly is possible at all levels back to and including 13/14. The regenerate at 17/18 lacks normal anal or buccal sculp- turing distally and, since nephropores are unrecognizable, pro- vides no external indications as to whether it is of cephalic or caudal organization. Abnormal metamerism and hypomery at anterior levels, in certain other earthworm species studied by the author, resulted from unfavorable conditions either in the external or internal environments. Accordingly, metamerically normal regenerates at 2 BREVIORA NO. 61 all levels to 13/14 and equimery at least to 5/6 can be expected in D. ruhida when conditions are more favorable. The level at which the indeterminate monstrosity was devel- oped is close to, if not actually in, a region where regenerative capacity in E. foetida (cf. Gates, 1949-1950) is bipotential. Monstrosities such as the one under consideration often are pro- duced in morphogenetie regions of dual capacity. However, the data now available as to the results of posterior amputation in D. ruhida scarcely warrant anticipation of heteromorphic tail regeneration in that species. Presumably then, a regenerate de- veloping at 17/18 in optimal conditions will be normally cephalic. Especially noteworthy, even in these few instances, is absence of any indication of decline in number of segments regenerated as level of amputation moves posteriorly (cf. Gates, 1949, p. 137). Table I Head regeneration in Dendrohaena ruhida •-> 11 o 2 ■sSs ! 2 » 1 a; 1 2/3 1 2 19 o 3/4 1 3 19 Metamerism still indistinct in regen- erated left halves of iv-vi. 3 5/6 3 4 32 Eegenerate metamerism nearly nor- mal. 4 5/6 4 4 32 Regenerate metamerism nearly nor- mal but 6/7 abnormal in regen- erated right halves of vi-vii. 5 6/7 3 4 32 Metamerism nearly nonnal. 6 6/7 4 -3 19 Metamerism probably was develop- ing abnormally. ' 7/8 - 4 37 Metamerism not quite normal. Ee- generated ventral half of viii ab- normal. 8 7/8 3 3 19 1956 REGENERATION IN DENDROBAENA n — 1- G fi £.2 1° a 1 Davs of regener tion 9 7/8 3 + 3 19 ] 10 8/9 4 3 19 1 11 8/9 4 4 32 1 12 8/9 4 4 32 !; 13 8/9 4 -3 19 I 14 11/12 5? . 4 32 I 15 13/14 •')? -r, 32 I Proximal portion of regenerate too short for a normal segment. Metamerism apparently developing abnormally in regenerated riglit halves of ix-x. Regenerate segments ii-iv not quite nomial. Segment iv too large. Excised dor- sal half of ix had been regenerated. Metamerism apparently normal. Regenerate metamerism abnormal. Metamerism in regenerated left halves of xiv-xvi probably develop- ing abnormally. Ifi 13/14 o-O ' -3 3(1 Metamerism rather indistinct in re- generated portions of xiv-xv, ven- tral half of xiv and a small portion of XV. 17 17/18 4-0 i -3 32 Regenerate terminating distally in rounded knob like a rudimentary prostomium but towards the ven- tral rather than the dorsal side. Except as indicated otherwise above, excisions appear to have been transverse and along intersegmental furrows. Stages. 1. Prostomium aJid buccal invagination as yet unrecognizable. 2. Prostomium and mouth developed. 3. Intersegmental furrows demarcate regenerate into segments. -3. Furrows still indistinct. 4. Pigment present but obviously different from that of substrate. 5. Setae and nephropores recognizable. (i. Pigment now like that of substrate, external stigmata of regen- eration, except for typical metamcric anonuilies, uni-ecogniz- able. NO. 61 Posterior regeneration in D. rubida All anus of more or less normal appearance had been acquired, at end of a July-August period of nineteen days, by the six worms from which a posterior portion of the intestinal region (cf. Table II) had been excised. Healing probably had been enteroparietal. The new anal region had not yet been delimited from the last substrate segment, by development of an interseg- mental furrow, in Nos. 4-6 where little or no indication of re- organization is externally recognizable. A very small anal re- gion, in No. 2, is demarcated by an intersegmental furrow and again there are no other external indications of reorganization. A terminal portion of No. 1, presumablj^ comprising only the last segment, at time of preservation was being reorganized. Completion of the process apparently under way probabh^ would have resulted in development anteriorly of eight new setal fol- licles with setae and appearance of an intersegmental furrow demarcating a terminal anal portion from a metamere with the usual stigmata of regeneration. Reorganization, instead of re- generation, would then have been recognizable externally only if the original nephropores had been retained in the smaller daughter segment or if some of the pigment had escaped lysis. Reorganization, accordingly, had been most drastic and had taken place most rapidly at the anteriormost level of amputation. Evidence as to tail regeneration has been sought in several hundred individuals of D. rubida that have been available from various states and from other countries. Many of these worms certainly are posterior amputees and some of the others, obvi- ously brevicaudate, very probably are. No tail regenerates were found. Externally recognizable indications of reorganization were noted but rarely, and in each case there could have been produced, in addition to the anal region, only one or two new segments. Homomorphic tail regeneration, from a growth zone of rapid segment production (Gates, 1948) obviously does not, usually, follow posterior amputation in this species. 1956 REGENERATION IN DKNDROBAENA Table IT Kesults of posterior amputation in D. ruhida m2 1 42/43 .' Teiniiiial sul)strate segment has lost nuuh pigment, setae and nephropores. 2 54/55 ? An intersegmental furrow delimits small white an;il region from last substrate segment. 3 65/66 12 Same as in No. 2. 4 80/81 6 Anal region not delimited from last siilv stiate segment which still has setae and nephropores. 5 85/86 7 Anal region not delLmited from last sub- strate segment where some of the setjie still are present. 6 87/88 10 Anal region not delimited. Time allowed for regeneration, 19 days. Uiseussioii llomomorphic tail regeneration does not necessarily follow immediately after amputation and may not begin until nine months later. The process gets under waj' in some earthworm species, regardless of time of amputation, only when the "inter- nal environment ' ' permits. Differences in that environment may well be responsible for discordant results obtained by different investigators who have used the same species. As the importance of the unknown factors of the internal environment of earth- worms seems not to have been appreciated, the following in- stance is worthy of record. Individuals of Eisenia foctida, in the author's earlier studies (Gates, 1049-1950). after r(Miioval of the posterioi- portion at 6 BREVIORA NO. 61 levels behind 40/41, almost always regenerated promptly and in all seasons of the year. Several dozen specimens of that species, apparently in good condition, from three localities, were deprived of their tails in the region between 40/41 and 50/51 on three occasions in the last five years, to provide material for a school demonstration. All of the worms survived the operation. Not one showed any indication of formation of new segments during a period of several weeks though the external environment was, so far as could be determined, the same as before. Summary Cephalic regeneration in an anterior direction can be expected, in B. ruhida in optimal conditions, at all levels back to 17/18, with equimery back to 5/6. Caudal regeneration in a posterior direction, from a growth zone of rapid segment formation, can- not be expected ordinarily if at all. Instead, the terminal sub- strate segment may be reorganized into an anal region and one or two metameres with some of the stigmata of regeneration, the reorganization possibly being more drastic and more rapid when at more anterior levels. EEFERENOES Gates, G. E. 1948. On segment formation in normal and regenerative growth of earthworms. Growth, 12: 166-180. 1949-1950. Regeneration in an eartliworm, Eisenia fuetida (Savigny) 1826. I-III. Biol. Bull., 96:129-139; 98:36-45; 99:425-438. 1953. On regenerative capacity of earthworms of the family Lumbii- cidae. Am. Midland Nat., 50: 414-419. 1954. Anterior regeneration in a sexthecal species of lunibricid earth- worm. Breviora, 27:1-5. E V I O R A Muiseitami of Comparative Zoology Cambridge, Mass. October 3, 19o() Number 62 A THIRD LEAF-N08ED 8PECIES OF THE LIZARD GENUS AXOLIS PROM SOUTH AMERICA By Jameh a. Peters J>i(i\vii T'liivcisity, Providence, Bliode IsL-iiul and (tustavo Orces-V. Escnehi J'olyteeiiica Xacioiinl, Qiiitd, Kcunaor It gives us considerable pleasure to be able to announce the discovery of another specimen of the leaf-nosed anoles. It is the first time that leaf-noses have been known to occur outside of the Amazon Basin. The locality from whence it came is on the Pacific side of the Andes in the hot and humid "Choco" portion of northwestern Ecuador. While the new individual is obviously closely related to the holotype of the species described by Myers and Carvalho (1945), the differences are sufficiently great that we feel justified in calling- it a new species. Since the original specimen was collected, the Escuela Poly- tecnica has sent several other collectors into the area and the .senior author made several trips to the type locality during the summer of 1954 when he obtained comparative material of other species of the genus Anolis.^ but no one has yet been able to obtain additional representatives. Nothing is known as yet as to the habits of these unusual animals or the use to which the rostral appendage is put. l'"i<'lil wurk in Kcuailor d\ii-iiis' l!l-">4 liy tlic senior author was suppo rant from the PtMirose Fund of the Anieriean Pliilosoidiical Society. 2 HREVIORA NO. 62 Akolis proboscis, new species Type. Museum of Comparative Zoology Xo. 54800, a mature male from the neighborhood of Cunuco, a small town at 1200 meters elevation, five kilometers northwest of Mindo, on the south bank of the Rio Mindo, a northern tributary of the upper Rio Blanco, in Piohincha Province, Ecuador, collected by An- tonio Proano, during April 27-29, 1953. Diagnosis. This new species is distinguished from all other species but one in the genus Anolis by the presence of an elon- gated, flesh}" appendage, which is about as long as the head, on the snout. The only other species with a similar appendage is A7iolis pliyllorhinus Myers and Carvalho, from which it differs in having a serrated edge on the rostral appendage, 10-11 upper labials to a point below the center of the eye, 9-10 lower labials, and a median dorsal row of scales which is produced into a serrated dorsal crest, as wtII as other characters mentioned below. Description. Snout with an elongate, fleshy appendage pro- jecting anteriorly, with a distinct downward curve ; length of appendage approximately equal to length of head, measured from tip of rostral to ear opening. Appendage arises approxi- mately half way between eye and nostril, on midline ; strongly serrate along upper margin to level of rostral, serration less marked but still serrate to tip, both above and below; nine scales wide at level of rostral, tapering to three at tip ; rather fleshy and rounded at snout, tapering to thin and vertically com- pressed at tip. Scales on dorsum of head smooth, pavimentose, irregular. No distinct depression in frontal region, frontal ridges almost obsolete, faintly indicated from middle of supraorbital semicircles to frontal area, where they end at level of anterior margin of eye. Canthus rostralis not strongly angulate but rather rounded, with 5-6 canthal scales from nostril to eye ; area between canthi occupied by irregular, smooth scales, 8-9 in a straight line between canthi at level of origin of nasal appendage and approximately the same number between the posterior ends of the canthi, due to gradual increase in size of scales in frontal area. Scales of supraorbital semicircles large, somewhat scari- fied, slightly angulate; not in contact on midline, separated by two rows of smaller, pavimentose scales. Supraorbital scales in contact with semicircles, no granular scales separating them; 1956 LEAF-NOSED ANGLE FROM SOUTH AMERICA 3 three or four scales in center of sni)raorbital area are markedly larger than others, but no rows or series are formed. Inter- parietal large, considerably larger than ear opening; separated from supraorbital semicircles by two rows of scales. Rostral completely horizontal below appendage, extending well beyond lip line. Nostril sunk in a single scale. Two scales between canthus rostralis and upper labials anteriorly, four or five posteriorly, no distinguishal)le horizontal rows of loreal scales. Last scale in canthus rostralis followed by a single super- ciliary, wliich is in turn followed by the granular area of eyelid. Granules of eyelid continuous with smaller scales of supraocular region; these two areas are not separated by a series of super- eiliaries. Single row of three or four enlarged subocular scales from loreals to granular postocular region and between the granules of eye and the upper labials ; terminal scale of this row enters the labial row on both sides, and on one side extends to the lip line. Temporal region with man}- small, pavimentose scales, not arranged in rows or series. Ear opening quite small. Eleven upper labials on left side, ten on right to below center of eye, one additional labial behind eleventh on left, one right sul)- ocular enters labial row, followed by granular scales. First labial on both sides wedged between rostral and second labial, lowei- edge only partially on lip line. Mental almost completely divided on midline, suture incom- plete anteriorly only. Nine lower labials on left, ten on right. Seven sul)labials on each side, anteriormost in contact with lower lal)ials, })osterior four or five separated from labials by smaller scales; sublabials reduced in size posteriorly, gradually merge with other chin scales, which are small, numerous, and pavi- mentose or occasionally granular. Body very slightly compressed, vertebral angle comparatively sharp. Dorsal crest of raised denticulations runs down spine from head onto tail, Avith 85 spines from back of head to level of posterior insertion of hind limbs. In several instances adjacent denticulations are fused, with sutures still visible. Scales on dorsum of body pavimentose, quite small, smooth, no enlarged paravertebral rows. Lateral scales approximately equal in size and appearance to dorsals. Granular scales, i.e., scales with raised centers rather than a fiat surface, are present on neck, shoulders. 4 BREVIORA NO. 62 and above hind legs. Ventral scales considerably larger than dorsals and laterals, smooth, for most part in imbricate rows. Scales on throat quite granular except on moderately developed gnlar fold, where the scales are larger and quite imbricate. Im- bricate scales of belly pass into pavimentose or even granular scales at level of pelvis, and remain so to anus. No enlarged preanal scales ; no femoral or preanal pores. Scales on all parts of limbs and feet subequal, pavimentose or slightly granular. Scales on dorsum of digits expanded, nmch wider than long, extend across dorsal surface of digit ; smaller scales on lateral edges of digits, which are dorso-laterally flat- tened ; expanded scales on ventrum of digits pavimentose at base of toe, lamellar on proximal part of toe. Lamellar counts are shown in Table 1. Tail very strongly compressed, vertebral angle sharp, with a prominent dorsal crest. Scales at base of tail smooth, pavi- mentose, and irregularly arranged ; gradually become larger, squarish, and arranged into vertical series, but still pavimentose at level of posterior end of hemipenial sheath. Two large, smooth scales form base of postanal hemipenial pockets. At this point the scale rows begin to become more prominent and imbricate, forming whorls; lowermost rows of scales begin to have keels at about % of length of tail, keels invade lateral rows until at i/. length of tail all rows are keeled; keels increase in prominence until at end of tail there are four pronounced edges formed by keels on four rows of scales. Table 1 Fuie Foot Hind Foot Left 5-11-16-17-10 4-11-19-29-18 Right 5-11-17-17-10 4-12-14-30-18 Scnle foinui'.ne for digits in liolotyiK". All enlarged scales on the venter of each digit were counted, beginning with the first distinctly enlarged, pavimentose scale at the base of the toe and continuing to the tip of the expanded lamellar portion. Scales on the most distal, curved phalan.x, which arises dorsally from the lamellar portion of the toe, are not included. Dorsal ground color (in alcohol) a dull gun-metal blue, with 1 !);■)() l.EAF-NOSKD ANOLE FROM SOUTH AMEHK'A 5 irregular, ])arallel, horizontal black streaks along vertebral mar- gin. These streaks vaguely line up and form a pair of lateral bands between the limbs. Dark brownish-black spot at shoulder. Jjimbs and lateral surfaces spotted with light yellowish-white ; limbs vaguely barred with darker blue ; skin in interstices be- tween scales of dorsum of digits light, giving appearance of banding on fingers and toes. Dorsum of head unicolor, as back ; temporal region and lips lighter, with faint reddish or purplish tinge and a marked light spot over the ear opening. Ventral surfaces generally lighter, M'ith vague reddish tints on chest and chin; belly stippled with light spots. Venter of limbs heavily spotted with white proximally, becoming totally white on foot. Tail alternately barred with gun-metal blue and blackish, bands approximately equal in width. Measurements (in millimeters). Total length (not including rostral appendage), 171; body length (rostral to vent), 74; tail length, 97; head length (rostral to ear opening), 23; head width at widest point, 11 ; length of rostral appendage, 23 ; length of fore limb, 31; length of hind limb, 42. Remarks. Since only a single male of A. proboscis is known, the possibility exists that the specimen belongs to a previously described Ecuadorian form, known either solely from females or from juvenile males, both of which might lack the rostral append- age. Since the species occurs in a very distinctive biotic region, the Choco of Ecuador, Colombia, and Panama, it is also possible that it might be the first Ecuadorian representative of a Choco species. There are seventeen species which have been described from type localities within the biotic area. Of the sixteen that are currently considered valid species, eight {chloris Blgr., eulaemus Blgr., festae Peracca, gracilipes Blgr., granuliceps Blgr., lotifrons Berthold, maculiventris Blgr., and peraccae l>lgr. ) have been previously recorded from Ecuadorian localities. While in the Choco area of Ecuador in 1954, the senior author collected four specimens of A. breviceps Blgr., three individuals of A. peraccae Blgr., and one of A. festae Peracca, at Hacienda Equinox, which is 30 kilometers northwest of Santo Domingo de los Colorados. A. breviceps had not been known from Ecuador. Seven additional species have been recorded in the literature as taken from Ecuadorian localities within the Choco area, although 6 BREVIORA NO. 62 their type localities are elsewhere. Two other taxons, aequatori- alis Werner and irregularis Werner, must also be considered, since they were described from Ecuador with no added informa- tion, and no additional specimens have been taken to establish a range for the species. There are sixteen species, including alhi Barbour, bitectus Cope, hreviceps Blgr., fasciatus Blgr. (with its synonyms elegans Blgr. and irregularis Werner), festae Peracca, fraseri Giinther (with its synonym devillei Blgr.), gracilipes Blgr., granuliceps Blgr., latifrons Berthold (with its synonym princeps Blgr.), lemniscatus Blgr., lemurmus Cope, macrolepis Blgr., maculi- ventris Blgr., notophoUs Blgr., palmeri Blgr., and peraccae Blgr., for which both males and females are known, and for which available specimens or descriptions provide many and obvious diiferences from A. proboscis. Females only are known of the species antonii Blgr. and ventrimacnlatus Blgr., but the descriptions given for the types shows that these species are quite different from A. proboscis. Only males of aequatorialis Werner, chloris Blgr., eulaemus Blgr., rosenbergi Blgr., and tropidogaster Cope have been described in the literature, but here again each appears to be quite distinct. The sex was not given in the original descriptions of binotatus Peters, gemmosus 'Shaughnessy, and lionotus Cope, but they are sufficiently well described to eliminate them as available names for A. proboscis. Myers and Carvalho (1945, p. 7) felt that the relationships of their new species were wdth the punctatus group of the genus Anolis, and this is undoubtedly true also of proboscis. They discussed A. punctatus, A. boulengeri, A. transfasciatus and A. nasofrontalis as probable members of the species group. They had no specimens of A. boulengeri available, and the specimens called A. punctatus were only tentatively identified as such. They had, however, excellent material of A. transfasciatus Amaral, a species they considered sufficiently similar to A. punctatus to suggest that the former might well be placed eventually as a subspecies or synonym of the latter. A single specimen of A. boulengeri 'Shaughnessy, collected in Santiago-Zamora Province, Ecuador, in the region between the Rio Pastaza and the Rio Santiago, is catalogued as No. 45776, in the Museum of Comparative Zoology. It provides an 1956 LEAF-NOSKD ANGLE FROM SOUTH AMERICA 7 opportunity to compare data for tliis species witli tlie notes made on other species by Myers and Carvalho. The specimen has, on its snout, a substantial protuberance which does not extend to any distance beyond the end of the jaw. This swelling- is without doubt the primitive condition that led eventually to the elonga- tion observed in phyllorhinus and proboscis. It is slightly more prominent than that of A. transfasciatus and presumably also of punctatiis. In addition to the presence of keels on the ventrals, A. houlengeri differs from both leaf -nosed species and trans- fasciatus in the presence of keels on the enlarged supraocular scales. The suboculars and the lowest row of loreals form a straight series in both houlengeri and phyllorhinus, while in transfasciatus and proboscis the subocular series turns upward around the eye, not forming a linear series with the loreal row. There is no middorsal scale row in houlengeri. Two rows of scales, slightly larger than the other dorsals, occupy the vertebral line. There is no midventral row of scales. All dorsal, lateral, and ventral scales have keels which can be seen if examined under sufficiently high magnification. This is not true of p^'o- boscis or phyllorhinus, and apparently not true of transfasciatus. It appears, then, that A. houlengeri is quite distinct from all these species, and presumably, from A. punctatus as well. Verifi- cation of the latter assumption must await direct comparison of the two species. Cope (1876, 1). 1(J5) described t