Pearce-Sellards Series 43 CARROLLA CRADDOCKI A New Genus and Species of Microsaur from the Lower Permian of Texas Wann Langston, Jr. E. C. Olson April 22, 1986 TEXAS MEMORIAL MUSEUM. T HE UNIVERSITY OF TEXAS AT AUSTIN Wann Langston, Jr., is Director of the Vertebrate Paleontology Laboratory,Texas Memorial Museum, and Professor of Geology at The University of Texas at Austin. Everett C. Olson is Professor Emeritus in the Department of Biology, the University ofCalifornia, Los Angeles. The Pearce-Sellards Series is an occasional, miscellaneous series of brief reports of Museum and Museum-associated field investigations and other research. All manuscripts are subjected to extramural peer review before being accepted. The series title commemorates the first two directors of Texas Memorial Museum, both now deceased: Dr. J. E. Pearce, Professor of Anthropology, and Dr. E. H. Sellards, Professor of Geology, The Universityof Texas at Austin. © 1986byTexasMemorialMuseum The University of Texas at Austin All rights reserved Printed in the United States ofAmerica CONTENTS Abstract 1 Acknowledgements 1 Introduction 2 Systematics 3 Description 3 Relationships 12 Habits 14 Geographic Distribution 16 Conclusion 17 Literature Cited 18 Abbreviations 30 LIST OF FIGURES Fig. 1. Carrolla craddocki, n. gen. and sp. TMM 40031-54. Stereograms of A) skull, dorsal view; B skull, attached lower jaw and stapes, ventral view;C lower teeth, Ungual aspect 4 Fig. 2. Carrolla craddocki, n. gen. and sp, TMM 40031-54. Reconstruction of skull A) dorsal view; B) ventral view; C) posterior view 7 Fig. 3. Carrolla craddocki, n. gen. and sp. TMM 40031-54. Right posterolateral part of skull, lower jaw and stapes as preserved. 10 ... Fig. 4. Carrolla craddocki, n. gen. and sp. TMM 40031-54. A) Reconstruction of left mandibular ramus, lingual view. B) Composite drawing of the presumed lingual surface of a lower tooth 11 LIST OF TABLES Table 1. Measurements of the skull and jaws of Carrolla craddocki 2 ABSTRACT The specimen, comprising a skull and jaws, is assigned to the suborder Microbrachomorpha and tentatively to the family Brachystelechidae on the basis of the structure of the temporal and occipital regions. It is unique among known microsaurs in the possession of marginal teeth with longslender hollow bases and bifurcated crowns and the apparent absence of palatal dentition. Carrolla craddocki, new genus and species, from Lower Permian (Wolfcampian) Belle Plains Formation in Archer County, Texas,is described and figured. If correctly assigned to the Brachystelechidae, Carrolla is the first record of the family in North America, but it was probably not equivalent ecologi­cally to the roughly contemporaneous Brachystelechus of Europe. Carrolla is believed to have been a burrower in hard soils. Its diet may have com­prised soft-bodied subterranean invertebrates, but the functional significanceof its bicuspidate teeth is obscure. It is one of several unrelated microsaurs which developed cryptic behavior under competitive pressures from surface-dwelling reptiles in an increasingly harsh Early Permian environment in North America. ACKNOWLEDGEMENTS Science is indebted to Mr. Craddock for placing this and numerous of his other finds in the public trust. We are indebted to Drs. Robert L. Carroll,Donald Baird, and Angela C. and Andrew Milner for their helpful counsel concerning microsaurs. The doctors Milner first suggested that the new specimen was a microsaur and Dr. Carroll first noted its resemblance to Brachystelechus. Illustrations are by Mr. Michael W. Nickell. Doctors Carroll and John Bolt read early versions of the paper and their counsel is acknowl­edged with thanks. Financial support to the senior author from the University ofTexas Geol­ogy Foundation Owen-Coates Publication Fund is gratefully acknowledged. INTRODUCTION The specimen described in this paper consists of the skull and jaw, firmlyunited. It was collected from the surface in the vicinity of Tit Mountain in northwest Archer County, Texas, by Mr. Kenneth Craddock of Denton,Texas, who has over many years displayed an uncanny ability to ferret out minute specimens from the nodules that so often litter the surface of the Permian outcrops in north central Texas. Mr. Craddock generously donated the specimen to the Texas Memorial Museum on recognition of its con­siderable scientific significance. A full review of the Microsauria by Carroll and Gaskill (1978) draws attentiontotheremarkablediversityofthisorderofextinctamphibians and obviates any need to review the group in this report. In addition, Schultze and Foreman (1981) have reviewed the tuditanomorph microsaurs and analyzed their relationships cladistically. The new specimen adds one more distinctive genus to the known array, for it is quite unlike any other de­scribed genus in many particulars. Carroll and Gaskill recognized two suborders of Microsauria, Tuditano­morpha and Microbrachomorpha, the former comprising about twenty genera and the latter but four. The new specimen pertains to the Micro­brachomorpha, but can only be assigned tentatively to the family Brachy­stelechidae among the microbrachomorphs. TABLE 1 Measurements of the skull and jaws of Carrolla craddocki, in millimeters. Measurements are of the specimen as preserved except the width which is an estimate made to correct for the distortion ofthe left side. Midline length ofskull, basioccipital to snout 15.0 Length of skull from top of foramen magnum to snout 13.0 Width ofskull (corrected for distortion) 16.0 Maximum length of orbit 5.0 Maximum height of orbit 4.8 Width offoramen magnum 3.0 Distance between external borders of exoccipital condyles 4.8 Width of occipital plate 12.5 Length of pineal opening 0.7 Length of stapes (foot to head) 2.0 Length oflower jaw 12.0 SYSTEMATICS ORDER MICROSAURIA Suborder Microbrachomorpha Family Brachystelechidae (?) Carrolla n. gen. Diagnosis.—Skull length and width approximately equal. Marginal teeth with long hollow bases and enamelled bicuspid crowns with cusps aligned antero-posteriorly. Probably no palatal dentition. Sphenethmoid element ossified. Stapes massive, perforate, and inserting distally into a recess on the medial side ofthe quadrate. Carrolla craddocki n. sp. Holotype.—Skull and attached mandibles. TMM No. 40031-54. Locality.—Dissected flats about 100 feet above the base ofthe south side of a butte termed Tit Mountain, about three miles northeast of Dundee, in northwest Archer County, Texas. Horizon.—Belle Plains Formation, Wolfcampian, Lower Permian. Etymology.—The genus is named after Dr. Robert L. Carroll whose work with microsaurs in collaboration with Pamela Gaskill has greatly facilitated thisstudy.The species is named inhonorofMr. Kenneth Craddock. Diagnosis.—As for genus. For measurements see Table 1. DESCRIPTION Conditionofthe Specimen The general condition of the specimen is shown in Fig. 1. The skull and lower jaws are tightly locked together by matrix. The skull has been Bat­tened somewhat dorso-ventrally, but not to the extreme frequently seen in microsaur specimens. Crushing has skewed the skull to the right so that although the right side is present, it is not visible from above, whereas the left side is more broadly exposed than normal. The occipital plate has been somewhat flattened, which exaggerates the apparent length of the plate. The dorsal surfaces of the lower jaw are obscured, having been forced up under­neath the edges of the upper jaw. Excepting a few lower teeth, the dentition is not well preserved. The surface of the bone has a polished appearance, which suggests some abrasion or weathering. Some of the sutures are distinct, accentuated by a thin layer of red iron oxide between the bones. At places, notably in the palate and the area of the middle ear, separation and displacement of the elements occurred during fossilization. Some of the dermal roofing bones Fig. 1.Carrollacraddocki,n.gen.andsp.TMM40031-54.Stereogramsof A) skull, dorsal view; B) skull, attached lower jaw and stapes, ventral view; C) lower teeth, lingual aspect. C is x 15. have been damaged and are partly broken away, especially on theright side of the skull roof and in the temporal region. In many places where the bone has been lost, an accurate natural mold of the deep surface is preserved bythe matrix. The articular portion of the right mandible is missing. Remains of a few upper teeth are embedded in the matrix, but most of the teeth have been destroyed. Four lower teeth are exposed deep within the palatal area on the left side, appressed against the underside of the displaced palatal bones. Dorsal and Lateral Skull Surfaces (Figs. IA, 2A) The dermal elements are very thick relative to the size of the skull and in places they are joined by strong sutures with blunt interdigitating tonguesand grooves. The skull appears to have been massive, akinetic and capableof withstanding substantial stresses. Some sculpturing is evident on the frontals and on the jugal, beneath the orbit. Small foramina are present in theroofing bones at various places (Fig. 2A). The gross features of the dermal skull roof are shown in the reconstruc­tion (Fig. 2A) and photographs (Fig. 1A). An effort has been made in preparing thereconstructions to restore the original proportions. Paired parietals form the posterior part of the skull platform and carry backward onto the sloping occiput, partly covering the unified occipital plate. A strong transverse ridge separates the dorsal platform and occipital surfaces of the parietals. Anterior to this ridge, in the forward half of the parietals, is a well developed pineal foramen. Latero-posteriorly each parietal forms an elongated lappet posterior to the postfrontal and postorbital bones (Fig. 2a, pa(?ta)). This lappet is in the area where a tabular might be expected and although no separation can be found, there is sufficient dam­age to this part of the specimen that the presence of a tabular cannot be ruled out. The paired frontals form a broad, largely featureless plate. Laterally, each frontal extends into a narrow contact with the dorsal orbital margin, here separating the post-and prefrontals. The orbit is relatively very large, as is usual in small skulls. Within it lies a single, well preserved palpebral cup (Fig. 2A, ppc) and some fragments of bone which suggest the possible presence ofsclerotic plates. Interpretation of the rather small posterior elements of the skull, exclu­sive of the occipital area, is uncertain. The squamosal appears to have been relatively large lying largely postero-ventral to the postorbital and making broad contact with the parietal lappet. Its lateral and posterior contacts with the quadrate are, however, largely destroyed. An elongated jugal extends forward from the squamosal to form the lower rim of the orbit, beneath which it exhibits the only distinct sculpture pattern visible on the skull. Ventral to the jugal and extending to the level of the posterior margin of the orbit is the tooth-bearing maxilla. It is a moderately deep bone which makes contact with the premaxilla at the level of the posterior margin of the external naris, below the septomaxilla. The anterior margin of the orbit is formed by a moderately large pre­frontal and a dorso-ventrally narrow lacrimal. The lacrimal passes forward to the external naris, but is excluded from this opening by the septomaxilla, a tiny cup-shaped bone which forms the posterior margin of the external naris. The ventral and anterior borders of the external naris are formed by a large, tooth-bearing premaxilla. The united premaxillae project forward above the tooth row anteriorly forming a slightly reflected muzzle. Posterior to the premaxillae, and forming the anterior dorsal surface of the skull are the paired nasal bones. Damage and some distortion have made their out­lines somewhat uncertain, but the structure and relationships shown in Fig. 2A are probably not seriously in error. These bones are remarkablyshort and wide. Oto-occipital and Basicranial Complex (Figs. 2, 3) The oto-occipital and basicranial complex forms a single structure with the constituent elements indistinguishably fused. Included in this complex and forming much of it, as exposed, is the dermal parasphenoid bone. The occipital portion of the complex forms an “occipital plate” which incorpo­rates the occipital elements and passes laterally into the otic elements, of which the long, posterior portion represents the opisthotic. As preserved,the “occipital plate” makes up about one-third of the total skull length, in dorsal aspect, but this is excessive, it having been exaggerated by crushing.Dorsally the plate is overlapped extensively by the parietal and it extends forward beneath this bone nearly to the level of the pineal foramen. At the lower corners of the nearly equilateral, triangular foramen mag­num lie strongly developed occipital condyles. As preserved, remnants of a thick “odontoid” process of the atlas fill the posterior notch in the basi­occipital bone in the fashion characteristic of microsaurs. Enough of the condyles can be seen, however, to show that they are joined by a straplikesurface of the intervening basioccipital. Lateral to each condyle is a largejugular foramen and medio-ventral to this foramen is a pair of smaller openings. These presumably accommodated nerves IX and X, and XII,respectively. The otic elements of the complex are not recognizable as distinct ossifications, but their general positions are evident. Overlying the broad, lateral expanse of the parasphenoid is the opisthotic part of the periotic and, more anteriorly, what is presumably a prootic is visible. These elements contribute, perhaps with the parasphenoid, to formation ofa verylarge fenestra ovalis into which the expanded foot-plate of the stapes appearsto have fitted snugly. Fig. 2. Carrolla craddocki, n. gen. and sp. TMM 40031-54. Reconstruction of skull A) dorsal view; B) ventral view; C) posterior view. Location of some features in C is somewhat conjectural; the articular surfaces of the occipital condyles are not visible in the specimen. Crosshatched areas are obscured in the specimen. The parasphenoid dominates the complex ventrally and coossification with the other elements makes precise demarcation of the elements impos­sible. The broad posterior plate of the parasphenoid passes anteriorly into a long cultriform process whose anterior end is concealed by the tightly appressed lower jaws. Near the anterior margin of the plate, to each side of the midline, is a small foramen for the internal carotid artery, a feature present in some but not all microsaurs. Lateral to the foramen the para­sphenoid forms the ventral surface of the broad basipterygoid process. Dorsal to this extension and evidently solidly fused to it, the partially ossified basisphenoid contributes to the articular surface of the process. The basicranial articulation lies essentially in a horizontal plane and was a freely movable joint in life. Although possibly not visible from below in the uncrushed skull, a portion of the pleurosphenoid [basisphenoid of Romer (1950), in Paritylus]is exposed dorsal to the posterior part of the cultriform process in the specimen (Fig. IB). No sutures are evident between this element and the para-basisphenoid. Near the posterior margin of the exposed portion is a prominent foramen, probably for the passage of cranial nerve V. Antero­ventral to the foramen, at the base of the pleurosphenoid, is a rounded prominence which may be the base of the otherwise unossified epipterygoid.Anterior to the pleurosphenoid and well separated from it, the left wing of what is believed to be the sphenethmoid rises from the cultriform processof the parasphenoid to the inner surface of the vault of the skull. It is a featureless plate of bone with a somewhat restricted base, and is separated from the parasphenoid by a distinct suture (Fig. IB). The anteriormost portions are not visible. A sphenethmoid element has not been reportedbefore in microbrachomorphs, and was evidently not ossified in Microbrachis and possibly Hyloplesion (Carroll and Gaskill, 1978). Stapes and Quadrate (Figs. 18, 28, C, 3) Crushing has expanded the middle ear region so that the left stapes is well exposed. The element has been significantly displaced laterad from its inser­tion in the fenestra ovalis (Fig. 3), and may be rotated about 90 degreesaround its long axis. Its distal end, however, is only slightly displaced from its lodgement in the quadrate. The stapes is a stout bone with a broadly ex­panded, somewhat rectangular footplate, a short, thick shaft and a rounded, slightly expanded distal end, which in life rested freely in a well-formed stapedial recess of the quadrate. As preserved, a large stapedial foramen is visible near the ventral margin. The quadrate is a large bone with a strong ascending process. The ptery­goid process is short and massive, and is joined suturally to the quadrate ramus of the pterygoid. Above it is the specialized triangular recess into which fitted the distal end of the stapes. The articular condyle of the quad-rate is a simple anteroposteriorly rounded hinge that, as preserved, lies in the transverse plane of the skull. Its transverse diameter is about seven times itsanteroposteriordiameter(Fig. IB). Palate (Figs. 18, 2B) The paired bones of the palate are exposed on the left side of the skull. They are slightly broken and somewhat displaced. The lower jaw and some matrix,which cannot be removed, obscure the lateral edges of the pterygoid, pala­tine and vomer, but some additional details can be observed dorsally throughthe orbit. The parasphenoid is in natural position, but the other bones have been displaced upwards. In the reconstruction (Fig. 2B), an effort has been made to eliminate distortion, which involved not only the lateral and dorsal displacement of the elements, but also some elongation of the occiput which has tended to shiftthebasicranialcomplex somewhatposteriorly.Thearrangementofthe palatal bones as shown in the reconstruction conforms for the most part to the customary microsaur pattern. The interpterygoidal vacuity, however, is very large and the pterygoids failed to meet anteriorly, so that the vomer forms the anterior border of the vacuity on either side of the skull. The parasphenoid was described earlier and is the dominant element of the palatal complex. The pterygoid, which was not fused with the basi­cranium, has been removed from contact with the para-basisphenoid bycrushing. As in other microsaurs, the pterygoid lacks a transverse flange and has a smooth margin around the subtemporal fossa. Although the lateral portion of the bone is concealed, a small part of the palatal ramus can be seen through the orbit, giving some indication of the nature of the con­cealed portion. The quadrate ramus of the pterygoid is short and stout, broken and somewhat displaced in the specimen. It makes sutural contact with the pterygoid process of the quadrate. Posteriorly, the palatal ramus of the pterygoid contains a broadly vaulted longitudinal channel which ends opposite the basal articulation. A notch occurs in the margin of the palatal surface of the pterygoid and lateral to the notch the posterior edge of the bone sends a flange upward into the matrix that fills the space between the pterygoid and the stapes. This flange is confluent with the quadrate ramus of the pterygoid and likely formed the osseous antero-lateral wall of what­ever middle ear chamber was present (a middle ear cavity may, in fact, have been absent, as in apodans and salamanders [Carroll, pers. comm.]). A well developed ectopterygoid is present. It is quadrilateral in form and somewhat wider than long. It is displaced in the specimen to lie over part of the pterygoid. The postero-lateral edge is turned down and may have formed the anterior edge of the subtemporal fossa as the medial edge of the ptery­goid is similarly deflected. Anteriorly, the ectopterygoid appears to have been inserted into a groove on the ventral surface of the palatine and, as seen from above, through the orbit, it appears to have lacked contact with the maxilla. The palatine is partially hidden by the lower jaw. As exposed, it has a subtriangular outline, and is broken anteriorly. How it relates to the internal naris is uncertain as this structure is not visible in the specimen. The ventral surface of the palatine was partly covered by the ectopterygoid, as noted in the preceding paragraph. Little can be said about the vomer other than that it was a plate-like bone which had a long contact with the anterior edge of the palatal ramus of the pterygoid. The preserved portion ofthe medial edge is curved dorsally, suggesting that it abutted against the anterior portion of the cultriform process of the parasphenoid, but this cannot be confirmed in the specimen. Mandible (Figs. 18, 3, 4A) Although much of the lower jaw is preserved, it is not well exposed (Fig. 4A). It is about two-thirds as long as the skull. The symphysis is thick and the dentary must be deep to accommodate the long roots of the teeth. At least one splenial was present and like other bones in the skull and jaw it is remarkably thick for its size. The angular projects backward beneath the articular, but there is no retroarticular process. The submeckelian foramen is evidently surrounded by the angular. The prearticular is a triangular bone on the medial side ofthe jaw. Its superior edge is exceptionally elevated, so that in medial aspect its dorsal outline resembles that of the lateral edge of the adductor fossa in other microsaurs. Anteriorly, however, the superior edge of the bone is seen to pass medial to the tooth row so there can be little doubt thatit is theprearticular. Fig. 3. Canolla craddocki, n. gen. and sp. TMM 40031-54. Right postero­lateral part of skull, lower jaw and stapes as preserved, ventral view. Approxi­mately x 6. Dentition (Figs. IC, 4A, B) No dentition is present on the exposed elements of the palate. Comparable areas are consistently denticulate in other microsaurs, excepting possiblyGoniorhynchus. It is, of course, possible that teeth or a shagreen of denticles is present on the concealed portions of one or more of the palatal bones as inGoniorhynchus,but eveninthisunlikelyeventtheabsenceofanysortof of dentition on the major parts of the palatal bones is unique among micro­saurs. That some sort of shagreen existed in the roof of the mouth but was unattached to the bone is also possible, but we believe highly unlikely. Such a condition has been reported in Crinodon limnophyes by Carroll and Gaskill (1978), however. The marginal dentition of Carrolla is also unique within this group of animals. It comprises closely spaced, slender teeth with enamelled bifurcated crowns (Fig. 4A, B). The dental formula is not determinable, but at least the following teeth are present: Pmx 4 + Max 6 or 7/Dent. >4. Of the upperteeth only a few damaged crowns are seen, but three teeth are completelyexposed on the lingual surface of the left dentary, owing to loss of the splenial bone (Figs. IC, 4A). The crown of a fourth tooth is also present.The hollow pulp cavities of some of the bases have been exposed acciden­tally during preparation. The teeth have long bases and are attached to the dentary in acrodont or possibly pleurodont fashion. The bases are cylindri­cal toward the bottom, but toward the crowns they become somewhat flattened transversely, thus producing a spatulate appearance. The crowns of the teeth are distinguished from the bases by a subtle color difference which probably signals a change in density and the pres­ence of enamel. The crown forms about one-fourth of the total height of the tooth (Fig. 4B). There is no evidence of pedicellate structure. The crowns are divided into anterior and posterior cusps, which vaguely resemble the forked tongue ofa serpent. The posterior cusp is always smaller than the anterior one and is bent a little backward near the tip. There is some varia­tion in the degree of forking, and the disparity in the size of the anterior and posterior cusps is greater in the upper teeth than in the mandibular teeth. Presumably the bases of the upper teeth are shorter than those of the lower teeth inasmuch as the dentigerous bones are narrower dorso-ventrallythan the dentary. Fig. 4. Carrolla craddocki, n. gen. and sp. TMM 40031-54. A) Reconstruc­tionofleftmandibularramus,lingualview.Approximatelyx 6.B)Composite drawing of the presumed lingual surface of a lower tooth. The larger cusp is anterior. Greatly enlarged. RELATIONSHIPS Carrolla craddocki is assigned to the Microsauria on the basis of its temporal region, which lacks supratemporal and intertemporal bones, and probably also the tabular (whose presence or absence does not affect the assignment), its basioccipital-exoccipital complex, which evidently forms a broad cranio­cervical articular surface of the type described as “strap-shaped” by Carroll and Gaskill (1978), and the absence of an otic notch. Assignment to the Microbrachomorpha is based upon the presence of a broad parietal-postorbital contact, which would exclude any connection between the tabular (if present) and the postfrontal. The absence of a sepa­rate postparietal is also suggestive, in view of the absence of this bone in Brachystelechus and its small size in the other genera of the suborder. In many particulars, however, including the presence of an ossified spheneth­moid, Carrolla differs from the four other known members of the Micro­brachomorpha, suggesting, as Carroll and Gaskill (1978) noted, that this suborder may not, in fact, be a natural group (i.e., not monophyletic). Carrolla is more similar to Brachystelechus than to other microsaurs in the general proportions of the skull and the configuration of the postorbital region, in particular, the shape and relationships to adjacent elements of the parietal lappet and the absence of a postparietal. It must be stressed, how­ever, that the skull of Brachystelechus is very incomplete. The upper and lower jaws, and the dentition, the palate and some ofthe marginal elements of the skull are missing, and the braincase is unknown. Thus, although the bifurcated marginal teeth and very large interpterygoidal vacuities of Carrolla are not known elsewhere among microsaurs, the possibility exists that such may occur in Brachystelechus. Several clear-cut differences between the two genera do exist. Massive pitting occurs in several places on the skull surface of Brachystelechus, whereas in Carrolla the skull roof is marked only by a few tiny pore-like foramina and pitting is developed only on the jugal beneath the orbit. Although some abrasion of the dorsal surfaces has occurred, it does not appear that the pronounced pitting seen in Brachystelechus could have been present in Carrolla. The postfrontal is relatively small in Brachystelechus and relatively large in Carrolla. Carrolla also lacks an interfrontal bone, presentin Brachystelechus. The prefrontal in Carrolla either just reaches the poste­rior tip of the external nares or is excluded from it by the septomaxilla,whereas in Brachystelechus it entersbroadly into the narial border. Like other microsaurs Carrolla exhibits a large suite ofcharacters, one or another of which occurs in other genera, but which is not duplicated in extenso in any of them. Two cladistic analyses published by Schultze and Foreman (1981) show this general pattern as well. One analysis is of the families of Tuditanomorpha. In it, the common familial features are pre­ dominantly plesiomorphs, several autapomorphies occur in the familial lineages, and only four clusters oftwo or more synapomorphies are defined,these being common to the families Gymnarthridae and Goniorhynchidae.A total of 23 characters was used in this analysis. Among the Gymnarthri­dae, in the second analysis which employs a larger array of characters, synapomorphies unite only the four species of Euryodus. The distribution of characters emphasizes the great variety of microsaurs stressed by Carroll and Gaskill (1978) and also, of course, renders many of them useless for taxonomic purposes, given the current state of knowledge. Some examples in Carrolla are as follows: 1. The ectopterygoid in Carrolla does not appear to make contact with the maxilla as it does in most microsaurs. This condition, however, is also found in Hyloplesion. 2. Limited bilateral contact of the frontal with the dorsal border of the orbit occurs in Carrolla, Saxonerpeton and Hapsidopareion. Limited contact on one side but not on both, as figured by Carroll and Gaskill (1978), occurs in known individuals belonging to Hyloplesion and Odontopteron and broad contact is present in Llistrophus and Brachystelechus. In most micro­saurs, however, no frontal-orbital contact occurs. 3. The septomaxilla is moderately small and forms the posterior rim of the external naris in Carrolla. In most microsaurs the bone is relativelylarger and is incorporated into the narial opening, but Carrolla and Crino­don are similar. The condition in Brachystelechus is unknown. 4.Thepinealforamen islargeandliesintheanteriorhalfoftheparietalin Carrolla. The foramen is absent in some microsaurs, small in others, and large in a few. Its position relative to the anterior margin of the parietalvaries. Its size and position in Carrolla is comparable to that in Hyloplesion,Brachystelechus and Hapsidopareion. The size of the foramen in Carrolla is comparable to Crinodon, where, however, the opening lies in the posteriorhalfof the parietal. 5. A dorsally directed process of the posterior margin of the palatine ramus of the pterygoid forms an anterior dermal margin of the middle ear in Carrolla and Microbates, but is not developed in several other genera of microsaurs. The precise condition in several genera is not known. 6. The basioccipital and exoccipitals in microsaurs form a continuous articular “strap-like” structure met by the atlas. The precise structure varies widely among the genera. The general structure of Carrolla is similar to that found in Hapsidopareion and Rhynchonkos (Goniorhynchus in Carroll and Gaskill, 1978, Rhynchonkos, Schultze and Foreman, 1981) in which the exoccipitals form a strong “double” condyle, but with the basioccipital still forming a part of the articular surface. This appears to be a derived condition among the microsaurs. 7. The stapes, both in form and relative size, varies remarkably among the microsaurs. A large, oval footplate and a slender stem that projects laterally, or somewhat ventrally from it, represent the most common struc­ture. The stem may or may not be perforate. In some genera, such as Pan­tylus there are well developed dorsal and ventral processes. An accessoryvertical structure is found in both Pantylus and Rhynchonkos. The foot­plate in some genera is angular and in Pelodosotis complexly irregular. The stapes of Carrolla does not fit well into any of the general patterns. The subrectangular foot is rather like that of Microbrachis, but the more lateral portions are thick and the stout shaft has a large stapedial foramen, whereas in Microbrachis the stem is slender and the foramen is very small. This element, as much as any, indicates the immense variation within the order and the difficulties of systematic assessments. HABITS That Carrolla was an ecological “equivalent” to Brachystelechus from roughly contemporaneous deposits in Europe seems unlikely. Carroll and Gaskill seem ambivalent about the habitat of Brachystelechus, statingthat it, “. .may have been equally at home in the water as on the bank.” . (Carroll and Gaskill, 1978:201). Familial relationship in this instance is no bar to widely different habits, because the respective localities were on opposite sides of a massive land barrier, were at least 5,000 kilometers apart, and the difference in geological age of the deposits may be several million years, correlations between the Rotliegende and the Wichita Groupbeing somewhat imprecise. Absence of lateral line canals in Carrolla, while not conclusive, suggests terrestrial habits. The massive construction of the little skull, the thick and broadpalpebral bones, thebroadoccipitalsurface(forattachmentofpower­ful neck muscles) and the inferred blunt and protruding rostrum may be evidence for burrowing habits in Carrolla. Burrowing habits have been imputed to Ostodolepis and Pelodosotis because of their wedge-shapedrostra, among other things. Cans (1974) illustrates a variety of head shapesencountered in amphisbaenians which not only resemble Ostodolepis, but also come close to the profile inferred for Carrolla. The widely dispersed“nutrient” foramina in the dermal skull roof of Carrolla may be related to a thick integument that might also be expected in a burrower (Cans, 1974). The Lower Permian redbeds in northcentral Texas contain buried soil horizons and remnants of ancient duricrust zones which could have pre­sented difficulties to such small burrowing vertebrates as Carrolla. Perhapsin these situations the exceptionally massive stapes with its well developedquadrate articulation added to the strength of the head as a unit. It should,coincidentally, have functioned more efficiently as a hearing mechanism by wayofboneconductionin asubterraneanenvironmentthanintheair. The very limited fossil record of microsaurs can be interpreted as reflect­ing an evolutionary tendency within some groups to move from a predomi­nantly aquatic, to a more terrestrial, reptilomorph, and burrowing anguiformhabitus. (The Gymnarthridae, on the other hand, are believed to have become more aquatic during their evolutionary history, according to Carroll and Gaskill, 1978). Selection for increasingly cryptic behavior and, in the case of the Gymnarthridae, more aquatic habits, might have resulted from competition with surface-living reptiles toward the end of micro­saur existence. The unique dentition of Carrolla seems suitable only for tender food, probably some soft-bodied subterranean invertebrates. The anteroposteriorlybicuspid teeth double the number of effective tooth crowns in the short jaws. Bolt (pers. comm.) has pointed out that under certain conditions occlusion of these crowns should produce an efficient shearing action, but nothing is known about the actual pattern of occlusion in the presentinstance. Bicuspid teeth are present in some Pennsylvanian and Permian labyrinthodonts, but as in lissamphibians, the cusps in these animals are arranged normal to the long axis of the tooth row (Bolt, 1967, 1977, 1979). Larsen and Guthrie (1975) observed that in the existing tiger salamander (Ambystoma tigrinum) the teeth pierce prey only to the level of the largely horizontal Ungual cusps. This prevents the prey from being too securelyimpaled and permits the clean release of the food for transfer toward the throat. “Thus the teeth, although part ofa kinetic-inertial feeding system . . ’ exert static pressures and function largely to retain rather than to maim. (Larsen & Guthrie, 1975:146). The cusps in Carrolla, however, are relatively larger and less divergent than in the salamander, so Larsen and Guthrie’s explanation seems unlikely for the microsaur. The bicuspid teeth in Amphibamus grandiceps from Mazon Creek were interpreted as a juvenile dissorophid character by Bolt (1979). The possi­bility that Carrolla represents a larval phase of some other microsaur is diminished, however, if it was a terrestrial form. In any event, if microsaurs had a complex life cycle, as is postulated for dissorophid labyrinthodonts by Bolt (1977), or if the larval and adult stages were as morphologically dis­ parate as they are in lissamphibians, it is doubtful that the adult form could be identified from a larval specimen. Carroll and Currie (1975) have pointed out a number of resemblances the skulls of microsaurs, in particular Goniorhynchus, and the between living apodans. In Carrolla, absence of postparietals, if confirmed, and fusion of the otico-occipital components into an os basale are reminiscent of apodans, which also share with Carrolla a hypertrophy of the cranial roofing bones and a posterior location of the cranio-cervical joint relative to the quadrate. The strong articulation between stapes and quadrate is a further similarity to most apodans. On the other hand, the bones of the skull of Carrolla display inter­relationships and details characteristic of microsaurs and other Paleozoic amphibians, there being no obvious fusion of elements in the cheek regionand no evidence of the apodan specialization of the lacrimal. Pedicellate teeth, present in all apodans according to Carroll and Currie, are lacking in Carrolla. Thus, as with other microsaurs, Carrolla exhibits some close paral­lels to the Apoda, but until more conclusive evidence is obtained to support a close microsaur-apodan relationship, e.g., an apodan fossil record, the resemblances noted above may be interpreted conservatively as paralleldevelopments between animals with somewhat similar habits. GEOGRAPHIC DISTRIBUTION To judge from what is clearly a fragmentary fossil record of microsaurs,tuditanomorphs were more diverse than microbrachomorphs which until now have comprised only four monotypic families. Of these four, the Odonterpetontidae and, questionably, the Hyloplesiontidae have been reported from North America. Carrolla represents a third microbracho­morph in North America and, if correctly assigned to the Brachystelechidae,itis the first North American record of that family. The geographic distribution of microsaurs is spotty (Carroll and Gaskill,1978), conforming to the Laurasian Permo-Carboniferous distribution of coal-swamp and redbeds vertebrates in general (Panchen, 1977). Milner and Panchen (1973) and Panchen (1977) argue convincingly that distribution of Permo-Carboniferous tetrapods was strongly influenced by upland topog­raphy attending the Appalachian and Hercynian orogenies. Terrestrial animals were evidently better able to disperse between the European area southeast of the barrier and the American side on the northwest than the aquatic fauna which was apparently restricted by the barrier. Although the vagaries of discovery are everpresent (Panchen, 1973), it is interestingthatallmicrosaurfinds arewithinaboutthreedegreesofthe West­phalian to Lower Permian paleoequators (Scotese, Bambach, etal, 1979).If theknown distributionofmicrosaursonboth sidesofthehighland“barrier” bears any resemblance to their actual ranges, it is arguable that this is as much a reflection, initially, of climatic zonation, because early occurrences lie well within the late Carboniferous tropical rainy predicted by zone Ziegler, Scotese, et al. (1979). It is well to bear in mind that development of the Appalachian and Hercynian orogenies was not a sudden event. Such North American Permian microsaurs as Carrolla, Pantylus, Micra­roter, Ostodolepis, etc., like many other redbeds tetrapods, may be viewed as specialized survivors adapted to the harsher environments that succeeded Late Carboniferous equability in the tropics (Ziegler, Bambach, et al, 1981). CONCLUSION The small microsaur skull described in this paper reveals many structural details. Although somewhat distorted and damaged, it is in general among the best preserved examples of this group. The small size of most micro­saurs and the diversity of known forms suggests that many new taxa remain to be discovered. As it now stands, CarroUa adds one more distinctive genus to this confusing group. Perhaps, as more representatives are reported, consistent suits of characters will be found to unite the genera into readily defined families and higher taxonomic units. The work of Romer (1950), Carroll and Gaskill (1978), and the recent analysis by Schultze and Foreman (1981), have aided greatly in bringing order in the form of a functional classification. Nevertheless, the full diversity of the order Microsauria and its relationships to other orders of amphibians remains largely obscure. LITERATURE CITED Bolt, J. 1969. Lissamphibian origins: Possible protolissamphibian from the Lower Permian of Oklahoma. Science 166:888-891. Bolt, J. 1977. Dissorophoid relationships and ontogeny, and the origin of the Lissamphibia. J. Paleontol. 51:235-249. Bolt, J. 1979. Amphibamus grandiceps as a juvenile dissorophid: evidence and implications. Pp. 529-563 in: M.H. Nitecki, ed., Mazon Creek fossils. New York & London: Academic Press. Carroll, R.L., and P.J. Currie. 1975. Microsaurs as possible apodan ancestors. Zool. J. Linnean Soc. 57:229-247. Carroll, R.L., and P. Gaskill. 1978. The order Microsauria. Mem. American Philos. Soc. 126, 211 pp.Cans, C. 1974. Biomechanics, an approach to vertebrate biology. Phila­delphia and Toronto: J.B. Lippincott Co., x + 261 pp. Larsen, J., Jr., and D.J. Guthrie. 1975. The feeding systems of terrestrial tiger salamanders {Ambystoma tigrinum melanosticum Baird). J. Morphol. 147:137-153. Milner, A.R. 1980. The tetrapod assemblage from Nyrany, Czechoslovakia. Pp. 439-496 in: A.L. Panchen, ed., The terrestrial environment and the originoflandvertebrates.London &NewYork: AcademicPress. Milner, A.R., and A.L. Panchen. 1973. Geographical variation in the tetra-pod faunas of the Upper Carboniferous and Lower Permian. Pp. 353­368 in: D.H. Tarling and S.K. Runcorn, eds., Implications of continental drifttotheearthsciences,Vol. 1.London:AcademicPress. Panchen, A.L. 1970. Batrachosauria Part A; Anthracosauria. In; O. Kuhn, ed., Encyclopedia of Paleoherpetology. Stuttgart & Portland: G. Fischer Verlag, vii + 84 pp. Panchen, A.L. 1973. Carboniferous tetrapods. Pp. 117-125 in: A. Hallem,ed., Atlas of Palaeobiogeography. Amsterdam, London, & New York: Elsevier Scientific Publ. Co. Panchen, A.L. 1977. Geographical and ecological distributionof the earliest tetrapods. Pp. 723-738 in: M.K. Hecht, P.C. Goody, and B.M. Hecht,eds., Major patterns in vertebrate evolution. New York: Plenum. Romer, A.S. 1950. The nature and relationships of the Paleozoic micro­saurs. American J. Sci. 248:628-654. Schultze, H.-P., and B. Foreman. 1981. A new gymnarthrid microsaur from the Lower Permian of Kansas with a review of the tuditanonnorphmicrosaurs (Amphibia). Occas. Papers Mus. Nat. Hist. Univ. Kansas 91:1-25. Scotese, C.R., R.K. Bambach, C. Baron, R. Van der Voo, and A.M. Ziegler.1979. Paleozoic base maps. J. Geol. 87:217-277. Ziegler, A.M., R.K. Bambach, J.T. Parrish, S.F. Barrett, E.H. Gierlowski, W.C. Parker, A. Raymond, and J. Sepkosi, Jr. 1981. Paleozoic biogeog­raphy and climatology. Pp. 231-266 in: K. Niklas, ed., Paleobotany, paleoecology, and evolution, Vol. 2. New York; Praeger. Ziegler, A.M., C.R. Scotese, W.S. McKerrow, M.E. Johnson, and R.K. Bam­bach. 1979. Paleozoic paleogeography. Ann. Rev. Earth Planet. Sci. 7:473-502. ABBREVIATIONS a art angulararticular ba basicranial articulation bo ch basioccipital area choana d ect f dentaryectopterygoidfrontal ic iptj1 foramen for internal carotid arteryinterpterygoid vacuityjugallacrimal m maxilla n nasal oc ocp pa occipital condyle (exoccipital)oto-occipital-parasphenoid complex parietal pa(?ta) parietal lappet, possibly including tabular pal palatine part prearticular ps parasphenoid pm premaxilla po postorbitalpf postfrontalprf prefrontal ppc palpebral bone pt pterygoid q quadrate sm septomaxillasmf submeckelian foramen sp splenial sq squamosal st stapes v vomer x foramen for 10th cranial nerve (jugular foramen) xii foramina for xii (?xi) cranial nerve