PaleoBios, Volume 14, Number 3 A REVIEW OF THE LONCHORHYNCHINE TREMATOSAURS (LABYRINTHODONTIA), AND A DESCRIPTION OF A NEW GENUS AND SPECIES FROM THE LOWER MOENKOPI FORMATION OF ARIZONA Samuel P. Welles Museum of Paleontology University of California Berkeley, CA 94720 ABSTRACT The labyrinthodont superfamily Trematosauroidea is diagnosed. It includes only the family Trematosauridae, which includes the subfamilies Trematosaurinae and Lonchorhynchinae, each of which is diagnosed. The several previously described genera and species of the Lonchorhynchinae are reviewed in chronologic order. A new genus and species is described from a skull, lacking the otic region, from the Meteor Crater Quarry in the Wupatki Member of the Moenkopi Formation, Spathian, of Northern Arizona. Isolated fragments of skulls and jaws from Algeria, Australia, India, Madagascar, and Nova Scotia are diagnosable only to subfamily, but indicate a broad geographic distribution of Lonchorhynchinae and show its geologic range to be from the Griesbachian up into the Anisian. INTRODUCTION Labyrinthodont amphibians were a major element of the tetrapod fauna of the Permian and Triassic Periods, probably a main source of protein for the developing thecodont reptiles. This predator-prey relationship probably was a major factor in the extinction of the labyrinthodonts by the close of the Triassic Period, except in China, Mongolia and Australia. Labyrinthodonts were present on all continents, and evolved into forms with extremely different skull shapes. The normal skull was broad, low, and ovoid. The extremes were the very broad, short skulls of the brachyopids, and the very long, slender skulls of the trematosaurids. The family Trematosauridae was represented on all continents in the Early Triassic, and some 31 genera have been assigned to it at one time or another. These include Aphaneramma, Erythro- batrachus, Gonioglyptus, Halobatrachus, Hy- perokynodon, Icanosaurus, Ifasaurus, Indoly- rocephalus, Inflectosaurus, Laidleria, Loncho- rhynchus, Luzocephalus, Lyrocephaliscus, Lyro- cephalus, Lyrosaurus, Meyerasuchus, Micropo- saurus, Panchetosaurus, Peltostega, Platystega, Rasaurus, Rhytidosteus, Stoschiosaurus, Ter- treme, Tertremoides, Thoosuchus, Trematosaurus, Trematosuchus, Tupilakosaurus, Volgasaurus, and Wantzosaurus. Several of these genera have been removed from this family and placed in other families. These include: Icanosaurus to synonymy with Pelorocephalus in the Chigutisauridae, Lyrocephaliscus to the Lyrocephaliscidae, Pelto- stega and Rhytidosteus to the Rhytidosteidae, Thoosuchus to the Benthosuchidae, Tupilako- saurus to the Brachyopidae, and Volgasaurus to the Wetlugasauridae. Hammer (1987:75) listed the following genera as invalid: Gonioglyptus, Hyper okynodon, Ifasaurus, Indolyrocephalus, Laidleria, Lonchorhynchus, Lyrosaurus, Micropo- saurus, Peltostega, Rhytidosteus, and Tertrem- oides. The remaining genera constitute a very large family, and its division was indicated by Efre- mov (1933A:144), who proposed the evolutionary line Platyops stuckenbergi, P. watsoni, Loncho- rhynchus, and Aphaneramma, and stated (p. 146) that this line was separate from the Tremato- sauridae. Save-Soderbergh (1935:86, 131, 138, 200) enlarged and formalized this suggestion by dividing the Trematosauridae into the three sub- families Lonchorhynchinae, including Loncho- rhynchus; Lyrocephalinae, including Lyroce- phalus and Platystega; and Trematosaurinae, including Aphaneramma, Stoschiosaurus, Ter- trema, Trematosaurus, and Trematosuchus. Cos- griff and Garbutt (1972:15) did not use the name Lonchorhynchinae, but instead used Group 1, to Page 2 Lonchorhynchine trematosaurs PaleoBios include Aphaneramma, Erythrobatrachus, Gonio- glyptus, Stoschiosaurus, and Wantzosaurus. Hammer (1987:82) accepted the conclusion of Save-Soderbergh (1936:58) that Lonchorhynchus is a junior synonym of Aphaneramma. He there- fore considered the subfamily Lonchorhynchinae invalid, contrary to Stoll, et al., 1961, article 40, which states that a family-group name based on a junior synonym continues to be valid. He then divided the Family Trematosauridae into the subfamilies Aphaneramminae [Aphaneramma- tinae], including Aphaneramma, Erythrobatra- chus, Halobatrachus, Stoschiosaurus, and Wan- tzosaurus; and Trematosaurinae, including ln- flectosaurus, Lyrocephaliscus, Platystega, Ter- trema, Trematosaurus, and Trematosuchus. To the Lonchorhynchinae I now add ?Ifa- saurus, and the specimens described below as generically incertae sedis, and Cosgriffius campi. I further consider Halobatrachus to be a junior synonym of Aphaneramma. Lonchorhynchines probably evolved out of the Archegosauridae, and Platyops has often been proposed as a direct ancestor. However, an even longer-snouted archegosaur, Prionosuchus plummeri Price, (1948:10), is known from the Permian of northern Brazil. Lonchorhynchines have previously been found only in rocks of Scythian age, and they occur in both marine and continental deposits. Cosgriff (1984:42) noted that they were deltaic and estuarine piscivores. The earliest is Gonioglyptus longirostris from the lower Griesbachian Panchet beds of India. The next younger Dienerian Substage has yielded Aphaneramma kokeni from the Salt Range of Pakistan; A. sp., Ifasaurus elongatus and Wan- tzosaurus elongatus from the middle Sakamena beds of Madagascar; and Lonchorhynchinae, in- certae sedis., from the Arcadia Formation of Queensland. The next younger is Aphaneramma rostratum from the Smithian of Spitsbergen and the aberrant Erythrobatrachus noonkanbahensis from the Smithian of Western Australia. In the following discussion of the various specimens that can be assigned to this subfamily I follow a chronological sequence. Some previously described genera and species are valid, and some are here considered indeterminate. After describ- ing the valid species, I describe the skull from the Meteor Crater Quarry of Arizona that represents the new genus Cosgriffius. Last come the fragments from Algeria, Australia, India, and Madagascar. Of the new specimens described in this paper. the earliest is Lonchorhynchinae, incertae sedis from the Dienerian of Madagascar. Cosgriffius campi is from the Spathian of Arizona, while Lonchorhynchinae, incertae sedis, from the Den- wa beds of India, extends the range up into the Anisian Stage, as does the material from Nova Scotia. All lonchorhynchines have very delicate skulls, and are extremely rare in the fossil record. The known specimens indicate that they probably had a circumtethyan distribution, similar to that of the Late Triassic metoposaur- phytosaur fauna. For purposes of description I divide the skull roof as follows: prenarial, snout to line across front of nares; nasoorbital, front of nares to line across front of orbits; orbitopineal, front of orbits to front of pineal foramen; postpineal, front of pineal to back of roof. The term "index" in conjunction with a number is used to denote the size of a morphologic structure relative to the interorbital breadth of the species in question. The number is the length of the structure in question divided by the interorbital breadth. ABBREVIATIONS BMNH - British Museum of Natural History. GSI - Geological Survey of India. ISI - Indian Statistical Institute. MNHN - Museum National d'Histoire Natu- relle, Institut de Paleontologie. PU - Princeton University, now housed at the Yale Peabody Museum. QMF - Queensland Museum. SAM - South African Museum. UCMP - University of California, Museum of Paleontology. USGS - United States Geological Survey. WAM - Western Australia Museum. SYSTEMATIC PALEONTOLOGY Subclass LABYRINTHODONTIA Meyer, 1842C, p. 301 (Labyrinthodontes). Order TEMNOSPONDYLI Zittel, 1885B, p. 348. Superfamily TREMATOSAUROIDEA Watson, 1919B, p. 67 (Trematosauridae).1 l Although Shishkin (1950A:112) was the first to use the term Trematosauroidea, Watson had earlier (1919B:67) proposed the family name Trematosauridae, and as stated in the International Code (Stoll, et al., 1961, Art. 23(c)) "The priority of the PaleoBios Lonchorhynchine trematosaurs Page 3 1919B Trematosauridae Watson, p. 67. 1935 Trematosauroideae Save-Soderbergh, p. 90. 1946B Trematosauroideae Nilsson, p. 48, superfamily. 1980A Trematosauroidea, Shishkin, p. 112. 1984 Trematosauroidea, Cosgriff, p. 31. 1985 Trematosauroidea, Warren and Black, p. 314. Diagnosis: Skull narrow, breadth at quadrato- jugal about four times interorbital breadth, elon- gate, high posteriorly; sides straight or slightly concave; orbits far lateral; nares elongate; orbito- pineal zone of intense growth, with pineal fora- men far back in posterior half of postorbital region; intertemporal absent; prefrontal meets postfrontal; jugal enters orbit; quadrate and occi- pital condyles nearly aligned; anterior palatal vacuities divided; parasphenoid with long basal plate, extending back below occiput, and with long pterygoid suture; cultriform process very narrow; pterygoid not reaching palatine; inter- clavicle elongate; vertebrae rhachitomous. Includes only Family Trematosauridae. Family TREMATOSAURIDAE Watson, 1919B, p. 67. 1919B Trematosauridae Watson, p. 67. 1920C Trematosauridae Huene, p. 212. 1935 Trematosauridae, Save-Soderbergh, p. 90. 1947C Trematosauridae, Romer, p. 314. 1960A Trematosauridae, Shishkin, p. 130. 1972 Trematosauridae, Cosgriff & Garbutt, p. 6. 1980B Trematosauridae, Shishkin, p. 130. 1987 Trematosauridae, Hammer, p. 73 1987 Trematosauridae, Hellrung, p. 18 Diagnosis: As for Superfamily Trematosauroidea. Includes Subfamilies Trematosaurinae and Lon- chorhynchinae. Subfamily TREMATOSAURINAE (Tremato- sauridae) Watson, 1919B, p. 67. 1935 Trematosaurinae Save-Soderbergh, p. 86. 1972 Group 2 Cosgriff & Garbutt, p. 15. 1987 Trematosaurinae Hammer, p. 82. Diagnosis: Skull moderately narrow; prenarial name of a taxon in the family-, genus-, or species- group is not affected by elevation or reduction in rank within the group" (see also Art. 37). I have therefore credited the superfamily name to Watson. region short; nares lateral; choanae broad, imme- diately behind nares; interpterygoid vacuity moderately narrow; dentary symphysis short; interclavicle moderately narrow. Genera in- cluded: Hyperokynodon, Hndolyrocqjhalus, In- flectosaurus, Microposaurus, Platystega, Sto- schiosaurus, Tertrema, Tertretnoides, Tremato- saurus, Trematosuchus. Subfamily LONCHORHYNCHINAE Save- Soderbergh, 1935, p. 86. 1935 Lonchorhynchinae Save-Soderbergh, p. 86, 200. 1972 Group 1 Cosgriff & Garbutt, p. 15. 1987 Aphaneramminae Hammer, p. 82. Diagnosis: Skull very narrow, with extremely narrow rostrum and extremely elongate prenarial and orbitopineal regions; choanae very narrow, and far behind nares; interpterygoid vacuity extremely narrow and elongate; dentary symphysis very long; adductor fossa and posterior meckelian foramen very long; interclavicle extremely narrow. Genera included: Aphane- ramma(=Halobatrachus, =Lonchorhynchus), Cos- griffius, Erythrobatrachus, Gonioglyptus (=Pan- chetosaurus), ?Ifasaurus, Wantzosaurus. Gonioglyptus Huxley, 1865C, p. 5. 1865C Gonioglyptus Huxley, p. 5. 1879D Gonioglyptus, Lydekker, p. 17. 1881D Gonioglyptus, Lydekker, p. 176. 1882C Gonioglyptus, Lydekker, p. 26. 1887H Glyptognathus, Lydekker, p. 68. 1919B Gonioglyptus, Watson, p. 48. 1920C Gonioglyptus, Huene, p. 212. 1920F Gonioglyptus, Huene, p. 448. 1947C Gonioglyptus, Romer, p. 186. 1969A Panchetosaurus Tripathi, p. 3. 1985 Gonioglyptus, Warren & Black, p. 307. 1987 Gonioglyptus, Hammer, p. 74. Type species: G. longirostris Huxley. Horizon: Deoli Stage of the Panchet Series. Age: Early Griesbachian. Distribution: Known only from northeast India. Diagnosis: A lonchorhynchine; length of rostrum unknown, but probably moderately long; naso- orbital region : interorbital breadth index 329; skull wide at middle of orbits, skull breadth: interorbital breadth index 236; sculpture mode- rately deep; supraorbital canal curves sharply medially rather far in front of orbit; inter- Page 4 Lonchorhynchine trematosaurs PaleoBios pterygoid vacuity 13 mm in front of the orbit (interorbitai breadth index 87); vacuities widely separated anteriorly, their anterior ends moderately sharp; mandible flat ventrally, with glenoid oriented anteromedially; posterior meckelian fenestra elongate and far in front of glenoid; retroarticular process long and narrow. Although Warren & Black (1985:308) did not allocate Gonioglyptus to a family, it has a very slender rostrum and is here included in the sub- family Lonchorhynchinae. Hammer (1987:74) considered it a nomen dubium. Gonioglyptus longirostris Huxley, 1865C, p. 1. Literature, Horizon, Age, Distribution and Diagnosis: As for genus. Type: GSI2135, the rostrum of a small skull. Hypodigm: Type and specimens referred by Huxley, and by Tripathi (1969A:7). Locality: Deoli Stop, Damodar Valley, Purulia District, West Bengal, India. Discussion: Associated with the rostral piece were the back of the left ramus of the lower jaw, another fragment, an incomplete interclavicle, and vertebrae. The beautiful illustrations in Huxley (1865C, pi. 6) show the sculpture of the type to be quite sharp, and the canals distinct. Huxley restored the rostrum as very short, but comparisons with subsequently described taxa indicate that it was much longer. The supraorbital canal leads directly toward the center of the orbit until at a distance of 13 mm from the orbit it curves sharply medially so that the left and right canals are but 3 mm apart. The infraorbital canal is not visible in dorsal view. In lateral view the suborbital bar is high and the jugal canal is at mid-height. In ventral view the interpterygoid vacuity extends forward 13 mm beyond the orbits (index relative to interorbitai breadth 88), their tips 10.5 mm apart. Referred material: Huxley (1865C:5, pi. 6, fig. 2) described the back part of a mandible. This was described again by Tripathi (1969A:2). It is the proper size for the skull, and could well be the same species. Lydekker (1879D:17, pi. 3, fig. 14, 15) described a fragment of jaw as Gonioglyptus, sp. nov. It is indeterminate. Lydekker (1882G27, pi. 1, fig. 1) described a tiny piece 38 mm long as part of a lower jaw, making it the type of his new genus and species Glyptognathus fragilis. This has been listed by many workers, and was finally described in detail by Tripathi (1969A:8) and identified as the palatine of Gonioglyptus longirostris. However, if, as described by Tripathi, the tusk pits are side by side, the specimen must be the tip of the dentaries because all tusk pits on labyrin- thodont palates are fore and aft. Only on the tip of the lower jaw do they become situated trans- versely. Lydekker (1882027, pi. 1, fig. 6) described a symphysial fragment, and this was redescribed by Tripathi (1969A:8), who referred it to Gonio- glyptus, its number GSI 4295. However, this is not a lonchorhynchine as the rami are divergent and the symphysis is shallow. Tripathi (1969A:7, pi. 2, fig. 4, 5) made the narial region of a trematosaur the type of Pan- chetosaurus panchetensis, placing it in the sey- mouriamorphs. This is evidently the narial region of a lonchorhynchine trematosaur, as indi- cated by Cosgriff (1984:36), who suggested that it might pertain to Gonioglyptus longirostris. It seems that Tripathi mistook the nares for the orbits, hence his concern over the absence of inter- pterygoid vacuities. I have tried, unsuccessfully, to adapt his descriptions of the skull bones to my interpretation. However, if his "orbits" be nares, his "parasphenoid" a vomer, his "paired cavi- ties" of his fig. 3, choanae, etc., the fragment is of a small form similar to Gonioglyptus longirostris and is a piece of the rostrum anterior to that of the type. It is slightly smaller than the type, but could well be the same species. Gonioglyptus huxleyi Lydekker, 1882C, p. 26. 1882C Gonioglyptus huxleyi Lydekker, p. 26, pi. 1, fig. 5, a 1919B Gonioglyptus huxleyi, Watson, p. 48, indeter- minate 1920F Gonioglyptus huxleyi, Huene, p. 449, incertae sedis 1947C Pachygonia, Romer, p. 187. 1969A Indolyrocephalus huxleyi Tripathi, p. 14. 1985 Gonioglyptus huxleyi, Warren & Black, p. 305, indeterminate Type: Posterior part of left mandible, GSI 4297. Horizon: Deoli Stage, upper Panchet Series. Age: Early Griesbachian. Distribution: Known only from Deoli, Purulia District, West Bengal, India. Discussion: This specimen was redescribed in detail by Tripathi (1969A:14) and is obviously not a trematosaur, as was first suggested by Romer PaleoBios Lonchorhynchine trematosaurs Page 5 (1947C:187). Tripathi placed it, tentatively, in the species Indolyrocephalus huxleyi. Aphaneramma Woodward, 1904D, p. 173. (=Lonchorhynchus Wiman, 1910). 1904D Aphaneramma Woodward, p. 173, pi. 12. 1910A Lonchorhynchus Wiman, p. 37. 1915A Aphaneramma, Wiman, p. 17, also Loncho- rhynchus. 1916B Lonchorhynchus, Wiman, p. 220. 1917A Aphaneramma Wiman, p. 230, also Loncho- rhynchus. p. 238. 1919B Aphaneramma, Watson, also Lonchorhynchus. 1920C Aphaneramma, Huene, p. 212, also Loncho- rhynchus. 1920F Aphaneramma, Huene, p. 443, also Loncho- rhynchus. 1933A Lonchorhynchus, Efremov, fig. 15. 1935C Lonchorhynchus, Bystrow, p. 118. 1935 Aphaneramma, Save-Soderbergh, p. 85, 140, also Lonchorhynchus. 1936 Aphaneramma, Save-Soderbergh, p. 58. 1937C Aphaneramma, Save-Soderbergh, p. 190. 1940G Aphaneramma, Efremov, p. 22. 1942 Aphaneramma, Nilsson, p. 48, ftn. 91. 1943A Aphaneramma, Nilsson, p. 230, 236, 237, 254. 1943B Aphaneramma, Nilsson, p. 5. 1947C Aphaneramma, Romer, p. 185. 1965 Aphaneramma, Cosgriff, p. 89 1966A Aphaneramma, Lehmen, p. 138. 1969A Aphaneramma, Cosgriff, p. 81. 1972 Aphaneramma, Cosgriff & Garbutt, p. 16. 1984 Aphaneramma, Cosgriff, p. 38. 1985 Aphaneramma, Warren & Black, p. 307. 1987 Aphaneramma, Hammer, p. 74, 80,82. Type species: A. rostratum Woodward. Age: Smithian. Distribution: Known from Spitsbergen and India. Diagnosis: (Based largely upon referred specimens), Lonchorhynchinae with extremely long prenarial region; extremely long nasoorbital region (index relative to interorbital breadth 302); very long naris (index 44); very long, and large orbit; very long orbitopineal region (index 290); skull very broad at mid-orbits (index 232); jugal forming entire lateral wall of orbit; quadratojugals diverging at 22°; skull very broad posteriorly (index 359); supraorbital, infra- orbital, and temporal canals deeply incised at junction; choana moderately far behind snout (index 424); choana far in front of interpterygoid vacuities (index 117); orbits far in front of vacuities (index 180). / / / 1 I 1 n 1 1 I 1 1 *¦" I / /I / / 1 ; / U * f \ / / 1 / ' / 1 / 1 1 1 1 I I J___I 5cm Figure 1. Aphaneramma rostratum Woodward, type, BMNH R3180-2. Dorsal view of skull with latex cast of postorbital region in place, and rostral steinkcrn fitted to contour, ioc-infraorbital canal; jc-jugal canal; n- nasal; po-postorbital; pof-postfrontal; prf-prefrontal; st-supratemporal; tc-temporal canal. Page 6 Lonchorhynchine trematosaurs PaleoBios Aphaneramma differs from Gonioglyptus in having a much gentler medial curvature of the supraorbital canal as it bends between the orbits, and in having this canal form an arch around the orbit rather than running parallel to the midline between the orbits; relative to the interorbital breadth, the nasoorbital region is shorter (index 302 vs. 329); the orbits are much farther in front of the interpterygoid vacuities (index 180 vs. 108); the choanae are closer to the vacuities (index 117 vs. 171). Aphaneramma rostratum Woodward, 1904D, p. 173. Literature: As for genus. Type: BMNH R3180-2, the internal cast of the orbital region of a skull, plus a small piece of the postorbital region with impression of the bone, and a piece of the rostral cast, lacking the tip, which does not contact the main piece, but probably is the same individual. Hypodigm: Type and the many specimens from Spitsbergen described by Wiman (1910 - 1917) and Save-Soderbergh (1935, 1936). Horizon: Fish Horizon, Sticky Keep Formation. Age: Smithian. Locality: Mt. Sticky Keep, Ice Fjord District, Spitsbergen. Diagnosis: As for genus. Discussion: Aphaneramma rostratum was based on five pieces of hard, black mudstone, to which some bone adhered. Several vertebrae were associated with the type by Woodward, but Nilsson wrote (1943A:251) that Stensio had informed him these were from a higher horizon, and are ichthyosaurian. The skull pieces were figured by Woodward (1904D, pi. 12), and Save- Soderbergh (1936:60) selected the steinkerns of the back of the skull (Woodward, 1904D, pi. 12, fig. 1, 2) as the type. In his restoration of the skull Woodward placed the rostral piece far forward, making a very long snout. Save- Soderbergh (1935, fig. 56) made the curvature of this fragment fit that of the orbital region, resulting in a much shorter snout. This was done from Woodward's illustrations plus the redescription by Wiman (1915A:l7). I, at first, also restored Aphaneramma with a relatively short snout, but correspondence with Franz-Josef Lindemann, at Oslo, who is restudying all of the Spitsbergen material as well as more that he collected, convinced me that Aphaneramma and Lonchorhynchus are synonymous, and that the rostrum is indeed very long. Save-Soderbergh (1936:60) wrote that the postorbital fragment (Woodward, 1904D, fig. 4) could not be proved to belong to the same specimen as the type, and he went on to refer this fragment to his Aphane- ramma sp. 1. However, Dr. Alan Charig kindly allowed me to examine the material, and I found that the postorbital piece actually has a good matrix fit with the type steinkern. Mr. Linde- mann wrote me that he had confirmed this fit. Woodward (1904D:173) described the skull as very little depressed, with the occiput sloping backward. He further described the quadrate as high and narrow, and sheathed internally by a great vertical plate of the pterygoid. His Figure 2 shows a high (45 mm) cheek region. The post- orbital piece (Fig. 1) has a sculpture of small, rather deep pits, and an almost equiangular junction of the supraorbital, temporal, and jugal canals. The canals are deep and sharply incised. Save-Soderbergh reversed his earlier view that Aphaneramma and Lonchorhynchus were distinct, and in his 1936 paper (p. 58), and again in 1937 (p. 190), placed the latter in synonymy with the former. In 1936 he went on to refer all of the Spitsbergen forms with elongate rostra to Aphaneramma. In this he distinguished two species, which he described as sp. 1 and sp. 2. It might be noted that his sp. 2 includes the type of L. oebergi, and this trivial name could well have been used. However, Nilsson (1942:58, ftn.) placed all of this Spitsbergen material in a single species, A. rostratum and I concur, especially after the convincing new finds by Lindemann. It would be of little use for me to describe these steinkerns in detail since Lindemann is restudying all of the material. Aphaneramma ocella (Meyer) Huene, 1920F, p. 443. 1848E Labyrinthodon (Trematosaurus) ocella Meyer, p. 469, nomen nudum 1855A Labyrinthodon ocella Meyer, p. 140, pi. 61, fig. 1, 2 1864A Trematosaurus, Alberti, p. 238. 1919B not Trematosaurus, Watson, p. 41. 1920F Aphaneramma ocella Huene, p. 443. 1947C Trematosaurus, Romer, p. 175. 1965 Labyrinthodon ocella, Welles & Cosgriff, p. 34, nomen vanum PaleoBios Lonchorhynchine trematosaurs Page 7 Type: Back of skull. Hypodigm: Type only. Locality: Bemburg, Germany. Horizon: Upper middle Buntsandstein. Age: Scythian. Diagnosis: This, as early suggested by Alberti (1864A:238), and again by Watson (1919B:41), and by Romer (1947G175), is probably a distorted specimen of Trematosaurus brauni. Aphaneramma kokeni (Huene, 1920F, p. 451). 1920F Gonioglyptus kokeni Huene, p. 451. 1947C Gonioglyptus kokeni, Romer, p. 186. 1972 Gonioglyptus kokeni, Cosgriff & Garbutt, p. 6. 1987 Halobatrachus kokeni Hammer, p. 80, 82. Type: Postrostral part of skull at University of Tubingen. Hypodigm: Type only. Locality: Chideru, the Salt Range, Pakistan. Horizon: The Prionolobus beds, Mittiwali Mem- ber of the Mianwali Formation. Age: Dienerian. Diagnosis: An Aphaneramma, differing from A. rostratum in that, relative to interorbital breadth, the orbits are smaller (index 78 vs. 110); the breadth at mid-orbit is less (index 205 vs. 232); the otic notch is slightly closer to the orbit (index 385 vs. 409); the breadth at the quadrato- jugal is greater (index 347 vs. 330); the inter- pterygoid vacuity is nearer the orbit (index 72 vs. 180). Discussion: This skull is so similar to A. rostra- tum that I transfer it to the same genus. Aphaneramma, sp., Lehman, 1966A, p. 21. 1966A Aphaneramma, sp. Lehman, p. 21, fig. 9, pi. 7D. 1972 Aphaneramma, sp., Cosgriff & Garbutt, p. 7, not a valid taxon. 1974 Aphaneramma, sp., Lehman, p. 166. 1985 Aphaneramma, sp., Warren & Black, p. 307, indeterminate Type: A skull roof fragment. Horizon: Middle Sakemena beds, upper Andava- koera Formation. Locality: Antsaba, Madagascar. Age: Early Dienerian. Discussion: I concur with the cited authors that the specimen is indeterminate. The type of A. rostratum lacks comparable parts. Compared with specimens referred to A. rostratum the posterior ends of the prefrontal are narrower, and the supratemporal projects farther forward between the postfrontal and postorbital. The sculpture is of radiating grooves and ridges, rather than pits. The small remaining segment of the infraorbital canal does not run so postero- medially, but is almost parasagittal. Even though indeterminate, this specimen could well be included in Aphaneramma, as A. sp, as was done by Lehman. Associated with this specimen were fragments of an elongate rostrum, which indicate only the presence of a trematosaurid. Wantzosaurus Lehman, 1961, p. 114. 1961 Wantzosaurus Lehman, p. 114. 1972 Wantzosaurus, Cosgriff & Garbutt, p. 16. 1974 Wantzosaurus, Lehman, p. 166. 1987 Wantzosaurus, Hammer, p. 74, 80. Type species: Wantzosaurus elongatus Lehman. Horizon: Middle Sakamena beds, upper Andava- koera Formation. Age: Early Dienerian. Distribution: Known only from northwest Mada- gascar. Diagnosis: A lonchorhynchine with short pre- narial region (index relative to interorbital breadth 260); long nasoorbital region (index 609); very short postpineal region (index 128); broad back of skull (index 556); interpterygoid vacuity extending well in front of orbits (index 141). Comparisons: Compared with Gonioglyptus, and relative to interorbital breadth, the nasoorbital region is longer (index 609 vs. 329); the breadth at mid-orbit is greater (index 342 vs. 236); the choana is farther in front of the interpterygoid vacuity (index 183 vs. 171); the vacuity extends farther in front of the orbit (index 141 vs. 108); the sculpture is not so sharp; the supratemporal canal is not so curved laterally in front of the orbit, nor does it have the sharp anterior bend. Compared with Aphaneramma, and relative to interorbital breadth, the nasoorbital region is greatly elongated (indices 609 vs. 302); the orbitopineal region is slightly longer (index 308 vs. 290); the breadth at the nares is greater (index 88 vs. 48); and the skull widens more at mid-orbit (index 342 vs. 232); it is much broader at the quadratojugals (index 556 vs. 330); the choana is farther from the interpterygoid vacuity (index 183 vs. 117); the vacuity does not extend so far in front of the orbit (index 141 vs. Pages Lonchorhynchine trematosaurs PaleoBios 180); the angle of divergence behind the orbits is 14° vs. 12° ; the sculpture is not so deep; the sensory canals are quite similar, except that the supraorbital canals diverge less sharply in front of the orbits, and are more nearly parallel between the orbits. Wantzosaurus elongatus Lehman, 1961, p. 114. Literature: As for genus. Type: A juvenile skull. Hypodigm: Type only. Locality: Near Ambilobe, near Madiromary, northeast Madagascar. Horizon, Age, and Diagnosis: As for genus. Discussion: The type skull has the relatively large orbits of a juvenile, so that comparisons with more mature specimens of other taxa is of minimal value. However, this is the first of the lonchorhynchines to have the tip of the snout preserved. It is a fairly short-snouted form, especially in the prenarial region, while the nasoorbital region is very long. Presumably the snout would elongate as the individual matured. Indolyrocephalus Tripathi, 1969A, p. 14. 1882C Gonioglyptus, Lydekker, p. 26. 1887H Gonioglyptus, Lydekker, p. 68. 1920F Gonioglyptus, Huene, p. 449. 1947C Pachygonia Romer, p. 187, 213. 1969A Indolyrocephalus Tripathi, p. 14. Type species: Indolyrocephalus huxleyi (Lydek- ker) Tripathi. Horizon: The Panchet Series. Age: Early Griesbachian. Indolyrocephalus huxleyi (Lydekker) Tripathi, 1969A, p. 14. Literature, Horizon, and Age: As for genus. Discussion: Lydekker (1882G26) named a second species of Gonioglyptus, G. huxleyi, on the articular region of a lower jaw. He later thought (1887H:68) that it was close to Trematosaurus. Watson (1919B:48) and Huene (1920F:449) both considered it indeterminate. The specimen, as Romer pointed out (1947C.-213) is not a trematosaurid, and he placed it in the genus Pachygonia. Hammer (1987:74) also considered it indeterminate, and that it could be tremato- saurid, as do 1. Erythrobatrachus Cosgriff & Garbutt, 1972, p. 7. 1965 Tertrema, Cosgriff, p. 89. 1972 Erythrobatrachus Cosgriff & Garbutt, p. 7. 1985 Erythrobatrachus, Warren & Black, p. 307. 1987 Erythrobatrachus, Hammer, p. 75, 80. Type species: Erythrobatrachus noonkanbahen- sis. Horizon: Blina Shale. Age: Smithian. Distribution: Known only from Western Austra- lia. Diagnosis: A lonchorhynchine with short, mod- erately broad rostrum; orbits far apart; orbital region bulging laterally; interpterygoid vacuities extending slightly in front of orbits (index 41), and with rounded anterior ends; choanae far apart, with rounded anterior ends, and far in front of the interpterygoid vacuities (index 166). Comparisons: Erythrobatrachus differs from Gonioglyptus in having the orbits farther apart, and the face concave in front of the orbits; relative to interorbital breadth the breadth at mid-orbits is about the same (index 231 vs. 236); the choanae are slightly closer to the inter- pterygoid vacuities (index 166 vs. 171); the vacuity does not extend so far in front of the orbit (index 41 vs. 108), and is more rounded anteriorly. The palatines do not extend back along the parasphenoid, and the parasphenoid extends far forward between the vomers. Compared with Aphaneramma, the orbits bulge more laterally, so the face is more concave in front of the orbits. Relative to interorbital breadth the breadth at the quadratojugals is less (index 291 vs. 330); the choanae are farther from the interpterygoid vacuities (index 166 vs. 117); the vacuity is much closer to the front of the orbit (index 41 vs. 180). The vacuity is much smaller, and is narrower (Length : Breadth index 538 vs. 767). The angle of divergence behind the orbit is 6° vs. 12°. Compared with Wantzosaurus, the skull is more concave in front of the orbits, and the orbits bulge more laterally. The interorbital breadth is much greater (34 vs. 13 mm) and the following indices reflect this. Relative to this breadth, the orbit is smaller; the breadth at mid-orbits is less (index 231 vs. 342); the breadth at the quadratojugal is less (index 291 vs. 556); the choana is nearer the interpterygoid vacuities (index 166 vs. 183); the vacuity does not extend so PaleoBios Lonchorhynchine trematosaurs Page 9 far in front of the orbit (index 41 vs. 141); the vacuity is much smaller and narrower (L : B index 538 vs. 600). The angle of divergence of the postorbital wall is 6° vs. 12°. Erythrobatrachus noonkanbahensis Cosgriff & Garbutt, 1972, p. 7. Literature, Horizon, Age and Diagnosis: As for genus. Type: Internal cast of middle of skull, WAM 62.1.46. Hypodigm: Type and internal cast of right posterolateral region, WAM 71.6.222, and cast of prechoanal rostrum, WAM 62.1.50, also additional material described by Cosgriff & Garbutt (1972). Locality: UCMP V6044, Noonkanbah Station, West Kimberley District, Western Australia. 124' 45'E;18* 20'S. See Cosgriff & Garbutt (1972) for detail. Cosgriffins, new genus Type species: Cosgriffius campi, new species Horizon: Wupatki Member of the Moenkopi Formation. Age: Spathian. Distribution: Known only from the type locality. Diagnosis: A lonchorhynchine with greatly elongate prenarial and orbitopineal regions; indices, relative to interorbital breadth, 672 and 480+); interpterygoid vacuities end below front of orbits; choanae far in front of interorbital vacuities (index 308). Supraorbital canals diverge anteriorly near the orbits, run very near the midline and are parasagittal between and behind the orbits before they diverge gently posteriorly. Jugal canals rise far (four orbital lengths) behind the orbits. Compared with Gonioglyptus, relative to interorbital breadth the nasoorbital region is longer (index 464 vs. 329); the choana is farther from the interpterygoid vacuity (index 308 vs. 171); and the vacuity ends below the front of the orbit rather than far in front of the orbit (index 0 vs. 108). The supraorbital canals are closer to the midline between the orbits, and diverge more abruptly in front of the orbits. Compared with Aphaneramma, relative to interorbital breadth the orbitopineal length is much greater (index 480 vs. 290); the choana is much farther in front of the interpterygoid vacuity (index 308 vs. 171); and the inter- pterygoid vacuity ends below the front of the orbit instead of far in front (index 0 vs. 180). The supraorbital canals curve laterally more abruptly in front of the orbits, and approximate the infraorbital canals much nearer the orbits. The supraorbital canals run parasagittally between the orbits, very close to the midline and they diverge posteriorly about one orbital length behind the orbits, rather than being convex between the orbits, and diverging there. The jugal canals curve upward about four orbital lengths behind the orbits, instead of below the back of the orbits. Compared with Wantzosaurus, relative to interorbital breadth the prenarial region is much longer (index 672 vs. 260); the nasoorbital region is shorter (index 464 vs. 609); the orbitopineal region is longer (index 480+ vs. 308); the choana is farther from the snout (index 840 vs. 538); and the interpterygoid vacuities end below the front of the orbits instead of far in front of them (index 0 vs. 141). The supraorbital canals are close together and run parasagittally between and behind the orbits, and they diverge farther back. They diverge more abruptly in front of the orbits. The jugal canals curve upward far behind the orbits instead of immediately in front of them. Compared with Erythrobatrachus, relative to interorbital breadth the quadratojugal is farther from the orbit (index 652 vs. 386); the choanae are farther from the snout (index 308 vs. 166) and the interpterygoid vacuities end below the front of the orbits rather than a little way in front of them (index 0 vs. 41). The choanae are farther from the interpterygoid vacuities (index 308 vs. 166) and the vacuities are pointed anteriorly rather than rounded. There are large vomerine tusk pits and tooth pits. Etymology: Named in honor of the late Dr. John Cosgriff, who contributed much to our understand- ing of the labyrinthodonts. Cosgriffius campi, new species Type: UCMP 112135, a skull, lacking the posterior end. Hypodigm: Type only. Locality: Meteor Crater Quarry, UCMP V3835, a low butte on the west side of the road into Meteor Crater, just north of the road leading east to Meteor Crater Ranch. NE1/4 of SE1/4, Sec 11, T19N, R12 1/2E, Coconino Co., Arizona, as shown on the USGS Meteor Crater Quad., 7.5 Minute Series (Topographic. Photorevised 1983.) (Fig. 2). Page 10 Lonchorhynchine trematosaurs PaleoBios Horizon: A local fresh-water limestone in the Wupatki Member of the Moenkopi Formation, about 58 feet above the Kaibab Limestone. Diagnosis: As for genus. Etymology: Named in honor of the late Professor Charles L. Camp, who furthered and participated in the work by both Cosgriff and myself on the labyrinthodonts. Occurrence: The Meteor Crater Quarry is a unique remnant of a small freshwater lake or pond in a vast expanse of early Triassic floodplain deposits known as the Moenkopi Formation. In the general region of the Little Colorado Valley this formation is divided into three members: the lower Wupatki Member, the middle Moqui Member, and the upper Holbrook Member. The quarry lies near the base of the Wupatki Member, about 18 meters above the Permian Kaibab Limestone. The section at the quarry is only about 3 meters thick, a small butte immediately west of the road, capped by a cross-bedded sandstone 1 meter thick. Its base is a light gray limestone about 30 cm thick at the east end of the quarry, and 56 cm at 25 meters to the west, where our excavation ended. This limestone is peculiar in that it splits into two layers. The lower layer has footprints of the reptiles Chirotherium and Rotadactylus on its surface, while the upper layer has bones of the labyrinthodont Paroto- suchus peabodyi in, on, and just above its top. Here also lay the skull of Cosgriffius campi. Description: The skull is slightly crushed down and to the right, and is badly shattered, with innumerable small cracks that make it very difficult to trace sutures. Many of the cracks are filled with a black material, probably manganese oxide, that also forms deposits on the bone surfaces. The preserved length is 411 mm, the last 17 mm having lost the roof bones. The posterior break is at a 17° angle to the midline so that the left margin extends 10 mm farther back than the right. The skull is 32 mm broad at the back of the nares, 64 mm broad at the back of the orbits, and 120 mm at the broken back end. Dorsal View: (Fig. 4A) The sculpturing is deep, the resulting corrugation greatly strengthening the bones, and the sensory canals are deeply incised. The tip of the rostrum curves ventrally and is 11 mm high, lowering to 8 mm in the midline at 16.5 mm from the tip. The nares are long and narrow, 19 by 6 mm. The orbits are 25 mm long and 18 mm wide, with raised margins. The parietal foramen is more than 100 mm behind the Figure 2. Location of Meteor Crater Quarry. cm ________red-chocolate ms up to ss cap 6 yellowish ms chocolate ms fish-juvenile Parotosuchus skulls 25 mottled gray and blue msandsh 18 8 grayms -----------------bone concentration------ gray Is 33 -----------------footprint level------------- gray Is 23___________quarry floor___________ Figure 3: Section at Meteor Crater Quarry. PaleoBios Lonchorhynchine trematosaurs Page 11 com -px 5 cm Figure 4A. Cosgriffius campi, type, UCMP 112135. Dorsal view of skull showing sutures and sensory canals. com- antcrior commissure; en-external naris; f,-frontal; ioc-infraorbital canal; mx-maxilla; n- nasal; p- parietal; po- postorbital; pof-postfrontal; prf- prefrontal;px- premaxilla; qj-quadratojugal; st-supratemporal. -$fc 5 cm Figure 4B. Cosgriffius campi, type, UCMP 112135. Ventral view of skull, ch-choana; cp-cultriform process; ipv- interptcrygoid vacuity; ect-ectopterygoid; mx-maxilla; pal-palatine; px- premaxilla; v-vomer. PaleoBios Lonchorhynchine trematosaurs Page 13 orbits, there being no sign of it on the bone or matrix. The extreme elongation of the skull has resulted in sutures with very long anteroposterior interdigitations that are difficult to trace, especially in the narial region. Premaxillae: The two form an enlarged semicircle 28 mm wide at the downcurved tip, then narrow posteriorly to 17 mm at 20 mm back. From here they expand gradually to a breadth of 29 mm in front of the nares. The total length of each is 190 mm to where their tips extend between the maxillae and the prefrontals. They form the roof of the rostrum for the anterior 102 mm, and from here on back each is divided by the nasal into medial and lateral projections. The medial projection of the left premaxilla extends back medial to the nasal, but the suture enters the suborbital canal 25 mm behind the tip of the nasal, where it is lost. The lateral projection extends back behind, and lateral to the naris, lying upon the maxilla and below the prefrontal. On the left side just behind the tip is a large foramen, 12 mm long, its lateral wall missing. Its inner wall is 8 mm high (9 mm on the slant), smoothly rounded, and has a cavity dipping posteromedially. This foramen is evidently for a large tusk at the symphysis of the dentary. On the right side, 19 mm from the tip of the snout, is a much smaller foramen, 3x1 mm. It is 4.5 mm from the midline, and its lateral wall is 2 mm wide. A large groove, 34 mm long, continues posteriorly from this foramen. The groove is 2 mm wide and 1.5 mm deep anteriorly, and lies immediately lateral to the sensory canal. The disparity in size of the left and right foramina indicates that there was a rapid resorption and redeposition of bone in response to the changes in the lower tusks. Nasals: These are extremely long (152 mm), with the prenarial part almost as long as the postnarial part. The left nasal begins 63 mm in front of the naris, where it inserts as a very narrow sliver into the premaxilla. The nasal widens gradually posteriorly and forms the anterodorsal and dorsal borders of the naris. It is 11 mm wide in front of the naris, narrows to 8.4 mm at the center of the naris, then widens to 10 mm at 25 mm back. It curves around the back of the naris, reaching slightly (2 mm) forward laterally. It narrows gradually to 8 mm at 42 mm behind the naris. Here it divides into a short (6 mm) medial prong, and a long (28 mm) lateral prong that extends between the frontal and the prefrontal. The supraorbital canal crosses over onto the prefrontal from 25 to 29 mm behind the naris. The right nasal begins 58 mm in front of the naris, and the area behind the naris seems to be a separate bone, the septomaxilla. Septomaxilla: This is indicated on the right side only, by a wedge reaching 20 mm behind the naris. It is 8 mm wide, and is concave anteriorly where it forms the posterior border of the naris and sends processes forward along both medial and lateral edges of the naris. These sutures cannot be confirmed on the left side, where the area seems to be all nasal. Frontals: The left frontal is 147 mm long, meeting the nasal at 55 mm in front of the orbit, and extending 71 mm behind the orbit. The right frontal extends 69 mm in front of the orbit where it interdigitates with the nasal. It continues behind the orbit for 70 mm. Each frontal is 5 mm wide between the orbits, and narrows anteriorly. The supraorbital canal enters the frontal from the postfrontal from 17 mm to 24 mm behind the orbit. Parietal: This is represented only by the ante- rior 24 mm of the left side, and 30 mm of the right. The left sends a prong forward between the postfrontal and frontal, and a lesser prong into the frontal. The right has a sharp anterior end medial to the frontal. Maxilla: Little can be seen of the maxilla in dorsal view until 15 mm in front of the naris, where it forms about 3 mm of the roof. The right maxilla begins 61 mm in front of the naris, at the top of the infraorbital canal. The suture with the premaxilla extends 20 mm posteroventrally across the canal where the suture curves onto the palatal surface. Behind the naris it forms more of the roof lateral to the nasal (or septomaxilla) until some 20 mm in front of the orbit where it bends down along the prefrontal. It underlies the jugal at the center of the orbit, and continues on back at least 40 mm, where the bone is lost. Prefrontal: This forms the anterior third of the orbital border, meeting the postfrontal at nearly the center of the orbit. The suture zigzags medially for 4 mm, where the prefrontal sends a slender (3 mm) projection back 12 mm between the frontal and postfrontal. The anterior orbital border is everted, and the center of ossification is 10 mm in front of the orbit. The bone is at least 25 mm wide at the orbit, and narrows rapidly anteriorly, punching out between the maxilla and the nasal at about 73 mm in front of the orbit. Postfrontals: This is at least 104 mm long, reaching to 67 mm behind its center of ossifica- Page 14 Lonchorhynchine trematosaurs PaleoBios tion, where the bone is broken away. This center is 23 mm behind the orbit. The left postfrontal is 16 mm wide where it forms the posterior border of the orbit. Behind this it narrows suddenly to 11 mm, and then broadens gradually to 16 mm. The supraorbital canal enters rather abruptly onto the postfrontal, and disappears at 45 mm behind the orbit. The right postfrontal has a nearly uniform width of 20 mm. Its canal leaves the frontal and enters the postfrontal at a very slight angle, from 17 to 24 mm behind the orbit. It can be traced to 44 mm behind the orbit, where it fades out. Postorbital: The left postorbital is 63 mm long at its midline, reaching back to the jugal flexure, above the jugal. Its center of ossification is 40 mm behind the orbit. The anterior edge forms the upturned posterior border of the orbit, and has six pits with diameters of 3 mm. In front of the center, the grooves and ridges become very elongated. The anterior grooves are very close together, while those running posteriorly are farther apart and diverge slightly. The postorbital is 14 mm broad anteriorly and 25 mm posteriorly. Its posterior end is concave around the front of the supratemporal, and it sends a slender prong back 78 mm from the orbit, between the supratemporal and the postfrontal. It also sends a ventral prong 70 mm from the orbit above the jugal. Jugal: This bone begins 2 mm in front of the orbit, between the prefrontal and the maxilla, where it is 1 mm high. It becomes 5 mm high at 20 mm back, where it forms the posteroventral quarter of the orbital border, becoming 10 mm high. The jugal continues posteriorly below the postorbital, the suture fused, but traceable by an angular unconformity of the grain of the two bones. It extends to 82 mm behind the orbit where it ends outside the quadratojugal. The jugal is deeply incised by the infraorbital canal for its entire length of 100 mm to where the canal sends a branch up and back, then forward along the suture with the quadratojugal and supratemporal. Quadratojugal: This is low and triangular, beginning as a point below and inside the jugal, and enlarging to a height of 11 mm at 35 mm back, where it is broken and lost. Its anterior dorsal tip forms the floor of the upturning infraorbital canal. Squamosal: This bone is represented only by its anterior tip, which extends anteriorly between the supratemporal and the quadratojugal. Supratemporal: This begins with a broad doubly convex anterior end butting into the back of the postorbital and above the posterior tip of the jugal. Its lower edge lies above the anterior tip of the squamosal. It is 26 mm broad above the jugal flexure. The lower convex end carries the jugal canal forward and up to continue back as the supratemporal canal. The canal also doubles forward and curves sharply back in the supratemporal, as a second canal above and parallel to the first. Each is 10 mm wide. I am not sure of the distinction between postorbital and supratemporal but the suture seems to run around the front of the canal as shown. Sculpturing: The sculpturing of the roof is deep and elongate. There are prenarial, nasoorbital, and orbitopineal zones of intense growth. The downcurved tip of the premaxilla has 24 small pits on each side, facing anteroventrally and radially. A band of unsculptured bone 4.3 mm high runs around the ventral tip of the snout. On the left side this band goes back 19.3 mm. On the right side the band can be followed for 13.5 mm to where it suddenly ends in an upcurved edge of the bone. At 12 mm above the midline tip is a vague horizontal groove that is probably a poorly developed anterior commissure. Behind this is a raised patch of sculpture 8 mm wide that narrows posteriorly into two ridges between the supraorbital canals. The premaxilla does not show a center of ossification as most of its sculpture behind the tip consists of long grooves and ridges. The sculpture of the other bones is normal for lonchorhynchines. Sensory canals: The anterior commissure, as just noted, is a poorly developed groove curving back parallel to the lower edge of the tip of the snout to about 5 mm. Its floor is not smooth, but is crossed by four low ridges. At each end it joins the enlarged tip of the supraorbital canal. The supraorbital canal begins as a ventro- laterally expanded opening that narrows posteriorly to 2.5 mm as it enters between two prominent ridges on the tip of the snout. It then narrows posteriorly to 2 mm and is about 1 mm deep as it runs straight back on the premaxilla and the nasal. Its inner ridge is but 1.5 mm from the midline anteriorly, and is only 2.7 mm from the midline between the nares. Behind the nares it curves laterally and is 10 mm from the midline at 30 mm behind the nares. Here it is just lateral to the center of ossification of the nasal. It then enters the tip of the prefrontal and continues back parallel to the midline for 50 mm to where it PaleoBios Lonchorhynchine trematosaurs Page 15 bends medially between the orbits and enters the frontal, where it comes to 2 mm from the midline. Here its inner wall is a very sharp ridge 1.5 mm wide and 2 mm high. The canal then curves laterally near the center of ossification of the postfrontal where it enters the postfrontal. Its inner wall is very low here and is 8.5 mm from the midline. The canal becomes very shallow and broad and is lost at about 40 mm behind the orbit, where a sharp ridge appears at the center of the end of the canal. On the left side there is a shallow connection with the infraorbital canal from 23 to 28 mm in front of the orbit. On the right this connection is broken by a sharp ridge above the lower canal, but the supraorbital canal sends off an anterolateral branch at 15 mm in front of the orbit. The infraorbital canal is first evident on the side of the premaxilla, 52 mm behind the tip of the snout. It runs back along the ventral edge of the premaxilla for another 56 mm, varying in width from 2 to 3 mm. It then enters the maxilla, and at 70 mm from the front of the maxilla it turns gently dorsally to rise to 9 mm at 25 mm in front of the orbit. Here, on the left side only, it appears to have a broad (5 mm) junction with the supraorbital canal, although this might be an artifact of preparation. It then curves down gently below the orbit, entering the tip of the jugal and descending gradually to 2 mm above the ventral edge of the jugal at 22 mm behind the orbit. At 75 mm behind the orbit it curves posterodorsally and broadens as it rises, extending to 87 mm behind the orbit between the supratemporal and quadratojugal- At 81 mm from the orbit it curves sharply forward around the back of the jugal in a jugal flexure to 64 mm from the orbit at 16 mm above the ventral border. Here it curves back in a broad arc to continue posterodorsally and slightly medially as a broad (4 mm) temporal canal, broken away at 87 mm behind the orbit. At the anteromedial apex of the arc, 20 mm above the ventral edge and 67 mm behind the orbit, it gives off a medial branch. This arches forward to 64 mm from the orbit, then curves sharply back parallel to the outer canal. It is 3 mm wide anteriorly and 5 mm at its broken posterior end. These two branches thus form a double temporal canal. Ventral view: (Fig. 4B) Each vomer, palatine, and ectopterygoid bears a tusk pit depression. In each pit are two tusk alveoli, one behind the other, which housed alternating tusks. The most prominent feature of the front of the palate is a high midline keel, formed by the premaxillae and the vomers. It begins 16 mm from the tip, behind the 6 mm excavation behind the semi- circular inner alveolar border. Here, at the front of the large left foramen, the keel is 9 mm wide. It narrows to 4 mm at 27 mm from the tip, then widens gradually to 13 mm at 172 mm, where it broadens around the vomerine tusk pits to 19 mm at 200 mm from the tip. This keel separates into narrow double keels, from 25 mm to 65 mm, and at 36 mm from the tip the slender tip of the vomer inserts between the premaxillae. At 117 mm the suture between the premaxillae and the vomers runs laterally into the alveolar area, and from here on back the vomers form the entire keel. Anteriorly, the side walls of the keel are concave vertically, so that at 130 mm from the tip the ventral breadth is 10 mm, and it narrows dorsally to an irregular 5 mm at the level of the alveoli. The alveoli thus extend into the keel, giving its dorsal edge a wavy antero-posterior outline. At 70 mm from the tip a cross section along a break shows a vertically oval cavity (Fig. 5A). Premaxillae: The two bones form a semicircular tip, its anterior edge curving to a height of 10 mm. At 9 mm from the tip a semicircular ventral ridge forms the inner wall of the four anterior alveoli. This ridge projects 1.5 mm ventrally, and behind it is a midline excavation 6x7 mm. The left premaxilla has a large foramen from 17 to 25 mm from the tip. Its side wall is broken away, but its inner wall is 8 mm high and has a slight concavity directed anterodorsally. This large foramen occupies an area corresponding to that of the small right foramen plus part of the small right fifth alveolus. The alveoli are all obscured by matrix and displaced teeth. The right premaxilla has two small fora- mina that lead dorsolaterally into the dorsal base of the semicircular ridge. This ridge continues posterolateral^ to form the posteromedial wall of the fourth alveolus. This wall is also the anterolateral wall of the foramen for the dentary tusk. The lateral wall of the premaxilla is raised from the second to the sixth alveolus, 5 mm dorsal to the midline keel. It then drops to an irregular 3 to 5 mm above the level of the keel. There are 23 rather large and irregular alveoli, their inner borders indenting the keel, their lateral borders sharp and scalloped. The anterior four radiate inward from the outer edge of the tip of the snout, decreasing rapidly in size. The first is 6 mm wide at the front, and its anterior wall is high. Its posterior Page 16 Lonchorhynchine trematosaurs PaleoBios wall is low and forms part of the semicircular ventral ridge. The second alveolus is 6.5 mm posteromedially, while its anterolateral edge is only 3 mm. Its anterolateral wall is very low, and slants posteriorly, so the tooth must have pointed slightly backward. The third alveolus is very shallow, its anterolateral edge much smaller than that of the second. It is triangular, with a rounded anterolateral edge, its apex posteromedial, ending against the semicircular ridge. The fourth alveolus is an equilateral triangle with bulging sides, its lateral edge very low. Its inner angle is against the back of the semicircular ridge. Its posteromedial border is the continuation of the ridge that separates the alveolus from the small foramen for the dentary tusk. The fifth alveolus is 6.7 mm behind the fourth, separated by a small foramen. A rounded bar 2.5 mm wide runs posterolateral^ and separates the alveolus from the foramen. This alveolus is 4.5 mm in diameter, with an extremely low posterior wall, and a wide lateral border. The sixth is 6 x 9 mm, with a very thin lateral wall, and is typical of the following alveoli, except for its extremely low anterior wall. Its roof is concave, as are those of all the others. Many of these roofs show radiating patterns, especially peripherally, matching the infoldings of the dentine of the teeth. The following alveoli are each about 6 mm broad, and vary in length from 6 to 9 mm. Vomers: These begin 36 mm from the tip of the snout as an extremely slender sliver, wedged between ridges of the premaxillae, the whole forming the midline keel. Here the vomer is a single bone, no midline suture is visible. This suture becomes evident at 73 mm from the tip of the snout. From 143 mm to 180 mm (left), 186 mm (right), each vomer becomes grooved longitudi- nally near the midline, thus forming a small secondary midline keel within the main keel. The grooves end at the tusk pits, but the secondary keel is 235 mm long, diminishing posteriorly, to end opposite the back of the choanae. The width across the two vomers gradually broadens from 11 mm at 117 mm from their tip, to 13 mm at 172 mm. They are broadest (19 mm) at 200 mm, and they narrow to 16 mm at 212 mm, where each extends slightly along the lateral wall of the choana, above the maxilla. The vomers are compressed by the choanae to a combined width of 8.4 mm. This width increases dorsally, especially posterodorsally, as the vomer curves dorsolaterally to close the choana posteriorly and form the roof of the postchoanal groove. A sharp lateral keel, more prominent on the right, develops at 167 mm from the snout tip and curves out around the tusk pits, dying out at the inner front of the choana, lateral to the row soc 5cm Figure 5. Cosgriffius campi, type, UCMP 112135. A, cross-section of rostrum 67 mm from tip; B, cross- section through choanae. ch-choana; mx-maxilla; px- premaxilla; soc-supraorbital canal; v-vomer. of small vomerine teeth. The vomerine tusk pits begin 178 mm (left) and 177 mm (right) from the snout tip, and each is 16 mm long. Each anterior tusk alveolus is 9 mm long and 5 mm wide, the two separated by the 1 mm vomerine midline keel. The posterior alveoli are each 8 mm long, the left situated 1 mm behind the right. The left is 5.5 mm wide, the right 5 mm, with the vomerine lateral keel crushed slightly in below it. The radially patterned remnant of a tusk base is irregularly placed in each alveolus. The left vomer widens gradually to 10 mm at the front of the interptcrygoid vacuity. The right vomer is 11 mm wide here, and it divides, sending a slender process back 80 mm along the ventrolateral edge of the parasphenoid, and a lateral process back 50 mm along the outer wall of the vacuity, where it ends between the ectopterygoid and the anterior tip of the pterygoid. The vomer thus forms the entire anterior wall of the vacuity (Fig. 4B). The two PaleoBios Lonchorhynchine trematosaurs Page 17 vomers begin to separate 10 mm behind the front of the vacuity, revealing the cultriform process. An irregular row of small teeth, each about 1.5 mm in diameter at the base, begins immediately behind each tusk pit and extends between the choanae along each side of the midline. The right row begins 15.8 mm in front of the choana. Its anterior alveolus is displaced medially, causing a sinistral flexion of the midline. The second is broken away, so only its base remains. Behind this is a single tooth, then an empty alveolus. Next is a row of 7 teeth, 15.7 mm long, ending 3.7 mm behind the front of the choana. The teeth in this row are inclined medially, possibly by crushing. Behind this row is an open groove 6.1 mm long for two teeth, as shown by the indented border. This is followed by a single tooth, then two broken bases, and a single tooth with its posterior edge 26 mm behind the front of the choana. The entire row is thus 42 mm long and includes about seventeen teeth. Palatines: The left palatine has two anterior prongs. The lateral prong begins at the outer roof of the choana as a thin vertical plate against the maxilla, but its anterior tip is missing. This prong widens to 3 mm at the back of the post- choanal groove and has a rounded ventral surface. The medial prong extends forward against the vomer, thus forming the posterior wall of the choana and the roof of the post- choanal groove. The palatine broadens to 8 mm at the front of the tusk pit, and 10 mm at the center of the pit. Its suture with the vomer continues posteriorly parasagittally for 28 mm to the meeting with the ectopterygoid. The suture with the maxilla curves gently laterally around the tusk basin and continues posteriorly in a straight line to the ectopterygoid. The post- choanal groove continues back in the palatine and vomer to end at the inner front of the anterior palatine tusk pit. The bone immediately in front of the pit bears small alveoli but no tooth bases are evident. The tusk pit begins 45 mm in front of the vacuity and is 14 mm long. It contains two alveoli, the anterior being 6 by 7 mm, the posterior 7 by 7 mm, its anterolateral corner extended. It is followed by a 2 mm wall. Behind this is a row of three alveoli 18 mm long, the first two 5 mm in diameter, the last 6 by 4 mm, ending against a prong of the ectopterygoid. The suture with the ectopterygoid is deeply interdigitating, with three prongs projecting into the ectoptery- goid. The lateral prong is 6 mm long, the two inner prongs slightly shorter. The medial prong ends 10 mm in front of the interpterygoid vacuity. The right palatine differs from the left in lacking an anterior medial prong. It begins at the center of the lateral wall of the choana as a thin vertical plate against the maxilla. The suture with the maxilla runs 4 mm posteriorly along the wall of the choana, then turns abruptly laterally as the palatine increases in width posteriorly, becoming a rounded bar 5 mm wide that forms the lateral wall of the postchoanal groove. At 10 mm behind the choana the suture with the vomer curves abruptly medially as the palatine broadens to 8 mm to accommodate the tusk basin. The postchoanal groove disappears at the front of the tusk pit. This pit begins 48 mm in front of the interpterygoid vacuity and is 14 mm long. The front tusk alveolus is 7 by 4.5 mm, the rear 5 by 6 mm. Both are shallow, the rear one extreme- ly so. At 4 mm behind the pit there follows a series of nine small teeth, 2 to 3 mm in diameter, ending 11 mm in front of the vacuity. The palatine-ectopterygoid suture is deeply inter- digitating, as on the left side. Ectopterygoids: The left ectopterygoid is 9 mm wide. Its alveoli begin 6 mm in front of the interpterygoid vacuity, and are quite small, the five anterior ones occupying a row 12 mm long and 2 mm wide. These are followed by a small tusk pit 10 mm long that held two tusks, each 3.5 mm wide. Behind this, the row of small alveoli continues on back to the break. The right ectopterygoid is 9 mm wide anteriorly, and continues back to 24 mm behind the front of the vacuity. Here the suture with the vomer curves in and up into the vacuity, and the ectopterygoid takes over the rim of the vacuity. The bone is broken off at a length of 70 mm, and there is no indication of the pterygoid. The tusk pit is 11 mm long and begins at a level 1 mm behind the front of the vacuity. There are no alveoli in front of the tusk pit, and the entire tooth row is 78 mm long back to the break. Cultriform process: This is quite narrow, and is eroded along its ventral surface. It and the sphenethmoid dip to the left as the skull roof was crushed slightly to the right. The process is hidden for 10 mm anteriorly by the conjoined vomers and is broken off 124 mm behind the front of the left interpterygoid vacuity. At 77 mm behind the front of the left vacuity the process is only 1 mm wide ventrally. Sphenethmoid: This lies above the cultriform process as far forward as the front of the interpterygoid vacuities, where it is covered by Page 18 Lonchorhynchine trematosaurs PaleoBios inaccessible matrix. It forms an open trough with sides 3 mm high anteriorly, its dorsal edge wavy. At about 53 mm back the sphenethmoid rises to 5 mm, its dorsal edge straightens, and it continues back to the break. The section at the break (Fig. 5) shows a solid base 5 mm high, above which its walls spread to form a trough 4 mm high, which opens to a dorsal width of 9 mm. Palatal openings: The foramen in the premaxilla for the left dentary tusk has lost its lateral wall. Its inner wall is 8 mm long dorsally, 10 mm ventrally, slightly concave anteroposteriorly and dips slightly medially, so a shallow concavity runs posteroventromedially. Its posterior wall is vertical. The right side has a much smaller foramen for the dentary tusk. The foramen is 4 by 2 mm and begins 17.2 mm behind the tip of the snout. It is oriented posterolaterally and opens onto the dorsal surface, the front of the dorsal opening situated 2.5 mm posterolateral to the front of the ventral opening. Choanae: These begin 210 mm behind the tip of the snout and open posteroventrally. Each is 11 mm long dorsally, and 2 mm wide, widening to 4 mm ventrally. The inner wall of the choana is 4 mm high, is vertical anteriorly, and dips 135* medially posteriorly. The lateral wall is formed by the vomer, the rounded medial edge of the maxilla, and the anterior tip of the palatine. The choana continues posteriorly as a shallowing postchoanal groove in the area between the vomer and the palatine (left), or in the vomer (right). Interpterygoid vacuities: These are sharply pointed anteriorly, and are asymmetric because of the crushing. The left extends 3 mm in front of the right. They are broken off posteriorly, but the left is at least 130 mm long. It is 28 mm wide, the right 23 mm, so each was probably 25 mm wide originally. Teeth: The teeth were weakly attached and set in shallow alveoli, but there does not seem to have been a finished base to a tooth. Each grows into the bone, and the bone evidently sent prongs up into the tooth, so that when shed or broken out, the bone often retains a radiating pattern that corresponds to the infoldings of the dentine. The anterior teeth are the largest, with lengths up to 20 mm and diameters of 6.4 mm. A dislocated tooth from near the front of the left premaxilla is 6.4 mm in diameter, and has fifteen vertical striations in 4.1 mm. Subfamily LONCHORHYNCHINAE, Indeterminate Species A 1909G Trematosaurus kannemeyeri Broom, p. 270. 1919B Trematosaurus kannemeyeri, Watson, p. 41. 1920F Aphaneramma kannemeyeri Huene, p. 452. 1925A Trematosaurus kannemeyeri, Haughton, p. 249. 1937C Trematosaurus kannemeyeri, Broili & Schroeder, p. 115. 1947C Trematosaurus kannemeyeri, Romer, p. 176. 1958 Rhytidosteus, Kitching, p. 81. 1965 Trematosauridae, not Trematosaurus, Cosgriff, p. 85. 1972 Dubious taxon, Cosgriff & Garbutt, p. 6. Type: SAM 1329, the orbital region of a fairly large skull. Horizon: Cynognathus Zone. Age: PSpathian. Locality: Orange Free State, South Africa. Description: This skull fragment has the very slender shape of this subfamily, with widely separated orbits, and the frontal excluded from the orbit. It is of a large individual, with an interorbital width about twice that of Cosgriffius campi. Diagnostic parts are lacking. Species B 1966A Ifasaurus elongatus Lehman, p. 132, fig. 8, pi. 5B. 1972 Ifasaurus elongatus, Cosgriff & Garbutt, p. 7. 1974 Ifasaurus elongatus, Lehman, p. 166. 1985 Ifasaurus elongatus, Warren & Black, p. 308. 1987 Wantzosaurus elongatus Hammer, p. 75. The species Ifasaurus elongatus was based on the orbital and postorbital part of the interior of a skull roof from the Middle Sakemena Beds at Anjavimalai, Madagascar. This is of Dienerian age and is certainly trematosaurid, but is based on inadequate material and is indeterminate, as indicated by previous investigators. Lehman thought that it had an elongate rostrum, but not so long as in Aphaneramma. The face narrows rapidly in front of the orbits, and this whole region, including the orbitopineal triangle, is similar to that of Aphaneramma rostratum, so it is possible that this is a lonchorhynchine. PaleoBios Lonchorhynchine trematosaurs Page 19 Species C 1971A Labyrinthodonte a museau allonge Lehman, p. 87, pi. 5B, C. This is the tip of the snout, and a fragment behind it (Fig. 6A, B), MNHN Zar 3,4. It was found near Zarzaitine about 5 km south of Gour Laod, easternmost Algeria. The horizon is the base of the Zarzaitine series of Spathian or Anisian age. The specimens were sent to Dr. Michael Morales of the Museum of Northern Arizona, Flagstaff, for study, and he kindly forwarded them to me for description in this paper. They are the tip of the mandible and the choanal region of the rostrum. They are labyrinthodont and their slenderness places them in the subfamily Lonchorhynchinae. The mandibular fragment includes the anterior tip, and is the first lonchorhynchine to have this part preserved. The fossil is 66 mm long and would be 44 mm broad at the expanded tip, except that the left anterolateral edge is missing. The breadth decreases to 24 mm at 45 mm behind the tip. In dorsal view there is a semicircular row of alveoli around the anterior periphery. Behind the three anterior alveoli is a depression 9 mm long, with a transverse groove 2 mm wide across its back, the groove suddenly angling anterolaterally at 5 mm from the midline. On the left side the tusk pit is 16 mm long. The tusk alveolus in this pit is probably the posterior one; it is 11 by 9 mm and 8.5 mm deep, extending down to within 2 mm of the ventral surface. Behind the pit is a 3 mm alveolus, followed by two deep empty alveoli 5 by 5 mm and 7 by 6 mm. Behind these is an eroded alveolus 8 mm wide. On the right side the bone curves dorsally, and at 11 mm from the tip forms the anterior edge of the large tusk pit. This pit is triangular, with rounded apices. It is 21 mm long and 18 mm posterolaterally. Its anterolateral side is slightly convex. Of the two tusks in this pit, the anterior is 10 mm in diameter, and it has thirty- one invaginations of the dentine in half of its circumference, with twenty-two in 10 mm. This is on a broken surface near the base, indi- cated not by the invaginations, but by the dentine blocks, each with a pit (or a vertical canal) near its periphery. The pulp cavity is 0.5 x 1 mm. The second tusk is 11 x 12 mm, its broken top 8x7 mm, with a 2 mm pulp cavity. It has thirty invagi- nations in half of its circumference, with the ¦.¦ft:tia ' '*¦*«'4 ' *•¦ ? | r<2\ 5cm Figure 6. Lonchorhynchinae, incertae sedis, type, MNHN Zar 3. Tip of chin in dorsal view (A) and ventral view (B). Referred choanal fragment MNHN Zar 4 in dorsal view (C) and ventral view (D). same spacing as in the first tusk, seen at 5.5 mm above the base. In posterior view the bone is 13 mm high and has a rounded dorsal ridge 8 mm broad. This ridge is indented along its midline and has a groove along each side between the ridge and the alveolar border. On the right side at 3.5 mm above the base a canal runs forward into the bone. The opening of this canal is 2 x 3 mm and angles 135 ° laterally. There is no corresponding canal on the left side. The choanal part of the rostrum, Zar 4 (Fig. 6C,D), was associated with the type, and probably belongs to the same individual. It is 64 mm long in the midline, 23 mm wide anteriorly and 33 mm posteriorly. In dorsal view it is flat across the top, its sides bending down lateral to the central patch of sensory pits. These pits are irregularly arranged and the patch narrows posteriorly from 16 mm to only 4 mm between the nares. The anterior pits are from 2 to 2.5 mm wide, and they decrease to 0.5 mm posteriorly. They lie in very shallow grooves. The suture between the premaxilla and nasal is clear on the left, beginning 5 mm from the midline and slanting in to 2 mm at 20 mm back. It then curves Page 20 Lonchorhynchine trematosaurs PalcoBios out to enter the naris 6 mm from the midline at 37 mm from the anterior break. The right naris begins 5 mm behind the left. In ventral view the prominent feature is a keel that is 8 mm wide between the choanae. This extends to 6 mm below the roof anteriorly, increasing gradually to 12 mm posteriorly. It has a cylindrical base between the nares, and above this it pinches to a vertical bar 5 mm high and 4 mm wide. The right choanal roof begins at the front tip and rises gently posterodorsally to the naris. The choana is 5 mm wide dorsally and 9 mm ventrally. Its roof, from 7 mm back, is covered by a thin lamina that seems to be a separate element, possibly the septomaxilla. Species D 1985 Trematosauridae, incertae sedis Warren, p. 293. This specimen is the vomerine region of a skull (QM F12272) and is about 19 mm wide at the vomerine tusks, considerably smaller than Cosgriffius campi at 31 mm. It is of an extremely slender-snouted labyrinthodont, and is evidently the first lonchorhynchine from eastern Australia. It comes from the Arcadia Formation of the Rewan Group, of Griesbachian-Dienerian age. The locality is 72 km northwest of Rolleston, southeast Queensland. Species E the Satpura Basin, easternmost Madhya Pra- desh, India. These have been given the catalogue number ISI A33, and were kindly forwarded to me for description. This specimen was mentioned by Chatterjee and Hotton (1986:163) as Denwa- saurus, a nomen nudum. The anterior piece (Fig. 7A) is 32 mm long in the midline dorsally, and 27.5 mm ventrally. It is uncrushed and is relatively quite massive, 27.2 mm broad near the tip, and narrowing to 24.9 mm at 20 mm back. In dorsal view it has a slight anterior bulge, which is 18 mm long in the midline and is slightly asymmetrical, reaching farther anterolaterally on the right. It continues back on each side as a pointed projection, the two 1985 Trematosauridae, incertae sedis Warren, p. 294, fig.l. This consists of three pieces of the rostrum of a labyrinthodont (QMF12271) from the Glenidial Formation of the Clematis group of Scythian age. The locality is the northeastern side of the Carnarvon Range, near Moolayember Dip, southeastern Queensland, 25° 10'S; 148* 31'E. As indicated by Warren, these fragments are from a very slender snouted labyrinthodont, and this extreme slenderness is known only in the Lonchorhynchinae. There is a deep, rounded midline keel, and the sensory canals are deep, but the material is indeterminate. Species F Dr. Sankar Chatterjee collected the tip of the rostrum and the second piece immediately behind it from the Denwa Formation (Anisian Age) of 5cm Figure 7. Lonchorhynchinae, incertae sedis, type, ISI A33. Tip of rostrum in dorsal view (A) and ventral view (B). Referred fragment in dorsal view (C) and ventral view (D). dr-dorsal ridge; vk- ventral keel tips 14.5 mm apart and extending postero- ventrally into the divisions of the supraorbital canals. Behind the bulge is a midline ridge 10 mm wide, its lateral walls descending 5 mm to form the inner walls of the supraorbital canals. Lateral to the canals the surface bends ventrally to form a quarter circle. The entire surface is pitted and each pit has a tiny foramen at its base. The pits open radially. The supraorbital canals branch at 21 mm from the tip, the medial branches running 45° PaleoBios Lonchorhynchine trematosaurs Page 21 anteromedially to meet in the midline at 90* and form the posterior wall of the anterior bulge. The lateral branches extend anteriorly and somewhat ventrally along the tip of the snout, curving down at its anterior end to 6 mm above the edge of the bone. There is no indication of an anterior com- missure around the ventral tip of the snout. Possibly the junction of the two medial branches of the supraorbital canals, behind the anterior bulge, served this function. In ventral view, (Fig. 7B) there is an anterior transverse row of five alveoli, each with a tooth base. Behind this is a semicircular ventral ridge, highest at the midline, and curving to 6 mm from the midline to end inside the third alveolus on the left side, the fourth on the right. Behind this is a crescentic depression 4 mm wide in front that curves back inside the alveoli, and becomes only 1 mm wide at the fourth alveolus. It continues back to the third alveolus, medial to the alveoli, at the top of the midline keel, and is pierced by very tiny irregular foramina. The keel begins 10 mm from the tip of the snout, behind the crescentic depression, becomes 9 mm wide at 17 mm and pinches to 6 mm at 24 mm. It then broadens to 8 mm at the posterior break. The tooth row is 6 mm anteroposteriorly at the front, and increases to 7 mm transversely at the broken end. The tooth bases extend deeply into the bone, nearly to the dorsal surface, and they all point downward and outward. All have been broken off, and there seems to have been a great deal of dissolution of the dentine. The left anterior tooth base is oriented slightly to the left, and is but 2.5 mm wide. The second is triangular, 5 mm across its rounded anterolateral edge, and narrows posteromedially. The third is 3 by 6.4 mm. The fourth has a smaller, circular base, 4.5 mm across, and is against the lateral wall. The fifth is similar. The sixth is 4.7 mm anteroposteriorly, and the alveolus is eroded to a depth of 3.5 mm where the radial folds of the dentine are still present. The seventh is the longest at 5.4 mm. On the right side the first tooth is nearly in the midline, and the second is a pinched remnant 1.5 mm wide laterally, and decreasing medially. The third is a triangular base at the antero- lateral edge of the bone, its diameter 4.7 mm. The fourth is but 3 mm long. The fifth is 4 mm wide. The sixth is 3.2 mm wide. The seventh is a circular alveolus 4.4 mm wide and 4.3 mm deep. The eighth retains the base of a tooth 5.8 mm long, broken off just below the edge of the bone. The medial edge of this tooth slants 45* ventrolaterally and has vertical grooves about 0.1 mm apart, but irregularly spaced. Each ridge between the grooves has, 0.1 mm from its periphery, a very small vertical canal. These canals are thus arranged in an ellipse, and there are evidently twenty-four on this tooth. The canals on other teeth become larger toward the base and are formed as cavities at the convergences of the dentinal folds, just within the peripheral mass. The dentinal folds continue in a meandering pattern nearly to the small central pulp cavity. The very slender rostrum is indicative of a lonchorhynchine trematosaur. Species G Dr. Donald Baird kindly sent me casts of several specimens from the Lower Economy fault- block, now included in the Wolfville Formation, of Nova Scotia. He wrote (21 Feb. 1991) that the specimens are now in the Yale Peabody Museum collections and are numbered as follows: PU 21690, juvenile skull table. PU 21691, midsection of left mandible. PU 21692, posterior end of left mandible. PU 23686, left premaxilla fragment with narial margin. Dr. Baird further wrote that the specimens are practically indistinguishable from a cast of Cosgriffius campi. I therefore include them in the Lonchorhynchinae, and they are the first of the subfamily to be reported from Nova Scotia. SUMMARY The slender-skulled Triassic labyrinthodonts are all included in the Family Trematosauridae, which is divided into the Trematosaurinae, with a moderately slender rostrum, and the Loncho- rhynchinae, with an extremely slender and greatly elongate rostrum. The Lonchorhynchinae includes the previously known genera Aphane- ramma, Erythrobatrachus, Gonioglyptus, ?Ifa- saurus, and Wantzosaurus, plus the new genus Cosgriffius, and the generically indeterminate specimens described in this paper, three of which are new. Stratigraphic occurrences of Lonchorhynchinae are shown in Fig. 8. Measure- ments of taxa discussed in the text are in the Appendix. Page 22 Lonchorhynchine trematosaurs PaleoBios Lonchorhynchinae, 1.5. Nova Scoiia Lonchorhynchinae, i.s. India z Cosgriffius campi, THIA Arizona "Trematosaurus " kannemeyeri, SPA S. Africa Aplianeramma rosirauim. Z Spitsbergen [ITHI/\ Erythrobatrachus noonkanbahensis, W. Australia is Lonchorhynchinae, i.s. DISS Queensland Lonchorhynchinae, i.s. < Madagascar £ Aphaneramtna kokeni, Pakistan Sp. A., Madagascar —. Z Aplianeramma sp., '¦$¦ 02 Madagascar 3 DIENI Lonchorhynchinae, i.s. Queensland Lonchorhynchinae, i.s. Madagascar Lonchorhynchinae, i.s. T Queensland :hiai < Gonioglyptus longirostris GRIESB India Figure 8. Stratigraphic occurrences of Lonchorhynchinae ACKNOWLEDGMENTS First of all, I thank Franz-Josef Lindeman of the Paleontologisk Museum, Oslo, Norway, for his many suggestions, and especially for the fine cast of the new specimen of Aphaneramtna rostra- turn that he collected in Spitsbergen. The anony- mous reviewers of the paper made valuable suggestions, while the careful review by M. A. Shishkin of the Paleontological Institute, U.S.S.R. Academy of Sciences improved it greatly, and I am deeply indebted to him. Figure 6 was drawn by the Museum Staff Artist Patricia Lufkin; the others were drawn by the author. Dorothy DeMars kindly retyped the manuscript. This is contribution No. 1562 from the University of California Museum of Paleontology. REFERENCES Alberti, F. 1864. Ueberblick uber die Trias, mit Berucksichtigung ihres Vorkommens in den Alpen. Stuttgart: Cottas. xx + 1-353,7 pis. Broili, F. and J. Schroeder. 1937. Beobachtungen an Wirbeltieren der Karrooformation. xxvii. Ueber einen Capitosauriden aus der Cynognathus-Zone. Sitz.-Ber. Bayr. Akad. Wiss. Munchen. 1937:(2), 97-117,9 figs. Broom, R. 1909. Notice of some New South African fossil amphibians and reptiles. S. Afr. Mus., Ann., 7:270-278.1 fig. Bystrow, A. 1935. Morphologische Untersuchungen der Deckknochen des Schadels der Wirbeltiere. I Mitteilung. Schadel der Stegocephalen. Acta Zool., 16:65-141,36 figs. Chatterjee, S. and N. Hotton III. 1986. The paleoposition of India. Jour. Southeast Asian Earth Sciences, 1(3):145-189. Cosgriff, J. 1965. A new genus of Temnospondyli from the Triassic of Western Australia. Roy. Soc. W. Australia, 48(3)65-90,13 figs. — 1969. Blinasaurus, a brachyopid genus from Western Australia and New South Wales. Roy. Soc. West. Austral., Jour., 52:65-88,11 figs. — 1984. The temnospondyl labyrinthodonts of the earliest Triassic. Jour. Vert. Pal., 4(l):30-46, 46 figs. — and N. Carbutt. 1972. Erythrobatrachus noon- kanbahensis, a trematosaurid species from the Blina shale. Roy. Soc. West. Austral., Jour., 55(1 ):5- 13,7 figs. Efremov, I. 1933. Ueber die Labyrinthodonten der U.d.S.S.R. II. Permische Labyrinthodonten des frueheren Gouvernments Wjatka. Trav. Inst. Pal., Acad. Sci. URSS., 2:117-163,20 figs., 5 pis. — 1940. (Preliminary description of the new Permian and Triassic Tetrapoda from USSR.) Trav. Inst. Pal., Acad. Sci. URSS., 10(2):1-140, 25 figs, 14 pis. (Russian). Hammer, W. 1987. Paleoecology and Phylogeny of the Trematosauridae. pp. 73-83, 5 figs., in: McKenzie, 1987. Haughton, S. 1925A. Investigations in South African fossil reptiles and Amphibia. 13. Descriptive catalogue of the Amphibia of the Karroo System. Ann. S. African Mus., 22:227-261,19 figs. Hellrung, H. 1987. Revision von Hyperokynodon keuperinus Plieninger (Amphibia: Temnospon- dyli) aus dem Schilfsandstein von Heilbronn (Baden-Wiirttemberg). Stuttgarter Beitrage zur Naturkunde, Ser. B (Geologie und Palaontologie), 136:1-28,11 figs. Huene, F. 1920. Systematische und genetische Be- trachtungen uber die Stegocephalen. Zeit. Indukt. Abstamm. Vererbungsl., 13:209-212. — 1920. Gonioglyptus, ein alttriassischer Stego- ccphale aus Indien. Acta Zool. 1:433-464.14 figs. PaleoBios Lonchorhynchine trematosaurs Page 23 Huxley, T. 1865. Indian pre-Tertiary Vertebrata. On a collection of vertebrate fossils from the Panchet rocks, Ranigunj, Bengal. Pal. Indica. (4)1:3-24, 6 figs., 6 pis. Kitching, J. 1958. A new small stereospondylous labyrinthodont from the Triassic beds of South Africa. Pal. Africana, 5:67-82,4 figs. Lehman, J. 1961. Les stegocephales du Trias de Madagascar. Ann. Pal., 47:109-154, 22 figs., 20 pis. — 1966. Nouvcaux stegocephales de Madagascar. Ann. Pal, 52:117-140,9 figs., 3 pis. — 1971. Nouveaux vertebres fossiles de la serie de Zarzaitine. Ann. Pal., Vertebr.,57(l):71-113,11 figs., 10 pis. — 1974. A propos des Trematosaures de Madagascar. C. R. Soc. Geol., France, 6:166-167. Lydekker, R. 1879. Indian Pre-Tertiary Vertebrata. Fossil Reptilia and Batrachia. Pal. Indica, (4)1(3):1- 36,6 pis. — 1881. Note on some Gondwana Vertebrates. Rec. Geol. Surv. India, 14:174-178. — 1882. On some Gondwana Labyrinthodonts. Rec. Geol. Surv. Ind., 15:24-28,1 pi. — 1887. The Fossil Vertebrata of India. Rec. Geol. Surv. India, 20:51-80. McKenzie, G., ed. 1987. Gondwana Six: Stratigraphy, Sedimentology, and Paleontology. Amer. Geophys. Union, Geophysical Monogr., 41. Meyer, H. 1842. [Letter on Mesozoic amphibians and reptiles.] Neues Jahrb. Min. Geol. Pal., 1842:301-304. Nilsson, T. 1942. Sassenisaurus, a new genus of Eotriassic stegocephalians from Spitsbergen. Bull. Geol. Inst. Univ. Uppsala, 30:91-102, 7 figs. — 1943. Uber einzige postkraniale Skelett-reste der triassischen Stegocephalen Spitzbergens. Bull. Geol. Inst. Univ. Uppsala, 30:227-272,16 figs, 4 pis. — 1943. On the morphology of the lower jaw of Stegocephalia with special reference to Eotriassic stegocephalians from Spitsbergen. 1 Descriptive part. K. Svenska Vetenskapsakad, Handl., 20:1-46, 25 figs., 9 pis. — 1946. On the genus Peltostega Wiman and the classification of the Triassic Stegocephalians. K. Svenska Vetenskapsakad., Handl., (3)23:3-55, 23 figs. Price, L. 1948. Un anfibia laberintodonti da formacao Pedro de Pogo, Estado do Maranhao. Serv. Geol. Min. Brasil, Bol., 124:1-32,3 pis. Romer, A. 1947. Review of the Labyrinthodontia. Bull. Mus. Comp. Zool. Harvard, 99:3-368, 48 figs. Save-Soderbergh, G. 1935. On the dermal bones of the head in labyrinthodont stegocephalians and primitive Reptilia with special reference to Eotriassic stegocephalians from East Greenland. Meddel. om Groenl., 96(3):1-211,68 figs., 15 pis. — 1936. On the morphology of Triassic Stego- cephalians from Spitsbergen, and the inter- pretation of the andocranium in the Labyrin- thodontia. K. Svenska Vetenskapsakad, Handl., (3)16(1):1-181,70 figs., 22 pis. — 1937. On the dermal skulls of Lyrocephalus, Aphaneramma, and Benthosuchus, Labyrin- thodonts from the Triassic of Spitsbergen and N. Russia. Bull. Geol. Inst. Univ. Uppsala, 27:189-108, 12 figs. Shishkin, M. 1960. A new Triassic trematosaurid Inflectosaurus amplus. Pal. Zhurn., 1960(2):130- 148,10 figs. (Russian). — 1980. (The Luzocephalidae, a new family of Triassic labyrinthodonts). Pal. Zhurn., 1:104-119, 3 figs. (Russian). Stoll, N., et al. (eds.). 1961. International Code of Zoological Nomenclature adopted by the XV International Congress of Zoology. London: Internat. Trust Zool. Nomencl., xviii + 1-176. Tripathi, C, 1969. Fossil labyrinthodonts from the Panchet Series of the Indian Gondwanas. Pal. Indica, N.S., 33:1-43,3 figs., 4 pis. Warren, A. 1985. Two long-snouted temnospondyls (Amphibia, Labyrinthodontia) from the Triassic of Queensland. Alcheringa, 9:293-394, 3 figs. — & T. Black. 1985. A new Rhytidosteid (Amphibia, Labyrinthodontia) from the Early Triassic Arcadia Formation of Queensland, Australia, and a consideration of the relationships of Triassic Temnospondyls. Jour. Vert. Pal., 5(4):303-327, 14 figs. Watson, D. 1919. The structure, evolution and origin of the Amphibia—the "Orders" Rhachitomi and Stereospondyli. Phil. Trans. R. Soc, (B)209:l-72, 2 pis. Welles, S. & J. Cosgriff. 1965. A revision of the labyrinthodont family Capitosauridae and a description of Parotosaurus peabodyi, n. sp. from the Wupatki member of the Moenkopi formation of northern Arizona. Univ. Calif- Publ. Geol. Sci., 54:1-148,48 figs., 1 pi. Wiman, C. 1910. Ein paar Labyrinthodontenreste aus der Trias Spitzbergens. Bull. Geol. Inst. Uppsala, 9:34-40,2 figs., 1 pi. — 1915. Ueber die Stegocephalen aus der Trias Spitzbergens. Bull. Geol. Inst. Uppsala, 13:1-34,10 figs. — 1916. Neue Stegocephalenfunde aus dem Posidonomyaschiefer Spitzbergens. Bull. Geol. Inst. Uppsala, 13:309-222, 4 figs., 2 pis. — 1917. Ueber die Stegocephalen Tertrema und Lonchorhynchus. Bull. Geol. Inst. Uppsala, 14:229- 240,8 figs., 3 pis. Woodward, A. 1904. On two new Labyrinthodont skulls of the genera Capitosaurus and Aphaneramma. Proc. Zool. Soc. Lond., 1904(2):170- 176,1 fig., 2 pis. Page 24 Lonchorhynchine trematosaurs PaleoBios APPENDIX Measurements and indices of lonchorhynchine trematosaurs. Indices are based on the measurement divided by the interorbital distance. Estimates are marked by asterisks. Cosgriffius Gonioglyptus Aphaneramma A. kokeni Wantzosaurus Snout to naris Index Length of naris Index Nasoorbital Index Length of orbit Index Orbitopineal Index Internares Index Interorbits Breadth at back of nares Index Breadth at mid- orbit Index Orbit to back of quadratojugal Index Breadth at quadratojugal Index Snout to choana Index Front of choana to interpterygoid vacuity Index Interpterygoid vacuity in front of orbit Index Erythroba- trachus 168 mm 34 mm 672 260 18 mm 11 mm 72 85 116 mm 46* mm 79 mm 464 329* 609 24 mm 33 mm 39 mm 24 mm 30 mm 96 110 78 188 88 120+ mm 84 mm 140 mm 40 mm 480+ 290 286 308 17 mm 6 mm 68 43 25 mm 14 mm 30 mm 49 mm 13 mm 34 mm 31 mm 11 mm 124 88 55 mm 33 mm 63 mm 84 mm 44 mm 78* mm 220 236 232 172 342 231* 138+ mm 85 mm 165 mm 72 mm 131* mm 652+ 283 337 556 386* 99 mm 170 mm 72 mm 100 mm 330 347 556 291 210 mm 24 mm 70 mm 840 171 538 77 mm 14 mm 34 mm 24 mm 56 mm 308 108 117 183 166 0 mm 54 mm 35 mm 18 mm 14 mm 180 72 141 41 •