sign in sign up
<<
Home , Captorhinidae, Captorhinus, Euryapsida, Labidosaurus, Protorothyris, Sauropsida

AMERICAN MUSEUM

Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET NEW YORK, N.Y. 10024 U.S.A. NUMBER 2688 OCTOBER 30, 1979

EUGENE S. GAFFNEY AND MALCOLM C. MC KENNA A Late Captorhinid from Rhodesia IITVWW"Wkq..-, if AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2688, pp. 1-15, figs. 1-6, tables 1, 2 October 30, 1979

A Late Permian Captorhinid from Rhodesia

EUGENE S. GAFFNEY' AND MALCOLM C. MC KENNA2

ABSTRACT Two partial from the late Permian torhinidae is a monophyletic group (possibly Madumabisa Mudstone in the Middle Zambezi Basin including ) with these derived characters: of Rhodesia belong to the captorhinid Pro- downturned premaxilla, ectopterygoid and tabular tocaptorhinus. Heretofore, Protocaptorhinus has absent, medial process of jugal. Protocaptorhinus been known only from the Early Permian of , differs from Romeria in having a shallow median whereas Africa has yielded only one other cap- parietal embayment and differs from remaining cap- torhinid, Moradisaurus, from . The Cap- torhinids in lacking a retroarticular process. INTRODUCTION In May 1976, Mr. Brian Hosking and the in Kuhn (1969), but the papers that are the junior author made a small collection of therap- most pertinent to this study are Clark and Car- sid and captorhinid skulls from a late Permian roll (1973) and Heaton (1979). locality in Rhodesia. The two captorhinid spec- The junior author provided the geologic and imens are of particular interest because they locality information (as well as the specimens), provide the second record of this group from and the senior author is alone responsible for Africa. The captorhinid skulls have been identi- the morphology and discussion sections. fied as Protocaptorhinus, a form previously known only from the Early Permian of North ACKNOWLEDGMENTS America. The is usually placed We are particularly indebted to Dr. Robert in the Captorhinomorpha, an avowedly para- Carroll, McGill University, Montreal, and Drs. phyletic group that is generally considered to Malcolm Heaton and Robert Reisz, Erindale contain the ancestors of later . This University, Toronto, for sharing their speci- situation has prompted a phylogenetic study of mens and unpublished data, and for examining the Captorhinidae using shared derived char- our material and providing us with some very acters (see Gaffney, 1979, for a discussion of useful ideas. Dr. Heaton gave us free access to the methodology used here). his thesis on Eocaptorhinus which, at that time, Previous work on captorhinids and cap- was still unpublished. The very difficult prep- torhinomorphs in general has been summarized aration of the specimens was accomplished by

'Associate Curator, Department of Paleontology, American Museum of Natural History. 2Frick Curator, Department of , American Museum of Natural History.

Copyright (©) American Museum of Natural History 1979 ISSN 0003-0082 / Price $1.30 2 AMERICAN MUSEUM NOVITATES NO. 2688

Mr. Gil Stucker using an air abrasive machine. Rhodesia. Fossils occur in sideritic concretions Mr. Chester Tarka and Ms. Lorraine Meeker forming a lag concentrate. produced the illustrations. HolUZON: Middle Madumabisa Mudstones, We also thank Professor G. Bond and Mr. K5d of Bond (1973); late Permian (Dzhulfian) in Brian Hosking of the Geology Department, age. The captorhinids were found associated University of Rhodesia, for their hospitality and with therapsids, most of which are endothio- logistic support of the junior author, both in donts similar to Emydops (C. B. Cox, personal Salisbury and in the field. Dr. Raath, formerly commun.). Director of Museums and Monuments of Rho- COLLECTORS: Malcolm C. McKenna and desia, now at the Bernard Price Institute, Brian Hosking, May 1976. Johannesburg, generously arranged for permis- sion to collect and for the loan of specimens DESCRIPTION for study. Dr. C. B. Cox, University of London, King's College, identified the associ- Among the captorhinid taxa used for com- ated anomodont material. parisons here, and Eocaptorhinus are by far the best known, being represented by ABBREVIATIONS a number of well-preserved skulls. Price (1935) INSTITUTIONS and Fox and Bowman (1966) have produced OU, University of Oklahoma descriptions of Captorhinus, and Heaton (1979) QG, National Museums of Rhodesia has described Eocaptorhinus in even greater ANATOMICAL detail. Considering that the two taxa differ only ang, angular pfr, prefrontal in the number of tooth rows (providing that one art, articular pm, premaxilla accepts the hypothesis that they are different at bo, basioccipital po, postorbital all), this type is as well known as any den, dentary ps, parasphenoid Recent form. Romeria and Protocaptorhinus, fr, frontal pt, pterygoid however, are known from three and two skulls, ju, jugal qj, quadratojugal respectively, all of which lack significant por- la, lacrimal qu, quadrate mx, maxilla sm, septomaxilla tions of the palate (Clark and Carroll, 1973). na, nasal sq, squamosal The currently known material of Romeria and pa, parietal st, supratemporal Protocaptorhinus and the excellent specimens pal, palatine vo, vomer of Captorhinus and Eocaptorhinus were avail- pf, postfrontal able for comparison with the Rhodesian Pro- tocaptorhinus during the course of this study. SYSTEMATICS In many cases comparisons had to be restricted to Captorhinus and Eocaptorhinus because of FAMILY CAPTORHINIDAE the poor preservation of Romeria and Protocap- Protocaptorhinus Clark and Carroll, 1973 torhinus specimens. In the following descrip- tion, Protocaptorhinus always refers to the Protocaptorhinus, sp. North American material, the Rhodesian skulls SPECIMENS: QG 1105, anterior portion of are not identified as Protocaptorhinus until the skull; QG 1106, posterior portion of skull (see Discussion section. figs. 1-3). LOCALITY: Small patch of badlands lying at SKULL the base of a large hill just south of the con- fluence of the Sengwa and Lutope rivers, PREMAXILLA: The premaxilla is best seen on 18°07' S. latitude, 28°12' E. longitude'; Gokwe the right side of QG 1105 where only a portion Tribal Trust Area, Middle Zambezi Basin, of the internarial region is badly damaged. The preserved areas agree closely with Captorhinus 'UTM grid reference PL283009 on 1:50,000 series of and Eocaptorhinus as described by Fox and Rhodesia, sheet 1828A. Bowman (1966) and Heaton (1979). Although 1979 GAFFNEY AND MC KENNA: CAPTORHINID 3 all the premaxillary teeth are damaged to some SEPTOMAXILLA: Although one was presum- extent, it does seem clear that each bone has ably present, the septomaxilla is not visible. four teeth. Protocaptorhinus, Captorhinus, and LACRIMAL: The lacrimal is preserved best on Eocaptorhinus have four or five premaxillary the right side of QG 1105 and extends from the teeth and a distinct size gradation in these teeth anterior edge of the orbit to the posterior edge with the anteriormost being longest and widest of the nares. It compares closely with Pro- and the posteriormost being shortest and small- tocaptorhinus, Captorhinus, and Eocap- est. The Rhodesian specimen is consistent with torhinus. these characters but the anteriormost tooth does NASAL: The limits and dorsal surface of the not seem to be relatively as large as that seen right nasal are clearly preserved in QG 1105 in the North American taxa. although the area bordering the nares is some- MAXILLA: Both maxillae are preserved in what damaged. The left nasal is broken and QG 1105, but the left maxilla is somewhat crushed but agrees in morphology with the crushed. The external limits and shape of the right. The only distinction between the nasal of maxilla agree with that seen in Protocap- the Rhodesian skull and Captorhinus is the torhinus, Captorhinus, and Eocaptorhinus. The slightly more extensive prefrontal contact and anterior limit of the maxilla forms part of the less extensive lacrimal contact in the former. narial margin, there is a mid-length dorsal PREFRONTAL: The prefrontal in QG 1105 is swelling, and a posterior jugal contact; all as in badly damaged on the left side, but complete the New World forms. The mid-length swelling on the right. It forms the border of the orbit coincides with the position of the largest max- anterodorsally and extends forward to separate illary teeth, some of which may become dis- the nasal and lacrimal for much of their tinctly larger than the other maxillary teeth. lengths. It agrees in detail with Protocap- The Rhodesian form, however, has maxillary torhinus, Captorhinus, and Eocaptorhinus. teeth that are somewhat more similar to each FRONTAL: The frontal is best seen in QG other in size and conspicuously enlarged ca- 1105 but fragments of it as well as the fronto- niniform teeth are absent. Bolt and De Mar parietal suture are preserved in QG 1106. The (1975) noted some variation in this feature in frontal agrees closely with Protocaptorhinus, Captorhinus. Captorhinus, and Eocaptorhinus. The lower jaws have not been removed from PARIETAL: QG 1106 has a complete right the skulls of the Rhodesian specimens but it is parietal and portions of the left preserved but possible to determine that a single row of teeth the posteromedial surface is damaged making is present as in Protocaptorhinus and Eocap- the midline parietal suture and the pineal fora- torhinus. men impossible to determine. The parietal in The number of maxillary teeth is in doubt the Rhodesian form differs slightly from Pro- because of the presence of the lower jaws, but tocaptorhinus, Captorhinus, Eocaptorhinus, on the right side 19 maxillary teeth are visible, and romeriids in apparently having a medial whereas the damaged left side shows evidence constriction about midway along its length. of a minimum of 12 teeth. Heaton (1979) re- Also, as restored here, the proportions of the ported 17 to 22 teeth in Eocaptorhinus, and parietal are somewhat closer to those seen in Clark and Carroll (1973) stated that Protocap- Romeria (Clark and Carroll, 1973) in that the torhinus had 18 to 22. Although the Rhodesian length and width are about the same rather than material is not well enough preserved to resolve the length being slightly greater than the width the matter, it does seem that as many as five as in Captorhinus. However, this difference gaps may be present in the right tooth row that may very well be within the margin of error in may represent damaged or lost teeth or replace- the restoration due to preservational artifacts ment pits. Twenty-one tooth positions are indi- and to the absence of a distinct midline suture cated in the restoration as a conservative to maintain control of symmetry. The posterior estimate. margin of the parietal lacks the occipital em- 4 AMERICAN MUSEUM NOVITATES NO. 2688

FIG. 1. Protocaptorhinus sp., QG 1106, Late Permian, Rhodesia. Posterior skull portion in dorsal (upper left), ventral (upper right), and right lateral (lower) views. bayment seen in Romeria and has the truncated right side of QG 1106; QG 1105 retains only the morphology seen in Protocaptorhinus, Cap- anterior regions of the two elements. The bones torhinus and Eocaptorhinus. as preserved agree closely with the postorbital : The occiput is known only in and postfrontal of Protocaptorhinus, Cap- QG 1106 and although the presence of torhinus, Eocaptorhinus, and Romeria. is clear their shape is not. What is JUGAL: The jugal is present on the right side present is consistent with Protocaptorhinus, of both skulls; QG 1105 lacks the posterior Captorhinus, and Eocaptorhinus. margin of the bone while QG 1106 lacks the SUPRATEMPORAL: The area where the su- anterior process. As in Romeria, Protocap- pratemporal lies is badly broken. The lateral torhinus, Captorhinus, and Eocaptorhinus the limits are clear but the medial ones are not. anterior process of the jugal tends to separate Although the occipital area is poorly preserved, the lacrimal and maxilla for a short distance. there is no indication of a tabular. The posterior, expanded plate of the jugal also POSTORBITAL AND POSTFRONTAL: Both agrees with those taxa. bones are preserved in their entirety only on the SQUAMOSAL: The squamosal is present only 1979 GAFFNEY AND MC KENNA: CAPTORHINID 5

FIG. 2. Protocaptorhinus sp., QG 1105, Late Permian, Rhodesia. Anterior skull portion in dorsal (upper left), ventral (upper right), and right lateral (lower) views. on the right side of QG 1106. The bone agrees sian form is somewhat convex upward rather with that element in Protocaptorhinus, Cap- than straight as in Eocaptorhinus. torhinus, and Eocaptorhinus except for the slight medial extension along the parietal su- PALATE AND BRAINCASE ture. As mentioned under Parietal this distinc- tion in shape of the parietal-squamosal suture The palate and braincase in both Rhodesian between New World forms and the Rhodesian specimens are poorly preserved; the general form may be due to errors in the restoration. features are determinable but many sutures and The posterior flange of the squamosal is par- other structures are dubious. Furthermore, the tially preserved on the right side but so badly palate in the two known specimens of North damaged that its original extent cannot be seen. American Protocaptorhinus is even less infor- QUADRATOJUGAL: As in Protocaptorhinus mative, restricting comparisons to Romeria, and Romeria the quadratojugal of the Rhode- Eocaptorhinus, and Captorhinus. In all deter- 6 AMERICAN MUSEUM NOVITATES NO. 2688

The primary area of systematic interest in the lower jaw is posterior to the quadrate artic- ulation. Captorhinus and Eocaptorhinus (Group 2 in fig. 6) characteristically possess a well- developed retroarticular process that is not found in Protocaptorhinus and Romeria. QG 1106 has this region well preserved and clearly shows that a retroarticular process was absent. DISCUSSION It is apparent that the Rhodesian skull is nearly identical to a form described by Clark and Carroll (1973) from the Lower Permian of Texas as Protocaptorhinus pricei. As can be seen in table 1 the Rhodesian form and Pro- tocaptorhinus do not differ in any features usu- ally deemed significant in captorhinomorph systematics and, because of this, the Rhodesian skull is here identified as Protocaptorhinus sp. In order to ascertain just what this might mean it is necessary to inquire into the phylogenetic relationships of Protocaptorhinus and the Cap- FIG. 3. Diagram showing extent of overlap and torhinomorpha in general. portions represented by the two skulls of Protocap- The group of amniotes arrayed under the torhinus from Rhodesia. heading Captorhinomorpha are characterized by the possession of features usually considered minable features of the palate the Rhodesian plesiomorphic for Amniota. Few, if any, authors have for strict of forms agree with Eocaptorhinus. Two impor- argued tant features diagnostic of the Captorhinidae of Captorhinomorpha, and the general consensus of most fossil is that the group Heaton (1979) are the medial jugal process and workers includes the ancestors of all , by defini- the absence of the ectopterygoid. What appears tion if not by fact. For example, according to to be a medial jugal process is present in QG Carroll and Gaskill sub- 1106 but its limits are not clear. The presence (1971, p. 450): "The order a central po- or absence of an ectopterygoid cannot be deter- Captorhinomorpha occupies mined. sition in the phylogeny of reptiles. Within the group are the ancestors of most, if not all, more advanced members of the ." The Captorhinomorpha is one of the classic "primi- The anterior portions of the lower jaws are tive" or "ancestral" groups set up for supposed preserved in QG 1105, whereas the right side of taxonomic convenience that has actually im- QG 1106 has the posterior half present. The peded the development of ideas about the rela- lower jaws are tightly clenched and crushed tionships of the contained taxa by covering dorsally preventing any view of the teeth but them with a cloak of taxonomic respectability. the broken section on QG 1105 shows that a The lower level taxa referred to the Cap- single tooth row was present at that position. torhinomorpha have been relatively consistent As in other members of Group 2 (fig. 6), the from the time of Watson's (1917) first delimita- prearticular forms a horizontal, medially di- tion of the assemblage, that is, Captorhinus rected process. In the Rhodesian material, and other taxa that are morphologically similar. however, the process ends in a small nubbin Watson diagnosed Captorhinomorpha as rather than a flat plate. "cotylosaurs [i.e., ] with an obliterated 1979 GAFFNEY AND MC KENNA: CAPTORHINID 7 otic notch and vertically placed quadrates" (p. long and gradual of the typical cap- 172). These features are found widely torhinid features from the romeriid pattern." throughout the Amniota and are presumably Heaton (1979, p. 81) stated: "The sequence of primitive for that group or a large subsection of Romeria-Protocaptorhinus-Eocaptorhinus-Cap- that group. torhinus is interpreted as a single continuous Within the Amniota strict monophyly for phylogenetic lineage. . . ." This type of ap- Synapsida (Theropsida), Diapsida, and Testudi- proach to phylogeny reconstruction has been nes is assumed for the purposes of this discus- effectively criticized (Engelmann and Wiley, sion, but even ignoring fossil groups such as 1977; and more particularly McKenna, En- , does not aid in the development of gelmann and Barghoom, 1977) and Heaton's restricted outgroups or a higher level phy- phylogeny (which is virtually the same as Clark logenetic system within the Amniota. In order and Carroll's) is reformulated here in more ob- to test rigorously hypotheses of relationship for jective terms as a potentially testable hypothe- forms placed in the Captorhinomorpha some S1S. well-tested higher level system is needed within Heaton's phylogenetic tree is one of several Amniota, but it is not available. The closest that are consistent with a single cladogram (fig. thing to it is Goodrich's (1916) and 6). Testing this cladogram using unique derived Theropsida, each of which are characterized by character distributions (synapomorphies or de- a uniquely derived aortic arch condition. The rived characters) is difficult, because of the Sauropsida would contain Testudines and Diap- higher level phylogeny problems indicated sida as the of Synapsida. However, above, but not impossible. Actually, a greater Parsons's (1959) demonstration that all living difficulty is one of diagnosing the basic taxa amniotes, except turtles, have a Jacobson's used in such a study because of Carroll's and organ (as delimited by him) and that there is no Heaton's adherence to the stratophenetic evidence that non-amniotes or turtles ever had method. In this method, the determination of one, contradicts the Sauropsida and supports generic and specific level taxa generally follows the idea of turtles as the sister group of a the "recognition" of phylogenetic sequences in monophyletic Diapsida plus Synapsida. More the rock column and the resultant taxa are not work on this problem clearly is necessary and intended to be strictly monophyletic nor distin- until we have some well-tested higher level guishable solely by morphology in the absence phylogenies within the Amniota, phylogenetic of stratigraphic information. work within such paraphyletic groups as Cap- In the cladogram presented here, Group 1 is torhinomorpha will be very tentative at best. effectively the Amniota, although quite a few Nonetheless, some ideas can be formulated so taxa are missing. Amniota has the following that they are more susceptible to tests using synapomorphies: pterygoid flange and astraga- synapomorphies. lus (see Gaffney, 1979, for discussion). Nearly Clark and Carroll (1973) and especially all the missing taxa would be exclusive of Heaton (1979) presented a phylogeny of Pro- Group 2; only turtles might be members of tocaptorhinus and its near relatives. They uti- Group 2. The Captorhinomorpha is considered lized a systematic methodology that is near that by Clark and Carroll (1973) to consist of the characterized by Gingerich (1976) as "strato- "Romeriidae" and Captorhinidae.' phenetic." In this methodology, similar mor- The two best known romeriids are Pal- phologies are arranged stratigraphically and eothyris from the Middle of connected using usually implicit rather than ex- plicit criteria, to form what are interpreted as 'Heaton and Reisz (MS.) diagnose the ancestor-descendant lineages. Clark and Carroll "Romeriidae" (which they call by its more apt name, presented such a phylogeny (1973, fig. 21) and Protothyrididae) and the Captorhinidae. They attempt to remarked (ibid., p. 392): "Although much diagnose these taxa using primitive and advanced features more remains to be done in preparing and illus- and the reader should refer to their paper because it pre- trating the material, it definitely demonstrates a empts much of the above discussion on the "Romeriidae." 8 AMERICAN MUSEUM NOVITATES NO. 2688

FIG. 4. Protocaptorhinus sp., Late Permian, Rhodesia. Reconstruction of skull based on QG 1105 and QG 1106.

Nova Scotia and from the Lower niota. Although Carroll is the leading contem- Permian of Texas and West Virginia, recently porary student of the "Romeriidae" and has described by Carroll (1969) and Clark and Car- produced excellent descriptions of members of roll (1973). According to Clark and Carroll the group, he has not published a diagnosis of (ibid., p. 400): "Of all known groups of Pal- this family, nor has Heaton, who recently al- eozoic reptiles only romeriids are sufficiently tered the contents of the family by removing generalized to be ancestral to any of the subse- Romeria from it. Romer's (1956) diagnosis quent lineages." Within the romeriids, Pal- rested primarily on Romeria and Protorothyris. eothyris and Protorothyris are primitive in If we use the taxa placed in the "Romeriidae" nearly all characters used in captorhinomorph by Carroll and Baird (1972) and Clark and systematics and, to put Carroll's ideas in cladis- Carroll (1973) but minus Romeria, we seem to tic terminology, they may be hypothesized as have the following features necessary for inclu- the plesiomorphic sister taxon to nearly all Am- sion in the group: 1979 GAFFNEY AND MC KENNA: CAPTORHINID 9

fr

-pm

FIG. 5. Captorhinus aguti, Early Permian, Ft. Sill, Oklahoma, University of Kansas 9978.

1. Characters derived for Amniota such as a trans- gin of skull forms nearly straight line in dorsal verse pterygoid flange. view. 2. Large squamosal broadly covering posterior 8. Maxillary tooth row straight in lateral view. margin of cheek, forming an upright or nearly 9. Jugal lacks medial process. vertical skull margin, no otic notch. 10. Moderate bilateral parietal embayment along 3. Small, splintlike supratemporal usually inserting posterior margin. into a posterolateral lappet of the parietal, not 11. Tooth row single in upper and lower jaws. contacting postorbital. 12. Retroarticular process of lower jaw absent. 4. Tabular and postparietal small, restricted to oc- 13. Quadrate low, largely covered externally by cipital surface. squamosal and quadratojugal. 5. Lacrimal extends from nares to orbit. 6. Ectopterygoid present. All of these characters seem to be primitive 7. Lateral deflection of maxilla absent, lateral mar- for Amniota and none is unique to the 10 AMERICAN MUSEUM NOVITATES NO. 2688

I

1 1 2 I CAPTORHINIDAE I I I I I65

I 'b ~ 4 S" i 4/* I 6 I 0 I w8Wis°WX / k~~~~~~~~~~~~~~~~~~~~~c

FIG. 6. Cladogram of Captorhinidae. Numbers refer to groups discussed in text. Asterisks indicate taxa that occur outside North America (Protocaptorhinus also occurs in North America). , although a distinct genus, fits in the cladogram at the same position as Protocaptorhinus.

"Romeriidae," which is not surprising for an "continuum." Clark and Carroll (1973, p. 399) intentionally paraphyletic group. believed that "This point of division has, Group 2 is the Captorhinidae of Heaton however, the lamentable taxonomic implication (1979), which equals the Captorhinidae plus of removing the type genus from the family Romeria of Clark and Carroll (1973). This dif- "Romeriidae". Since the name Romeriidae has ference in familial content lies in the fact that long been associated with the phylogenetically Group 2 is more readily characterized and dis- most important family of reptiles and tinguished than the Captorhinidae of Clark and also honours one of this century's greatest con- Carroll (1973) and, to the stratopheneticist, it is tributors to vertebrate paleontology, an alterna- delimited by a marked break in the perceived tive point of separation should be considered." ,,gdnoif)~~~~~~~~~~~~~~~~~~~~~a

E<~ < 0

PC U'UTSQpoqu -a U U' U C

0 u .0 a .0

~~~ Cu~~~

Cu ~~~~~~~~~~~~~~0 'Cu~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~b

I. 12 AMERICAN MUSEUM NOVITATES NO. 2688

In a cladistic classification such a situation many captorhinomorphs as well as an ex- would not be a problem; by choosing the cor- tremely detailed description of this form. rect level for its application, 90 percent of am- Eocaptorhinus differs from Captorhinus solely niotes could be placed in the "Romeriidae" in in the possession of a single row of teeth rather order to make it monophyletic. In any case, than multiple rows as in Captorhinus. Bolt and Heaton makes his division excluding Romeria De Mar (1975, p. 829) have shown that at the from the "Romeriidae", and this is followed Fort Sill, Oklahoma, locality some specimens here because Romeria has the characters of have a single tooth row on one side and a Group 2. multiple row on the other and, therefore, ques- Group 2, the Captorhinidae, has these fea- tion the validity of this character in diagnosing tures that we interpret as synapomorphies: a new . Heaton (1979) has discussed this problem and argues that Eocaptorhinus is a 1. Downturned premaxilla valid taxon. The question really revolves 2. Medial process of jugal 3. Ectopterygoid absent around paleontologic interpretations of the bio- 4. Tabular absent logic species concept and, frankly, there is no real solution; rather the problem comes from A downtumed premaxilla occurs in various "seeing" too much in the rocks. Eocaptorhinus groups but the only group in which it is diagnosable using objective criteria; whether appears to be primitive is the turtles. Pely- or not it is a "true" biologic species cannot be cosaurs, some therapsids, and turtles have a answered. On the other hand, if one chooses medial jugal process but only turtles lack an not to recognize Eocaptorhinus, then Group 4 ectopterygoid. The presence of a tabular seems and 5 are merged. to be primitive for as it occurs in Group 5 has multiple tooth rows usually and , but it is absent accompanied by some expansion of the tooth in euryapsids. have a well-developed bearing elements. Captorhinus aguti is well tabular, but turtles do not. For the purposes of known through the work of Price (1935) and this study, I will accept Group 2 as a mono- Fox and Bowman (1966). phyletic group that may include turtles. Group 6 contains the captorhinids with four Group 3 is characterized by the absence of or more tooth rows, and a distinctly large max- the distinctive bilateral parietal embayment seen illary and mandibular tooth bearing surface. in Romeria and Protorothyris and instead has a This group is of particular interest here because straight posterior margin or a margin with a it includes all of the previously known extra slight median concavity rather than bilateral North American records of captorhinids. Unfor- ones. Group 3 contains as sister taxa to Group tunately, many of the included taxa are based 4 Protocaptorhinus and Labidosaurus. Pro- on fragmentary material and none of them can tocaptorhinus is a monospecific genus de- be considered well known. Captorhinikos is the scribed on the basis of two skulls by Clark and best known in the literature (Olson and Barg- Carroll (1973) with additional data by Heaton husen, 1962) and Labidosaurikos is represented (1979), whereas Labidosaurus although known by a well-preserved skull (UO 3-1-S2) so that for many (Williston, 1910, 1917) has the concept of this group presented here is never been thoroughly described and compared. based primarily on these two taxa. Group 4 is advanced over the other taxa by The North American members of Group 6 the possession of a well-developed retroarticu- include Labidosaurikos, its first-named member lar process, a feature that has appeared many (Stovall, 1950). Labidosaurikos differs from times in history but, given the mono- Captorhinus in being larger, having six maxil- phyly of Group 2, only once among these taxa. lary tooth rows and five mandibular tooth rows, Eocaptorhinus, the sister taxon to Group 5, has and having a less well-developed retroarticular recently been described by Heaton (1979) in a process, although one is present in contrast to work that provides comparative morphology for Protocaptorhinus. Captorhinikos (Olson, 1954, 1979 GAFFNEY AND MC KENNA: CAPITORHINID 13

TABLE 2 Distribution of Captorhinidaea

North America Russia Rhodesia Niger Late Permian Protocaptorhinus Moradisaurus presumed (Tatarian/Ochoan) captorhinid Middle Permian Kahneria Hecatogomphius (Kazanian, Kun- Rothia gurian/) Early Permian Captorhinikos (, Labidosaurikos SakmarianlLeonar- Labidosaurus dian, Wolfcampian) Captorhinus Eocaptorhinus Protocaptorhinus Romeria aRomer (1973, p. 161) referred to a possible captorhinid from Tasmania, but in the absence of confirmation, I am ignoring it.

1970; Vaughn, 1958; Olson and Barghusen, agree in detail with Eocaptorhinus and Cap- 1962) is known from some good material but torhinus. In particular the strong lateral cheek its differentiation from Labidosaurikos (as well deflection and the shallow median parietal em- as other Group 6 taxa) is principally on the bayment suggest identification with Group 3 basis of tooth arrangement and shape (Olson but in the absence of information on other char- and Barghusen, 1962). Captorhinikos has a acters I am only identifying it as a captorhinid larger retroarticular process than Captorhinus (Group 2). and, in general, seems to agree with that form The first discovered non-North American except in tooth row number. captorhinid was Hecatogomphius (Vyushkov Kahneria (Olson, 1962) and Rothianiscus and Chudinov, 1957; Olson, 1962), based on a (Olson and Beerbower, 1953; Olson, 1962, lower jaw with five tooth rows that was found 1965) are Middle Permian Group 6 captorhinids in Zone II of the Russian Permian (Late Per- known from somewhat damaged skulls that do mian). Later, Taquet (1967, 1969, 1972; not allow full comparison with other taxa. Both Ricqles, 1969) announced the discovery of have four or five tooth rows and the typical Moradisaurus from the Late Permian of Niger expanded palatal and mandibular surfaces. based on a lower jaw that differs from all other Kutty (1972) announced the presence of a captorhinids in having nine tooth rows. Al- captorhinomorph from the Permian of India though more material of Hecatogomphius has associated with endothiodonts. On the basis of not been forthcoming, Moradisaurus is also the endothiodonts Kutty suggested a correlation represented by good skull (Ricqles, 1969) and with the and Kistecephalus zones postcranial material as yet undescribed. of , which would be Late Permian Table 2 summarizes current knowledge of in age. Although the stratigraphic data leave captorhinid distribution. Although all but one something to be desired, the Indian form is (Protocaptorhinus) of the post-Early Permian placed in the Late Permian in the absence of captorhinids are members of Group 6, this one contradictory information (table 2). The 9 cm.- exception is enough to contradict the coinci- long captorhinomorph skull was not described dence of phylogeny with stratigraphy in this and only the dorsal view has been published as case (fig. 6). Other than hypothesizing the yet but the figure clearly shows features that widespread nature of primitive as well as ad- 14 AMERICAN MUSEUM NOVITATES NO. 2688 vanced captorhinid taxa, there seems to be little Gingerich, Philip D. of significance in the distribution as known at 1976. Cranial anatomy and evolution of early present. Tertiary Plesiadapidae (Mammalia, Pri- mates). Univ. Michigan Papers on Pal- LITERATURE CITED eont., no. 15, pp. 1-141. Goodrich, Edwin S. Bolt, John R., and Robert DeMar 1916. On the classification of the Reptilia. Proc. 1975. An explanatory model of the evolution of Roy. Soc. London, vol. 89, pp. 261-276. multiple rows of teeth in Captorhinus Heaton, Malcolm aguti. Jour. Paleont., vol. 49, no. 5, pp. 1979. Cranial anatomy of primitive captorhinid 814-832. reptiles from the early Permian. Okla. Bond, G. Geol. Surv. Bull., vol. 127, pp. 1-84. 1973. The palaeontology of Rhodesia. With a Heaton, Malcolm, and Robert Reisz section on the palynology of the Middle MS The interrelationships of captorhinomorph Zambezi Basin, by Mrs. R. Falcon. Bull. reptiles. Rhodesia Geol. Surv., no. 70, pp. 1-121. Kuhn, Oskar Carroll, Robert L. 1969. Cotylosauria. In Kuhn, 0. (ed.), Encyclo- 1969. A middle Pennsylvanian captorhinomorph, pedia of Paleoherpetology, part 6, Stutt- and the interrelationships of primitive rep- gart, Gustav Fischer Verlag, pp. 1-88. tiles. Jour. Paleont., vol. 43, no. 1, pp. Kutty, T. S. 151-170. 1972. Permian reptilian fauna from India. Na- 1977. Patterns of evolution: an ex- ture, vol. 237, no. 5356, pp. 462-463. tended example of the incompleteness of McKenna, Malcolm C., George F. Engelmann, and the fossil record. In Hallam, A. (ed.), Steven F. Barghoorn Patterns of evolution, as illustrated by the 1977. [Review of] Cranial anatomy and evolu- fossil record. Developments in paleontol- tion of early Tertiary Plesiadapidae (Mam- ogy and stratigraphy 5. Amsterdam, New malia, Primates). Syst. Zool., vol. 26, no. York, Elsevier Sci. Publ. Co., pp. 2, pp. 233-238. 405-437. Olson, Everett C. Carroll, Robert L., and Donald Baird 1954. Fauna of the Vale and Choza: 9. Cap- 1972. stem-reptiles of the Family torhinomorpha. Fieldiana:Geol., vol. 10, Romeriidae. Bull. Mus. Comp. Zool., no. 19, pp. 211-218. vol. 143, no. 5, pp. 321-364. 1962. Late Permian terrestrial , Carroll, Robert L., and Pamela Gaskill U.S.A. and U.S.S.R. Trans. Amer. Phil. 1971. A captorhinomorph reptile from the lower Soc., n.s., vol. 52, prt. 2, pp. 1-224. Permian of . Jour. Paleont., vol. 1965. New Permian vertebrates from the Chick- 45, no. 3, pp. 450-463. asha Formation in Oklahoma. Oklahoma Clark, John, and Robert L. Carroll Geol. Surv., Circ. 70, pp. 1-70. 1973. Romeriid reptiles from the lower Permian. 1970. New and little known genera and species Bull. Mus. Comp. Zool., vol. 144, no. 5, of vertebrates from the lower Permian of pp. 353-407. Oklahoma. Fieldiana:Geol., vol. 18, no. Engelmann, George F., and Edward 0. Wiley 3, pp. 359-434. 1977. The place of ancestor-descendant relation- Olson, Everett C., and Herbert Barghusen ships in phylogeny. Syst. Zool., vol. 26, 1962. Permian vertebrates from Oklahoma and no. 1, pp. 1-11. Texas. Oklahoma Geol. Surv., Circ. 59, Fox, Richard C., and Merton C. Bowman pp. 1-68. 1966. Osteology and relationships of Cap- Olson, Everett C., and James R. Beerbower torhinus aguti (Cope) (Reptilia: Cap- 1953. The San Angelo Formation, Permian of torhinomorpha). Univ. Kansas Paleont. Texas, and its vertebrates. Jour. Geol., Contr., art. 2, pp. 1-79. vol. 61, no. 5, pp. 389-423. Gaffney, Eugene S. Parsons, Thomas H. 1979. Tetrapod monophyly: A phylogenetic anal- 1959. Nasal anatomy and the phylogeny of rep- ysis. Bull. Carnegie Mus. Nat. Hist., vol. tiles. Evolution, vol. 13, no. 2, pp. 13, pp. 92-105. 175-187. 1979 GAFFNEY AND MC KENNA: CAPTORHINID 15

Price, Llewellyn I. tile captorhinomorphe (Cotylosaurien). C. 1935. Notes on the braincase of Captorhinus. R. Acad. Sci., Paris, vol. 268, pp. Proc. Boston Soc. Nat. Hist., vol. 40, no. 779-781. 7, pp. 377-386. 1972. Un example de datation et de correlation 1937. Two new cotylosaurs from the Permian of stratigraphique base sur les cap- Texas. Proc. New England Zool. Club, torhinomorphes (Reptiles cotylosauriens). vol. 16, pp. 97-102. Mem. B.R.G.M., France, no. 77, pp. Ricqles, A. de 407-409. 1969. Un reptile fossile africain approche-t-il les Vaughn, Peter P. continents? Atomes, no. 265, pp. 1958. A specimen of the captorhinid reptile Cap- 3 18-319. torhinikos chozaensis Olson, 1954, from Romer, Alfred S. the Hennessey Formation, lower Permian 1956. The osteology of the reptiles. Univ. Chi- of Oklahoma. Jour. Geol., vol. 66, no. 3, cago Press, 772 pp., Chicago. pp. 327-332. 1973. Permian reptiles. In Hallam, A. (ed.), At- Vyushkov, B. P., and P. K. Chudinov las of Paleobiogeography, Amsterdam, 1957. The discovery of Captorhinidae in the up- New York, Elsevier Sci. Publ. Co., pp. per Permian of the USSR. [In Russian] 159-167. Doklady Akad. Nauk USSR, vol. 112, Stovall, J. W. no. 3, pp. 523-526. 1950. A new cotylosaur from north central Okla- Watson, D. M. S. homa. Amer. Jour. Sci., vol. 248, pp. 1917. Sketch classification of the pre- 46-54. tetrapod vertebrates. Proc. Zool. Soc. Taquet, Philippe London, 1917, pp. 167-186. 1967. Decouvertes paleontologiques recentes Williston, Samuel W. dans le nord du Niger. In: Problemes ac- 1910. The skull of Labidosaurus. Amer. Jour. tuels de Paleontologie (Evolution des Anat., vol. 10, no. 1, pp. 69-84. vertebres). Coll. Internatl. C.N.R.S., no. 1917. Labidosaurus Cope, a lower Permian 163, pp. 415-418. cotylosaur reptile from Texas. Jour. 1969. Premiere decouverte en Afrique d'un rep- Geol., vol. 25, no. 4, pp. 309-321.

777' WM -N,7iRI!frMv' `L' II :'