UNIVtRSlTv Cp ILLINOIS I 5RARY AT URBANA-CHAMPAIGN L IOLOGY CO CO /&£^<-*x~*yw FIELDIANA Geology Published by Field Museum of Natural History Volume 37, No. 3 June 30, 1977 Cacops (Amphibia: Labyrinthodontia) From the Fort Sill Locality, Lower Permian of Oklahoma John R. Bolt Assistant Curator, Fossil Reptiles and Amphibians Field Museum of Natural History The Library of the MRRU1978 ABSTRACT at Urbana-ChamP«* The armored dissorophid (Super family Dissorophoidea ) labyrinthodont amphi- bian Cacops aspidephorus is unusual in having a large otic notch closed posteriorly by the tabular. Cacops was previously known only from the "Cacops Bone Bed," Lower Permian of Texas. Poor preservation makes this material difficult to study. Excellently preserved, though disarticulated, Cacops material has now been recov- ered from the Fort Sill fissure fills, which are probably very close in age to the "Ca- cops Bone Bed." Identification of the Fort Sill material as Cacops is based on pala- tines (primarily), armor scutes, and quadrates; the latter are here described for the first time from Fort Sill. The Cacops quadrate resembles that of other dissorophoids in having a posterodorsal process, which is unusual in the marked anterior expan- sion of its dorsal end. Comparison with other dissorophoids having a closed otic notch shows that Cacops is not unique in this anterior expansion of the process. Orientation of the process can apparently be used to distinguish trematopsids with a slit-like, closed (by the tabular) otic notch, from dissorophids. At least one such trematopsid occurs at Fort Sill, and resembles Cacops in anterior expansion of the process. Ontogeny of Cacops is discussed, and it is concluded that: 1. Closure of the otic notch by the tabular occured early in ontogeny; 2. Exclusion of the laterally-ex- posed portion of the palatine from the orbital rim, a feature unique to Cacops among known dissorophids, also developed early in ontogeny; 3. Growth-stages of Cacops available at Fort Sill were all probably post-metamorphic. All were dentitionally normal labyrinthodonts, with no trace of the "lissamphibian" dental features found in the related Doleserpeton. This may be due, however, to the fact that Doleserpet on is represented by less-mature specimens than is Cacops. INTRODUCTION Cacops aspidephorus is a large armored dissorophid amphibian. Dissorophids belong to the labyrinthodont Superfamily Dissoro- Library of Congress Catalog Card Number: 77-76917 Publication 1264 61 Uci^QG]£ 62 FIELDIANA: GEOLOGY, VOLUME 37 phoidea, consisting of Dissorophidae, Doleserpetontidae, and Trem- atopsidae (Bolt, 1969), plus Branchiosauridae and Micromelerpe- tontidae (Boy, 1972). All definitely identifiable Cacops material comes from the Lower Permian "Cacops Bone Bed" (hereafter CBB), Baylor County, Texas. The CBB was exhausted long ago, according to Williston (1911); in any case, it is now inaccessible beneath Lake Kemp. Cacops material from the CBB includes three complete and sev- eral partial skulls. The skull is unusual in having the large otic notch closed posteriorly by a ventral extension of the tabular ( fig. 1). This tabular extension meets a tall postero-dorsal process from the quadrate, and overlaps part of the lateral face of the pro- cess. Presence of a postero-dorsal process (hereafter abbreviated PDP) on the quadrate is common, perhaps universal, in dissoro- phoids (Bolt, 1969). In Cacops, however, the PDP is markedly ex- panded anterodorsally (figs. 1, 3, 4), a feature not known in any other dissorophoids. This unusual otic region and the skull as a whole, have never been well described due to the hard matrix of the CBB and the near-impossibility of tracing sutures in the available material. Two localities other than the CBB have been mentioned as pos- sibly producing Cacops. Olson (1956, p. 320) tentatively referred two humeri from, his locality FA to Cacops. This identification can- not be confirmed; the humeri are probably dissorophid, but generi- cally diagnostic characters in dissorophid humeri are not known at present. I have suggested the possible presence of Cacops at the Fort Sill locality, near Lawton, Oklahoma, based on scutes (detailed description and comparison in Bolt, 1974a) and a highly distinctive type of dissorophid palatine (briefly described in Bolt, 1974b). Posi- tive identification was not possible at that time; possible ontoge- netic changes in taxonomically important characters precluded defi- nite identification using armor alone, even though the armor resem- bles that of Cacops quite closely. Evidence for assignment of the palatines to Cacops was only indirect, and thus inconclusive, because of the difficulty of tracing sutures in then-available CBB specimens. In an effort to find suitable comparative material, preparation of the last of Williston's CBB collection was undertaken during this study. Several useful specimens of Cacops (among other genera) were discovered, including two separated palatines and a quadrate. Most of the CBB material has now been at least rough-prepared. BOLT: CACOPS FROM LOWER PERMIAN 63 PDP 0 5 cm Fig 1. Cacops aspidephorus from the Cacops Bone Bed, FM UC 649. Outline drawing of skull in left lateral view; based on right side, reversed. Area within otic notch ( shaded) is roughly cone-shaped, with base of cone directed laterally. Sutures within the otic notch are unknown, but most of the area anterior to the postero- dorsal process of the quadrate is occupied by the squamosal. Abbreviations: EN, external naris; PDP, postero-dorsal process of quadrate; T, tabular; X, location of distal end of stapes. In addition to previously-reported material, possible Cacops quadrates have now been identified at Fort Sill. This paper com- pares the CBB palatines and quadrates with those from Fort Sill, and concludes that the evidence supports identification of the Fort Sill specimens as Cacops. Although all Fort Sill Cacops material recovered to date is disarticulated, it is very well preserved and covers a considerable size-range; it can thus make a significant contribution to our knowledge of osteology and ontogeny in Cacops. As a dissorophoid, Cacops is part of a radiation that may have given rise to some or all of the living amphibians (lissamphibians; this is a collective term which does not imply monophyly). The pos- sibility of relationship was originally supported mainly by the pres- sence of bicuspid, pedicellate (sensu Parsons and Williams, 1962) marginal and fang teeth in Doleserpeton (Bolt 1969). This genus, sole member of the Doleserpetontidae, is known only from Fort Sill. Similar dental features have recently been found in some dissoro- phids (Bolt, in press), and in Doleserpeton itself at least some of the palatal denticles are now known to be pedicellate (Bolt, ms.). A survey of dissorophoids for such "lissamphibian" characters is desirable, but very difficult with most types of preservation. The fine preservation of Fort Sill specimens has made it possible to study the palatine denticles of Cacops; and, as explained below, the morphology of marginal teeth can at least be inferred. 64 FIELDIANA: GEOLOGY, VOLUME 37 "Skull length" in this paper refers to length as measured in the midline and in the plane of the skull table. Measurements were taken between the tip of the snout and the posterior end^ of the postparietals on the skull table. My thanks to J. Hopson, who critically read the manuscript, and to Z. Jastrzebski, who did the drawings. I thank Dolese Brothers Company for permission to collect in the Richards Spur Quarry. MATERIALS Specimen repositories are indicated as follows: AMNH— Ameri- can Museum of Natural History; FM — Field Museum of Natural History; KU— Kansas University Museum of Natural History; MCZ— Museum of Comparative Zoology, Harvard University. The Fort Sill locality is a commercial limestone quarry in SW% section 31, T4N, R11W, Comanche County, Oklahoma. Quarry operations periodically expose clay-filled fissures, some of which contain great numbers of small tetrapod bones. The fissure fills are discarded as waste by the quarry operators, and most specimens recorded from Fort Sill have come from the clay dumps. The age of the fissure fills is unquestionably Lower Permian, and probably close to that of the Arroyo Formation of Texas (Olson, 1967). Bones from Fort Sill are usually disarticulated, but otherwise excellently preserved. The CBB is on Indian Creek in Baylor County, Texas. The fauna is of Lower Permian age; according to Olson (1958), the CBB is in either the upper Arroyo formation or the lower part of the Vale. The original size of the pocket was about 6 ft. by 10 ft., and 2 ft. thick (Williston, 1910, 1911). MORPHOLOGY AND COMPARISON Palatine— Dissorophid palatines from Fort Sill, tentatively re- ferred to Cacops, were described and figured previously (Bolt, 1974b, text-figure 8); the figure is reproduced here in Figure 2. These palatines have a laterally-exposed portion (LEP), as do most (perhaps all) dissorophoids. A unique feature is exclusion of the LEP from the orbital rim. This presumably occurred by means of a secondary union of the lacrimal and jugal dorsal to the LEP. Two partial skulls from the CBB (FM UR 2431 and UC 900) show a LEP shaped like that in Figure 1, with a sutural separation between the LEP and bone dorsal to it. The orbital rim is not preserved in BOLT: CACOPS FROM LOWER PERMIAN 65 PTC PTC? Fig 2. Right palatine of Cacops, from the Fort Sill locality; FM PR 1034. Top, in dorsolateral view; arrow indicates anterior. Bottom, in palatal view. Large denticles occur medial and anterior to the fang teeth; small denticles (indicated only schema- tically) cover most of the palatal surface. Abbreviations: CH, area bordering choana; EC, contact area for ectopterygoid: LEP, laterally-exposed portion of palatine; PF, area occupied by fang tooth (two such areas are shown); PTC, contact area for pterygoid; VC, contact area for vomer. either case, but the LEP was clearly excluded from it. Similarly, two parallel cuts 5 mm. apart through the orbital floor of FM UR 2430 show a LEP outlined by sutures and separated from the orbital rim by bone. Finally, a separated palatine (FM UR 2433) recently recovered from the CBB collection very closely resembles the Fort 66 FIELDIANA: GEOLOGY, VOLUME 37 Sill palatines in shape of the LEP and its apparent exclusion from the orbital rim. A second Cacops palatine from the CBB (FM UR 2432) is exposed only in ventral view. Thirteen Cacops-like palatines are presently known from Fort Sill. The largest, FM PR 1034, is complete (fig. 2) and about 3.5 cm. long. This is about the same size as the separated palatines from the CBB; exact comparison is impossible, since the thin anterior and posterior portions of both CBB specimens are apparently some- what damaged. Most of the Fort Sill specimens are broken, some consisting of little more than the LEP. The smallest measurable specimen is about 2.3 cm. long, and is considerably smaller in all dimensions than the CBB palatines. Some of the less-complete specimens were much smaller, probably about a centimeter long. The LEP is similarly-shaped in all Fort Sill palatines, and was clearly not exposed within the orbit. Most of the ventral surface of each Cacops-like palatine from Fort Sill is covered with monocuspid denticles (fig. 2); preparation has necessarily removed denticles from the CBB specimens, so no com- parison is possible. No trace of pedicely is visible in the Fort Sill specimens. This was confirmed by examining a small palatine with the scanning electron microscope. The denticles are largely intact, but show no trace of pedicely. The palatal surface shows clear contact areas for the ectoptery- goid (presumably) and vomer (fig. 2). At least the posterior part of the pterygoid contact area is coarsely striated; the original descrip- tion (Bolt, 1974b) concluded that the pterygoid did not extend much beyond that area, and thus did not reach the vomer. This conclusion now seems less probable; the very thin medial border of the palatine suggests that it may have been supported along its length by a nar- row anterior extension of pterygoid. The CBB specimens are not well enough preserved to provide a check on this suggestion. If Cacops does have a pterygoid-vomer contact, it may not be the only dissorophid with this primitive condition. Only in Tersomius tex- ensis is it certain that there is no pterygoid- vomer contact (and even here there is uncertainty as to the exact position of all sutures around the pterygoid— cf. Carroll, 1964; Bolt, 1974b). The anterior relationships of the pterygoid are obscure in all other described dissorophids. Quadrate— A total of 14 quadrates has been found which prob- ably pertain to Cacops; all have a strong postero-dorsal process BOLT: CACOPS FROM LOWER PERMIAN 67 H 5mm Fig. 3. Cacops quadrate from Fort Sill, FM PR 1032, medial view in plane of pos- tero-dorsal process. Abbreviations: AB, points between which (maximum) process length was measured; CD, points between which basal length was measured;/? , the process angle, is measured between the posterior border of the postero-dorsal pro- cess, and the line formed by the junction of contact areas for the pterygoid (medi- ally) and the squamosal + quadratojugal (laterally). (PDP). In the largest of the quadrates, the PDP is markedly expanded anteriorly (fig. 3). Some expansion is visible even in the smallest examples (fig. 4). This type of expanded PDP can be seen in the Cacops skull ( FM UC 649) figured by Williston ( 1910), and in a separated quadrate from the CBB (FM UR 2434) which is about 68 FIELDIANA: GEOLOGY, VOLUME 37 0 3mm Fig. 4. Cacops quadrate from Fort Sill, FM PR, 1033, medial view in plane of postero-dorsal process. the same size as that in the skull. The Fort Sill quadrates are consid- erably smaller with, for example, a process length in the largest (FM PR 1032; see fig. 3) of 1.2 cm. as compared to about 2.1 cm. in FM UR 2434. Similarly, FM PR 1032 has a maximum basal length (see fig. 3) of about 1.5 cm., compared to 2.1 cm., for FM UR 2434 from the CBB. Both the CBB and Fort Sill quadrates have a process BOLT: CACOPS FROM LOWER PERMIAN 69 angle (see fig. 3) of about 75-85°, as measured in FM PR 1032 and 1033 and UR 2434. The quadrate is sufficiently well preserved in a number of dis- sorophoids, to establish the presence of a PDP. It is usually not possible to be certain whether the process is expanded as in Cacops, and no similar examples have been reported. Such an expansion is clearly absent in Doleserpeton which is, however, a much smaller animal than Cacops and has an open otic notch. Cacops is more ap- propriately compared with other large dissorophoids having a closed otic notch. There are only three such species: Longiscitula houghae, Dissorophus multicinctus (both dissorophids), and Trematops milleri (a trematopsid). These are also the only dissoro- phoids known to have a closed otic notch, with the exception of the Russian dissorophid Zygosaurus lucius which I cannot consider because there is no useful published information on it and the only known specimen has been lost. The three dissorophids (including Cacops ) have very large otic notches closed by a tabular extension which, in lateral view, stands at about 90° to the skull table. Long- iscitula (FM UR 430; skull length about 12 cm.) has a PDP, and there is some indication of expansion, but the area is very poorly preserved. Dissorophus multicinctus (MCZ 2122-1; skull length about 13.5 cm.) has an expanded postero-dorsal process, but the quadrate is so damaged that the process angle and length cannot be obtained. The process in D. multicinctus seems much thicker than that of Cacops, and not as high. The latter feature could be partly due to breakage; however, D. multicinctus has a considerably lower skull than Cacops, and therefore a lower PDP would be expected. The process angle in Longiscitula and D. multicinctus is probably close to that of Cacops, since all have a nearly-vertical tabular extension. A high process angle may be characteristic of dissoro- phids, based on these genera plus Tersomius texensis (unexpanded, relatively low PDP on small skull associated with holotype, AMNH 4719; process angle about 85°, like that in Doleserpeton). In con- trast to the dissorophids, Trematops milleri (fig. 5) has a slit-like otic notch and the posterior tabular extension is at about 45° to the skull roof. The quadrate in such trematopsids should be readily distinguishable from that in all dissorophids, based only on process angle. This seems to be the case: the process angle in T. milleri FM UC 1760 is about 35°. There is apparently no anterior expansion of the process in T. milleri. Some trematopsids may, however, have an expanded process: a single quadrate from Fort Sill (KU 34652) O Tlfl CO fl) -6 S g c 2 S og - 2 ftfe IS A" b 08 c as . 01 CD .1 * S5 53 .2 co •Cog ^ ° s .s * £ 9 cu -Q | S . oo -d b <*h « * 0 S. s 0 cu w •B j§ ^ > *j co § « ° 1 8 S 3 Ih u O ft o © S ft LOO B 2 § 2 •§ , ft ft.e K S cu 3 C co « -a -3 "2 fa £ 1 i o a cr 70 BOLT: CACOPS FROM LOWER PERMIAN 71 has an anteriorly-expanded PDP with a process angle of 35°, which suggests trematopsid rather than dissorophid affinities. DISCUSSION AND CONCLUSIONS Identification of Cacops. — I believe that the palatine provides convincing evidence for identification because it is a unique type on the basis of comparison with most other dissorophoids ( the rela- tionships and shape of the LEP cannot be determined in some species). The armor is also good evidence, because it is reasonably distinctive and can be compared with that of other armored dissoro- phids (cf. Bolt, 1974a). The quadrate provides the weakest evidence due to lack of comparative information for other dissorophoids, although so far as known it resembles that of Cacops most closely. The quadrate region of dissorophoids and labyrinthodonts in gener- al will be treated in greater detail in a future paper. In addition to the comparisons given above, another line of rea- soning suggests that the isolated bones described are best assigned to Cacops: The palatine and quadrate are easily distinguished from those of the unarmored Doleserpeton, the most common dissoro- phoid at Fort Sill. The next most common dissorophoid elements (although actually rare) are those here assigned to Cacops. Other dissorophoid armor scutes, quadrates, and palatines are extremely rare in collections made to date. It is thus likely that the elements assigned to Cacops at least belong to the same species. Identifica- tion of the Cacops palatine makes it virtually certain that the species is Cacops cf. C aspidephorus. Ontogeny of Cacops— The Fort Sill specimens, with the exception of the largest palatine, represent animals considerably smaller and therefore probably younger than those known from the CBB. Just as clearly, the Fort Sill animals were not larval-sized; as a rough estimate, skull lengths probably lay in the range from about 12.5 cm. (the size of most CBB specimens) to 6 or 7 cm. The shape and (especially) the relationships of the LEP are po- tentially valuable taxonomic characters, but LEP ontogeny must be better known before they can be used with confidence. It is inter- esting that over a considerable size range, these characters are stable in Cacops. It is therefore most unlikely that any of the numerous dissorophid species with skull lengths of about 6 cm., and a LEP which participates in the orbital rim, are juveniles of Cacops. 72 FIELDIANA: GEOLOGY, VOLUME 37 As noted above, a closed otic notch is restricted to the larger dis- sorophoids. (Not all large dissorophoids appear to have closed otic notches, but this requires further study.) This may simply reflect failure to find juvenile skulls of the species with closed otic notches. On the other hand, it seems at least as likely a priori that otic-notch closure may have occurred rather late in ontogeny; some of the spe- cies with open otic notches may thus be represented only by juve- niles. Two types of evidence suggest that Cacops, at least, devel- oped otic notch closure at a rather small skull size: direct evidence comes from FM UR 2435, a partial left half of a Cacops skull from the CBB discovered in the course of this study. The tip of the snout is damaged in this specimen, but the position of the orbit and the posterior end of the postparietal is clear. Skull length can be esti- mated as about 6 cm., or no more than half the size of the Cacops skulls described by Williston (1910) which were until now the only ones available. This skull is thus smaller than that of most other dissorophoid species (see text-figure 1 in DeMar, 1968; also Olson, 1941). Although the quadrate is missing in this specimen, the pre- sence of a long ventral extension from the tabular indicates that it almost certainly had a closed otic notch. Indirect evidence comes from the Fort Sill Cacops quadrates. All have a rugose area, V- shaped in outline, on the lateral face of the PDP. This area begins at the dorsal border of the process and runs less than one-third of the way down; it probably marks the contact of the tabular extension with the quadrate. Presence of this area in even the smallest quad- rates, thus suggests that the otic notch was closed even in small Ca- cops specimens. Presence of pedicellate denticles in Doleserpeton may be transi- tory, and characteristic of a juvenile, early post-metamorphic stage (Bolt, in press). The non-pedicellate palatine denticles of Cacops do not indicate whether or not that genus passed through a similar stage, in as much as the available palatines are all probably from more mature individuals than are the Doleserpeton specimens. Isolated marginal tooth-bearing bones of Cacops are not identifi- able at present in the Fort Sill material. It is virtually certain, how- ever, that at least some of the labyrinthodont maxillae, premaxillae, and dentaries at Fort Sill pertain to Cacops. None of the labyrintho- dont teeth (except those of Doleserpeton) at Fort Sill are either bi- cuspid or pedicellate. Thus Cacops was probably a dentitionally normal labyrinthodont, in the growth-stages preserved. This does not exclude the possibility of a more or less Doleserpeton-\ike mar- BOLT: CACOPS FROM LOWER PERMIAN 73 ginal dentition at younger stages as observed in Tersomius and "cf.Broiliellus sp." (Bolt, in press). REFERENCES Bolt, J. R. 1969. Lissamphibian origins: possible protolissamphibian from the Lower Per- mian of Oklahoma. Science, 166, pp. 888-891. 1974a. Armor of dissorophids (Amphibia: Labyrinthodontia): an examination of its taxonomic use and report of a new occurrence. Jour. Paleontol., 48, pp. 135- 142. 1974b. Evolution and functional interpretation of some suture patterns in Paleo- zoic labyrinthodont amphibians and other lower tetrapods. Jour. Paleontol., 48, pp. 434-458. In press. Dissorophoid relationships and ontogeny, and the origin of the Liss- amphibia. Jour. Paleontol. Boy. J. A. 1972. Die Branchosaurier (Amphibia) des saarpfalzischen Rotliegenden (Perm, SW-DeutschlanH). Abhandl. hess. L.-Amt. Bodenforsch, 65, pp. 1-137. Carroll, R. L. 1964. Early evolution of the dissorophid amphibians. Bull. Mus. Comp. Zool., 131, pp. 161-250. DeMar. R. E. 1968. The Permian labyrinthodont amphibian Dissorophus multicinctus, and adaptations and phylogeny of the family Dissorophidae. Jour. Paleontol., 42, pp. 1,210-1,242. Olson. E. C. 1941. The family Trematopsidae. Jour. Geol., 49, pp. 149-176. 1956. Fauna of the Vale and Choza: 1 1. Lysorophus: Vale and Choza. Diplocaulus, Cacops and Eryopidae: Choza. Fieldiana: Geol, 10, pp. 313-322. 1958. Fauna of the Vale and Choza: 14. Summary, review and integration of the geology and the faunas. Fieldiana: Geol., 10, pp. 397-448. 1967. Early Permian vertebrates. Okla. GeoL Survey, Circular 74, pp. 1-111. Parsons. T. S., and E. E. Williams 1962. The teeth of Amphibia and their relation to amphibian phylogeny. Jour. Morphol. 110, pp. 373-390. WlLLISTON, S. W. 1910. Cacops, Desmospondylus; new genera of Permian vertebrates. Jour. Geol., 21, pp. 249-284. 1911. American Permian vertebrates. Univ. Chicago Press, Chicago. 145 pp.