<<

AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2882, pp. 1-19, figs. 1-14, table 1 June 24, 1987

The in

EDWIN H. COLBERT'

ABSTRACT The prolacertid reptile, Prolacerta broomi, hith- skeleton, including hind limbs and some caudal erto known from the Zone of the vertebrae of prolacertilian relationships, likewise Middle Beaufort beds ofSouth , is described would seem to represent an individual somewhat from specimens including two with jaws, smaller than the South African . The pres- other fragments with skulls and jaws, and various ence of Prolacerta broomi-in addition to a ly- postcranial elements, found in the Fremouw For- dekkerinid amphibian Cryobatrachus kitchingi, the mation ofthe Transantarctic Mountains. The Ant- Lystrosaurus murrayi, Lystro- represent individuals generally some- saurus curvatus, Lystrosaurus mccaigi, and Myo- what smaller than those known from , saurus gracilis, the procolophonid with slightly fewer teeth in the upper and lower trigoniceps, and the theriodonts lio- jaws, and with other minor differences. These vari- rhinus and Ericiolacerta parva-corroborates the ations are, however, probably attributable to age close relationships between Antarctica and South differences and perhaps geographic separation Africa and their probable connection during Early rather than to any specific distinctions. A partial Triassic time.

INTRODUCTION The collection of tetrapods secured were found during the second field season, from the Lower Triassic Fremouw Forma- and were collected in the vicinity ofthe junc- tion ofAntarctica during the austral summers tion between McGregor and Shackleton gla- of 1969-1970 and 1970-1971 contains var- ciers, at about latitude 85°13'S and longitude ious remains representative of the diapsid 174°30'E. The fossils were collected by Dr. reptiles, these being specifically prolacertili- James W. Kitching of the Bernard Price In- ans and thecodonts. All of these specimens stitute for Palaeontological Research, Uni-

I Curator Emeritus, The American Museum of Natural History, New York City; Professor Emeritus, Columbia University in the City of New York; Honorary Curator of Vertebrate , The Museum of Northern Arizona, Flagstaff. Copyright © American Museum of Natural History 1987 ISSN 0003-0082 / Price $2.25 2 AMERICAN MUSEUM NOVITATES NO. 2882 versity of the , Johannesburg, are prevalent in the vicinity of McGregor and his associates, namely John Ruben, then and Shackleton glaciers. ofthe University ofCalifornia, Berkeley, and In the previous papers of this series it has Thomas Rich, then ofthe American Museum been shown that the fossil tetrapods from the of Natural History, New York. are closely related to At this point I wish to acknowledge advice those occurring in the Lower Triassic Lystro- and comments from Dr. Robert Carroll of saurus Zone, Middle Beaufort beds of the the Redpath Museum, McGill University, Karroo Series in southern Africa. Montreal; from Dr. Peter Galton, ofthe Pea- body Museum, Yale University, New Haven, DESCRIPTIONS OF FOSSILS Connecticut, and the University of Bridge- port, Bridgeport, Connecticut; and from the SYSTEMATICS late Professor Georg Haas, The Hebrew Uni- CLASS REPTILIA versity, Jerusalem, Israel. Consultations with these authorities were most helpful; it must SUBCLASS DIAPSIDA be emphasized, however, that all conclusions ORDER EOSUCHIA reached in this paper are solely the respon- sibility of the author. FAMILY PROLACERTIDAE The drawings were made by Miss Pamela Comments: The family Prolacertidae was estab- Lunge; the photographs by Mr. Mark Mid- lished by Parrington in 1935, in his original de- dleton, both ofthe Museum ofNorthern Ar- scription of Prolacerta broomi, new and izona. species. He said about the higher taxonomic re- This work has been supported in part by a lationships of the genus and species: grant from the National Science Foundation, Though most closely allied to the Eosuchia No. DPP 75-23126. Prolacerta cannot, I think, be included among them, since they are the most primitive diapsids and the group from which all later, spe- STRATIGRAPHIC RELATIONSHIPS cialized, forms may have evolved. It is excluded from the Squamata by the presence of part of For the benefit of those unfamiliar with the quadratojugal and the fact that the quadrate previous papers in this series describing is fixed. [But the quadrate is not fixed, as was Triassic tetrapods from Antarctica, a few re- shown by Camp in 1945 and by Gow in 1975.] marks will be included here concerning the The proper place seems to be in a new family stratigraphic relationships of the Fremouw of the , for which the name Pro- Formation. This, the lowest unit in the Ant- lacertidae is proposed. (Parrington, 1935: 205) arctic Triassic sequence, overlies the In his paper describing in detail the of Buckley Formation, from which it is sepa- Prolacerta, Camp (1945: 95) stated that for Pro- rated by a disconformity, and in turn it is lacerta the "concert of characters seems to place it in the (=Eosuchia), and within overlain by the Falla Formation and above the family Protorosauridae." that the Prebble Formation, both of Triassic The family Prolacertidae was placed within the age. The Prebble Formation is separated by order Eosuchia by Romer in 1956 and 1966, and a disconformity from overlying vol- by Piveteau in 1955. It was included within the canic rocks, the Ferrar Dolerite, which may Lacertilia by Tatarinov (in Rozhdestvenskii and also be intruded into the Triassic sediments. Tatarinov) in 1964. In 1956 von Huene included The Fremouw Formation in the Transantarc- Prolacerta in his underorder Prolacertilia, but did tic Mountains may reach thicknesses of 650 not designate a family to contain the genus. m. Kuhn-Schnyder (1962: 132) described a skull of The sediments ofthe Fremouw Formation and showed its close similarities to are cyclic in nature, ranging from the skull ofProlacerta. He felt, however, that the coarse, con- relationships of the higher taxa of lepidosaurians glomeratic channel deposits at the base ofthe were too uncertain to warrant a definitive position cycle to fine floodplain deposits at the top. for Prolacerta and its relatives. "The relationships The specimens described in this paper were between Eosuchians, Archosaurians and Prolacer- collected from the fine-grained sediments that tilians are by no means clear. Therefore we at pres- 1987 COLBERT: PROLACERTA 3 ent classify only the Prolacertilians with the Lep- temporal arch incomplete, the jugal termi- idosaurians." nating posteriorly in a point that does not Pamela Robinson, in her paper of 1967 on "The reach the quadrate. Quadratojugal persisting, of the Lacertilia," placed the origin of but very slender. Supratemporal fenestra this order among eosuchian diapsids such as Pro- lacerta. However she did not attempt formal des- rather large. Nasals extensive, overlapped by ignations of the higher taxa of these reptiles. maxillae. Numerous palatal teeth on ptery- In his definitive study ofProlacerta, Gow (1975: goid and palatine. Lower jaw slender, prob- 1 18) established a new order, Parathecodontia, for ably with a ligamentous symphysis. Teeth the family Prolacertidae-probably to be included pleurothecodont. Five premaxillary teeth, within the Lepidosauria. His justification for set- more than 20 maxillary teeth extending to a ting up this new order of reptiles was as follows: position beneath front of orbit; more than 20 "It is thus clear that Prolacerta is representative dentary teeth. Vertebrae amphicoelous. Post- of an as yet poorly known group of cranial skeleton slender; bones frequently prearchosaurian thecodonts which in all proba- hollow or internally cancellous. Pectoral gir- bility gave rise to the later Macrocnemus and . These can then be regarded as a dle with expanded, very thin scapulocora- sterile group which paralleled the Squamata in the coid, heavy clavicle and long interclavicle; loss of the lower temporal arcade." pelvic girdle with broad puboischiatic plate. Although the Prolacertidae probably do repre- Limbs slender; hind limb much longer than sent a group separate from both the eosuchians forelimb. Ankle with large astragalus-calca- and the lacertilians in the strict sense of the two neum. "Hooked" fifth metatarsal, with fifth terms, it seems premature to set up a new reptilian digit offset from the other digits. Typical rep- order for the three genera involved. Since Prola- tilian phalangeal formula. certa seemingly is not directly ancestral to the true , perhaps the best course at the present time is to retain the family Prolacertidae in the Eosu- Prolacerta broomi Parrington chia, realizing that a satisfactory definition of the Eosuchia is not now possible. Prolacerta broomiParrington, 1 935: 197-205, figs. This is substantially the procedure followed by 1-3, pl. 9. Evans (1980). She recognized the subclass Diap- TYPE: A skull in the Museum of Zoology, sida to contain three infraclasses, namely Eosu- Cambridge University, from Harrismith, Or- chia, Archosauria and Squamata, and she placed ange , South Africa. the Prolacertidae along with 14 other families AND Lower Trias- within the Eosuchia. Moreover, she indicated the HORIZONS LOCALITIES: Eosuchia as equally ancestral to the other two in- sic; Middle Beaufort beds, Lystrosaurus Zone, fraclasses. She admited (1980: 257) that: "This is , South Africa; Fremouw an artificial arrangement since the Eosuchia be- Formation, from localities in the Transant- comes a receptacle for a wide range ofgenera, but arctic Mountains, Antarctica. it has the advantage of leaving the two probably DIAGNOSIS: See generic diagnosis above. monophyletic groups-the Archosauria and the Squamata-as independent infraclasses. The term Lepidosauria becomes (? temporarily) redundant. ANTARCTIC SPECIMENS UNDER CONSIDERATION Prolacerta Parrington, 1935 The Antarctic specimens of Prolacerta broomi were found at several localities, Prolacerta Parrington, 1935: 197-206. namely: Halfmoon Bluff, Shackleton Glacier, TYPE SPECIES: Prolacerta broomi Parring- near its junction with McGregor Glacier, ton. Central Transantarctic Mountains; Mt. Ken- HORIZON AND LOCALITY OF TYPE SPECIES: yon, about 10 km due east of Halfmoon Bluff; Lower Triassic, Middle Beaufort beds, Lys- Thrinaxodon Col., about 12 km east and trosaurus Zone, Katberg Formation, South slightly north of Halfmoon Bluff; Kitching Africa. Ridge, on the western side ofShackleton Gla- GENERIC DIAGNOSIS: Small lepidosaurs, cier, about 14 km directly west of Halfmoon with a narrow skull, a somewhat elongated Bluff; also Graphite Peak, approximately 108 preorbital , and a large orbit. Lower km west by north of Halfmoon Bluff. 4 AMERICAN MUSEUM NOVITATES NO. 2882

TABLE 1 Prolacerta broomi (measurements of skull in millimeters) Type BPIb AMNH AMNH (Parrington, UCMPa 2675 9520 9521 1935) 37151 2676 Skull pmx.-quad. 29 35c 57 54 68 pmx.-ant. bord. orbit (A) 13.5 _ 29 28 35 ant. bord. orbit-quad. (B) 15.5 - 28 26 33 Ratio A/B 0.87 1.03 1.07 1.06 a University of California Museum of Paleontology, from Camp, 1945. b Bernard Price Institute, from Gow, 1975. c Estimated measurement.

All of the Antarctic specimens of Prola- From Mt. Kenyon certa broomi are preserved as rather fragile AMNH 9558. A partially preserved skull fossils in blocks offine or siltstone and isolated disarticulated skull elements with matrix. In many of these blocks the fossils lower jaws; anterior fragments of right and are remarkably abundant, but extremely frag- left dentaries with teeth; various articulated mentary. Consequently, the descriptions that vertebrae and limb bones; preserved in a follow are based on a few selected specimens. heavy block of matrix. From Halfmoon Bluff From Thrinaxodon Col. AMNH 9520. A somewhat distorted skull AMNH 9552. Postcranial bones in matrix. with lower jaw, in the rock, with the left side AMNH 9557. This species? and top of the skull visible. Humerus, bone AMNH 9521. A skull with lower jaw, in fragments. the rock. The specimen is crushed dorsoven- From trally and many of the bones are incomplete Kitching Ridge or out of place. AMNH 9513. Fragment of small maxilla AMNH 9522. Numerous bone fragments with five teeth. of the skull and postcranial skeleton, includ- AMNH 9535. A few small vertebral frag- ing an interorbital region (the frontals) seen ments, in series. in ventral view, and a fragmentary maxilla AMNH 9564. Fragmentary cervical ver- with five teeth. tebrae. AMNH 9526. Fragment of a mandibular AMNH 9568. A block containing a right ramus with teeth; also a palatine with several pterygoid seen in dorsal view, two cervical alveoli, as well as various skull fragments. In vertebrae, and six posterior caudal vertebrae. addition, some postcranial bones, including AMNH 9573. Part of a scapulacoracoid, metatarsals and phalanges. two humeri, and other bones. AMNH 9551. Humerus, and various bone AMNH 9502. A slab containing a left pre- fragments (from Sentinel Hill, near Half- maxilla with two teeth, some skull fragments, moon Bluff). including what appears to be a part of the AMNH 9561. Ten articulated presacral anterior region of the skull, a left pterygoid, vertebrae, with ribs; also caudal . an elongated bone that may be a palatine, a Various other bones, including paired ischia large isolated tooth, an incomplete pelvic gir- (seen ventrally), part of a femoral shaft, a dle with both ilia, and fragments of other fragment ofa scapulacoracoid with the prox- elements, both hind limbs, the right one with imal end ofcrushed humerus, and ajaw frag- a complete articulated pes, part of a sacrum, ment with four alveoli. five incomplete posterior presacral vertebrae, AMNH 9574. Three articulated vertebrae some with articulated holocephalous ribs, and and limb bones. 15 caudal vertebrae with chevrons. 1987 COLBERT: PROLACERTA 5

From Graphite Peak AMNH 9560. A possible skull fragment, with fragments of the cranial roof. Jaw frag- ment with six teeth. Several vertebrae, in- cluding a series of about eight very fragmen- tary caudal vertebrae. A hind limb, and other fragments of limb bones. DISCUSSION In the execution of this study, close com- parisons have been made between Prolacerta from Antarctica and similar fossils from the Karroo sequence of Africa, notably Prola- certa broomi, Pricea, Palaeagama, Pali- guana, Saurosternon, Helosaurus, Galesphy- rus, and . These several forms have been described in papers by Broom and Rob- inson (1948), Camp (1945), Carroll (1975a, Fig. 1. Prolacerta broomi Parrington. AMNH 1975b, 1976a, 1976b, 1977), Gow (1975), 9520, skull and mandible. Left lateral view, x 2. and Parrington (1935). Other papers bearing on this study have been the classic classifi- arctica is significantly smaller than Prolacer- cation of lizards by Camp (1923), the anat- ta from Africa. The two Antarctic skulls, omy of the head of Ctenosaura by Oelrich AMNH 9520 and 9521, which are closely (1956), and Cruickshank's 1972 paper on comparable in size, are only slightly more proterosuchian thecodonts. than half as large in linear dimensions as is the type skull, based on Parrington's 1935 THE SKULL AND DENTITION restoration of the type, and the same size The description ofthe skull, jaws, and den- relationships pertain in a comparison of the tition that follows is based largely on AMNH Antarctic specimens with the skull described 9520, with additional information from by Camp in 1945. They are even relatively AMNH 9521, when such is available and smaller when compared with the skull de- pertinent. Remarks concerning other speci- scribed and figured by Gow (1975), being less mens listed above as belonging to the species than half as large as this specimen. will also be included, particularly in consid- A certain size range is evident in Prolacerta ering the postcranial skeleton. broomi. It would appear however, that the In this connection it should be pointed out largest elements in the Antarctic fossils do that Prolacerta broomi occurs in considerable not significantly overlap, either by direct abundance within the Fremouw Formation measurement or by extrapolation, the small- of Antarctica, a marked contrast to its rela- est comparable elements among the African tive scarcity within the Lystrosaurus Zone of fossils. Although this contrast in size might South Africa. The difference may represent possibly be a manifestation of a specific sep- the accidents ofcollecting, but in view of the aration between the African and Antarctic extensive exploration that has been carried fossils, it is here proposed that (in light ofthe on in the Lystrosaurus Zone over a period of close morphological similarities involved) the a century or more, as compared with the nec- size differences may be attributed to vari- essarily brief time that has been spent in the ability within a single species, and probably Fremouw Formation, the difference probably for the most part to growth factors. There is indicates environmental conditions in Ant- reason to think that some of the Antarctic arctica that favored the relative abundant material, because of its small size, especially burial and fossilization of Prolacerta, which as exemplified by AMNH 9520 and 9521, seemingly was not the case in South Africa. may represent juvenile specimens. It would appear that Prolacerta from Ant- There also are some differences in skull 6 AMERICAN MUSEUM NOVITATES NO. 2882

_.~~~~~~~~~~~~~~~~~~~~~~.

Fig. 2. Prolacerta broomi Parrington. AMNH 9520, skull and mandible. Left lateral view, x 4. proportions. That part of the Antarctic skull possibility that this fossil represents a young (AMNH 9520) anterior to the front of the individual. orbit is relatively shorter than the region pos- A key diagnostic feature of Prolacerta is terior to the antorbital margin, as compared the incomplete jugal bar. In African speci- with specimens from Africa, as described by mens the jugal is a crescentic-shaped bone in Parrington (1935), Camp (1945), and Gow lateral view, bounding the lower border of (1975). The comparatively shorter preorbital the orbit and terminating in two posterior skull region in the Antarctic specimen is per- points, one directed dorsally to embrace the haps a function of the relatively large orbit ventral prong ofthe postorbital bone and the in this specimen, which in turn indicates the other directed posteriorly toward the quad-

Fig. 3. Prolacerta broomi Parrington. AMNH 9521, skull and associated bones. Dorsal view, x 2. 1 987 COLBERT: PROLACERTA 7 rate-quadratojugal. The bone in the Antarctic Prolacerta, AMNH 9521 drawn in figure 4, is similar in its shape and relationships to the same element in Prolacerta broomi from Af- rica, especially as figured by Gow in 1975. In 9521 this respect the prolacertilian nature of the AMNH Antarctic form is well established. Fig. 4. Prolacerta broomi Parrington. AMNH Naturally there is a gap between the pos- 9521, left jugal. Lateral view, x4. terior point of the jugal and the quadrate, so that the quadrate, thus released from any as a lacrimal foramen. It is quite possible that contact with a jugal bar, may have been a there is another foramen in a slightly more movable bone with its distal end enabled to anterior position; if so, the Antarctic speci- rock back and forth by virtue of an active men would resemble Prolacerta broomi as articulation between the proximal end ofthis described by Gow in having two such foram- bone and the squamosal. Whether the skull ina. in Prolacerta broomi was truly streptostylic The frontals form a narrow cranial roof is a debatable point, as will be noted below. between the large orbits. These are elongated In African specimens the quadratojugal is bones, reaching from about opposite the an- a small splintlike bone attached to the ante- terior borders of the orbits to their probably rior border of the quadrate; the preservation sinuous posterior terminations, more or less of the Antarctic materials is not sufficient in opposite the posterior borders of the orbits. this area of the skull to determine whether Of course the frontals join the parietals in the same is true. this latter region, where the cranial roof is The nasal opening in the Antarctic skull somewhat expanded by the increased width (AMNH 9520) would appear to be rather ofthe parietals as compared with the frontals. elongated, with a narrow median bridge be- A long posterior process or wing extends back tween the two apertures, formed by the con- from each parietal to articulate with the squa- fluent premaxillae. In front ofthe nasal open- mosal, the limits and shape ofthis latter bone ing is a small foramen in the premaxilla as not being discernable. The ventral surface of in some modem lizards, evidently for the the frontals may be seen in AMNH 9522, a passage of the medial ethmoidal nerves. It is block containing a miscellany of broken probable that the maxilla is excluded from bones. In AMNH 9520 there is an elongated the nasal opening by the premaxilla, but this bone behind and parallel to the posterior pro- cannot be conclusively established on the ba- cess of the left parietal; this bone may be sis of AMNH 9520. The maxilla and nasal identified as a displaced paroccipital process are large, and the nasal is overlapped by the of the opisthotic. It seems evident that there deep vertical wall of the maxilla. is a pineal foramen in AMNH 9520. In the The prefrontal has a narrow posterior point descriptions of Prolacerta broomi by Par- extending back over the anterosuperior bor- rington and by Camp no such foramen was der ofthe orbit. The exact dimensions ofthis perceived, but Gow's description (1975: 102) bone cannot be ascertained because of poor clearly indicated a pineal foramen in one preservation of its more anterior region. It specimen. As this author remarked, "B.P.I. possibly was expanded anteriorly to border 2675 is the only Prolacerta skull with a sub- the posterolateral edge of the nasal, thus in- stantial pineal opening midway along the pa- tervening between this latter bone and the rietal suture; in the other described speci- lacrimal. The lacrimal in turn is not well de- mens there is no sign of this opening; this fined, but it would seem to be a small bone must be taken to be a variable character." between the prefrontal above and the pos- In the typical Prolacerta broomi a very terior part of the maxilla below. It may form small, splintlike supratemporal participates a small segment of the anterior border of the in the articulation between the parietal wing orbit. There is apparently a small foramen and the squamosal. The Antarctic materials immediately in front ofthe lower anteroven- are not preserved sufficiently to determine tral border of the orbit, this being identified whether the same is true. 8 AMERICAN MUSEUM NOVITATES NO. 2882

A large postorbital forms much ofthe pos- have questioned this; for example, Carroll terior border of the orbit in AMNH 9520. (1977: 387) has stated that: "The squamosal This bone does not show the triradiate struc- retains an extensive ventral process, how- ture that is typical ofthe postorbital in many ever, greatly limiting the motility ofthe quad- reptiles; it is evident that the posterior prong rate." of the bone has been broken away. Dorsally So the matter stands. In this connection it this bone comes to a point and the anterior can only be said that the upper end of the edge ofthe bone beneath this point articulates quadrate in the Antarctic specimen would for some distance with the posterior edge of seem to indicate a bone capable of move- the postfrontal. The ventral termination of ment. the postorbital appears to be blunt, but this The distal end of this bone is not well pre- probably is a result of damage to the end of served; it would appear that there probably the bone. The remnants of a postfrontal may was a transversely broad articular surface. A be seen along the posterodorsal border ofthe rather poorly defined bone occupies the place orbit in this specimen. of the left quadrate. Its distal end appears to AMNH 9560 is a block containing various lie within the articulating surface of the ar- skeletal elements, among them a part of the ticular. The bone is anteriorly convex, and orbital region of a skull with a postorbital, the anterior surface shows a series of dentic- some bone fragments that may be ofthe fron- ulate points, possibly for contact with a quad- tal area, and a piece ofmaxilla with four teeth. ratojugal. The postorbital of this specimen preserves Several fragmentary bones from the ven- slightly more of its middle portion than is tral part of the skull are present in certain present in AMNH 9520, to indicate that t:he blocks ofAntarctic matrix containing the fos- bone is somewhat more robust than the post- sil remains ofProlacerta. AMNH 9522 con- orbital of the African fossils as illustrated by tains a partially preserved parabasisphenoid Camp (1945) and by Gow (1975). element. This element is characterized by its The jugal is not clearly discernable in broad posterior region, with strong basipter- AMNH 9520. Fortunately a reasonably well- ygoid processes, for articulation with the preserved right jugal is present in AMNH pterygoids. The parasphenoid rostrum ex- 9521 and, as has been described in a preced- tends anteriorly to a broken termination ing paragraph, the bone does not make con- which exposes the hollow cancellous interior tact with the quadrate. of the bone. It seems probable that the com- A broad and robust bone is situated within plete rostrum extended forward as a very thin, the supratemporal fenestra of AMNH 9520, bladelike process. contiguous to the posterior process ofthe pa- In the block numbered AMNH 9568 there rietal. This element is here interpreted as the is a right pterygoid of an individual consid- displaced right quadrate, its posterior con- erably larger than the two skulls, AMNH 9520 cave surface lying uppermost, exposed to and 9521. The bone is exposed in dorsal view view. What is here taken to be its proximal so that pterygoid teeth are not evident except end takes the form of a rounded articular at its posterior end, where breakage reveals surface, which might have afforded an active, four small, pointed teeth. In shape the bone movable articulation with the squamosal. resembles the pterygoid in Prolacerta as fig- Gow was convinced that the skull of Prola- ured by Gow; it points anteriorly and broad- certa was streptostylic, stating that (1975: ens posteriorly into a robust pterygoid 112), "the quadrate is free to move antero- flange, the lateral edge of which would have posteriorly as a result ofthe loss ofthe lower limited movement of the lower jaw during temporal bar." closure. The quadrate process is widely ex- In the original description of Prolacerta panded. broomi, Parrington did not express an opin- A small fragment in block number AMNH ion as to possible streptostyly in this reptile, 9526 may be a part of a palatine containing but in his paper of 1945, Camp listed a a row of very small alveoli along one edge. "streptostylic quadrate" as one of the diag- The lower jaw as seen in AMNH 9520 is nostic characters ofProlacerta. Other authors slender, and there is little dorsal expansion 1987 COLBERT: PROLACERTA 9 in the coronoid region. Anteriorly the dentary broomi as exemplified by AMNH 9520 and shows a short symphyseal region; its surface 9521 seems to have four fewer teeth, above is marked by a number of small foramina in and below, than does the African variant. line beneath the anterior dentary teeth. Per- This seems reasonable in view of the some- haps there was a ligamentous symphysis. It what shorter preorbital region ofthe Antarc- would appear that the border between the tic form as compared with its African coun- dentary and the surangular and angular bones terpart. is located approximately beneath the middle In the original description (1935: 198) of region ofthe orbit. Posteriorly there is a large Prolacerta broomi, Parrington characterized retroarticular process, forming much of the the teeth as being "small, slightly com- articular bone. pressed, and hooked." Likewise, Camp (1945: The dentition is pleurothecodont and the 11) described the teeth as "small, slightly teeth are held in the alveoli by bone of at- compressed and recurved at their tips." Gow tachment, nicely seen in AMNH 9520. The (1975: 100) characterized the teeth as "lat- maxillary teeth extend back to a point be- erally compressed, with sharp unserrated neath the anterior part of the orbit, a resem- edges and are recurved." The same applies blance to South African specimens. to the teeth of Prolacerta from Antarctica, In block number AMNH 9558 there are except that in some specimens, notably two associated dentary fragments from the AMNH 9520 and AMNH 9521, they are not symphyseal region and the rami immediately compressed but rather are rounded or ellip- posterior to the symphysis, the left dentary tical in cross section. The teeth in the Ant- containing eight teeth, the right containing arctic fossils are entirely coated by enamel, seven teeth, all somewhat laterally com- as is true of the African specimens. pressed and most of them separated by in- A left premaxilla, nicely preserved in tervening sockets, or perhaps very large re- AMNH 9502, shows the lower border of a sorption pits. The mesial ventral borders of rather large nasal opening, and posteriorly an these sockets or pits extend down on the me- ascending sutural border for articulation with dial surface ofthe dentary. The first five teeth the maxilla. The first and third premaxillary in the right dentary are regularly arranged, teeth are present, the space between them with an empty socket or pit between each of being occupied by the socket for the second them. There is some confusion as to the next tooth, and the space behind the third tooth two teeth, in part because of the manner of showing a socket for a fourth member of the preservation, but the last four teeth, like the premaxillary series. Nutrient foramina are anterior five, are regularly arranged with in- present on the surface of the bone above the tervening pits or sockets. These pits may be alveolar edges, the one above the alveolus for equated with the medial resorption pits de- the fourth tooth being rather large. The teeth scribed by Gow in Prolacerta broomi. are laterally compressed, posteriorly re- In AMNH 9520 the premaxillae are bro- curved to some degree, and without serra- ken and distorted to such a degree that the tions. Behind the premaxilla is a rather large number of teeth cannot be satisfactorily de- isolated tooth, evidently from the maxilla. It termined. Both Camp (1945) and Gow (1975) too lacks serration. indicated five premaxillary teeth in Prola- The apparent difference between the pre- certa broomi. The premaxilla ofAMNH 9502 maxilla, AMNH 9502, as well as the denta- reveals only four teeth, as noted below. The ries, AMNH 95 58, and some other Antarctic number of maxillary teeth cannot be deter- specimens, such as AMNH 9520 and 9521, mined in this specimen, but in AMNH 9521 so far as lateral compression or rounded cross there are at least 21 maxillary teeth. In this sections of the teeth are concerned, perhaps same specimen there are seemingly 23 den- is not so great as it first appears. AMNH 9502 tary teeth. In Prolacerta broomi Gow, as represents an individual considerably larger mentioned, has determined S premaxillary than either AMNH 9520 or AMNH 9521 teeth and 25 maxillary teeth, making a total and comparable in size to the adult Prolacerta of 30 teeth in the upper dentition. He indi- described and figured by Camp and by Gow. cates 27 dentary teeth. Thus Prolacerta Might there be an ontogenetic or a positional 10 AMERICAN MUSEUM NOVITATES NO. 2882

difference involved? As for the latter possi- bility, AMNH 9521 does show positional dif- ferences, the more anterior maxillary teeth being rounded in cross section, the more pos- terior teeth showing some lateral compres- sion. Other preserved parts ofthe skull in AMNH 9502 are ofan equivocal nature, and difficult to interpret. As noted, there appears to be the anterior part of a snout, which, if the iden- tification is correct, is narrow and elongated. _AIVIIN~MH Y)WZ J In this respect it shows some resemblance to the snout ofProterosuchus, but it is so poorly Fig. 5. Prolacerta broomi Parrington. AMNH preserved that no definite comparison is here 9502, left premaxilla. Lateral view, x 4. implied.

F

ART

Fig. 6. A. Prolacerta broomi Parrington. Restoration of skull from Antarctica in left lateral view, x 4. B. Prolacerta broomi Parrington. Restoration of skull from South Africa, in left lateral view, after Gow, x 1.5. Abbreviations: ANG, angular; ART, articular; DEN, dentary; F, frontal; J, jugal; L, lacrimal; MX, maxilla; N, nasal; P, parietal; PMX, premaxilla; PO, postorbital; POF, postfrontal; PRF, prefrontal; Q, quadrate; QJ, quadratojugal; SUR, surangular; SQ, squamosal. 1987 COLBERT: PROLACERTA I I

On the basis of the information available as described above, a restoration ofthe skull of Prolacerta from Antarctica, and a similar restoration of Prolacerta from Africa, are shown in figure 6.

THE AXIAL SKELETON Several blocks of matrix from the Fre- mouw Formation contain various elements ofthe postcranial skeleton ofProlacerta. Cer- vical vertebrae are seen in AMNH 9558 with 9 vertebrae, and AMNH 9564 with 2 ver- AMNH 9568 tebrae; 10 dorsals and a caudal are exposed in AMNH 9561, 2 cervicals are exposed in Fig. 7. Prolacerta broomi Parrington. AMNH AMNH 9568, 3 dorsals in AMNH 9574, 9568, cervical vertebra. Left lateral view, x 4. while in AMNH 9502 there are 5 incomplete posterior dorsals and 15 caudal vertebrae. The from Antarctica, cervicals, as is the case in the African fossils, here identified as ofProlacerta, are elongated, but differ in that the spines are slanted pos- perhaps to about the same degree as would teriorly to a slight degree. Gow showed the seem to be characteristic ofProlacerta broomi, dorsal vertebrae as having vertical spines. but with taller spines. The skull length ofPro- When seen in ventral view, as shown by lacerta as illustrated by Gow (1975), is about AMNH 9561, the dorsal centra are "pinched" equal to the articular length offour cervicals; laterally so that each vertebra has a pro- as illustrated by Camp (1945) to the articular nounced ventral ridge. The rib facets are not length of about five cervicals. Unfortunately prominent and it appears that in the more there is no specimen in the Antarctic mate- posterior ribs the facets had merged so that rials with an associated skull and cervical ver- these ribs were essentially holocephalous. tebrae, but it would appear that the ratio of There are well-developed intercentra in the skull length to cervical length, as based on a dorsal series. comparison of AMNH 9520 and AMNH The postenor presacral vertebrae, which 9521 with AMNH 9558 (probably repre- are seen in AMNH 9502, have strong zyg- senting a larger individual than represented apophyses, and on each vertebra a rather tall, by the skulls) is on the order of about four strong spine rising from the posterior part of cervicals. The cervicals in the Antarctic ma- the neural arch, its postenror edge being above terials have, as mentioned, somewhat taller the posterior zygapophyses. Each of these spines than those from Africa, and these vertebrae is characterized by a single anterior spines are relatively long, there being but lit- facet on each side ofthe neural arch, directly tle distance between the anterior and poste- beneath the anterior zygapophyses, for artic- rior ends ofspines ofadjacent vertebrae. One ulation with a holocephalous rib. of the two cervicals in the block numbered The imperfectly preserved sacrum of this AMNH 9568 has a very long spine, the an- specimen shows at least one broad sacral rib, terior end of which projects toward the an- distally expanded, evidently for a junction terior zygapophyses, an indication that this with a second sacral rib. It indicates that the is the third vertebra in the cervical series. The sacrum was rather broad in this reptile. cervicals seen in AMNH 9558 do not show The considerable series of caudal verte- the low, lateral horizontal ridge that char- brae, so nicely preserved in AMNH 9502, acterizes comparable vertebrae from Africa, are seen in ventral aspect, so it is not possible but the two vertebrae ofAMNH 9568, which to determine the height or strength of the are exceptionally well preserved, do have this neural spines. The centra are expanded at feature. their articulating ends, the medial part ofeach The dorsal vertebrae are shorter than the vertebra being contracted, thus giving to each 12 AMERICAN MUSEUM NOVITATES NO. 2882

I-_It Fig. 8. Prolacerta broomi Parrington. AMNH 9561, a block containing a series ofarticulated presacral vertebrae with ribs and associated bones. Oblique ventral view, x 1.

centrum something ofa hourglass shape. The the space between the anterior face ofthe first transverse processes are long. Elongated vertebra carrying a complete chevron and the chevrons, distally expanded, are articulated posterior face of the immediately preceding between most of the exposed caudal verte- vertebra. brae. The broken base ofa chevron occupies THE APPENDICULAR SKELETON- FORELIMB Elements ofthe pectoral girdle and the fore- limb are visible in several blocks, namely, AMNH 9526, 9551, 9557, 9558, 9561, and 9573. The upper part ofwhat may be a right scap- ulacoracoid is present in block AMNH 9573. The vertebral edge appears to be broken in a straight line across the blade ofthe bone, while the posterior edge of the bone descends in a gentle, sweeping curve. Anteriorly the bone is greatly expanded, very much as illustrated by Gow in his figure of the scapulacoracoid of Prolacerta broomi. Ventrally the bone is poorly preserved. Humeri are preserved in each ofthe blocks numbered AMNH 9551, 9561, 9557, and 9573. In 9551 and 9557 only the distal half of the bone is preserved. In 9561 the bone is crushed and only partially preserved. In 9573 AMNH 9551 there are two humeri, one fairly well pre- served, the other consisting ofa partial shaft. Fig. 9. Prolacerta broomi Pamngton. AMNH In describing the humerus of Prolacerta 9551, partial left humerus. Dorsal view, x 4. broomi, Gow observed (1975: 109) that the 1 987 COLBERT: PROLACERTA 13

Fig. 10. Prolacerta broomi Parrington. AMNH 9502, caudal vertebrae with chevrons, part ofa tibia anid a left pes, x 1.4. bone is "primitive and simple, though the AMNH 9551 and 9573 the humeri are dis- eratepicondylar foramen has been lost; there tally expanded to an even greater degree than is a deep ectepicondyla groove." So far as shown by Gow. Numbers AMNH 9552 and they are preserved, this description applies 9573 are considerably wider distally than is tc the bones from Antarctica. In blocks 9557, so the possibility exists that the latter 14 AMERICAN MUSEUM NOVITATES NO. 2882

F .1-1~.i '- ' Y

-.4 'if "' 14""'INr .-*, 'P. .0 \1:1_- ., --jI. .- ... >& .1- vg. if

-. w ...Y.

AMNH 9502

Fig. 1 1. Prolacerta broomi Parrington. AMNH 9502, a partial counterpart ofthe slab shown in figure 10. Caudal vertebrae and tibia below, hind limb parts of pelvis and presacral vertebrae above, x 1.4. is not to be identified as Prolacerta. The two An elongated, slender bone in block num- humeri to be seen in the block numbered ber AMNH 9558 is here considered as prob- AMNH 9573 are robust bones, but in each ably a radius. One end ofthe bone is crushed the ends are too crushed to yield useful in- to such a degree that it is not possible to formation about the humerus. determine satisfactorily its length or shape. 1 987 COLBERT: PROLACERTA 15

It is essentially a straight, slender bone, the ends of which are slightly expanded. The length of the bone would appear to be some- what less than the estimated lengths of the humeri described above, which would be in accordance with the relationship between hu- merus and radius-ulna in Prolacerta.

THE APPENDICULAR SKELETON-HIND LIMB Block number AMNH 9561 contains a part of the mid-shaft of a femur. This is a large bone, somewhat curved, its diameter being closely comparable to the diameter of the fe- mur ofProlacerta broomi as figured by Gow. It is also a large hollow bone, the walls ofthe shaft being comparatively thin. Our knowledge of the hind limb in Pro- lacerta broomi from Antarctica is best ob- tained from the partial skeleton, AMNH 9502. The ilium is rather tall, its anterior border being quite straight; this is in contrast with the ilium ofProlacerta as figured by Gow. It would appear that the superior border of the bone forms a sweeping curve, extending back and down, so that there is considerable posterior extension. The ventral part of the ilium where it articulates with the pubis and ischium is only partially preserved, a rather large acetabular depression is evident. Both the pubis and the ischium are large AMNH 9502 and platelike. The pubis terminates ventrally Fig. 12. Prolacerta broomi Parrington. AMNH in a rather straight, thickened, and somewhat 9502, left pes. Dorsal view, x 3.4. everted border. The pubis is not sufficiently well preserved to indicate the presence or ab- sence of an obturator foramen. The tibia is a stout, straight bone, equal to The two ischia preserved in the block num- the femur in length. This contrasts with the bered AMNH 956 1 are platelike, each resem- hind limb ofProlacerta as described by Gow, bling in shape the ischium of Prolacerta in which the lower bones are somewhat lon- broomi as figured by Gow. There is a ven- ger than the femur. The fibula is of course trally directed, low ridge on the medial sur- much smaller in diameter than the tibia, yet face ofthe bone, and posteriorly it terminates the shaft of this bone is strong and rounded acutely. in cross section. The femur ofAMNH 9502 is a strong bone, A nicely articulated pes completes the curved posteriorly at its distal end. Although specimen here described. There are two large somewhat crushed proximally, it would ap- proximal bones in the tarsus. One of these, pear that there is a large internal trochanter, a large, rounded element obviously is the cal- separated from the head of the bone by an caneum. A transversely elongated bone sit- intertrochanteric fossa. The proximal artic- uated immediately medial to the calcaneum ular end of the femur appears to have been is here interpreted as a large astragalus (or rounded. The shaft of the bone is smooth, perhaps a fused astragalus and centrale), ar- without muscle scars. ticulating with both the tibia and fibula, much 16 AMERICAN MUSEUM NOVITATES NO. 2882

B

Fig. 13. A. Prolacerta broomi Parrington. AMNH 9502, left pes. Dorsal view, x 3.4. B. Prolacerta broomi Parrington. Pes, after Gow. Dorsal view, x 3.4. Abbreviations: Ast, astragalus; Cal, calcaneum; Cen, centrale; 1, 2, 3, 4, tarsals. as Gow has shown for Prolacerta. Gow fourth metatarsal, and to a larger degree with showed a small centrale in contact with the the proximal end of the fifth metatarsal. astragalus in Prolacerta, but such an element Proximally it articulates with both the as- is not discernable in the specimen at hand. tragalus and the calcaneum. Yet, generally speaking, the disposition ofthe There are five well-developed metatarsals, proximal tarsals in the present specimen ac- progressively increasing in length from the cords closely with the arrangement seen in first to the fourth member of the series. The Prolacerta. fifth metatarsal is approximately equal to the There are four distal tarsal bones, the first first in length. There are two phalanges in the three articulating with the proximal ends of first digit, three in the second, four in the the first three metatarsals. The fourth distal third, and four in the fifth. Only the two large tarsal is larger than its fellows, and is related proximal phalanges are preserved in the in small part with the proximal end of the fourth digit, but it seems quite probable that 1987 COLBERT: PROLACERTA 17

the right pubis is also visible. The femur, its head in articulation with the pelvis, is a long, slender bone, slightly expanded proximally and distally. It evidently represents an indi- vidual considerably smaller than those rep- resented by AMNH 9502 and by the humeri, described above. The tibia and fibula are in articulation with the femur. The distal ends ofthese bones are missing as are any elements of the tarsus. As in AMNH 9502, the tibia and fibula are approximately equal to the fe- mur in length. The metatarsals are present although incompletely preserved. They do show, however, an increase in length from the first through the fourth metatarsal, as would be expected. The fifth metatarsal can be seen as a short, broad bone, being partic- ularly wide at its proximal end, but it is so incompletely visible that no exact determi- nation can be made as to its shape. Poorly preserved articulated phalanges are to be seen distal to the second and third metatarsals. These show the expected numbers of three and four, respectively. Fig. 14. Prolacerta broomi Parrington. AMNH 9560, right hind limb with fragments ofthe pelvis CONCLUSION and parts of the pes, x 1. The discovery of Prolacerta broomi in the Lower Triassic Fremouw Formation of the Transantarctic Mountains forms one more there were five present. Therefore there is link in the paleontological evidence showing every reason to think that the phalangeal for- the very close paleozoological relationships mula of this foot was characteristically rep- between Antarctica and southern Africa at tilian, namely 2-3-4-5-4. the beginning of Triassic time. These rela- The fifth metatarsal merits special atten- tionships are exemplified by the presence in tion. As noted, it is about equal to the first Antarctica ofthe labyrinthodont amphibians metatarsal in length, but it is very broad, and Austrobrachyops and Cryobatrachus, the for- divergent from the rest of the foot. It does mer comparable to the South African Cy- not, however, have the hooked shape so char- nognathus Zone brachyopid Batrachosuchus, acteristic ofthecodont reptiles; rather its me- the latter to the South African Lystrosaurus dial edge is curved. The same is true of the Zone lydekkerinids and Lim- lateral edge, which forms a strong curve lead- noiketes (Colbert and Cosgriff, 1974). The two ing to a broad flange lateral to the proximal continental are further linked among articulation ofthe bone. In a general way this the reptiles by the specific identities between bone resembles the fifth metatarsal of Pro- , procolophonids, , and lacerta, as figured by Gow. It is distinguished, scaloposaurians. Thus Lystrosaurus murrayi however, by its very strong proximal artic- and Lystrosaurus curvatus (Colbert, 1974), ulation. Lystrosaurus mccaigi (Cosgriff et al., 1982), An articulated right hind limb, with an in- Myosaurus gracilis (Hammer and Cosgriff, complete foot and a very fragmentary pelvis, 1981), Procolophon trigoniceps (Colbert and is present in block AMNH 9560. The pelvis Kitching, 1975), Thrinaxodon liorhinus (Col- is seen in ventral view, with both ischia poor- bert and Kitching, 1977), and Ericiolacerta ly preserved though, visible. A small part of parva (Colbert and Kitching, 1981), found in 18 AMERICAN MUSEUM NOVITATES NO. 2882 the Fremouw Formation, extend the ranges Colbert, Edwin H., and John W. Cosgriff ofthese species into East Antarctica from their 1974. Labyrinthodont amphibians from Ant- original distributions in the Lystrosaurus Zone arctica. Ibid., 2552:1-30. W. other sca- Colbert, Edwin H., and of the Karroo beds. In addition, 1975. The Triassic reptile Procolophon in Ant- loposaurian reptiles enlarge the relationship, arctica. Ibid., 2566:1-23. although not down to the generic or specific 1977. Triassic reptiles from Antarc- levels. Prolacerta broomi does, however, add tica. Ibid., 2611:1-30. to the record still another genus and species 1981. Triassic scaloposaurian reptiles from common to the two regions. One can only Antarctica. Ibid., 2709:1-22. conclude that this reptile, like the other tet- Cosgriff, J. W., W. R. Hammer, and W. J. Ryan rapods from the Fremouw Formation, indi- 1982. The Pangaean reptile, Lystrosaurus cates most clearly that there was a close li- maccaigi, in the Lower Triassic ofAnt- gation between East Antarctica and southern arctica and South Africa. J. Paleontol., 1-385. the range of the Lystrosaurus 56(2):37 Africa, with Cruickshank, A. R. I. fauna extending across what was once a uni- 1972. The proterosuchian thecodonts. In K. fied geographic region. A. Joysey and T. S. Kemp (eds.), Studies in vertebrate evolution, pp. 89-116. Edinburgh: Oliver and Boyd. Evans, Susan E. LITERATURE CITED 1980. The skull ofa new eosuchian reptile from Broom, Robert, and J. T. Robinson the Lower Jurassic ofSouth Wales. Zool. 1948. Some new fossil reptiles from the Kar- J. Linn. Soc., 70:203-264. roo beds of South Africa. Proc. Zool. Gow, C. E. Soc. London, 118(2):392-407. 1975. The morphology and relationships of Camp, Charles Lewis Youngina capensis Broom and Prola- 1923. Classification of the lizards. Bull. Am. certa broomi Parrington. Paleontol. Af- Mus. Nat. Hist., 48(11):289-481. ricana, 18:89-131. 1945. Prolacerta and the protorosaurian rep- Hammer, W. R., and J. W. Cosgriff tiles. Am. J. Sci., 243:17-32. 1981. Myosaurus gracilis, an anomodont rep- Carroll, Robert L. tile from the Lower Triassic of Antarc- 1 975 a. The early differentiation ofdiapsid rep- tica and South Africa. J. Paleontol., tiles. Colloques internationaux du 55(2):4 10-424. Centre National de la Recherche Scien- Kuhn-Schnyder, Emil tifique, No. 218. Problemes actuels de 1962. Ein weiterer Schadel von Macrocnemus Paleontologie (Evolution des Ver- bassanii Nopsca aus der anisischen Stufe tebres), 433-449. der Trias des Monte San Giorgio (Kt. 1975b. Permo-Triassic "lizards" from the Kar- Tessin, Schweiz). Palaontol. Z. H. roo. Paleontol. Africana, 18:71-87. Schmidt-Festband: 110-133. 1 976a. Galesphyrus capensis, a younginid eosu- Oelrich, Thomas M. chian from the Cistecephalus zone of 1956. The anatomy ofthe head of Ctenosaura South Africa. Ann. South African Mus., pectinata (Iguanidae). Misc. Publ. Mus. 72(4):59-68. Zool. Univ. Michigan, 94:1-122. 1976b. Eosuchians and the origin of archo- Parrington, F. R. saurs. In R. Ont. Mus. Life Sci. Misc. 1935. On Prolacerta broomi, gen. et sp. n., and Publ., Athlon-Essays on Paleontology the origin of lizards. Ann. Mag. Nat. in Honour of Loris Shano Russell, pp. Hist., ser. 10, 16:197-205, pl. 9. 5 8-79. Toronto: Royal Ontario Mu- Piveteau, J. (ed.) seum. 1955. Traite de paleontologie, vol. 5. Paris: 1977. The origin of lizards. In S. Mahala An- Masson et Cie, 1 1 13 pp. drews, R. S. Miles, and A. D. Walker Robinson, P. L. (eds.), Problems in vertebrate evolu- 1967. The evolution of the Lacertilia. Colloq. tion, pp. 359-396, Linnean Soc. Symp. Int. Cent. Natl. Recherch. Sci., 163:395- Ser. no. 4, 1977. 407. Colbert, Edwin H. Romer, A. S. 1974. Lystrosaurus from Antarctica. Am. Mus. 1956. Osteology of reptiles. Chicago: Univ. Novitates, 2535:1-44. Chicago Press, xxi + 772 pp. 1987 COLBERT: PROLACERTA 19

1966. . Chicago: Univ. von Huene, F. R. Chicago Press, ix + 468 pp. 1956. PalIiontologie und Phylogenie der nied- Rozhdestvenskii, A. K., and L. P. Tatarinov (eds.) eren tetrapoden. Jena: Verlag, xii + 716 1964. Fundamentals of paleontology. Am- pp- phibia, Reptilia, Aves. Moscow: Nauka, 722 pp. (in Russian).