24 <16P NH | Loe — THE UNIVERSITY OF KANSAS : Lae mr MUSEUM OF NATURAL HISTORY No. 85 The Amphibamidae (Amphibia: Temnospondyli), with a Description — of a New Genus from the Upper _ Pennsylvanian of Kansas ) Eleanor Daly LAWRENCE : | 17 February 1994 PUBLICATIONS . The University of Kansas Publications, Museum of Natural History, beg t 1946, was discontinued with volume 20 in 1971. Shorter research papers fori above series are now published as The University of Kansas Museum of Natural I Histe Papers. The University of Kansas Museum of Natural History Miscellaneous u regarding style and submission procedures betas. Teauscane ‘omg - subjected to critical review by intra- and extramural specialists; final accept an of the Director. ey Institutional libraries interested in exchanging publications may obtain the Ocea and Miscellaneous Publications by addressing the Exchange Librarian, The Univ Library, Lawrence, Kansas 66045-2800, USA. Individuals may purchase sepz Office of Publications, Museum of Natural History, oe University of Kansa 66045-2454, USA. : THE UNIVERSITY OF KANSAS MUSEUM OF NATURAL HISTORY MISCELLANEOUS PUBLICATION No. 85 hy February 1994 The Amphibamidae (Amphibia: Temnospondyl.), with a Description of a New Genus from the Upper Pennsylvanian of Kansas ELEANOR DALY Mississippi Museum of Natural Science 111 North Jefferson Street Jackson, Mississippi 39201 USA Museum OF NATURAL HISTORY DycHE HALL Tue UNIVERSITY OF KANSAS LAWRENCE, KANSAS MISCELLANEOUS PUBLICATIONS Editors for this issue: Robert M. Mengel, Richard F. Johnston, and Larry D. Martin Managing Editor: Joseph T. Collins Assistant Managing Editor: Kate Shaw Miscellaneous Publication No. 85 Pp. iv + 1-59; 40 figures; 4 tables Published 17 February 1994 ISBN: 0-—89338—046—6 © 1994 py MuseuM oF NATURAL HISTORY DycHE HALL THE UNIVERSITY OF KANSAS LAWRENCE, KANSAS 66045-2454, USA PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS CONTENTS TENT RCD ONT sO PR ct cs ee tee ee” AG NO NUON Yan Ch es l MEME EAMVIETON LAB YRINTHODONT .22...2.:.c--cccccccce\cceeecseetl Bl eae one FAOIKENOWIEEIDGMENTS .....ccccecsestessonseseescctsecnecesntccceeeeetenes : Seeing Ayes RR oe ie Cee D) BATS RMS NV PIC NOIDCONINS ocr esse ante ona sctormnsedereee oceteosacdleedandeceesaiaddedsanuasusaceoe thee hweemeeeet¥ittes chs ssouusaliecccte. 3 Smee eel vay Ab @ ale NE Es ON MOT OG ras cssdtcinccacecacsonsansnacececessesnnonsestaccpasddeasedoscaccessseeescoceccedaeieosast 3 SOT so AMMO I Dy electron san oee meaner et acinar dals sad aw sh dieaied eahitaee deme ee dus ceiidaied Mase swasnaldesvaveeedideedeecse 3 Sri cram leiiMmaro spore yAiecrean ae weet ee ere es cee ccceeee nace eeuise sor dauer nace deessee Miacnnedadsebuburdenanedeceueeuears 3 SUpemeamailly sD SSOROMMOMd CAM eeiere ecco a eee tennis clsceh ceecindes soap cndasside Geum ssaectetaanaddedacuseecae: 3 Fei caunri aya /-Netan fo nll Aa Gl ACS aero oe eras Meas cr ccs ce need er cen ta. dec caae wae isc dioadnndesaateacs osu seeacbod don cade becucesees 3 BT FO iS TU S prereset eee cece ae Rene entra UE Gtcindecauiesuutinne cavevansedectoae See : FRE WAS CECT ACTA O SUSE ee eeerecarns tee Al ucnsseantecel eso couse ce h vet vieGiie GonteueneSetioee decks 3 TRASKOIMOTIONIC SOVOTUS. g.ctaus eepasarcect oe necae cea REG See nOr kc Bee ern eS oer 3 /EOSCODUIS; THEW SIS NUIS Rp coscapeon ops Sob SIRE Sane ne ete NCEE eT eC ree ee ere 3 FS ays OLO) Say prersemeeee oerceem aenea one se ee Acodeey sUaalgnctecteehscassanchere spe ieseascenerecs 3 | DITBNETIVOSTIS eoeatigroudtotcanonce eee e ARCA Macc eS SnACR re nn coe San te Mt seca Baran ere sare eee ee 3 OS GOPUS MOCK ANGE, MIC WES DE LES ox. conce avocm ech -sneeceeneroctiecastnaieecnuaatoevee sadctineautarte - Aa MONIMI COU eee te ets tenn eco cesea seer ennetccrenanesaeuns cnet onedcesaanstecsawen sete 4 Ts OSU ON) [ce pete see rece ea er aerate ret nee es Oe oes sean cnabes aatusanaanacseaseoasebenes 4 egellr cl [Oe Sieeree arte me mtn mt MUN cane OS er RI eee ed all yond ede ead = DIOS CIAT ONT ONT sacs season BasncaeaascGCeson see tet se sear See eee arenes eer nee 5 Grete Tallisie ANU Spenser ere creer ete aoe neces sot caida daamer eae os sciee se'sisdeusennte cadences 5 Steer ea Nec a eee oe aes Seeds Sutic essa eo Sis nanseusadoesueSeavonde Mec ceecawecacees 5 J COSRISIETE NWN ance cacoaen den scceGe ion noua one ae Inns SES ashe ate ARE cr ene ee te teen eee een eee 10 WSIS OTANS asa caaace aad cogetaacach bab tacidocee oaBNaS Ce Conen Gane ERAT ere ee ene ena een 11 RUT Sie sete secre eet seme eat cece Cenc ee ateicar Sec eeene, «ca deadenavnicdess iSeapaseeeswes 14 Fe Ct rect le ried pp esse rte ee ses ee ga ac rose Ste elec dev aa abuso Seu de 15 Relwacsoimdlle weer: ceca ree acne sosboSbonkposoacobseoe Socssacoseeecetedong Hoots Ren hS Re © es nt Lge eee cm nese eres ce ate eS Senin coe aeii ce BS A deaicasd te necnctatu ned andatnciesste 19 Pe inti lita to eterno ee OS, ox. ee ec were ire sc MARS Ss Eas du, dal nonewdledae Cae tewcues 20 JADU DGTATS (COTS MRSS) scence santos Sessoosonsey oo eee eee ee eee eee 24 RAIDS sesetacanncadsadcseseceenoaotbancne cnnaediesontedindvensecthcless«ad teen 30 Git les occ ccscsosvacersecdscscesnsdeenseeobgvacdectecrecsdeccsesedn er 32 LAIMNIDS Sonncdnesedundnsnsssedecsecduusoeudisnesaseacdeveuetsorndeecssceeene ean EZ Dermal Structures 2... cbccc.. duscatoveeeecesbole ripe eeeene err 33 Tersomius CaS@ VOVO oicc.ccc.Giccsshassasceassvucecconapesecctece coveneesteeeenee 34 SKUMD sats coed sadecckcceavedeacaseecctinsss sesdoadeld cesaseesesostaeeeee anes i LOWED JAW soioooaceclcacsccdaeatennessqasssssosensnsootesoocecacs soeeeeeeen eee = =e aaa 39 AAXIALSKCICIOM cece cere are ener Siveetesesesenea Jee eeee-e a ee 39 Girdles and [MDS ..ii.ot co... ccesceedereoocectesecootseeceenevsneet eee er 42 The Family Amphibamidae .2...c.0...ccoscsatieccsecssstonnecsosesocneehvoacsepastousosssareseeseneees tee aan 45 The Dissorophoidea NOW .iii.cc0ieiidlecendescoonsssssatecanessusssdonsacsescaasedeeset ue nesne saa =a 47 DISCUSSION ver cccciecediescicecesteubestannessacheucnadencnsesuesnunsoentoreneeeeeteecuccesteessee tet === Dy SUMMARY wicccscecscsed.pusssdctanasescceissennnveeueutnn laqeeau sone seeeueceeueeien: occrsvedu: aeth eee rrr 55 LITERATURE CITED (eines ciececiacacocacocawadece cee edeeeseneraledeuseentee ty eateser eee eee rr 35 INTRODUCTION A previously unknown Pennsylvanian dis- sorophoid is part of the fossil fauna of Hamilton Quarry, eastern Kansas (University of Kansas Verte- brate Paleontology KU-GRN-O1); Greenwood County, ca. 2 mi E of Hamilton, (Sec. 5, T24S, R12E). The fossil-bearing limestone is a channel deposit in or has been eroded into the Topeka Lime- stone of the Shawnee Group, Virgilian Stage. Its relative time relations with other well-known Penn- sylvanian vertebrate fossil localities are shown in Figure | (compiled from: Bell, 1944; Branson, 1962a, hIG ol 9O2e; Carroll, 1967a, 1967b; 1969: Copeland, 1957; DeMar, 1970; Eagar, 1964; J. T. Gregory, 1950; Moran, 1952; Olson, 1946; Panchen and Walker, 1961; Peabody, 1952; Reisz, 1972; Romer, 1952; Vaughn, 1969, 1972; and Wanless, 1962). Walter Lockard discovered the first vertebrate specimen, an acanthodian fish, at the quarry site in 1964 while looking for invertebrate fossils. In 1969, Thomas E. Bridge (Department of Geology of Em- poria State University) began to take field parties to the quarry. Since then, fossil material from Hamilton Quarry has become more widely studied. The first published report was an abstract of a paper read at a Geological Society of America meeting (Bridge et al., 1972). In 1976, the fauna and flora of the quarry were the subject of a symposium at the 108th annual meeting of the Kansas Academy of Science (Trans. Kansas Acad. Sci. 79:99-102). In 1988, another Hamilton Quarry symposium was held at the 22nd annual meeting of the south-central section of the Geological Society of America, with a day-long field trip to the quarry area. The comprehensive field trip guidebook (Mapes and Mapes, 1988) is the best source of information on Hamilton Quarry and includes a complete history of work at the quarry (Bridge and Mapes, 1988). The fauna reported from Hamilton Quarry in- cludes tetrapods, acanthodian fish, sharks, lungfish, eurypterids, arachnids, insects, myriapods, crusta- ceans, bivalves, crinoids, bryozoans and fusilinids. By far the most common vertebrates are an acanthodian, Acanthodes bridgei (Zidek, 1976), and the labyrinthodont described herein. The fish re- mains often are whole bodies, and the tetrapods often are found as ‘articulated groups of bones, sometimes as nearly complete skeletons. These speci- mens are either flattened onto a bedding plane in a layered brownish-gray to buff limestone, or pre- served three dimensionally in a massive gray lime- stone. THE HAMILTON LABYRINTHODONT This labyrinthodont belongs to the Dissoro- phoidea, a group of rhachitomous temnospondyls known mainly from the Lower Permian of the south- central and southwestern United States. This super- family is fairly well known, with more than 20 named genera. In general, dissorophoids are small to medium in size and their skeletons are well ossified overall with relatively long slim limbs and short vertebral columns. The interclavicles and clavicles of dissorophoids are small and their femora bear large adductor flanges. Typically, their skulls are short-faced with large orbits and high otic notches. Distinguishing marks of smaller scale are not lack- ing. Initial assignment of the Hamilton labyrinth- odont to a dissorophoid family did not present any problem. The gilled Micromelerpetontidae and Bran- chiosauridae (Boy, 1972) were clearly not nearest relatives. Of the terrestrial families, the Trematopi- dae were easily dismissed also. (The usual spelling of this family name, Trematopsidae, is not properly formed (from the genitive singular of the Greek ops, opos) and should be corrected according to the Code (Milner, 1978)). The Trematopidae are distinguished by antorbital openings that communicate with the external nares and by a somewhat squared snout (Berman et al., 1985). The Hamilton dissorophoid does not show these features. The best match was with the characters of the family Dissorophidae. Most notably, the squamosal of the Hamilton laby- rinthodont bears a semilunar flange that projects ventrally into the otic notch (DeMar, 1968). The dissorophids are best known for their dorsal dermal armor plates. Their skulls usually are orna- mented with characteristic superficial ridges and exostoses. The Hamilton dissorophoid lacks these striking features, but they are lacking also in Amz- phibamus of the Pennsylvanian, Tersomius of the Lower Permian and Micropholis of the Lower Trias- sic, all heretofore considered dissorophids. These 7 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 PENNSYLVANIAN Meee MIDCONTIN.| APPALACH. -Hamilton Pittsburgh Falmouth Garnett, Howard Pitcairn CONE= MAUGHIAN McCOR- FOViiSse MICKIAN ean VILLEAN Rise le unarmored, normal-skulled genera have always been placed at the base of proposed dissorophid phyloge- nies because of their apparently unspecialized na- ture and the great age of Amphibamus (Westphalian D equivalent). Boy (1985) recommended removal of these atypi- cal genera from the Dissorophidae, and suspicion is growing among other recent workers (Berman et al., 1987) that they cannot be included in any reasonable LOCALITIES UPPER CARBONIFEROUS Kounova-]- - --- SiiEs PHANIAN Nyrany- Florence- PICTOUAN Fenton- CUMBER- miles a Newsham- Airdrie- Joggins- Aime Time relations chart of well-known fossil vertebrate localities of the Pennsylvanian and Upper Carboniferous. definition of the family. To test the idea that these unspecialized dissorophoids and the Hamilton dissorophoid are members of a long-lived distinct group, specimens of the North American Amphiba- mus and Tersomius species were assembled for reconsideration. As a result of this study and an extensive literature survey, this hypothesis is sup- ported and familial reassignments are proposed. ACKNOWLEDGMENTS I would like to express my gratitude to Larry D. Martin, William E. Duellman and Hans-Peter Schultze of The University of Kansas for their advice and criticism during the production of this paper. Thanks are also due to those institutions that lent their specimens for study. The author has profited from conversations with Andrew R. and Angela C. Milner. The holotype was acid-prepared by Edward O. Wiley, The University of Kansas, and Donald Baird of Princeton University made a cast of the type of Amphibamus lyelli for me. The SEM (scanning electron microscope) photomicrographs were made by Lorraine Hammer of The University of Kansas. This paper was improved by the comments of two anonymous reviewers. A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) 3 ABBREVIATIONS Used in figures: Seen Beet cat meade tet casa septomaxilla Fe cc cetaokd vonaccovs aduduetenes frontal I scoanaponcecsznsobectssenentce costo Sqvemoss J eat Oe Rect noaatue een een supratemporal Os TS el et eo ie AT Ses coh, a tiems tiie Falak econ s LaDULAr IRR cheaters rcdocnccsssesessneuetes lacrimal MD ee er ane Nici ine tee ae titers Litised sais Noha than ides 3h vomer INR ceo sacictucebecananactsedencs nasal Institutional acronyms are the following: Ameri- PPO ee ee eck hvasiteeessanesesneentn’ parietal can Museum of Natural History (AMNH), Field Rg PRR eee oceisasweresarsesseseeds palatine Museum of Natural History, University of Chicago IPE coos uccSCeSSCCTE INE eee eee postfrontal (FMNH UC or UR), Emporia State University (HQ), NN ARs ca ssecack sccaneasiens premaxilla Museum of Natural History, The University of Kan- TO) cocccb se iar nana eee eee postorbital sas (KUVP), Museum of Comparative Zoology Pere ce Secceessossenseccdcies postparietal (MCZ), University of California at Los Angeles NRG eo esens rte sa.initenanacettonne prefrontal (UCLA VP), United States National Museum of Oe sede selblcaheccosdcaptecease quadrate Natural History (USNM), and Yale Peabody Mu- (QU) cose ceoseeeeceneeine meee seer quadratojugal seum (YPM). SYSTEMATIC PALEONTOLOGY Class Amphibia Order Temnospondyli Superfamily Dissorophoidea Family Amphibamidae Type genus.—Amphibamus Cope 1865. Revised diagnosis.—Dissorophoid temno- spondyls, distinguished as follows: 1. Pleurocentra extend ventrally to meet, or almost meet, below the notochord; 2. Ribs short posterior to the shoulder re- gion, such that the rib series lacks a thoracic basket. In addition to these apomorphic characters, a constellation of primitive ones is usually present, as follows: small external nares, stapedial foramen, moveable basal articulation of the braincase, iliac blades curving posteriorly, overlapping bony scales. The short ribs of amphibamids are distinct from those of trematopids in that they are not flattened or overlapping in the thoracic region (Williston, 1909; Case, 1911). The reduced axial rib pair of bran- chiosaurs (Boy, 1978) is absent in amphibamids. Taxonomic note.—Peliontidae Cope 1875 is the oldest family name attached to the type genus. The combination Pelion lyelli was created by Cope (1868) for Wyman’s (1858) Raniceps lyelli because the generic name was preoccupied, and he created the family name soon after. Pelion Cope 1868 1s a junior homonym of Pelion Kirby 1858. According to Ar- ticle 39 of the Code, a family name is invalid if the name of its type genus is a junior homonym, so Peliontidae should be replaced. The family name Amphibamidae Moodie 1910 is used instead. Eoscopus, new genus Etymology.—From Greek eo “dawn” and sko- pos “watcher.” _ Diagnosis.—Amphibamid dissorophoid with a relatively narrow interorbital bridge; eight pairs of shoulder ribs that begin on the axis, with the first five pairs more robust; a single spatulate pair of sacral ribs; a deep contact between pelvis halves for the anterior two thirds of their length, followed by an interdigitation; terminal phalanges knobbed. 4 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 Eoscopus lockardi, new species Named in honor of Walter Lockard. Diagnosis as above. Taxonomic note.—This name was published in Daly (1976) without a taxo- nomic description, thereby creating a nomen nudum. The correct date of publi- cation is therefore that of this paper. Holotype-—KUVP 80408. Skull and anterior vertebrae, part and counterpart (Fig. 2). Paratypes.—Listed below. All skulls are dorsoventrally flattened and split through the dorsal bones unless otherwise indicated. KUVP 47270. Posterior dorsal part of a skull and a long series of vertebrae. KUVP 47272. Skull fragment, including a set of circumorbital bones and a palpebral cup. KUVP 49491. Presacral vertebral col- umn, pelvis, and femur. KUVP 50000. Premaxilla and maxilla. KUVP 80409. Skull and anterior verte- brae. KUVP 80411. Skull, presacral vertebral column, girdles, and femora. KUVP 80412. Posterior part of whole animal, with tail and lower part of one leg missing. KUVP 80413. Skull, part and counter- part. HQ 14. Partial skull, including orbits and anterior portion. HQ 15. Partial skull, lacking part of right side. HQ 122A-B. Skull and anterior verte- brae, part and counterpart. HQ 123A-B. Pelvis and one hind leg, part and counterpart. HQ 127. Half pelvis and one hind leg. HQ 247. Pelvis of young individual (pu- bes lacking). HQ 250A-B, D-I. A largely complete specimen, preserved in the round. Most of the skull and tail are missing, as well as the distal part of one hind leg, the interclavicle, and cleithrum. HQ 252. This specimen is the counterpart of KUVP 80412. Fig. 2. Eoscopus lockardi, gen. et sp. nov. The holotype, part and counterpart, x 1.5. KUVP 80408A is at the top and KUVP 80408B is at the bottom. A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) 5 HQ 254A-B. Partial foot, part and counterpart. HQ 255B. Partial skull, posterior half. HQ 258A-—B. Skull, part and counterpart. HQ 260. Skull. HQ 424A-B, HQ 428. Partial skull, some vertebrae and limb bones. HQ 431. Vertebrae, ribs, femur, tibia, and fibula. HQ551A-C. Nearly complete skeleton with partial body shadow, some scale impressions, and dark eyeball remains; part and counterpart. HQ 598A—D; HQ 599A-D. Partial skeleton, in- cluding a laterally compressed skull, vertebrae in line, and scattered limb bones. HQ 641. Scapulocoracoid and humerus. HQ 716A-B. Partial skull, part and counterpart. HQ 720. Disrupted skull and complete vertebral column. Description General features.—The size and proportions of Eoscopus are easily observed in specimen HQ 551A, B, and C (the poorly preserved remains of a nearly whole animal). The head of HQ 551 is nearly half as long as the trunk (3.5—7.5 cm). The head of HQ 598 and 599 is two-fifths the length of the trunk (3.5 to 8.5—9.0 cm). This is also the proportion of KUVP 80411 (3.3-7.5 cm). The tail of HQ 551 can be seen on part B as an impression on the stone, when the light is almost parallel to the surface. It is estimated to have been 5.7 cm long when complete, about three fourths as long as head and body together. Individual HQ 252- KUVP 80412 has both trunk and tail; the tail is about four fifths as long as the trunk (6.3—7.5 cm). The hindlimb of HQ 551 is approximately 4.5 cm long (three fifths of trunk length), and that of HQ 252-KUVP 80412 is about 5 cm long (two thirds of trunk length). It seems that the individual with a longer tail also has longer limbs. HQ 250, a larger animal with a trunk 12—13 cm long, has a hindlimb that is three fifths of trunk length. The humerus of HQ 250 is essentially the same length as the femora (2.6, 2.7 cm), and its ulna is the same length as the tibia (1.6 cm), but the radius is shorter (1.4 cm). In general the proportions of the limb segments are primitive; thus the distal parts are shorter than the proximal ones and the radius is 0.54 of the length of the humerus (Hotton, 1970). Scale impressions are faintly visible in the dark- ened areas around some of the bones of HQ 551. Apparently Eoscopus was completely covered in overlapping, rounded scales, arranged in vertical rows around the body. They are clearest at the wrists and ankles, which are unossified in this individual. No dermal armor or ossicles have been found asso- ciated with any specimen, except in the eyelids and palate. Some of the specimens were prepared for study by embedding them in plastic and removing all of the matrix with 10% acetic acid. After this treat- ment, it was possible to see the internal structure of the long bones and the neural arches, because the articular surfaces were cartilaginous. The interior of the round bones is filled with a meshwork of bony tubules. Skull.—Nearly all of the skull specimens are flattened dorsoventrally; those that are preserved in the round are fragmentary. The dermal bone pattern is shown in the reconstruction (Fig. 3). Measure- ments were made of the distance from the ventral edge of the quadratojugal to the lateral edge of the supratemporal to approximate the greatest height of the skull. This distance is about one fourth to one third of the midline length, indicating a rather low skull. Skull measurements are presented in Table 1. Because the bones are split in most Hamilton specimens, the outer surfaces must be observed on the acid-prepared skulls KUVP 80408 and KUVP 47270. Eoscopus possesses the usual ridge-and-pit surface ornamentation of the outer skull bones. Lateral line canals are absent, as are the ridges and exostoses seen on the skulls of most dissorophids. The sutures are unfused. The most noticeable feature of these skulls is the relatively large size of the orbits, and the accompa- nying narrowness of the interorbital bridge. Orbit length varies from 0.30—0.42 of the total length in the midline of the skull. There does not seem to be any associated trend with increase in size. A ridge, borne on the prefrontal, frontal, postfrontal, and postorbital bones borders each orbit around the medial half of its circumference. The eyes of Eoscopus were covered with large palpebral cups, composed of a mosaic of small, cornered pieces of bone (KUVP 80413). The intact cups closely approximate the prefrontals, frontals, UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. Eoscopus lockardi, gen. et sp. nov. Reconstruction of the skull in dorsal, ventral and lateral view, xX 2. Jeno 3), A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) Table 1. Selected skull measurements of Eoscopus lockardi in cm. Specimen Skull Muzzle Orbit Table Bridge Table Skull Postorb. Height/ number length length length length width width height bar length length KUVP 47270 — — — — 1.41 JED 2d) i — KUVP 80408 3205) 1.0 LZ LSD 0.65 1.85 1.05 OFS 0.28 KUVP 80409 3.8 1.4 1.4 oe 0.42 — = — KUVP 80411 Jed 1.06 eS 0.85 0.54 LGB) 0.95 0.45 0.29 1.28 2.42 KUVP 80413 3:65 Ihodl 1.45 ileal 0.65 2.05 1.05 0.6 0.29 HQ 14 — Le 1.6 — OST — — — — H@-15 so) 2.0 2.0 1.8 LS) — IESFeStan lkO) OW HQ 122 dae Ite 1.6 eZ 0.62 Dies 1.0 0.7 0.26 HQ 258 325 teil 1.0 ZS 0.5 — — OWS — HQ 260 3.85 SVU) 1.45 ILS) 0.6 ZANS 1.0 0.65 2. HQ 598-599 3.4 1235 ON95 0.95 0.45 — OW, 0.65 oD and postfrontals and leave a crescentic area uncov- ered on the lateral side of each orbit. Many other specimens have partially preserved cups, and in all cases the bony pieces have the same size and shape. Ocular plates were present in the eyeballs (KU VP 47272 and HQ 551). It is not possible to say how many made up a complete ring but 15 plates is the minimum (KUVP 47272). The infilling of the canal is a prominent marker of the lacrimal bone in many specimens. The foramina of the orbital surface are preserved only on KUVP 80408. This acid-prepared specimen shows three foramina in each orbit, arranged linearly on the left lacrimal and triangularly on the right. The contact between the prefrontal and the palatine expected in a dissorophoid is internal to the orbital surface of the lacrimal. For confirmation of this, it is possible to observe the size and shape of the disarticulated palatine in the right orbit of KUVP 80408. The dorsal part of the palatine extends from a short rostral process of the prefrontal to contact the jugal at the midventral border of the orbit. The external exposure of the palatine probably is confined to the anteroventral rim of the orbit. There is no external exposure of the ectopterygoid. HQ 14, 122A, and 258B have the bones of this area in articulation. In the intact skull there is sometimes a definite comering anterodorsal to each orbit, as the roof passes from top to side. This, together with its inturning into the front wall of the orbit, gives the prefrontal an inflated appearance in some cases. In some specimens the prefrontal is only slightly curved, whereas in others it is cupped. It is best observed on HQ 14, 122A, and KUVP 80408 and 80413. An- other visible cornering usually is present at the anterior apex of the otic notch, between the skull table and the cheek. The external nares are observable only on HQ 258B. They are small and anterolateral in position. On all other skulls they are concealed or obliterated by displacement of the surrounding bones. The septomaxilla is not observable on any specimen. A rostral fontanelle, bounded by the nasals and pre- maxillae, is present between the nares. This feature is best viewed on HQ 14, 15, and 260. The dorsal processes of the premaxillae are long and close to the nares. They fit into notches in the nasals, which are floored with bone. As a result, the dorsal processes cannot be viewed without remov- ing some of the nasal bone ventral to them. HQ 14 and 260 were prepared to expose the processes. The notch medial to each naris is a feature of the nasal and is exposed on KUVP 80408. It can be seen in outline on HQ 14 and 551C. An internal nasal flange was discovered in a few other dissorophoids (Bolt, 1974a). It is present in Eoscopus andis partially preserved on KUVP 80408, where it can be seen through a gap in the anterior portion of the palate. On the ventral surface of the nasal bone, the long flange extends anteromediaily from the apex of the prefrontal. A posterior part of the flange extends along the anteroventral edge of the prefrontal from the anterior corner to the curve near the orbit. The flange may include a lacrimal 8 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 contribution, but the ventral surface of the lacrimal bone is not preserved on KUVP 80408. ihe otic notch is lareessas expected ston dissorophoids, curving concavely ventrally from an anterior apex. Asupratympanic flange is present, but is preserved only on KUVP 47270 and HQ 598A. Its ventral edge has a squamosal semilunar flange, typical of dissorophids (DeMar, 1968). A supratym- panic shelf (Bolt, 1974c) extends from the skull table above the supratympanic flange. The area ventral to the shelf is unsculptured, as in trematopids (Berman et al., 1985) and (apparently) /ratusaurus (Gubin, 1980). Ventral to the shelf, near the end of the skull table, the supratemporal bears a ventral projection, a semilunar flange (Bolt, 1974c). There is a tabular contribution to the supratympanic flange also. It is difficult to determine whether the squamo- sal has a small contact with it or not. Only KUVP 80408 shows the condylar area well. The subotic parts of the squamosal and quadratojugal completely cover the posterior portion of the quad- rate. They are delimited from the lateral sculptured parts by a definite rim. The posteromedial process of the quadratojugal nearly encircles the dorsal process of the quadrate. The lateral portion of the condylar surface is borne by the quadratojugal cover in Eo- scopus. The dorsal process of the quadrate is a strong knob, wider than long, with a rounded top. Eoscopus possesses short tabular horns, which vary somewhat in shape and angle and can be seen on KUVP 47270, 80408, and 80413, and HQ 15 and 598A. The horns turn slightly ventrally from the plane of the skull table. On KUVP 80408 the horn has a light pitting on its outer side, rather than normal sculpturing. This is not true of KUVP47270, another acid-prepared specimen. The round parietal foramen is always present a short distance behind the orbits. This opening lacks a raised rim. The posterior border of the skull table is marked by a grooved ridge running from side to side. The occipital cover is not well preserved on any specimen. Some of it seems to be present on KUVP 80413 and HQ 15.A sheet of dermal bone probably extends to the foramen magnum and ventrally on both sides as two lobes, and then shortens laterally until it disappears into the skull table again at about the middle of the tabulars. The ventral surface of the _ dorsal part of the postparietals are exposed on HQ 250A. A round hole, located somewhat laterally, occurs in each bone. The holes may represent rugose attachment areas that were pulled off when HQ 250 was split into its parts. The marginal teeth are large and curve inward at the tips. Broken teeth have thin walls and large pulp spaces. Externally, labyrinthine grooving seldom is visible, but when larger teeth are chipped, ridges on their inner walls leave vertical stripes of dentine on the matrix infilling. Size and spacing of teeth is similar in upper and lower jaws. Counts of teeth and tooth spaces of selected upper jaws are presented in Table 2. Countable upper jaws of Eoscopus possess 40-65 teeth; the number varies with the length of the bones. The palatal view of the skull is dominated by the large interpterygoidal vacuities. Braincase bones cannot be seen through them, or posterior to the parasphenoid; these bones were unossified in all specimens. The bones surrounding the vacuities are covered with large curved denticles, and the vacu- ities themselves were covered with denticulated skin. The denticles are borne in small groups on bony platelets (HQ 15). The palpebral cup ossicles can be distinguished from the palatal platelets by shape: palpebral cup bones are cornered, whereas palatal platelets usually are larger and more rounded. A similar covering for the vacuities has been re- ported for the dissorophids Broiliellus hektotopos (Berman and Berman, 1975) and Kamacops acervalis (Gubin, 1980). Steen (1931) found them on Am- phibamus lyelli and Stegops divaricata, and Romer (1930) described them on Evpetosaurus,acolosteid. Fang-and-pit pairs are present in the positions typical of temnospondyls. The best prepared vomers, on KUVP 80408, are patchily set with strong re- curved denticles, which point toward the postero- medial corners of the bones. There are three pairs of fangs-and-pits on the right vomer and four on the left. The largest fang is located at the anterior end of the internal naris. Its recurved tip points posteriorly. Another occurs on the medial border of the naris. On the left vomer, another smaller fang is located on the border just posterior to it. The remaining fang lies in the middle of each vomer, between the naris and the midline of the palate. The smaller fangs are not as erect as the one at the end of the naris, and they are slightly more curved and point strongly in the same direction as the denticles. The palatal pattern of KUVP 80409, a second A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) 2) Table 2. Measurements in cm and tooth counts of upper jaw bones of Eoscopus lockardi. Left above, right below. Specimen Premaxilla Premaxilla Maxilla Maxilla number length count length count KUVP 50000 eS wk 4.89 44 KUVP 80408 OFA tL — — OWS 1 a — KUVP 80409 0.80 12 not 35 0.81 W2 measurable B37 KUVP 80413 0.74 WZ 2.62 3 HQ 14 0.70 10 27 oS 30 0.76 10 incomplete 30 HQ 15 Ite) 16 4.23 44 HQ.122A 0.70 10 2.86 34 OWS is 2.82 36 acid-prepared specimen, is somewhat different (Fig. 4). Only the fangs at the anterior ends of the internal nares are present. The remainder of the vomers is covered with a uniform carpet of denticles, so that it is certain that other fangs were absent. Either there is a great deal of individual variation in vomerine dentition, or some kind of polymorphism is present. Fig.4. Eoscopus iockardi, gen. et sp. nov. The palate as seen on KUVP 80409, x 2. It is possible that this specimen might represent another amphibamid taxon; however, it differs from KUVP 80408 only in this character. A large palatine fang-and-pit present at the pos- terior end of each naris 1s comparable in size to the anterior set. The tip of this fang points pos- teromedially. On KUVP 80408, a small fang is located medial to this larger one, but this is not the case with KUVP 80409. Like the palatine, the ectopterygoid bears a fang, the tip of which points pos- teromedially. The exact shape of the vomers is not preserved because of breakage anteriorly and posteriorly. Nevertheless, most of the vomer is visible. Posterior to the trans- verse posterior borders of the premaxillae is a rounded, denticle-free space. This probably is the internarial pit found in some other dissorophoids and which is flattened into the plane of the rest of the palate in both of these acid-prepared speci- mens. No internarial foramen has been observed on these specimens. Both the upper and lower sides of the ectopterygoid are observable on KUVP 80408. The contacts of its anterior end are visible on KUVP 80413. On the palatal surface a wide smooth groove extends anteriorly from the posterior edge and to an elongate foramen. In dorsal view, the foramen is larger and rounder than in ventral view, and an anteroposterior in- 10 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 ternal divider is visible. The medial border of the pterygoid is thickened, so that the interpterygoidal vacuity is rimmed with a prominent heavily den- ticulated ridge for about half its length. The ridge is low anteriorly, but enlarges posteriorly and finally transforms into the round, hollow projection of the internal process of the pterygoid. The denticles that cover the ridge continue onto this process for most of its length and pass posteriorly onto the quadrate ramus also. On KUVP 80408, denticles exist only near the internal process, but on 80409, they extend halfway to the squamosal contact. The internal processes of the pterygoid are pre- served on KUVP 80408, 80409 and 80413. Their open ends are confluent with V-shaped openings in the posteroventral sides of each process. In life, the pterygoid shells covered cores of palatoquadrate cartilage. A similar morphology was described for the dissorophid Ecolsonia (Berman et al., 1985), which has an ossified braincase. In this animal, a triangular posteroventral part of the basipterygoid process fits into the V-slot of the internal process of the pterygoid. | For the lateral half of its length, each internal process of the pterygoid is attached to the convex anterior portion of a dorsoventrally compressed lamina ascendens. This transverse sheet of bone, that normally is oriented vertically, 1s now curled due to flattening of the skull. Laterally it merges with the quadrate ramus just above the internal process of the pterygoid, while its free end extends a little further medially than the process. Anepipterygoid 1s present in the left interptery- goidal vacuity of KUVP 80408, lying lengthwise. It consists of a conical base, unossified ventrally, and a strong ascending process. The epipterygoid ossification could not have participated in the lat- eral side of the basal articulation, because the pterygoid completely surrounds that part of the palatoquadrate. The most prominent feature of the parasphenoid is its elevated, triangular denticle patch at the base of the rostrum. This area is preserved on KUVP 80408 and KUVP 80409, along with most of the rostrum. The apex of the triangular denticle patch projects well below the palate. From there the surface slopes posteriorly and at its base, bends to the level of the rest of the parasphenoid. Recurved -denticles carpet this triangular area closely and uniformly. The two foramina for the internal ca- rotid arteries are in the lateral faces of the denticle platform. They are invisible ventrally because the edges of the triangular area are laterally produced. On KUVP 80409, shallow grooves can be seen curving medially to approach the foramina at a 45° angle. Most of the slender parasphenoidal rostrum is present on KUVP 80409. As it approaches the vomers, two narrow ridges rise along its sides, creating a shallow ventral trough. The rostrum probably widens where it contacts the vomers. The anterior corners of the posterior plate are everted and convex, while the transverse anterior edge remains straight. Each protrusion once covered the posterior and ventral sides of a cartilaginous basi- pterygoid process. Unfortunately the posterior parts of the parasphenoid are not well preserved. A stapes is approximately in position on the right side of the parasphenoid of KUVP 80408. It is a robust, perforated rod with an enlarged proxi- mal portion. A second stapes lies inside the curve of the lamina ascendens on the left. Its unfinished footplate shows its tubular interior structure. KUVP 80409 has one stapes preserved, lying behind the skull (Fig. 4). - Lower jaw.—Skulls of Eoscopus usually are preserved with the lower jaw. In spite of the num- ber of jaws in the collection, the only adequate examples for description are the acid-prepared KUVP 80408 and 80409. The symphyseal region of both mandibles is entirely visible. The expected fang-and-pit pairs at the anterior ends of the bones are present. At the symphysis, the dentaries enlarge to double the contact area, resulting in a projection into the angle of the jaw. This strengthening must exist because the jaw is slender at the symphysis. The inner end of each dentary is hollowed ventromedially by the meckelian sulcus. The anterior splenial follows the sulcus and most likely participated in the symphy- sis, although all four of the preserved splenials are broken before reaching this point (Fig. 5). The posterior ends of the jaw are not as well preserved as the anterior ends and nothing can be said about the articulation or the adductor fossa. A single meckelian fenestra can be seen on the right side of KUVP 80408. It is a small, oval opening between the prearticular, angular, and postsplenial. A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) 11 The exterior surface of the left jaw is visible, and the angular bone is more heavily sculptured than the dentary. This is also true of Dissorophus and Broiliellus (DeMar, 1968). The angular in Foscopus lacks the ventral keel present in Dissorophus, how- ever. Sharp, recurved denticles cover the dorsomedial part of the jaw and extend from the region of the symphysis to the presumed position of the adductor fossa. The coronoid bones that bear the denticles cannot be described because of their position. The contact between the two splenials is well preserved on KUVP 80409. Vertebrae.—Six specimens have complete se- ries of presacral vertebrae, lying in line and with the arches more or less in position. These are HQ 551, 598-599, and 720, and KUVP 47270 and 80412. Despite their completeness, none of these is unobscured in all parts of its column, and only KUVP 47270 can provide a vertebral count. Most likely, KUVP47270 has 24 presacral vertebrae (Fig. 6). Itis impossible to count the postsacral vertebrae of either KUVP 80412 or HQ 551, but there must have been at least 50. Caudal vertebrae diminish Fig. 5. Eoscopus lockardi, gen. et sp. nov. Reconstruction of the lower jaw in ventral view, x 2. Inset, an enlarged view of — the anterior end of a half jaw. rapidly in size posterior to the ribs. The presacral neural arches are best observed on the acid-prepared specimen KUVP 47270 and on HQ 250F. In lateral view, the neural arch of each middorsal vertebra shows a long posterior sweep behind the pedicel. The neural spine surmounts this unsupported posterior region. The shapes of the neural spines vary regionally; on the first seven arches, the spines are square in profile, whereas for a typical middorsal arch each spine is a smoothly arched hump (Fig. 7). The anterior zygapophyses face anteriorly by 30-40°. The arches are not co- ossified in the midline, and the right and left halves are often shifted relative to each other. Internal views of the neural arches are available on KUVP 47270 and 49491. The vertebrae are exposed in ventral view, and the inside of the slightly spread arches are visible between intercentra. On the inner surface of each neural arch pedicel, Eosco- pus bears a pit, which must have lain opposite the spinal cord in life. On KUVP 47270, a large speci- men, the pits shown on arches 10—15 are deep and open, except on the 11th arch, in which the opening is half covered by bone so that it faces posteriorly. KUVP 49491, a much smaller specimen, has shallower pits surrounded by thin rims restricting the opening. The arch pedicels rest on triangular para- chordal processes, the apices of which lie halfway down the sides of the vertebral centra (HQ 250F). The posterior edge of each triangle is noticeably longer than the anterior one. Elongated transverse processes, corresponding to the joined rib heads, in- cline anteroventrally, parallel to the poste- rior edges. The articular surfaces face posteroventrally and laterally. They were finished in cartilage rather than bone, like the zygapophyses. Preservation of transverse processes depends greatly on the extent of ossification at death. In KUVP 47270, the upper borders of the rib facets appear to be the lower edges of the parachordal processes, while on KUVP 80408 transverse processes are not visible at all. Although many specimens have verte- brae, few have centra well-enough preserved to be readily interpretable. The intercentra are much more commonly observed than UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 2 ‘TX OLTLH dANM ULUNJOD FLIGS}IDA dy} JO SOP] ‘AOU “ds Jo “Uds “IPAPyO] SndodsOy *g ‘B14 A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) 13 Fig. 7. Eoscopus lockardi, gen. et. sp. nov. Reconstruction of a middorsal vertebra in lateral and ventral views. Based on HQ 250F, x 7. pleurocentra, which are less ossified, especially in younger individuals. The central elements are well displayed on HQ 250K, where a short segment of the middorsal column is preserved with all the elements well ossified and lying almost in their original posi- tions. The height of the intercentra is not great when compared to their other dimensions. Their posterior edges are angled where they lose contact with the pleurocenira and contact the neural arches. On a complete, well-ossified specimen, it is possible to distinguish the posterior face by the presence of this angle. None of the intercentra in the collection bears visible rib facets, with the single exception of one on HQ 250F. In this case, the facet is just above the posterior corner. The well-ossified dorsal pleurocentra of HQ 250F are large in lateral view. Their long dorsal contact may have been co-ossified above the notochord. Like the intercentra, the pleurocentra have corners at the level of the ventral apices of the parachordal processes. However, the pleurocentra have corners on the anterior edges rather than the posterior. Ven- trally, the pointed tips of the pleurocentra approach each other closely, usually preventing contact be- tween adjacent intercentra. The centra of Eoscopus resemble those of Micropholis in this respect (Broili and Schréder, 1937). Specimen KUVP 49491, which has been partially uncovered by acid treatment, re- vealed a damaged presacral column in ventral view, where the pleurocentra are clearly visible (Fig. 8). On KUVP 49491, pleurocentra are well pre- served in the round. These do not conform entirely with the pleurocentra described from HQ 250F; they lack the dorsal development that allows a long contact above the notochord. The pleurocentra are roughly rhombic with the anterior corner slightly ventral to the middle of their height. The posterior, rounded corner is high and the pleurocentra may have had slight contact above the notochord. The difference between these pleurocentra and those of HQ 250F may reasonably be ascribed to ontologic variation, because KUVP 49491 is a smaller indi- vidual. Pleurocentra are observable on KUVP 80408 also. The vertebrae of this specimen are laterally compressed, and only the pointed dorsal halves of the pleurocentra of one side are preserved as impres- sions. There is less dorsal development than in KUVP 49491. There is no dorsal, posterior corner, and the pleurocentra may not have touched above the notochord. The position of these vertebrae is farther anterior in the column, and this may influ- ence the size and shape of pleurocentra. KUVP 49491 affords a view of the inner or notochordal surfaces of the central elements. Andrews and Westall (1970:272) published a de- scription of this part of the osteolepiform fish Eusthenopteron that also applies to KUVP 49491. The inner surfaces of the central elements are made of unfinished bone of an unusually close texture, and are covered with irregular vertical ridging. KUVP 80409 and 47270 provide lateral views of the atlantal neural arch. This arch is two thirds as 14 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 Fig.8. Eoscopus lockardi, gen. et sp. nov. Femur, pelvis and lumbar vertebrae, KUVP 49491, x 2.5. long anteroposteriorly as those following it, and it exists as two half-arches, unfused to any central element. Transverse processes are absent. The ante- rior zygapophyses face anteriorly at about a 45° angle. The neural spine rises on a slope, so that its posterior corner lies just above the hindmost point of the posterior zygapophyses. The axis is nowhere well preserved, but what can be seen of it on KUVP 47270 and other specimens indicates that it is similar to the vertebrae posterior to it, except that the neural spine is anteroposteriorly enlarged. [t bears transverse processes and ribs, as can be verified from KUVP 80408. The sacral arch can be seen on KUVP 47270, with an associated sacral rib (Fig. 6). In isolation, it can be readily identified by its enlarged transverse processes and anterior zygapophyses. The arch immediately preceding the sacral arch is corre- spondingly marked by enlarged posterior zygapo- _ physes. The pedicels of the sacral arch are somewhat larger than those of other vertebrae, and show heavy vertical grooving that fans onto the anterior zygapophyses and the base of the neural spine. The neural spine is not preserved but it seems prob- able that it was higher and more square than those of its neighbors. Eight rib-bearing vertebrae fol- low the sacral. The last of these bears the first haemal arch, as in Eryops (Moulton, 1974). KUVP 80412 shows an array of haemal arches in lateral view. Unfortunately, the central elements are unossified. The haemal spines are robust and diminish in length posteriorly in a regular fashion. Four vertebrae with haemal arches terminate the series preserved on KUVP 47270. The haemal canal occupies about half the height of the whole structure (Fig. 9). The legs on either side of the canal are laterally compressed and are narrow when viewed from the anterior or posterior. The spine _ itself is round and heavy, and sock- eted at the end, indicating that it was continued by acartilaginous segment. The neural arches do not have lengthened neural spines in the tail. The five anteriormost caudal arches have spines with a low square profile. More posteriorly, the anterior corner disappears and the spines slope toward the highest point at the rear of each arch. The dorsal portion of each caudal neural arch is smaller relative to the parachordal parts than is true of the dorsal vertebrae, and the caudal neural arches are not lengthened behind the pedicels. Ribs.—All the ribs of Eoscopus are double- headed, with the heads conjoined with bone. They often show two articulating faces on the heads, meeting at a projecting apex at about 140°. The distal segments of the ribs are dorsoventrally flattened and have square ends. Regionalism is evident (Fig. 10). The first pair of ribs is borne by the axis. Length increases to the fourth pair, then diminishes through the eighth pair. Behind the shoulder ribs, a series of similar dorsal ribs extends to the sacrum; thus, there is no pectoral region in the rib series of Eoscopus. There is a single A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) Ee. 9: of a haemal arch in lateral and ventral views, based on KUVP ATD IO), KOs = sacral pair, and eight pairs of caudal ribs follow the sacrum. A complete series of shoulder ribs is displayed on KUVP 80411 and 80412. The anterior five pairs are preserved on one side of KUVP 80408, and a few lie near the disarticulated anterior vertebrae on HQ 122B. The shoulder ribs curve gently; the longest of them is about half again as long as the ribs following. They are narrowest about one third of their length from the proximal end. The ribs vary regionally, in that the sixth, seventh, and eighth pairs are shorter than the four preceding ones (Table 3). The fourth pair, or the third and fourth pairs are the longest. The larger ribs sometimes are visibly angled where the shaft rotates to the plane of the distal end. On KUVP 80412, the five anterior pairs are markedly more robust than the three following, and the second pair has very wide distal ends. The third rib pair has the second largest distal ends on this specimen. The dorsal ribs are all about equal in size. They are narrowest just distal to their heads, which are shaped like isosceles triangles. From their narrow- est diameter, the dorsal ribs expand evenly to their square distal ends, which are narrower than the joined heads. Examples of the massive sacral rib are available on KUVP47270, and HQ 250H, and 252. The sacral rib has a large oval head, a thick waist located at a third of its length, and a wide-bladed distal portion. That of KUVP 47270, the most complete available, is about 13 mm long, 5 mm wide at the head, and 9 mm at the distal end. The sacral arch bearing it is about 7 mm long. The head and the flat blade are Eoscopus lockardi, gen. et sp. nov. Reconstruction only slightly turned with respect to each other. Grooving similar to that on the arch pedicel runs longitudinally on the waist of the rib, spreading onto tiemblades Rhee wriby 1s slightly arched from end to end. The sacral rib of HQ 250H differs from the others is that it is curved from edge to edge (con- vex dorsally) also. A series of caudal ribs is visible on KUVP 47270, and three pairs are partly exposed on HQ 250H. The two anterior pairs are straight, strong, and square-ended; they are nearly as long as the sacral rib. The remaining six pairs curve poste- riorly. Of these, the middle four pairs are thick, square-ended, and extend two vertebrae posteriorly. Their two heads are delimited by an indentation in the bone bridg- ing them. The two last pairs of ribs are short and pointed, and their heads cannot be distin- guished. Pectoral girdle —Four speci- mens possess parts of the pecto- ral girdle. HQ 250E has most of a scapular blade, possibly part of another scapulo-coracoid, and most of aclavicle. HQ 551 has a pair of scapulocoracoids, a pair of clavicles, and a bone that may be the interclavicle. The third specimen, HQ 641, bears a scap- ulocoracoid with the scapular part preserved as an impression. KUVP 80411 has the inter- clavicle and possibly both clavicles. The scapula of HQ 250E is short and broad, and has superfi- cial sculpturing consisting of fine vertical grooves, each with a —_ N ITT TOT TOTTI TI ig. 10. Eo- scopus lockardi, gen. et sp. nov. Reconstruction of the rib series, based on KU VP 80412A andKU VP 47270, x 2. 16 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 Table 3. Measurements of anterior ribs of Eoscopus lockardi in cm. Rib KUVP 80408 KUVP 80411 KUVP 80412 pair left left right left right Ist 0.70 0.67 — 0.70 2nd 0.80 O72 0.73 0.74 3rd 0.88 0.80 O¥S 0.74 4th 0.88 0.83 0.77 0.78 Sth 0.73 OF. 0.80 0.74 0.72 6th —_ O72 0.72 0.68 0.64 7th — OFZ 0.70 0.66 _ 8th — 0.69 0.70 0.69 9th _ 0.50 0.50 0.55 — 10th — 0.54 = _ — 11th a 0.49 0.47 — — small pit or foramen at its ventral end. HQ 551A is a much smaller animal, and the scapulocoracoids are quite different in shape. They are waisted and symmetrical on either side, like bow ties. The best scapulocoracoid is that of HQ 641 (Fig. 11). The bone is crescentic, but the lower arm is truncated rather than pointed. At the transition from the scapu- lar portion to the coracoid portion of the bone, it bends at an angle slightly more than 90 degrees. Because the edge of the angle 1s not parallel to a line bisecting the crescent, but rises anteriorly, the cora- coid portion has more area than the scapular. A deep impression in the stone preserves clearly the thick, robust area of the supraglenoid buttress. The bone is much thickened below and behind the buttress, until near the posterodorsal corner. Although the coracoidal portion is preserved as bone, this lateral area was removed in order to demonstrate its thick- ness. The glenoid fossa is not preserved, nor are there any traces of foramina; however, a deep pit is present on the inner surface just below the glenoid area. Evidence for its existence was a smooth- topped stone projection, uncovered while removing the lateral part of the coracoidal plate. The projec- tion has been accidently removed. The pit may be the subscapular fossa, which ought to be present in this position. However, the pit is narrow and no foramina appear to open into it. The split interclavicle of KUVP 80411 (Fig. 12) presents the medial side of its exterior half. Its _ silhouette is approximately hexagonal, with the two anterolateral faces concave. The exterior anterior half bears grooves, which radiate anteriorly from the center of the bone. Thirteen of them can be counted as matrix stripes. Originally the anterior border of a /\ Oe -alel vd Fig. 11. Eoscopus lockardi, gen. et sp. nov. the scapulocoracoid and humerus of HQ 641, x 2.5. A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) 1) the interclavicle may have been scalloped or den- tate. The exterior surface is ornamented where it is not grooved or overlapped by the clavicles, as can be seen from the irregular appearance of the postero- medial portion of the bone. The clavicle of Eoscopus (Fig. 12) is long- stemmed and lacks surface ornament on its ventral expanded portion. The dorsal portion is well pre- served on HQ 250E and is as tall as the scapula lying slightly posterior to it. It curves smoothly where it rounds the anterior edge of the scapula (and presum- ably the ventral end of a cleithrum). A low keel arises near the ventral end of the stem, becoming most prominent just below the angle and descending toward the point of the leaf-shaped ventral plate. The angle between the stem and the plate is about 135°; in addition, the plate is rotated clockwise with respect to the slanting stem, so that it lay more or less horizontally in the body. An enigmatic feature of the clavicle of Eoscopus is the flat triangular process projecting from the anterior border of the stem, at about a third of its length. This process is visible on both the clavicles of HQ 551A, and 1s relatively larger in this smaller individual. A possibly homolo- gous anterior process was described for the clavicle of Dissorophus multicinctus by DeMar (1968). The cleithrum described for Branchiosaurus by Boy (1978) muchresembles a rotated Eoscopus clavicle, but note that in Branchiosaurus the triangular pro- jection is on the opposite (articular) side of the stem. Pelvic girdle —There are eight pelves preserved in the collection, the best of which is HQ 123A and B. All of the bone was removed from the part and counterpart of this specimen, and silastic casts made, which display the internal and exterior surfaces of most of a whole pelvis (Fig. 13). In the next best specimen, HQ 250H, there is an impression of the interior surface of a half-pelvis. Both of these pelves have the pubes, which were unossified in smaller specimens (KUVP 49491, 80411, and 80412, and HQ 127, 247, 252, and 551A). No puboischial suture line is visible on the two with complete ventral plates. Both acetabula can be seen on the cast of HQ 123B, and the upper part of another is exposed on KUVP 49491. The articulation surface is distinct, and has the expected posterodorsal indentation. An- other, smaller scallop is present anterodorsally, such that the surface has an upper lobe. The acetabulum has projecting rims dorsally and ventrally, but there is no supra-acetabular buttress. The iliac blade is straplike with a squared end, and curves posteriorly. Its anterior edge bears a Fig. 12... Eoscopus lockardi, gen. et sp. nov. Reconstructions of the interclavicle and clavicle, x 5. Left: the interclavicle, based on KUVP 80411. Right: the clavicle, based on HQ 250E. 18 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 prominent hump, such that the ilium is almost two- pronged in silhouette. At the hump, the dorsal edge of the iliac blade bends and then runs straight posteri- orly. The ventral, more robust edge curves more gradually. This feature of the ilium is well shown on the acid-prepared specimen HQ 252. The internal surface of the blade 1s faintly undulated, marking the sacral attachment. The cast of HQ 123 bears an additional deep groove through the dorsal edge of the iliac blade, running obliquely from the interior to the exterior. The groove pierces the crest immedi- ately posterior to the bend in the iliac blade. The groove 1s continued by three ripples in the internal surface. The suture line between the ilium and the ischium is marked on the internal surface of the pelvis by a slight ridge, accompanied by faint grooving perpen- dicular to it. On the external edge of the pubois- chiadic plate, ahump opposite the acetabulum marks where the bones meet. The suture line between the ilium and the pubis is not well marked; the ridge and the grooving diminish and disappear as they pass anteriorly. The midline joint is best observed on HQ 123A, in which the contact is exposed. The halves of the puboischiadic plate originally met at an angle slightly greater than a right angle. The union between the right and left halves was deep and indicated by a ridge on the interior midline along the anterior two thirds of the plates. Posteriorly, the contact becomes a peg-and-socket arrangement. The right plate bears a rounded projection, which originally fitted into a matching indentation on the left plate, at the posteriormost point of the- midline contact. The areas immediately anterior and posterior to the “peg” also interlock slightly. It is possible that a somewhat similar arrangement is present in the pelvis of Ecolsonia; Figure 12B in Berman, et al., (1985) is suggestive. On HQ 123, the pubes are sharply everted or downturned along their anterior edges. On the ven- tral half (HQ 123B) this eversion is apparent as a ridge running between the acetabula, starting just posterior to the anterior corner on one side and passing ina slight arc across to the other acetabulum. In the interior view of the pelvis the pubes meet at an angle anteriorly, but this is not reflected in the edge of the deflected portion. The everted anterior portion of the pubis must be the area of origin of the puboischiofemoralis internus muscle in Eoscopus. The cast of HQ 123B, showing the external side of the pelvis, has two large, round obturator fo- ramina in the expected position. They are not visible on HQ 123A ina position opposite that of the outer = Higa l3: internal surface. Right: the left external surface. Eoscopus lockardi, gen. et sp. nov. The pelvic girdle, from casts of HQ 123A and B, x 3. Left: the nght A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) WD openings. Instead, two large foramina are visible on the downturned, anterior borders of the pubes. Ap- parently the canals run anterodorsally to open ante- riorly. The pubes of Koscopus are similar to those described for Eryops (Case, 1911), considering their difference in width. Forelimb.—The humerus is moderately ex- panded at its ends and has a definite, if short, shaft with a round cross-section. Good examples of the humerus can be seen on HQ 250H (Fig. 14) and 641 (Fig. 11). Poor ones are present on HQ 551A. The plane of the distal end of the humerus is rotated about 45° with respect to the plane of the proximal end of HQ 250H. There seems to be less rotation in the humerus of HQ 641, but this may be the result of flattening. The ends of both bones are smoothly rounded without processes or condyles. A deltopectoral crest projects from the ante- rior edge of the humeral head at right angles to the axis of the bone. This feature is a little more prominent on the smaller humerus (HQ 641) than the larger (HQ 250H). The bone is otherwise fea- tureless except for fine lon- gitudinal grooves on the sur- face of the head. The distal portion of the forelimb is present on HQ 250H and 551A. The ulna bears an olecranon on the proximal end of HQ 250H; this is uncertain in the case of HQ 551A. The radius and ulna are equally thick and long if the olecranon is not considered. The radius and ulna bow slightly toward each other as they lie on HQ 250H. A keel runs down the poste- rior face of the ulna, from the proximal end of the olecra- non to the distal end. Nine wrist bones are present on casts made from HQ 250G and H. The wrists and ankles are unossified on HQ 551A. The nine bones probably do not make up the complete carpus, because Olson (1941) found eleven in the carpus of Acheloma cumminsi (as Trematops milleri). The element lying with the first metacarpal most likely is the first distal carpal, whereas the small nubbin between the last two metacarpals may be the fourth. The three bones just distal to the radius may be the ulnare, fourth centrale, and radiale, with the small intermedium considered missing. The other four carpals are interpreted as distals and centralia, with one of their number miss- ing. The four metacarpals can be observed on both Fig.14. Eoscopuslockardi, gen. et sp. nov. The forelimb, from HQ 250G and H, x 3. Above: humerus, ulna and position of natural molds. Below: radius, carpals, and metacarpals, from casts. 20 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 HQ 250G and H and 551A. The first metacarpal is the thickest and is about two thirds as long as the others. Metacarpals two through four become pro- gressively more slender. They are straight, except for the fourth, which bends slightly towards the third. The third metacarpal has a smaller distal end than the others. All the metacarpals are short. On HQ 250, only one phalanx is present, and on HQ 551A, both front feet are superimposed. Thus it is not possible to count the phalanges, but Eoscopus appears to have a rather blunt forefoot, with short digits. A terminal phalanx is unobscured on HQ 551A. It narrows abruptly from its articulation, as if to form a claw, but instead ends with a slight bulb. Hindlimb.—Good examples of the femur are with specimens KUVP 49491, and HQ 123A and B, 250H and I, 252, and 431. Three specimens have been prepared with acetic acid and the other two have had the damaged bones removed from the matrix and silastic casts made of the spaces where they lay to allow observations of their intact sides. Perhaps the best femur is that of the casts of the part and counterpart of HQ 123, which show both sides of the bone (Fig. 15). Hindlimbs exist on specimens HQ 127, 253, and 551, but these are not well pre- served. The proximal end of the femur, viewed end-on, has approximately the shape of a curved teardrop, with the tail at the posterior side of the head. Appar- ently there are no distinctive features of the upper, convex surface of the femoral head; the rugosities on the femoral cast of HQ 250I are artifacts of preparation and the groove on the cast of HQ 123B is absent from other specimens. There 1s a prominent internal trochanter distal to the end of the femoral head, projecting at almost right angles to the plane of the head. In well-ossified specimens, such as HQ 123 and 250, the internal trochanter is stepped below the proximal end and separated by a notch. This morphology was described for Caerorhachis (Holmes and Carroll, 1977) and Doleserpeton (as Cardiocephalus in Gregory et al., 1956; fide Bolt, 1969). The adductor flange continues the projection of the trochanter distally along the shaft of the bone. There is no posterior ridge and no fourth trochanter. On the two larger femoral specimens, the _ intertrochanteric fossa extends for about two fifths of the length of the bone. The adductor flange crosses the long axis of the bone distal to the intertrochanteric fossa, in passing from the internal trochanter to a terminus slightly closer to the poste- rior condyle. The flange diminishes in height dis- tally, and finally blends into the shaft at the apex of the triangular popliteal concavity. On the dorsal side of the distal end, the intercondylar fossa extends about a fourth of the length of the bone. In end view, the posterior condyle is square and the anterior one more rounded, except for a projection marking the anterior side of the popliteal space. The distal end of the femur is straight on the shaft, or nearly so, and Eoscopus was probably able to completely extend its leg. Good specimens of the tibia and fibula exist on HQ 123A and B, 252, and 431. Both bones have widely expanded ends, except for the distal end of the tibia, which is more modestly enlarged with a flat oval cross-section. As is typical in lower tetrapods, the tibia and fibula bow toward each other and the fibula is shorter than the tibia in most specimens. The fibula is nine tenths the length of the tibia in three measured specimens; however, in one case (HQ 123A and B) they are the same length. On both bones, the distal end of the bone is rotated about 45° with respect to the proximal end. This feature sup- plies half the torsion necessary to point the pes anteriorad. The expanded ends of the bones leave room for each other by lying at right angles; that is, the proximal end of the tibia is extended anteroposteriorly and that of the fibula mediolaterally, whereas the distal ends have the opposite relation. The proximal end of the tibia is bilobed in end view, with the posterior lobe smoothly rounded and the anterior lobe sharply angulated by a prominent cnemial crest. The crest extends distally along the proximal expansion of the bone, which is about a third of its length. Both inner and outer surfaces of the tibial head bear grooving radiating proximally. An unexpected feature of the tibia is a second crest running down the posterior or internal side of the shaft. This is revealed on the acid-prepared specimen HQ 431. The crest arises from the distal part of the head, and disappears at the distal expan- sion of the bone. A posterior crest has been described for the tibia of Cacops (Williston, 1910), but it may not be homologous. The crest of Cacops begins at the proximal end toward one side and has a small hump at its midlength. Possibly these crests increase attachment area for the interosseus cruris. In modern salamanders this muscle holds the tibia and fibula A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) Fig. 15. _ Eoscopus lockardi, gen. et sp. nov. The hindlimb, anterior view, HQ 123B. Right: posterior view, HQ 123A. from casts of HQ 123A and B, x 3. Left: 21 Mp UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 together and assists in transfer of weight between them (Schaeffer, 1941). The wide, flattened ends of the fibula have about equal areas in lateral view. It is not possible to observe them end-on in any specimen. The head of the fibula is flat or slightly convex in profile. Its distal end describes an S-curve, very noticeable in posterior view (HQ 123A), but more gentle in lateral or anterior view (HQ 252). This curved surface conforms to the articular surfaces on the intermedium and fibulare. The high medial corner of the distal end articulates with one side of the intermedium. A broad groove passes distally across the high corner (HQ 123A), possibly marking the passage of an artery connecting with the perforating artery which passes through the tarsus immediately distal to this point. An ossified ankle with the bones in position 1s available in the collection. HQ 123A and B are part and counterpart bearing a pelvis with the right hind limb, lacking only part of the foot. The incomplete remains of the bone have been removed and silastic casts made for study from the impressions on the two sides. The posterior side of the tarsus is dis- played on the cast of HQ 123A, and the anterior side on the cast of HQ 123B (Fig. 15). Part of HQ 250H also bears an ossified tarsus, which was prepared in the same way, but it is disarranged and incomplete. The tarsus was unossified in all other hindlimb specimens available. The tarsus of Eoscopus resembles that of Trematops, as described by Schaeffer (1941). The name 7rematops is now considered synonymous with Acheloma (Dilkes and Reisz, 1987) but it 1s convenient for Williston's and Schaeffer's name to be used in this section. The tarsus specimen was provided by E. C. Olson, then of the University of Chicago; it is not further identified. Apparently the pretarsale present on the medial side of the ankle of Trematops is lacking in Eosco- pus. This bone cannot be seen and there is no space for it to articulate with the distal end of the first centrale. On the whole however, it is appropriate to describe the tarsus of Eoscopus by comparing it with Schaeffer’s account. Gregory et al., (1923) described the carpus of Eryops as being originally arched, but flattened _ before fossilization; this also is the situation with the tarsus of HQ 123A and B. Its tarsal elements are a little hollowed on their ventral surfaces, indicating that the whole complex arched from side to side in life. Schaeffer (1941) did not mention such a feature for the tarsus of Trematops, perhaps because his specimen provides a view of the dorsal surface only. In that view it is not obvious that the tarsus of HQ 123 was originally arched; it seems only to have a rather broad foot. Individual tarsal elements differ in the details of their outlines from one surface to the other. This was true for Trematops also, and it occasioned some of Schaeffer’s corrections of Williston’s (1909) origi- nal description. The type specimen affords a view of the plantar side rather than the dorsal. A good example of the resulting disagreements 1s the case of the intermedium. The intermedium of Eoscopus and Trematops 1s a large bone between the tibia and the fibula, with a free proximal border that rises toward the high medial corner of the fibula. Its other borders contact the fibula, tibia, fibulare, and fourth centrale. Williston (1909) described the intermedium of Trematops without mentioning the tibia, but Schaeffer (1941) described and figured the bone with a large facet contacting the tibia, to which it was apparently ligamentously attached. The source of the error is revealed when both sides of a similar ankle are examined. The facet for attachment to the tibia is plain in anterior view (HQ 123B) but invis- ible in posterior view (HQ 123B). In both views, the intermedium has a definite rounded notch for the perforating artery in the border toward the fibulare. This must be a real difference between Eoscopus and Trematops, because neither Williston nor Schaeffer found it in the latter, although Williston thought that there was a gap between the intermedium and the fibulare. In plantar view, the fibulare of Eoscopus is nar- rowest at its distal end rather than at its contact with the fibula, as reported by Schaeffer for the dorsal view in Trematops. Otherwise it conforms to the description of Trematops, with the same exception as that noted for the intermedium. The proximal end in posterior view has a marked step cut into the medial corner. When the fibula, fibulare, and intermedium were naturally aligned a gap existed between them at this corner, such that it continued A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) 33 the space enclosed by the notch in the intermedium. On the dorsal side of the fibulare, a notch cut in the border matches that in the intermedium, such that the fibula does not participate in the rim of the passage on the dorsal side of the ankle. The tibiale is like that of Trematops, except that it is much narrower. It articulates with the first, second, and fourth centralia. The fourth centrale, the widest bone in the tarsus, most clearly shows the arch of the ankle in plantar view. An end view of this bone is possible on the cast of HQ 250H, in which the axis of the arch, the thickness of the bone, and its articular surfaces are evident. The fourth centrale resembles that of Trematops. Ventrally, the centrale is nearly rectan- gular, but larger on the tibial side, as shown in Williston; dorsally, its distal edge is V-shaped to provide two articulation surfaces, as described in Schaeffer. In Eoscopus, the fourth centrale contacts the tibia, as Schaeffer said it probably did in Trematops. The other three centralia differ in details of shape from the descriptions of Trematops. The first centrale of Eoscopus is narrow and elongated to fill the space between the tibiale and the first tarsale. The second centrale is almost square in anterior view, but a little narrower mediolaterally in posterior view, to con- form with the curve of the ankle. The third and smallest centrale deviates from a square only in that its proximal edge rises toward the fibular side in dorsal view, to follow the lower edge of the fourth centrale. The shapes of the first two tarsalia of Eoscopus closely match those depicted for Trematops by Schaeffer. The first tarsale is a little squarer than in Trematops. It is slightly hollowed dorsally, proximodistally rather than from side to side. Its ventral surface is not preserved. The second tarsale is shorter than that depicted for Trematops, and the smallest of the series in Eoscopus. The third tarsale of Eoscopus is more complex in shape. Its fibular edge is strongly concave, en- croached upon by the fourth tarsale. Articulation with the second tarsale is straight, but the proximal border is concave to articulate with the second centrale. The proximal edge is distal to the third centrale and rises toward the tibial side. There is a slight central depression in the dorsal surface of the third tarsale, whereas the ventral surface is a little hollowed proximodistally. Largest of the tarsaliais the fourth, as in Trematops. Unlike the relations described for Trematops, the fourth tarsale of Eoscopus does not contact the fifth metatarsal, or any metatarsal other than the fourth. Its greater width overlaps toward the third metatar- sal, but the distal corner of the third tarsale prevents contact. The tibial side of the fourth tarsale is a long convex curve that borders the third tarsale, third centrale, and fourth centrale. Contact with the fifth tarsale is short and straight. The proximolateral corner of the fourth tarsale appears truncate, be- cause the articulation with the fibulare is straight and diagonal, matching the inner of the two slanting faces on the distal end of the fibulare. The line of contact on the tarsale 1s everted, so that the edge is raised and the articular surface faces slightly dor- sally (covered in life by the fibulare). Inspection of Schaeffer’s Figure | shows that Trematops’ fourth tarsale is also like this. When seen in plantar view, the lateral side is quite different: the distolateral corner of the tarsale is rounded and everted, and the portion near the fibulare is slightly cornered. The fifth tarsale is narrow distally, to match the narrow-headed fifth metatarsal. It becomes wider proximally by enlarging laterally until it covers the remaining half of the fibulare. Its proximal face slants ventromedially, along with the lateral sloping facet of the distal end of the fibulare. The line of contact is everted on the fifth tarsale, as it is on the fourth. In plantar view, the fifth tarsale is not wider dorsally, everted, or hollowed. Consideration of all the tarsalia leads one to believe that the shape of the tarsus of Eoscopus is more complex than a simple arch. A dorsal concav- ity must exist at each side of the more distal part of the ankle. This interpretation accounts for the dorsal hollowing of the first tarsale, and the everted proxi- mal articular surfaces of the fourth and fifth tarsalia. The physical traces are considerably more marked laterally on the ankle because the long straight fibulare concentrates the curvature. The peculiari- ties of the plantar surfaces of the tarsalia are not readily explained. The metatarsals are best observed on the silastic casts of HQ 123A and B, which display them in dorsal and ventral views. KUVP 80412 and HQ 252 24 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 provide another view of the first two metatarsals, not wholly preserved on HQ 123. On these specimens the first metatarsal is thickest and about two thirds the length of the others, the fourth metatarsal is the longest, and the fifth has smaller ends than the others. The metacarpals are 70% as long as the metatarsals, as can be verified by making measure- ments of HQ 250. As is typical of temnospondyls, the hindfeet of Escopus are longer than its forefeet. The heads of the metatarsals are marked on both sides with a few short longitudinal grooves of vary- ing depth. As can be seen from either of the casts of HQ 123, the metatarsals are not quite straight, but are slightly skewed medially, as if the twist neces- sary to point the foot anteriorad could not be pro- vided entirely by the ankle. The phalangeal formula of the pes of Eoscopus is 2—2—3-4—3. This was determined by inspection of the acid-prepared specimen HQ 252 (Fig. 16). The same formula has been reported for Micropholis (Broiliand Schroder, 1937) and for Micromelerpeton and Branchiosaurus (Boy, 1972). DeMar (1968) said that the formula for Dissorophus was 2—34—4— 2, but his judgment may have been influenced by that given by Schaeffer for Trematops. The pes of Dissorophus, shown in DeMar’s Figure 17, can be interpreted more plausibly as 2—2—3-4—3. The pha- langeal formula of Escopus appears to be common among temnospondyls. As with the front toes, the terminal phalanges of the hind toes have slender, delicate ends. These are not clawlike, butare straight and knobbed rather than pointed. The terminal pha- langes of Branchiosaurus cf. B. petrolei and Micromelerpeton are known to have slightly en- larged ends (Boy, 1972). This toe form may be widespread. A reconstruction of the possible appearance of Eoscopus in life is presented in Figure 17. Amphibamus Cope 1865 The coal-swamp amphibian genus Amphibamus is composed of three known species from three different Westphalian D localities (see Fig. 1). Speci- mens of A. lye/li (Wyman, 1858) from Linton, Ohio, and A. grandiceps Cope 1865 from Mazon Creek Illinois, were examined. Amphibamus has been carefully described from detailed, articulated specimens. Wyman (1856) first published a preliminary note on this new “batra- chian” found at Linton. Many authors have dealt with the genus Amphibamus subsequently, includ- ing Cope, Moodie, Romer, Steen, J. T. Gregory, Watson, Carroll, Bolt, and A. R. Milner. Modern descriptions are Watson (1940, A. grandiceps), Gre- gory (1950: A. grandiceps), Carroll (1964, A. lyelli), Bolt (1979, A. grandiceps), Milner (1982, A. lyelli and A. grandiceps) and Hook and Baird (1984, A. lyelli). The two American species were studied from Fig. 16. Eoscopus lockardi, gen. et sp. nov. The ilium and hindlimb, HQ 252, x 2.5. A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) M5) Fig. 17. - Eoscopus lockardi, gen. et sp. nov. Reconstruction of appearance in life, based on HQ 551, x 0.875. casts of natural molds. The holotype of A. /yelli, AMNH 6841, was made available as a plaster cast from the original coal block, showing the dorsal side of the animal. A latex cast of AMNH 2002 shows a lateral view of the posterior body and anterior tail of a larger individual. USNM 4461, a block of coal bearing an impression of the posterior trunk and a hind leg, was loaned for study. The A. grandiceps specimens examined were three split Mazon Creek concretions, each showing most of a small individual in part and counterpart. They include YPM 794, the neotype (Fig. 18); YPM 795, the type of Mazonerpeton; and FMNH UC 2000, the type of Miobatrachus. The two Yale nod- ules were prepared for silastic casting by complete removal of their mineral infiiling, exposing clear impressions. Amphibamus specimens fall into size classes. The largest A. grandiceps (YPM 794) is about half the size of the smallest A. /yelli (AMNH 6841). All the nodule skeletons are of juveniles, although postmetamorphic (Carroll, 1964; Bolt, 1979). De- spite their age difference and their great similarity, the specimens represent two species. Differentiat- ing features include the number of teeth and the vertebral count (Carroll, 1964). Milner (1982) adds that A. grandiceps is smaller at metamorphosis, and that A. /yelli has a wider skull. His examination of a skull of a young A. J/yelli made his distinctions possible. | Because Amphibamus was previously described, _ my original intent was to only point out characters of taxonomic importance. However, a number of inter- esting features, previously unknown, were discov- ered and the following description has therefore been broadened. Skull.—The quadrate of Amphibamus grandiceps is much like that of Doleserpeton (Bolt, 1968). It is best viewed on the right side of the skull of FMNH UC 2000, where the lower jaw does not cover it (Fig. 19). The quadrate has a prominent dorsal process on its exposed medial corner, with a groove running up the posterior side and over the end (Bolt, 1979). This may be seen on the right side of YPM 794 also. The medial condyle is almost a mirror image of the dorsal process. In posterior view the quadrate is Y- shaped, with the arms facing medially, such that the articular surfaces face slightly outward. The quadratojugals of FMNH UC 2000 have a straight medial border with no sign of a posteromedial process on either side. YPM 794, however, shows this character on both sides. The posteromedial process in A. grandiceps covers part of the posterior side of the dorsal process of the quadrate, rather than embracing it as in Eoscopus. The quadratojugal does not contribute to the condylar surface, contra Watson (1940), but as observed by Bolt (1979). A ventral prefrontal process is visible in the right orbit of AMNH 6841. Other features of the orbital area are not certain. A laterally exposed palatine may be present in the ventral rim. Another character diagnostic of the Dissorophoidea, the supratym- panic flange, has never been observed on Amphiba- mus. The preservation of available specimens does 26 UNIV. KANSAS MUS. NAT. HIST. MISC. PUB. No. 85 Fig. 18. Amphibamus grandiceps Cope. The neotype, YPM 794, x 4. A cast of the dorsal impression. not permit inspection of the dorsal side of the otic notch. Dorsal processes of the premaxillae have been depicted for the American species of Amphibamus (Milner, 1982:fig. 3). Examination of the cast of AMNH 6841 revealed narrow, sharp processes on either side of the muzzle. On the two A. grandiceps skulls examined, the dorsal parts of the premaxillae are not preserved, but the slots they once lapped into _are observable on the nasal bones. On the left nasal of YPM 794 the dorsal process was leaf-shaped; on the right it had a medial point and a lateral lobe (Fig. 20). The left nasal of FMNH UC 2000 bears an impression similar to the left nasal of YPM 794 (the right is damaged). In FMNH UC 2000, the points of the dorsal processes were joined across the midline by thinner extensions of the premaxillae; possibly a juvenile characteristic. The form of the right pre- maxilla of YPM 794 has been described for Platyhystrix (Berman etal., 1981) and is depicted on A NEW AMPHIBAMID GENUS (AMPHIBIA: TEMNOSPONDYLI) ye) the right side of the muzzle of Micropholis (Broili and Schroder, 1937). A nearly complete palate is visible on YPM 794 and FMNH UC 2000. All of the dermal bones are covered with denticles except the anteriormost part of the vomers. Denticulated skin is absent, unlike the palate of A. /yelli (Carroll, 1964). Tubercle- bearing platelets described by Steen (1931, as Platyrhinops) were also absent. This does not mean that such a covering was not originally present, or that it might not appear later in ontogeny. The parasphenoidal denticle field continues an- teriorly on the rostrum, borne on a raised strip that conforms to the width of the bone. The denticle field terminates in a point. A. grandiceps specimens have a groove for the internal carotid that crosses the base of each basipterygoid process, resulting in sharp borders on the denticle field. Each groove becomes a canal into the braincase, just posterior to the basal articulation. An indentation in each side of the denticle field marks the position of these canals. YPM 794 bears a vomerine fang on the anterior end of one internal naris, but not on the other. A fang is present on each palatine and ectopterygoid. In contrast, FMNH UC 2000 has three small fangs or large denticles on the anteromedial rim of each internal naris. There are no palatine or ectopterygoid fangs. The row of palatine teeth described by Watson (1940) is present, as confirmed by Bolt (1979). There are additionally three more palatine teeth on the opposite side of the palate, close to the maxillary series. The palatine row included eight to ten teeth when intact, set along the contact with the maxilla. Change with age is the most plausible explanation for the difference between YPM 794 and FMNH UC 2000 (Bolt, 1979). Similar ontological change ap- parently occurs in A. lye/li as well (Milner, 1982). Details of the basal articulation are well pre- served on the two nodule skulls (Fig. 21). The internal process of the pterygoid and the basiptery- goid region of the parasphenoid are seen in near-life positions on the left side of FMNH UC 2000. The processes appear to be transverse cylinders with an oblique contact. The palate of YPM 794 affords separated views of the pterygoid and basipterygoid processes. On the right, the free end of the basiptery- goid process faces anteriorly and slightly laterally. It consists of a projecting rim enclosing an oval de- pression, indicating that it was unossified except for a bony cover. On the left, the internal process of the pterygoid is a narrow hollow cylinder laterally. The cylinder enlarges to match the wide oval of the opposite side of the joint. The concave internal surface of the internal process of the. pterygoid faces posteriorly. The epipterygoids of A. grandiceps are ossified with large ventral ends. As in Eoscopus, they could not have participated di- rectly in the basal articulation. A reconstruction of the basal articulation of A. lyelli, shown in Carroll’s Figure 22 (1964), re- sembles the morphology in A. grandiceps. The accompanying text stated that the relevant areas of the best available specimen were damaged and that the reconstruction was made using Tersomius from Archer City. Tersomius is depicted in Carroll’s Figure 4. Milner (1982) presented the palate of juvenile A. /yel/i with aconventional reconstruction of the articulation; 1.e., a right-angle rather than oblique contact. Thus the configuration of the basal articulation of A. /yell/i is still undescribed and unknown. ~ Vertebrae.—Carroll (1964) stated that AMNH 6841 has 25 presacral vertebrae and this has been verified. YPM 794, the neotype of A. grandiceps, has 20 presacrals. FMNH UC 2000 seems to have 20 presacrals, as Watson (1940) thought it might. For my count of FMNH UC 2000, I identified the sacral as the last vertebra with a wide pedicel (as did Watson), because I cannot discern sacral ribs on my cast. In general, the neural arches of Amphibamus are as described by previous authors. The atlantal arch of A. lyelli resembles that of Eoscopus, except that A. lyelli’s neural spine does not slant posteriorly as much and the arch is fused to acentral element. The atlantal arch of A. grandiceps is visible in lateral view on the ventral cast of YPM 794, and is not fused to its centrum, possibly due to its youth. The large parachordal processes of Amphibamus are much like those of Eoscopus. The transverse pro- cesses parallel the posterior edges, which are longer than the anterior edges. Only the dorsal portions of the transverse processes are preserved on FMNH UC 2000 and they are as Watson (1940) described them. The impressions of the nodule halves of YPM 794 were completely cleaned for this study. The ‘0007 ON HNWH 3U8re “p6L NdA ‘JOT ‘ssvo Jo sydersosorwiojoyd WAS “S| x ‘sjeseu oy) UO oe]]IxeW1d oY) 10J suOTssoidut oY “odoD sdaripunis snupqiyduy QZ “314 ‘Op X ‘IYSU oy} UO ssad0id [esIOp ‘QOOT ON HNWd Jo aespenb iY4311 oy) JO MOIA JOLIO\SOd & WYSTY “OT X “P6L WdA JO YOU ONO JYBL OY] JO MATA JOLIO}SOd & :1Ja'T ‘s}Svd Jo sydessosOIWOOYd PAS Ul aJespenb oyf *edoD sdaripupis snupqiyduy “6{ “314 a) CO fo) Zz ns — oO. s Z = = Z an 2 < Z Yn — 5 Y < Y Z,