56.81(115:76.

4)

Article XVII.- NEW OR LITTLE KNOWN REPTILES AND AMPHIBIA\NS FROM THE PERMIAN (?) OF TEXAS.'
By E. C. CASE.

REPTILIA.
Order COTYLOSAURIA.

SKULL OF Diadectes phtseolinus COPE.
In 1905 (Jour. of Geol., Vol. XII) the author gave a preliminary description of the Diadectidee based on two specimens in the University of Chicago. The first, No. 1075, is a nearly complete skeleton, but the lower jaws are in position and the skull is preserved in such a refractory matrix that only the external features could be made out. The second, No. 1078, is an imperfect skull. The specimen here described 2 is nearly perfect, lacking only the premaxillaries, the articular portion of the left quadrate and the lower jaws. The preservation is especially favorable as the skull is little distorted and the details of structure can all be made out. Upper surface. The posterior portion is marked bv a very coarse sculpture but this does not extend much anterior to the orbits. Beyond this the skull is comparatively smooth except the anterior end of the nasals which are marked by a series of pits arranged in an elongate sculpture. The sutures are so indistinct that the limits of the bones can not be made out. The upper portion of the skull is much narrower than the lower so that the distance across the posterior end of the upper surface is not more than half of that across the articular surfaces of the quadrate bones. The pineal foramen is "enormous" in size. Near the posterior end of the skull there are two deep pits in the position of the superior temporal foramina in the Diapsida; in other specimens I have thought that I detected perforations of the skull at this point but there are certainly none in this skull. The character seems to have been an extremely variable one.
1 The following descriptions were prepared in the course of work conducted under a grant from the Carnegie Institution of Washington and are published in advance of the final paper by arrangement with the Carnegie Institution and the American Museum of Natural History. 2 No. 4839, American Museum of Natural History. From the upper layer of conglomerate, in Archer County, Texas, a few miles south of the town of Dundee, discovered by the author.
163

164

Bulletin American Museum of Natural History. [Vol. XXVIII,

Posterior aspect. The occipital condyle is wider than long and is excavated on the upper border by the foramen inagnum so that it presents a broad heart-shaped outline. The articular face for the atlas is concave and there is a deep pit for the continuation of the notochord. The foramina for the exit of the posterior cranial nerves are not apparent, but a deep pit on

Fig. 1. Diadectes phaseolinus.
nat. size.

No. 4839, A. M. N. H.

Superior aspect of skull.

X i

either side of the basioccipital may have led into them. The sutures between the various bones forming the posterior plate of the skull cannot be made out, but it is apparent that the paroccipital bone was very short and joined, almost at once, the descending processes of the squamosal and the quadrate. I have described ' two perforations in the posterior portion of
1 Jour. of Geol., vol. XIII, p. 127.

The quadrate rises almost vertically.1910. X i nat.K ... . anterior border is attached to and somewhat overlapped by the bones of the postorbital region.'1 ":7 y . Diadectes phaseolinus.. Same skull as in Fig.' . 1.O w .' 11 /I.. . size. the '_:Z. 2.] Case. near the upper end the quadrate turns at a sharp angle and joins a process of the paroccipital.. in this specimen I find only two deep pits in this position. Posterior and lateral aspects. No trace of a columella auris was found and if present . Permian Reptiles and Amphibians of Texas. Fig. 165 the skull in the position of the post-temporal foramina and likened the condition to that of the turtles. The quadrate is quite similar to that of the turtles. Lateral surface of the skull.

Poste- . but one was probably present. in other specimens this region is perfectly smooth. The basioccipital joins the basisphenoid by an indistinct suture. 3. X i nat. [Vol. There is no presphenoid rostrum: a short and very stout process excavated longitudinally by a deep pit on its lower Fig. The basisphenoid ends abruptly in front just anterior to the basipterygoid processes. Same skull as in Fig. The opening of the auditory canal is not apparent but it must have been almost directly opposite the angle described. in the Pelycosauria and Rhyncocephalia generally. there are two foramina in this position for the passage of the internal carotid arteries. surface is attached to the anterior end of the basisphenoid. In other specimens of Diadectes I noticed no foramina in this position. Diadectes phaseolinus. size. XXVIII. Inferior aspect. 1.166 Bulletin American Museum of Natural History. Lower surface of the skull. The lower surface of the basisphenoid is penetrated in the median line by a single foramen. it must have been very short and stout.

these occupy the position of palatines and probably are such. but the vertical plates are curved so as to approximately follow the curve of the outer wall of the skull. In the Pelycosauria this is a single pit. or palatines (Case). Permian Reptiles and Amphibians of Texas. Anteriorly the pterygoids nearly. Posterior to their articulation with the basipterygoid processes of the basisphenoid the pterygoids pass backward as vertical plates to join the quadrate. Between the maxillary plates (Cope). 167 riorlv the basisphenoid widens and the lower edge is continued backward as a sort of flange overhanging the basioccipital. but not quite. Anterior to the otic region there is a very large foramen which probably gave exit to the seventh cranial nerve: it is likely that other nerves escaped through the same opening. the alisphenoids. but their position underlying the pterygoids is very peculiar. The brain-cavity shows the sharp downward bend posterior to the epiphysis mentioned by Cope in his description of a cast of the cranial cavity. but at the anterior end the plates are strictly vertical and touch at the upper edge a single vertical plate in the median line of the skull. In other specimens I have traced an incomplete ectopterygoid but none can be made out in this one. Longittudinal section of the skutll. probably following the line of the basioccipitalbasisphenoid suture. which sheltered the long ante. There are traces of a few elongate. This edge is excavated by a notch in the median line and the notch is divided by a narrow partition which is continued backward as a ridge on the lower surface of the basisphenoid and then divides.1910. conical teeth. attached to the under side of the roof above and uniting with each other and the basisphenoid below. This ends somewhat abruptly behind the point where it joins the basisphenoid. the walls of the brain case are formed by solid descending plates. It remains in contact . The anterior ends of the pterygoids are covered by plates attached to the maxillaries.] Case. The upper edges of the posterior ends flare out to join the pterygoids or palatines. irregularly arranged. probably the opening of the Eustachian canals. passing forward they articulate with the inner edges of the maxillaries or with the palatines.rior prolongation of the brain shown in Cope's cast. but extends upward until it unites with descending plates above. These are vertical plates of considerable height but very narrow on the palatal surface. there is left quite a space in which lie the prevomers. On either side of the dividing ridge there is a deep pit. the parasphenoid (ethmoid). Anterior to this foramen. they mav be traced forward as far as the anterior edge of the orbit. and this leaves no great space between them. meeting in the mid-line below and forming a long cavity open at the anterior end. on the lower surface of the prevomers. meet in the median line. Anterior to the basisphenoid and joining the short anterior rostrum is a single vertical plate.

pt. In the article in the 'Journal of Geology' the author suggested the possible relationship of Diadectes to the turtles on the following counts: . The anterior. basisphenoid.r. This plate is similar in all respects to the bone figured by Broom as the ethmoid in the skull of Lystrosaurus latirostris and by the author in the skull of Dimetrodon gigas. It is a direct continuation of the parasphenoid rostrum of the basisphenoid. The lower edge of its anterior half is in contact with the upper edges of the prevomers. right side. als. o. In the Pelycosauria there is a well developed presphenoid rostrum and in the Fig. it is united by cartilage to a median prominence on the lower surface of the nasals. a transverse section of the skull at center of orbits. opening for seventh.r. prevomer. seen from within. eth. toothbearing. Upper figure. paired elements are true prevomers. It is my belief that this plate is in part.r. right side. and beyond them. basisphenoid.168 Bulletin American Museum of Natural History. specimen described below. even if it is not a part of it. with these plates to their anterior end. alisphenoid. basioccipital. right side. section of pterygoid. derived from the parasphenoid and represents the beginning of the unpaired vomer as stated by Broom. and perhaps other cranial nerves. [Vol. at least. orbit. pv. XXVIII. mx. maxillary. Lower figure. the ethinoid is directly attached to its anterior end. v. al. alisphenoid. bo. ethmoid. Diadectes phaseolinus. bsp. Dimetrodon gigas. a longitudinal section of the skull. 4. ps.

in Archer Co. the ntumber and arrangement of the phalanges and the exact number of caudal vertebrae. The method of entrance of the internal carotids into the brain cavity. Versluys has recently shown the presence of a parasphenoid in Dermochelys. No. V. Texas. Mus. 169 I. Am. VI. RECONSTRUCTION OF Diadectes phaseolinu8 COPE. described above. VI.] Case. III. IV. lacking only the skull. Of these numbers. this one shows a short and degenerate process. Permian Reptiles and Amphibians of Texas. There is a separate parasphenoid (ethmoid. so the only points left in doubt are the structure of the carpus and tarsus.1910. V.. V. and VI must be modified. Hist. VII. IV. IV. The specimen was found in the upper conglomerate layer on the south side of Godlin Creek. This is wrong. notably in regard to the plates overlying the anterior ribs and in the length of the tail. II. in this there is a single large foramen. this renders the position and number pretty certain. In some other specimens there are no foramina perforating the lower surface of the basisphenoid. This reconstruction is based on specimen No. vomer) and distinct paired prevomers in the median line. The presence of paired descending plates from the roof of the skull anterior to the brain case. It consists of a nearly anatomically perfect skeleton. The position of all the presacral vertebrae has been determined by fits between the fragments with the exception of a single place. The degenerate palate and the disappearing transverse bone. the atlas and axis. . The lower jaw is in good preservation and shows that the animal belongs to the same species as the skull. In 1907 the author published a reconstruction of Diadectes based on the very perfect skeleton in the University of Chicago but the present specimen indicates several errors. 4839. Several specimens show no trace of the parasphenoid rostrum. The absence of prevomers and the presence of an anteriorly placed single vomer (parasphenoid). The absence of teeth on the palatines and pterygoids. 4684. Nat. a few caudal vertebr-e and many of the bones of the feet. The skeleton was worked out and reconstructed by Charles Falkenbach of the American Museum. by the author in 1908. The corrections do not modify the conclusions drawn in the first paper. The form and proportions of the present reconstruction are subject to little doubt. The form and relations of the quadrate. The absence of a parasphenoid rostrum on the basisphenoid bone.

_ . XXVIII.I co GO Q AS CQ 10 . [Vol. c3 x Gs ._ 0 0 w cN . .170 Bulletin American Museum of Natural History.

The bodies of these ribs are expanded into thin triangular plates. with the front edge straight and the posterior edge drawn out into a point which . The remaining cervicals and the dorsals are similar in form. Intercentra of small size occur in the specimen in Chicago but are not preserved in this specimen. 171 but if we add the axis and atlas there is one more vertebra than in the Chicago specimen. There are no true lumbars in the sense of any vertebrae being without ribs. The spines are all stout and low with slightly expanded. Between the fifth and the sixth vertebrie the first chevron appears and these continue to the extremity of the tail. The atlas probably bore a pair of slender ribs of considerable length. The ribs of the third.] Case. to avoid this there must have been a large preatlantal intercentrum. The neural arches are broad and strong. showing an approach to the triangular form of the three following ribs. not united into a sacrum. but with reduced zygapophyses and evidences of a very close union between them.. beyond this the ribs are anchylosed to the transverse processes and are rapidly reduced to a rudimentary condition. These latter strongly resemble the same elements in Dimetrodon. The axis resembles the succeeding vertebrie in general form but the spine is thinner and expanded somewhat antero-posteriorly. The anterior caudals. T'he ribs of the axis are short but broader. There are two sacral vertebree. It is not greatly higher than the spines of the other vertebrie. The atlas as shown by the Chicago University specimen number 1075. The ribs. so that if it articulated directly with the face of the atlas the skull would have been held at an angle to the vertebral column. The tail is as long as the presacral portion of the column. has a simple disk-like centrum with the neural arch composed of two separate halves. have free ribs. the zygapohyses are flat and there is a well developed hyposphene-hypantrum articulation throughout the presacral series. very rugose upper ends which are strongly suggestive of the presence of a row of dermal plates overlying the neural spines as in Pareiasaurus. The articular face of the occipital condyle is placed at a strong angle to the main axis of the skull. for the first presacral has a short rib codssified with the transverse process. but in common with all the rest of the ribs there is no division into capitulum and tuberculuin.1910. but the vertebrie do not assume the elongate form which indicates the presence of a very slender tail. however. fourth and fifth vertebrax show a well defined articular end with a distinct neck. The chevrons and the spines of the caudal vertebrae indicate that the tail was somewhat flattened vertically as in the alligators. Permian Reptiles and Amphibians of Texas. No trace of such plates has been observed in any specimen. but this is indicated very imperfectly in the specimen. up to the fourth or fifth.

Diadectes had practically no neck. plates which continue backward the protection of the thoracic region to a point opposite the posterior end of the scapula. In the reconstruction they have been placed as nearly in the correct position as possible. The anterior edge of the clavicle was so far forward that it was very close to the angle of the lower jaw. but this is shown to be an error. The following ribs are of normal form. In common with many of the Permian and Triassic reptiles. The limbs were exceptionally stout and short and the feet very wide and stumpy. the feet shortened and the claws blunted. XXVIII. and its alligator-like tail. seventh and eighth are overlain by thin. The broad. the closelv knit and well formed carpus and tarsus all indicate a terrestrial habit. with the bones of the girdle strongly interlocked. gradually increasing in size to the eighth or ninth and then decreasing. of low brain power and sluggish movement. narrow. the powerful limbs with well formed articulations and especially strong rugosities for muscular attachment. overlaps the succeeding rib. They occur in most of the reptiles of the Permian which are related in form or habit to the Diadectids. must have been sufficiently bizarre. The appearance of this creature. showing the presence of a broad. the animal was distinctly narrow chested. but not more so than that of many living reptiles. with its large head. blunt nail. this forms a strong protection for the anterior thoracic region.172 Bulletin American Museum of Natural History. The strong girdles. In the first restoration of Diadectes it was thought that the expanded anterior ribs were also plates overlying the ribs. narrow chest and round body. but the result is very doubtful. and it can only have offered a passive defense to such animals as the carnivorous Pelveosaurs. with its short stumpy legs and feet resembling those of a land turtle. There are only a few of the carpal and tarsal bones preserved. The pelvis is complete but has been badly crushed in fossilization so that it is much too narrow and the ischium and pubes are not horizontal as they were in life. short neck. Perhaps Heloderma would be the most similar among living forms in attitude and proportions. No trace of abdominal ribs have been found in any specimen of Diadectes but it seems probable that such elements were present. if the cylindrical tail were compressed into that of an alligator. The bones of the shoulder girdle are separated but show the same character as previously described. short feet with powerful spade- . The sixth. It is difficult to form any idea of the habits' except that the animal was herbivorous. The metapodial bones are all very broad and short and the terminal phalanges are spatulate and rugose. The ilium rose almost vertically and at right angles to the other two bones. [Vol.

Permian Reptiles and Amphibians of Texas. terrestrial herbivores. A small fragment of shale from the same region in which this specimen was found carries two footprints of a smaller animal. nor could they develop any speed even for short intervals. probably a Diadectid. Reconstruction of the skeleton of Duidectes phaseolinus. The attitude was habitually prone and they could not elevate the body on the limbs even as much as an alligator can. at least to the extent of excavating burrows Fig. . the Rhyncocephalian double arch can not be derived from the skull and the animal is otherwise too highly specialized. size. 6. Gopherus polyphenus. proportions. possibly fossorial in habit.] Case.. in' this only the impression of the claws and of a flat pad are shown. sluggish. 173 like terminal phalanges which were covered by powerful nail-like claws have strong resemblances to those of certain fossorial forms such as the gopher turtle of Florida. the type forms of the Cotylosauria. the suggestions previously made by the author that these animals ar ie nearest discovered forms to the ancestors of the turtles. The mounted skeleton bears out in form. X t nat. I believe these animals to have been harmless. for their protection. Finally: These Diadectids. attitude and probable habit.1910. can not be retained in the ancestral position in which Cope placed them. It seems probable that the foot rested fairly flat on the ground with a pad raising the carpus and tarsus slightly.

Three large vertebrae give certain characters stufficiently distinct to warrant the separation of the species. Cope Coll. across the proximal end. Height from the bottom of the posterior edge of the ceiitrum to the base of . No. including the sacral and a few presacral vertebrrn. An anterior caudal vertebra. No. These bones indicate an aniinal at least one third larger. No. Nat. however. The second femur is also crushed.5 mm. Nat. Hist. These are from the posterior portion of the column and are of a size to go with such femora as are indicated by the fragmentary specimens Nos. Nat. than the reconstructed specimen No. . . Hist. is represented by a crushed fragmenit which is 122. Cope Coll. 4684. The femur. 140. and a width across the transverse processes to the beginning of the ribs of 83 mm. the spine 146. 4361 and 4379. Am. . nov. 4392. XXVIII. . Hist. . [Vol. . 7. mmn.174 Bulletin American MIuseum of Natural History. Am. 58. Diadectes maximus sp. 4797.. Mus.. . and a length of 261 mm. et. Hist.. in its present condition it has a width across the distal end of 138. upper row. Am.. rTpe: A short series of vertebrce. Mus. Mus. 4361. (Fig. The vertebrae differ from other known forms in their greater size and in having the faces of the hyposphene and hypanitrum nearly horizontal instead of being inclined at an angle of nearly 450 to the vertical axis. NEW REPTILES.. . Nat. .. Diasparactus zenos gen. . This species is indicated by the not uncommon occurrence of especially large bones of this genus. sp. Mus. Width across the zygapophyses . Hist. From New Mexico.) This new genus and species of the family Diadectidoe is characterized by the small size of the centra compared to the height and spread of the neural . Am.5 . Muis. is not sufficient alone to indicate distinctioin among the reptiles. Cope Coll. Length of bottom line of the centrum . . Measurements. Am. 4838. . 48. Nat. Type: Three large lumbar vertebrae.5 Height of the posterior face of the centrum .5 mm. has a total height from the base of the centrum to the top of the spine of 182 mm. Cope Coll. No. Size. nov.

No. 4794. the upper end hardly . X i nat. Lower row: Diadectes maximus. 4392. M. never extending out beyond the edges of the zygapophyses and in most cases not reaching so far. The transverse processes are exceedingly short. 7... Upper row: Diasparactus zenos. On the left side. Fig.] Case. H. type vertebrT. type vertebra.1910. lateral and posterior aspects. The second presacral has a very short transverse process which is inclined very slightly forward at the lower end. Permian Reptiles and Amphibians of Texas. It is evidently a case of an abnormal support of the pelvis. 175 arches and by the short transverse processes. strong process with a large face for the sacral rib. No. The first presacral has only small transverse processes without articular facets for ribs. N. A. when viewed from the front or rear. anterior and posterior aspects. X 166 nat. A fragment of the vertebra succeeding has the appearance of carrying a large face on the right side. M. On the right side the transverse process is reduced to a mere rudiment. The whole vertebra is relatively very thin antero-posteriorly so that while it has the general form of all members of the family. size. H. size. there is a short. N. it looks much higher and wider and the small centrum gives it something of a high-shouldered kite shape. A. The last perfect vertebra in the series presents an abnormal condition.

Otenosaurus rugosus sp. nov. 4797. . 20. Among the fragments bearing the same number as the type specimen of Zatrachys apicalis are several spines of a Pelycosaurian reptile which indicate a new species of the genus Ctenosautruis. mm.. Fragments of scapulh and limb bones associated with the spines are typically.Pelycosaurian in form. No. In the anterior vertebrae the faces of the transverse processes become longer but renmain nearly vertical. The spines here described are not very long.easurements. Cope Coll. . 23. Hist.. [Vol. Matthew has located in the collection of the American Museum certain spines from Texas which are very similar to Ctenosaurus koeneni of von Huehne. 12. It is probable that this genus or something very similar to it existed in North America. Some of the spines are more slender and less expanded antero-posteriorly at the top than others but all bear the characteristic sculpture. The sides of the spine from the base to the top are covered with small irregular bosses similar to that on the skull of many amphibians. There are isolated vertebrae of Diadectid reptiles in the collection which exhibit the same relative shortness antero-posteriorlv as is shown in No. Transverse diameter of a centrum same centrum. New Mexico. the base is narrowed with almost equal anteroposterior and transverse diameters. . T'ype: A few neural spines. 63. but do not have the proportionately small centrum nor short transverse processes. 66. bottom line Antero-posterior diameter of same centrum Height of vertebra to base of spine Width across posterior zygapophyses Anterior-posterior width across zygapophyses . From Rio Arribas Co. 4785. XXVIII. Nat. Mus. In all. 37. the transverse process is so short as to be practically absent and the upper end of the articular face is on the posterior surface of the anterior zygapophysis. It may be that they indicate a new genus but we know so little of the osteology of these animals that it may well be that they are lumbar vertebrae of some form already described from the teeth and so no new iname is proposed. Am. The upper portion becomes more thin and is elongated in the antero-posterior direction. This genus has not previously been recorded from North America but Dr.176 Bulletin American Museum of Natural History. . overhangs the lower. In these respects it differs from Diadectes in which the articular face is inclined sharply downward and forward.

The quadrate is elongate and of peculiar shape. The premaxillaries were apparently large and . so it is impossible to give the exact length. No enlarged teeth in premaxillaries. . nov. Paratype: A skull. Lower edge of postorbital portion cut away by the development of a notch until the quadratojugal is lost and the prosquamosal greatly reduced. Locality uinknown. No. . Antero-posterior diameter at top Same at bottom . Teeth regularly decreasing in size anteriorly. . Hist. Mus. At first sight the skull resembles that described by Cope as Pariotichus but it is radically different in the posterior portion. 15. GYMNARTHRIA subord. There is a small but (listinct parietal foramen. . 177 mm. Distinguished from Lysorophus by the single occipital condyle. Teeth blunt cones with no indication of anterior and posterior cutting edges. The description given is taken from the type largely. The type skull is small. it articulates with the squamosal above and seemingly with the occipital plate (paroccipital + exoccipital + supraoccipital) behind. and character of the teeth. The upper portion of the skull is roofed over and there is no suggestion of temporal openings. Nine teeth in maxillary and three or four in the premaxillary. near the head of Coffee Creek. until the quadratojugal is lost. . Hist. 113. From Baylor Co. the prosquamosal doubtfully present. No. nov. Nat. Am. 30. Gymnarthrus willoughbyi gen. loss (?) of prosquamosal and quadratojugal. in a red-clay above the Wichita Conglomerate. and the quadrate fully exposed and perhaps movable. The posterior surface is somewhat injured and it is impossible to make out the form and relations of the various bones. total length slightly over 16 mm. Type: A skull. The lower edge of the postorbital region has been cut away in the manner of some turtles. 8D. nov. Quadrate freely exposed and perhaps movable. Permian Reptiles and Amphibians of Texas. 4892.] Case. Cope Coll. 4763 Am. et sp. .1910. Parasphenoid rostrum large. Nat. Mus. The relations of the bones of the upper surface is shown in Fig. but additional facts have been determined from the paratype. . Skull without teinporal openings. Height of a nearly complete spine . The premaxillaries are imperfect. Gymnarthridae fam.

the teeth on this bone sharing in the regular b/MIlll Y w \ a process i//gh f 1Iili' X 5 DM . The type skull shows two smiall openings between the orbits and nares which were supposed to be preorbital openings but as nothing of this kind appears in the paratype. 8. XXVIII. 1 type skull. its relations to the basisphenoid and quadrate are normal. and D superior aspects. The articular region is low and just anterior to this the upper edge rises in a prominent coronoid process. A. vi. The lower jaws lie in position between the upper jaws and the teeth. 45. cannot be seen. LI. [Vol.1 but this is described as having the J nat. but this is impossible. The last maxillary tooth is small.sll1>fl1o pR D :t diminution in size toward the anterior end llll:$AS C 1g Fig. Taf. and unites without suture with a strong parasphenoid process.178 sent back nares were Bulletin American Museum of Natural History. . size.' Paleontographica. The parasphenoid process is relatively the largest in any of the Permian reptiles recovered from Texas and in this 'Permische Stegocephalen und Reptilien aus Texas. the next to the last is the largest and the others diminish regularly toward the anterior end. No. s. they are probably accidental. The nearest approach to this form is the small amphibian skull described by Broili as Cardiocephalus sternbergii. the parasphenoid meets two diverging plates which are apparently the palatines but they are in large part covered by the lower jaws. N. On the lower surface the basioccipital occupies a small space at the rear. There are no cutting edges on the antero-posterior faces. of the skull. 4892.. Bd. There are no tusks or enlarged teeth on the premaxillaries. The lower edge of the pterygoids can be made out as a thin line in the matrix which fills the back part of the skull. fig. the animal approaches the interA lateral. It was thought at first that both the basisphenoid and the parasphenoid process constituted the parasphenoid bone and that the animal was an amphibian. between the nares to meet the nasals. The paratype shows that the pterygoids are large flat plates reaching the middle line. H respect. The different bones of the jaw cannot be made out. not widely separated as in amphibians. the basisphenoid is a large flat plate. At the anterior end. The anterior of good size and terminal in position. X mediate form between the amphibians'and reptiles. but it is apparent that the dentary takes part in the svmphysis. M.a Gymnarthrus willoughbi. 5.

This family approaches the Otocoelidte (now known to be a family of amphibians containing the genera Dissorophu. A. Hist. Cope mentioned that he considered as belonging to the genus Zatrachys all forms of amphibians with serrate edges of the skull and with expanded rugose terminations of the neural spine. differs from A. Order TEMNOSPONDYLI. AMPHIBIA. 179 skull complete. There is no evidence in the Cope Collection or elsewhere for the union of such forms in a single genus. From the upper conglomerate layer a few miles south of Dundee. crucifer. and by Broili as Aspidosaurus chiton. Otoca?lus and others). corresponding to the neural spines in number but distinct from them. and as there is no place for them. by the expanded tops of the neural spines which are marked on the upper surface by a rugose pitted area. This specimen is similar to those described by Cope as Zatrachys apicalis. . and by having the ribs flattened and expanded distally and overlapping to form a protection for the thoracic cavity. but differs in the fact that in the latter family the expanded spines are not rugose above and are overlain by a set of rugose plates. nov. Collection of 1908. Texas. by the intercentra with more or less prominent processes for the heads of the ribs. no parietal foramen. glascocki. (Zatrachys) crucifer Case. and A. Permian Reptiles and Amphibians of Texas. A. ASPIDOSAURIDME fam. The new species. As there seem to be several distinct genera with the expanded neural spines. teeth regularly diminishing in size anteriorly but with cutting edges. This family is characterized by the rhacitomous vertebrTe. (Zatrachys) apicalis Cope. by Case as Z. and lyra present.1910] Case. The genus Zatrachys was originally described from skulls alone and the species apicalis and crucifer were described from fragmentary dorsal spines. No. The apices of the spines touch but do not overlap. Type: A fragmentary skull and portions of the vertebral column. I propose a new family Aspidosauridae for Aspidosaurus chiton Broili. 4864. nov. Mus. Nat. The present specimen shows that the expanded neural spines are not coinected with a skull of the type of Zatrachys but with one more like that of Trimerorhachis. Archer Co. chiton Broili in having the expanded apices of the neural spines much larger and marked with a coarser sculpture. Am. Aspidosaurus glascocki sp. The intercentra have prominent processes on the sides for the heads of the ribs..

X i nat. Aspidosaurus glascocki. XXVIII. Type: A small skull. 4864. M.. A. Mus. et sp. glascocki they extend ouit from the sides of I- 71< -. size. In A. Am. N. chiton there are facets a little more prominent than in Eryops and Trimerorhachis. Tersomius texensis gen. 9. the intercentrum for two or three millimeters. The skull resembles that of Trimerorhachis as far as it is preserved. 4719. From a yellow clay near the upper layer of conglomerate. No. No. type vertebrm. on the south side of the Little Wichita River.. H. while in A. Nat. Hist. . Fig. lateral and inferior aspects. superior. [Vol. The animal was about 300 centimeters long.180 Bulletin American Museum of Natural History. in Archer Co. Collection of 1906. nov. Texas.

The orbits are large. but the tops of the spines are still concave and show the former presence of a considerable mass of cartilage. The neural spines show no indication of being divided. mosal (intertemporale) and the orbits are much larger and placed farther to the rear. A. Cope Coll. Hist.5. .. maxillary or the mandible. Interorbital width . No. Mus. 25. . . 181 The skull is flattened. No. there is no second prosquasis. The vertebrae are larger than any of the typical specimens. Type. The teeth are small and sharply conical with no enlarged ones visible in the . .] Case. The intercentra are marked on the lower face by deep pits lying on either side of a median keel. Permian Reptiles and Amphibians of Texas. . . 32. . 34. 4719. Compared with Trimerorhachi8 conangulus. There is no tabulare visible and it is ' probable that it was not present. Nat. . . 4577 Am. . The nares are small aind look almost \ & directly upward. Length on the median line Length of the lower jaw Width at the back of the skull .1910. The posterior edges of the upper ends of the intercentra are reflected forward and form a concave facet. . This specimen differs from the type specimen in such important respects that it is necessary to regard it as a new species. which it most Fig. Tersomius texenclosely resembles. type skull. 10. The position and relations of the various bones is shown in the figures. The pleurocentra are proportionately very small. having nowhere near so much of an arch as in the specimen described by Cope as Trimerorhachis conangulus. . Trimerorhachis alleni sp. mm. . Four dorsal vertebrae. extending so close to the edge that there is a very thin maxillary border. Measurements. . nov. From Texas. The spines are nmore erect and there are well formed posterior zygapophyses. 7. . four intercentra occupying 166 mm. .