€*4 O CO Cu I LU « GO « UNIVERSITY OF ILLINOIS LIBRARY AT UP3MA-CHAMPA1CN GFOLOGY k *' FIELDIANA Geology J^ss&aas; Published by Field Museum of Natural History Volume 30, No. 3 August 29 ,0?4 A Trematopsid Skull from the Lower Permian, and Analysis of Some Characters of the Dissorophoid (Amphibia: Labyrinthodontia) Otic Notch John R. Bolt Assistant Curator, Fossil Reptiles and Amphibians Field Museum of Natural History INTRODUCTION The partial skull (UR 2400) described in this paper came from the Lower Permian Fort Sill locality, Comanche County, Oklahoma. This locality has been described elsewhere (see Gregory et ah, 1956; Olson, 1967). Olson's faunal list contains eight genera, of which all but one (Xenacanthus) are small tetrapods. This list can be expanded to 11 by the addition of Basicranodon (Vaughn, 1958), Phlegethontia (McGinnis, 1967), and Doleserpeton (Bolt, 1969). As Olson notes, there are certainly a number of undescribed forms in the fauna, but the usually disarticulated condition of materials and dispersal of col- lections hampers study. With the exception of Doleserpeton, UR 2400 is the first example of even a partial labyrinthodont amphibian skull described from Fort Sill. The specimen is thus important in providing a guide to associa- tion of well-preserved but disarticulated labyrinthodont dermal bones from Fort Sill. Excellent preservation of this skull fragment makes it useful in understanding the morphology of other dissorophoid labyrinthodonts, which are mostly known from less well-preserved red-beds specimens. (The superfamily Dissorophoidea was originally composed of Dissorophidae, Doleserpetontidae, and Trematopsidae (Bolt, 1969). Boy (1972) has added the Micromelanerpetontidae and Branchiosauridae, neither of which will be discussed here due to unavailability of material). Finally, UR 2400 is of interest as a possibly new species of trematopsid. Library of Congress Catalog Number: 7U-82821 Publication 1189 67 i_r- I W ATM/ 68 FIELDIANA: GEOLOGY, VOLUME 30 I am grateful to H. Barghusen, R. DeMar, and J. Hopson for discussion and criticism. S. Grove did the figures. Abbreviations used in this paper for repositories of specimens are: AMNH— American Museum of Natural History MCZ— Museum of Comparative Zoology, Harvard University UC or UR— Field Museum of Natural History UT— University of Texas Memorial Museum. Materials Studied— All are from Lower Permian deposits. A com- plete list is given in Table 1. DESCRIPTION OF UR 2400 The roofing bones show well-developed, typically labyrinthodont sculpturing. No attempt has been made to reproduce this sculpturing exactly in the figures. The sculpturing suggests that UR 2400 was not a juvenile (cf. Bystrov, 1935), but gives no information as to maximum adult size. There is no development of bony knobs or ridges on the outer skull table (cranial ornamentation in the sense of Carroll, 1964). Presence or absence of ornamentation in a single specimen is probably of low taxonomic importance, however (cf. DeMar, 1968, on variation in this character). Some, perhaps all, sutures are strongly interdigitated. Small clusters of pyrite crystals (not shown) were scattered over the surface, especially on the outer skull table, but occupied only a small percentage of the total area. Although the pyrite tends to occur along suture lines and obscure short segments of them, virtually all sutures could be accurately determined. The skull fragment as preserved is shown in Figures 1-3. The tabulars are broken off posteriorly, but the natural borders of tabu- lars and postparietals have been preserved on the occiput. Several points which are not clear from the figures, or which will be discussed below, merit comment: 1. The bone around the orbits is thickened. This is achieved with- out producing a very noticeable circumorbital ridge. Thickening is greatest at the level of the postfrontal-postorbital contact. 2. The squamosal bears a thin, transverse process (FSQ in fig. 3) in front of the lamina ascendens of the pterygoid, and presumably made contact with it originally. 3. The postfrontal-postorbital contact is strongly bevelled in the orbital rim, with the postfrontal underlying the postorbital. 4. The lamina ascendens of the pterygoid is preserved on each side, apparently in situ as the orientation is identical on both sides. BOLT: TREMATOPSID SKULL 69 Each stands vertically and at approximately right angles to the midline. The lamina is paper-thin dorsally, and ends just short of a contact with the skull roof. 5. There is a pronounced thickened area or boss (B in fig. 3) on the ventral surface of each postparietal. Each boss lies just anterior to an occipital flap of the postparietal, of the type which in labyrinthodonts usually covers the upper part of the occipital aspect of the exoccipital. The boss is therefore almost certainly directly dorsal to the exoccipi- tal. Direct contact may have existed between boss and exoccipital; more likely, a supraoccipital or part of the opisthotic (cf . Bolt, 1969) intervened. 6. Just lateral to the boss on the postparietal is a process (partially preserved, and only on the left side) on the tabular. This process (OP in fig. 3), directed ventro-medially, may have made contact with the parocipital process of the opisthotic. There is no other indication on the tabular of a possible opisthotic contact. 7. The otic notch is deep. Its dorsal border is formed by a vertical flange composed of squamosal, supratemporal, and tabular. Such a flange will hereinafter be referred to as the supratympanic flange (SF in figs. 2-4). The squamosal and tabular meet below the supra- temporal, excluding it from the ventral margin of the flange. Extend- ing laterally above the flange, the skull table forms what may be called the supratympanic shelf (SS in fig. 3). The shelf is here defined as existing only if there is a supratympanic flange, or at least a semilunar flange on the supratemporal (see below). The shelf is damaged on the right side. The supratemporal extends into the supratympanic flange as a rounded projection. Such a projection of the supratemporal above the otic notch, whether or not associated with a supratympanic flange, is called a semilunar flange (SLF in figs. 2 and 4). The terminology adopted is based on, but different from, that of DeMar (1968). DeMar observed that dissorophids generally have a distinctively-shaped supratympanic flange, which he called a "half -moon-shaped flange" (see fig. 4). The same condition is re- ferred to here as a "supratympanic flange of dissorophid shape." DeMar (1968, p. 1212) noted that ". . . Trematops [milleri] does have a ventrally directed ridge on the dorsal margin of the otic notch which narrows it from above," but he did not name the flange in Trematops. 8. The antero-ventral wall of the otic notch is formed entirely by the squamosal. This part of the notch is roughly cone-shaped with the apex of the cone pointing medially. 70 FIELDIANA: GEOLOGY, VOLUME 30 F 1cm Fig. 1. Dorsal view of partial trematopsid skull, UR 2400. Abbreviations: F, frontal; M, matrix; P, parietal; PF, postfrontal; PO, postorbital; PP, postparietal; SQ, squamosal; ST, supratemporal; T, tabular. SYSTEMATIC POSITION OF UR 2400 There is little doubt that the specimen is a dissorophoid labyrin- thodont. There are no features inconsistent with this conclusion, and it is supported by the presence of a supratympanic flange and by the shape of the squamosal. These features do not indicate a particular position within the Dissorophoidea, however. The affinities of UR 2400 within the Dissorophoidea are examine J in more detail beiop . using quantitative and qualitative charact * s. An attempt to distinguish the three families within the Dissoro- phoidea on the basis of postorbital skull proportions was unsuccessful. The three types of measurements in Table 1 were graphically com- pared as follows: Lpb:Lt, L pb: W, and Lt:W. The three families could not be distinguished from one another nor from UR 2400 on this basis. This is interpreted as strengthening the argument for relation- ship of UR 2400 to the Dissorophoidea. These prop^rt" ~ns are prob- BOLT: TREMATOPSID SKULL 71 ably not unique to dissorophoids, however, even though apparently not common among rhachitomes. Text-figures in Moustafa (1955), for instance, suggest that Parioxys ferricolus (family Parioxyidae) resembles dissorophoids in the parameters measured for Table 1. Parioxys can, of course, easily be distinguished from dissorophoids when other skull proportions are considered. CHARACTER ANALYSIS The systematic position of UR 2400 can apparently best be deter- mined by use of qualitative characters. Comparison with other dis- sorophoids suggests several characters that might usefully be exam- ined (table 1). The states defined for each character will be analyzed below as primitive or derived, as advocated by Hennig (1966). Hennig's method of phylogenetic reconstruction requires that a number of characters be so analyzed. There are presently very few properly analyzed characters available for dissorophoids; thus no phylogenetic reconstruction is given in this paper. Boy's (1972) tentative analysis of sister groups within the Dissorophoidea is in- teresting but premature. The advantages and theoretical basis of Hennig's system have been treated in many recent publications (e.g., Brundin, 1966; Nelson, 1970; Cracraft, 1972; SchaefFer et al., 1973), and the arguments pro and con will not be repeated here. However, a note on methodology is necessary: This paper basically follows the practical procedure outlined by Marx and Rabb (1970, 1972). Criteria for determining primitive and derived character states are modified from those pre- sented by Marx and Rabb. Criterion 1 is uniqueness: character states occuring in the "descendent" group (in this case, dissorophoids), but not in the "ancestral" group (see below) are derived. Criterion 2 is relative abundance: character states found in both the "ancestral" and "descendent" groups are primitive. The superfamily Edopoidea is generally considered ancestral to the Eryopoidea, which in turn appears the most likely group of origin for the Dissorophoidea. Accordingly, the edopoids and eryo- poids have been taken as the "primitive" reference groups. In prac- tice, due to the lack of good descriptions of the region under consider- ation, three reasonably well-described genera will represent the two superfamilies. The Edopoidea are represented by Edops (Romer and Witter, 1942) and Dendrepeton (Carroll, 1967); the Eryopoidea, by Eryops (Sawin, 1941). Addition of a number of less completely de- scribed genera would strengthen the analyses only in appearance. CD O, V*"S co o ^5 fc ea & 9-.. UK'S _I_P- C^ <! C~ I <1 CD CO cfl 73 a. a S~ cu a c » a 3 s-< '3 o o .5 t- o J! 53 .3 Mh c CS 53 fH B c3 S II c<1 ^,i z 0 <d c +j Cu rj 1 1 £ § fe 01 -u £ c 15 ii 'S 03 «-, -tj 0?> yj -■ -c a c3 3 II pq < H 1 ' ^ I •• II 'El "Is- II |j|+++ 1 ++ + +s + + + ++r+s + o £ § CD.. . S| ■ ||+gT g ++++<+++ x+%+£ + V CD ^§ *£w co w 3 i- cd cr I" g^^+* g ++ + ++ + + + +++++ + r^ ft "«)H ©+» O0CO 00 ■<*"<* <I t> t> o I <J<Of- 00 ^-coco cm S ioio oo wt- fe» c- eo tj< I £t-eo co OOiOO rH 00 CO 05 IOCO <J 00 t> y-t I <I CD f\J <M i-l i-l rH COCO H COCO JgJ ^* *""• CN ' 'Z "* <M 00O CM COCM <3 CD CM CM I CD "^ I MH rH i— I CM ^ CM rH i-l I rt^rH Oi •*}< o. a, HH >> >> NN o o ^°o "o as* M ft «s a co co co *8 -i P« lo cd +? :* -^ Sqo>> oo a o «o o gt- ^ -g 0"5 j3 KrfS s >2So-g^yoo2 ^ja 8 gs •§ ■HHgcogN,go,g co .c* r^^ v co s sLSJ^)X?' BHH^ H H^_„-« ^ g S ' eg a^ w g .•§ S <*>Q}Ci>!lu ca nufc^, T3 S CM ' ' ^S^Si^S ■«■«■« -^t -as •is^scMe -S *= »N» .s .saS'ss eoeoeowH »- » u^v^ BcO- Sh S Jj § J h » u S BS siiis aoM8iSMS|oPiljl|i|S . co co co "33 •^^r-^-^i«H g1^ o >^>zr w>r-icooSo,2?'»'as»S^r 72 s> II 2 >>j= S. ;* to £& O 3 o ea-S 2 c ca s a *-■ «,= Q. a ••3 ■gji eS « tn C boo lis « 5. s O 3 00 + + + h t + + 50 15S oo a; oo o J3 "if 2oo Is e£ e - o a> IS I; o So 0) o 5 s a bo 3 O — ■5 ■u 2 3 c 8 = .2 -t-> Xi 0> 3 ..- 3 T5 a> £ -^ E o 2 £ -3 .2 hi o £ 2 - 3 -u .2 8 — 0, °o ■3-w S-w e« 3 *a is b _ p - C 0) OJS •2^ «§•« "m 3,2 jh^a 2 <« o'a « o> p c 3 00 3 o J3 T3 « 3 o 0 x> 0 O J3 0 a .2 3 O -t-> 3 O "5 "8 3 - 0 a Ph O 3 l-H 5 O 3 2 a> E a> 0) I E 3 a> £ 0) N-l U3 T3 3 - — 8 S£ -go •3 3 « w. 5 3 rf <" 3T> SB 0) '/. — O C8 as is 2=3 §3£3 S *Sa g S 3.2 I sSo IIP"! ^o-S^-2 o o JSJS o> o a) £ 3 .2^3 e o> s jMS? 3' 55 JS-^ v«.Bi 3* 3 j- c 3<~ o. 4) /i> t: t: *h 35? 73 74 FIELDIANA: GEOLOGY, VOLUME 30 1cm Fig. 2. Lateral view of partial trematopsid skull, UR 2400. Vertical shading in- dicates broken surfaces. Abbreviations: J?, possible fragment of jugal; SF, supra- tympanic flange, including part of squamosal, supratemporal, and tabular; SLF, semilunar flange of supratemporal. Other abbreviations as for Figure 1. The tabulation below includes most North American dissorophoid genera and species. Missing are Ecolsonia (Trematopsidae) and Amphibamus (Dissorophidae) . Amphibamus has recently been re- viewed by Carroll (1964), but the nature of the material severely limits interpretation of the region around the otic notch. This is unfortunate in view of Amphibamus' position as the oldest and generally most primitive dissorophid (Carroll, 1964). Particularly in the case of Doleserpeton, which is small and is also represented by more or less immature specimens, the effects of abso- lute size and of degree of maturity may be important. This is simply a cautionary note; nothing can be said about these problems at present due to the absence of proven growth series. All specimens are here treated as mature for purposes of analysis. Character 1: supratympanic flange— Four states: (a) absent; (b) intermediate; (c) present and trematopsid shape; (d) present and dissorophid shape. It must be emphasized that the difference in shape between the flanges of (c) and (d) is due to the presence (fig. 4) or absence (fig. 2) of a semilunar curvature of the squamosal along the ventral border of the supratympanic flange. In UR 2400 (fig. 2), the BOLT: TREMATOPSID SKULL 75 Fig. 3. Ventral view of partial trematopsid skull, UR 2400. Vertical shading in- dicates broken surfaces. Abbreviations: B, boss on postparietal; FSQ, flange on squamosal; LA, lamina ascendens; MB, markings associated with contact between braincase and skull roof; OP, process from tabular, presumably for contact with opisthotic; SF, supratympanic flange, composed of parts of squamosal, supra- temporal, and tabular; SS, supratympanic shelf. Other abbreviations as in Figure 1. flange ends posterior to the squamosal-tabular suture by rising to merge with the undersurface of the tabular. Particularly if the skull table were bent downward, this inflection could be mistaken for a semilunar curvature of dissorophid type. None of the three reference genera shows any sign of developing a supratympanic flange, nor indeed is there any known development of the flange in any other labyrinthodont, as noted also by DeMar (1968). Absence of the flange (a) is therefore primitive; the intermediate condition (b), seen only in the early dissorophid Tersomius, is reasonably interpreted as derived. Still more derived are states (c) and (d), but it is presently impossible to determine whether one was derived from the other. The apparently straight ventral edge of the weak flange in Tersomius provides some evidence that state (c) may be more primitive than (d), but this is inconclusive. Tentative phylogeny of these states is 76 FIELDIANA: GEOLOGY, VOLUME 30 thus: (a) ►(b): Character 2: squamosal-tabular contact— Two states: (a) absent and (b) present. State (a) applies to species which completely lack the contact. State (b) applies to species in which there is slight contact, or to species with a well-developed contact such as that in UR 2400 (fig. 2). All dissorophoids in which a squamosal-tabular contact is definitely present resemble UR 2400. It is uncertain whether or not contact is present in Tersomius texensis. Longiscitula houghae is de- scribed by DeMar (1966) as possessing state (a), but re-examination shows that state (b) is present. There is definitely a slight contact of squamosal and tabular in Eryops, and possibly in Edops and Dendrer- peton as well. State (b) is thus presumably primitive, state (a) derived. Character 3: semilunar flange of supratemporal— Two states: (a) absent and (b) present (SLF in figs. 2 and 4). The condition in Edops is uncertain; in Dendrerpeton and (particularly) Eryops, a more or less wedge-shaped flange is present. The available evidence thus suggests that state (b) is primitive, state (a) derived. Character U' supratympanic shelf— Three states: (a) absent, (b) intermediate, and (c) present. State (a) applies to specimens with no projecting shelf or ridge, or very slight development of a ridge. This state also applies to specimens (Dissorophus multicinctus) in which dermal sculpturing extends downward over the semilunar flange, with no development of a shelf. Apparent absence of the shelf may be secondary and due to breakage. As noted for UR 2400 in Table 1, the loss of a supratympanic shelf by breakage can significantly affect measurement of table width (W). Secondary loss of the shelf can also result in a significant change in shape of the skull table— compare right and left sides of UR 2400 in Figure 1. Skull- table shape is a potentially useful character, but must obviously be used carefully. A specimen lacking the shelf but showing signs of damage in the area is scored (-?); probably most of these were originally state (b). State (b) applies to specimens in which there is no shelf as such, but a distinct ridge of dermally-sculptured bone. State (c) applies to speci- mens such as UR 2400, where a shelf is present and its undersurface, where visible, is unsculptured. The semilunar flange of dissorophoids is presumably homologous to the more or less wedge-shaped supratemporal flange seen in the reference genera. Projection of the skull table laterally beyond the BOLT: TREMATOPSID SKULL SQ Fig. 4. Broiliellus olsoni, UT 3189-8. Left lateral view of part of skull of holo- type, to show dissorophid shape of supratympanic flange. Supratympanic flange, in- cluding all of tabular portion, is broken off posteriorly; flange is in plane of paper. Vertical shading indicates broken surfaces. Abbreviations: J, jugal; MX, maxilla; PO, postorbital; SC, semilunar curvature of squamosal within supratympanic flange; SF, supratympanic flange, including parts of squamosal, supratemporal, and tabular (tabular portion is broken off in this specimen) ; SLF, semilunar flange of supratemporal; SQ, squamosal. homologue of the semilunar flange would thus constitute a supra- tympanic shelf. A supratympanic shelf as thus defined apparently does not occur in the reference genera. Therefore state (a) is primitive; state (b) is reasonably considered derived, and state (c) more derived. Systematic results— XJR 2400 is probably a trematopsid, on the basis of the derived state (c) of character 1: supratympanic flange present and of trematopsid shape. The specimen cannot be definitely assigned to any trematopsid genus. However, in view of the general lack of morphological information on trematopsids, and the small portion of the skull preserved, a new taxon should not be established for this specimen. State (c) of character 1 apparently serves to define the Tre- matopsidae as a monophyletic group in Hennig's sense, i.e., a group all members of which share at least one derived character only with each other. This conclusion leaves out of consideration the possibly trematopsid shape of the flange in Cacops— the specimen is so poorly preserved in this area as to make determination of character state 78 FIELDIANA: GEOLOGY, VOLUME 30 very uncertain. DeMar (1968) suggested the presence of the dissoro- phid shape of supratympanic flange in most dissorophids as a unifying character of the family. This is correct in a sense; however, the absence of this character state in Tersomius, and possibly in Am- phibamus and Cacops, precludes using it to formally demonstrate monophyly in Hennig's sense. Similarly, none of the other characters analysed can be used to demonstrate monophyly of the Dissoro- phoidea or its constituent families. This should not be taken as a denial of monophyly, but simply as a statement that monophyly has not yet been demonstrated. Finally, this analysis suggests that it was incorrect to use the semilunar flange on the supratemporal to show relationship of Doleserpeton to dissorophids (Bolt, 1969). The semi- lunar flange is probably a shared primitive character, and as such should not be employed in phylogenetic analysis. REFERENCES Bolt J. R. 1969. Lissamphibian origins: possible protolissamphibian from the Lower Permian of Oklahoma. Science, 166, pp. 888-891. 1974. Osteology, function and evolution of the trematopsid (Amphibia: Laby- rinthodontia) nasal region. Fieldiana: Geol., 33, no. 2, pp. 11-30. Boy J. A. 1972. Die Branchiosaurier (Amphibia) des saarpfalzischen Rotliegenden (Perm, SW-Deutschland). Abh. hess. L.— Amt Bodenforsch., 65, pp. 1-137. Brundin, L. 1966. Transantarctic relationships and their significance as evidenced by chironomid midges, with a monograph of the subfamilies Podonominae and Aphroteniinae and the austral Heptagyiae. K. Svenska Vet. Akad. Handl., 11, pp. 1-472. Bystrov, A. P. 1935. Morphologische Untersuchungen der Deckknochen des Schadels der Wirbeltiere. Acta Zool., 16, pp. 65-141. Carroll, R. L. 1964. Early evolution of the dissorophid amphibians. Bull. Mus. Comp. Zool., 131, pp. 161-250. 1967. Labyrinthodonts from the Joggins Formation. Jour. Paleontol., 41, pp. 111-142. Cracraft, J. 1972. The relationships of the higher taxa of birds: problems in phylogenetic reasoning. Condor, 74, pp. 379-392. DeMar R. E. 1966. Longiscitula houghae, a new genus of dissorophid amphibian from the Permian of Texas. Fieldiana: Geol, 16, pp. 45-53. 1968. The Permian labyrinthodont amphibian Dissorophus multicinctus, and adaptations and phylogeny of the family Dissorophidae. Jour. Paleontol., 42, pp. 1210-1242. BOLT: TREMATOPSID SKULL 79 Gregory, J. T., F. E. Peabody, and L. I. Price 1956. Revision of the Gymnarthridae, American Permian microsaurs. Bull. Peabody Mus. Nat. Hist., 10 pp. 1-77. Hennig, W. 1966. Phylogenetic Systematics. University of Illinois Press, Urbana, 263 pp. McGinnis, H. J. 1967. The osteology of Phlegethontia, a Carboniferous and Permian aistopod amphibian. Univ. Calif., Pub. Geol. Sci., 71, pp. 1-49. Marx, H., and G. B. Rabb 1970. Character analysis: an empirical approach applied to advanced snakes. J. Zool., London, 161, pp. 525-548. 1972. Phyletic analysis of fifty characters of advanced snakes. Fieldiana: Zool., 63, 321 pp. Moustafa, Y. S. 1955. The skeletal structure of Parioxys ferricolus Cope. Bull. Inst. d'Egypte, 36, pp. 41-76. Nelson, G. J. 1970. Outline of a theory of comparative biology. Syst. Zool., 19, pp. 373-384. Olson, E. C. 1941. The family Trematopsidae. Jour. Geol., 49, pp. 149-176. 1967. Early Permian vertebrates. Okla. Geol. Survey, Circular 74, 111 pp. Romer, A. S. and R. V. Witter 1942. Edops, a primitive rhachitomous amphibian from the Texas red beds. Jour. Geol., 50, pp. 925-960. Sawin, H. J. 1941. The cranial anatomy of Eryops megacephalus. Bull. Mus. Comp. Zool., 86, pp. 407-463. Schaeffer, B., M. K. Hecht, and N. Eldredge 1973. Phylogeny and paleontology, pp. 31-46. In Dobzhansky, T., M. K. Hecht, and W. C. Steere, eds., Evolut. Biol, 6. Vaughn, P. P. 1958. A pelycosaur with subsphenoidal teeth from the Lower Permian of Oklahoma. Jour. Wash. Acad. Sci., 48, pp. 44-47.