ERYOPSID REMAINSREMAII\S FROM THE ,

BRAXTON COUNTY, WESTUTEST VIRGINIA

By

JameJames s L. Murphy Case WesternWestern Reserve University Cleveland, OhioObio

ABSTRACT

Well-preserved , pectoral girdle, limb and vertebral ele­ele- mentsrnents of an specimenspecirnen have been found in a roadcut near Sutton, Braxton County, West Virginia. The amphibianarnphibian remains occurred in a green siltstone of unquestionable mid-Conemaugh mid-Conernaugh age. Taxonomic ciif­cLif­ciif- ficultiesficultie s involving-theinvolving-involvingithethe genus GlaukerpetonGlauke rpeton RomeRomerRorner r and speciation within the genus Eryops are briefly discussed.discus sed.

INTRODUCINTRODUCTION TION

The amphibianarnphibian specimen described in this paper was discovered by the author in July, 1969, in a roadcut between one and 1.l. 1 I mile southeast of the southern end of the Elk River bridge at Sutton, Braxton County,Counfy, WestWe st Virginia. Hennen (1917) published a stratigraphic section measured along this highway (now U. S. Route 19)t9) by R. M. Gawthorp. Numerous changes in the path of the highway, uncertainties regarding the aneroid elevations cited inHennenrsin Hennen's descriptionde sc ription of the section, a strong down section dip component,cornponent, vagueness of the upper limittirnit of the described section and lack of key ormarkeror marker beds of distinctive lithology have rnademade reinterpretation of this section difficult. A second visit to the region was madernade in May, 1971,LJ7 I, to confirm the stratigraphic occur-occur­ rence of the .

STRATIGRAPHIC OCCURRENCE

The 545foot545 foot sequence measured by Gawthorp consists primarily of sandstone, siltstone and variegated shale and c1ay.clay. The only coal in the section (Brush Creek coal, elevation 950t)950') is no longer exposed. The "Ewing Limestone'rLime stone" can still be seen in the ditch on the westwe st sideofside of the road, a few feet below the 1200 foot contour. HennenrsHennen's identifica-identifica­ tion of this limestone with the Ewing Limestone Member is suspect, however, and the nodular limestone probably represents the Rock o 265

, RiffleRiffle RunRun LimestoneLimestone MemberMember associatedassociated withwith thethe HarlemHarlem underclay.underclay. HennenrsHennen'sHen nen's t'Pittsburgh"Pittsburgh redred shalerroccurringshale" occurring 1515 feetfeet belowbelow thisthis freshwaterfreshwater limestonelime stone probablyprobably representsrepre sents thethe RoundRound KnobKnob ShaleShale MemberMember (Pitts-(Pitts­ burghburgh redbedsredbeds ofof some authors).authors). InIn anyany case,case, thethe RoundRound KnobKnob ShaleShale MemberMember lies.,abovelielies s "above thethe EwingEwing LimestoneLimeLimestone stone MemberMembeMember r andand notnot belowbelow it.it. TheThe onlyonly bedbed higherhigher in the sectionsection thatthat cancan bebe deemeddeemed ofof anyany stratigraphicstratigraphic valuevalue is a thin,thin, impureimpure freshwaterfreshwaterfre shwater limestonelimestonelime stone referredreferred toto thethe Elk LickLick LirnestoneLimestoneLime stone MemberMember by Hennen. IfIf this'identificationthis identification isis correct, thenthen thethe overlyingoverlying redbedsredbeds (30(30 feetfeet inin thickness),thickness), rrmassive"massive sandstonerrsandstone" (10(10 feetfeet inin thickness)thickness) and ttgreenish-gray"greenish-gray shalettshale" 15(5 feetfeet inin thickness) represent respeciivelyrespectively thethe Morgantown Redbed and Morgan-Morgan­ towntown SandstoneSandstone Members. TheseThe se units are well exposedexposed inin thethe lowerlower part of the extensive roadcutroadcut at thethe toptop of thethe hiLl,hill, elevation of thethe Elk Lick Limestone MernberMember being approximately 1280 feet. The remains were found within one to twotwo feet of thethe top of the "massivettrnassive"mas si ve sandstonerrunit,sandstone" unit, elevation approximately 1340 feet,fee t, on thethe east side of the highway, four to five feet above the pavement.pavement. AcceptingAccepting the correla'tionscorrelatiocorrelationsn s and eelevationsl eevvationsations of Hennen and Gawthorp,G aawtwthorp,horp, this unit is the Morgantown SSandstoneandstone MemberMember of the mid portionp ortion of the Conemaugh Group. Although a fewt'ew frafragrnentaryfragmentaryg m entar y eryopsid rremainse mains have been ddes-des­e s­ cribed previously fr fromom ststratar ata oofo{f ththee ConemaughConernaugh Group (Case, 19081908;; Romer, 1952),L95Zl, refreferrablee rra ble to eitheithere r EryopsEryoPs orot Glaukerp9l3g!gjot,Glaukerpeton,et on, ththee pre­pre- sent specimen is bettbettereerr ppreseivedpreservedreserve d ththana n ppreviouslyr eviously described materimaterial;a l ; it is, in fact, the finest eeryopseryopsidryopsidid specimsPecimene n yet discoverdiscoveredeedd in rockstocks of this age in the AppalachiAppalachiana n BasiBasin.n .

PRELIMINARYPRE LIMINARY DESCRIPTIONDESC RIPTION

Preparation

Only a small portion of thethe leftleft sideside of thethe skull roofroof and rightright mandiblemandible were exposedexposed inin thethe siltstone siltstone matrix.matrix. The specimen waswas carefullycarefully prepared, largely largely with a a White air abrasiveabrasive unit,unit, byby Mr.Mr. PeterPeter Hoover,Hoover, ClevelandClevelClevelandand Natural ScienceScienceMuseum. Museum. AdditionalAdditional elementselements werewere discovereddiscovered duringduring the the coursecourse ofof thethe preparationpreparation workwork andand thesethese areare alsoalso notednoted below.beiow.

SkullSkull

(Plate(Plate 1,l, figuresfigurefigures s 1-3)1-l-3)3 )

TheThe skullskull roofroof waswas badlybadly crushedcrushed andand considerablyconsiderably distorted,distorted, particularlyparticularly onon thethe rightright side,side, makingrnakiog accurateaccurate measurementsmeasurements impos­imPos- siblesible (Plate(Plate 1,l, figurefigure 2).2). MaximumMaximum lengthlength ofof thethe skullskull isis estimatedestimated atat 200200 mmmrn fromfrom muzzlernuzzle toto thethe tiptip ofof thethe leftleft quadrate.quadrate. InterorbitalInterorbital width,width,

266266 o

a

<.

---~ thethe parameterparameter leastleast distorteddistorted byby crushing,crushing, isis 4545 mm.mm. TheThe specimenspecimen isis thr:sthus considerablyconsiderably srnallersmaller thanthan eveneven thethe holotypeholotype ofof EryopsEryops avinoffiavinoffi (Romer)(Romer) andand muchmuch smallersmaller thanthan E.E. me€acePh$megacephalus (Cope),(Cope), E.E. willistoniwillistoni (Moodie)(Moodie)(Moodie ) andand E. grandisgrandis (Marshjl(Marshf(Marsh~ t-[?TheThe u"tir"entire skull roofroof islrnarnentedis-ornamented E. "t"11 byby aa finefine reticulationreticulation oror pitting.pitting. ThereThere areare aboutabout 60-70pits60-70pits perper squaresquare inch,inch, asas countedcounted onon thethe rightright postfrontalpostfrontal atat midmid orbit.orbit. TheThe leftleft nostrilnostril lieslies aboutabout 2020 mmmm fromfrom thethe tiptip ofof thethe rnuzzle.muzzle. TheThe distancedistance betweenbetween thethe nostrilnostril and thethe orbitorbit (left(left side)isside) is approximate-approximate­ lyly 77 rnm.mm. TheThe medianmedian parietalparietal foramenforamen isis obscuredobscured byby crushingcrushing and over-over­ ridi.ngriding of thethe leftleft postparietal.postparietal. IndividualIndividual bonesbones of thethe dermal roofroof areare notnot always easilyeasily delineated, due to thethe crushed naturenature of thethe skull, faintnessfaintne s s of thethethe sutures,suture s, and difficultydifficulty of distinguishingdistinguishing post-mortem breaks fromfrom sutures.suture s. InIn some instancestheinstancesins tances the specimen has broken along sutures, thoughthough not toto such a degreedegree as toto indicateindicatindicatee thatthat thisthis is neces-neces­ sarily a consequence of imrnaturityimmaturity inin the individual.individual. The sutures, in sos o far as theythey can be discerned,discerned, do not differ materially from thethe pat-pat­ terntern described by Sawin (194(1941) l) for E. ryephalus..megacephalus. Unfortuhately,Unfortunately,Unfortunately, the area occupied by the interfrontal is not ex-ex­ posed. TheThe right anterioranterior portion of the skull, including the right pre-pre­ maxillary and nasal,nasaln asal,, appear to have been shoved posteriorly so thatthethatthat the right nasal completely overlaps the interfrontal. The interparietal and interfrontal suture ca-ncanca,n be traced easily enougheno'..lgheno'.lgh anteriorly to the point where it is overlapped by thethe dislocated right nasal. In photographs and eveneven upon cursory examinationexarnination of the specimespecimen,n, it appears that the median suture continues anteriorly, uninterrupted by an interfrontalinterfrontal elernent.element. Close inspection, however,however, suggests that the right nasnasala l has been pushed som.esome 15l5 t020to20 rnmmm posteriorlyand anundetermineddistanceanundetermineddistanc e sinistrasinistrally.lly. This dislocation is thought to be sufficient to cover the interfrontal. Even so, belief in the presence of an interfrontal in this specimen is necessarilyn ecessarily somewhatsornewhat subjective, based as it is upon the hypothetihypotheticalcal restoration of various dermal elementselernents to their original positions. ItIt can be argued that thethe right nasal only slightly overlapsoverlaps thethe leftleft nasal, that a medianrnedian internasal suture continues a!1teriorlya!lteriorlyanteriorly and an interfrontalinterfrontal elementelernent is absent absent.. AlAlthoughthough thethe nasofrontal marginrnargin isis readily discerned, there does not seemseern toto be a pronounced indentationindentation thatthat would provide space for the interfrontal; interfrontal; on thethe other hand, thethe llaterallaterala tera l margin of thethe nasal isis not nearly so sstraighttraight as thatthat indicatedindicated forfor "Glaukerpe"Glaukerpeton"rrGlaukerpetonuton" avinoffi Romer, inin which thethe interfrinterfrontalinterfrontalontal isis pre­pre- sumedsurned toio be absent. RemovalRernoval of ththee rightright nasal mayrnay bebe necessarynecessary toto settlesettle thisthis importantimportant point concerningconcerning thethe WestWest Virginia specimen.specirnen. PostPosteriorPosieriorerior elementselernents ofof thethe cranialcranial roofroof areare poorlypoorly preservedpreserved andand incompleteincomplete inin somesorne instances.instances. TheThe rightright temporalternporal cannotcannot bebe satis­satis- factol'ilyfactorilyfactorily located,located, andand thethe rightright quadratojugalquadratojugal isis badlybadly ccrushed,crushed, rushed, shovedshoved dorsallydorsally andand anteriorly,anteriorly, severelyseverely damagingdarnaging thethe rightright squamosal.squamosal. TheThe rightright postorbitalpostorbital isis almostahnost entirelyentirely isolated,isolated, havinghaving been been shovedshoved intointo thethe orbit,orbit, andand therethertheree isis aa widewide gapgap betweenbetween thethe rightright postorbitalpostorbital andamiand thethe

267267

------rightright postfrontal.postfrontal. TheThe postparietalpostparietal ofof the rightright sideside and thethe posteposterior rior partpart ofof thethe parietalsparietals havhavee not been identifiedidentified anandd areare apparentlyapparently entire­entire- ly,ly, oror inin largelarge part,part, missing.missinrrrissing.g. Ventrally,Ventrally, mostmost of thethe bonesbones ofof thethe palatepalate areararee well prpreserved.eser ved. (Plate(Plate 1,l, figurefigure 3)3) The brain case,casecase,, though dislocated, isis intact,intact, withwith thethe right stapesstapes nearly inin place.placep1ace.. ThThee anterioranterior extensionextension of thethe para­para- spenoidspenoid andand thethe sphenethsphenethmoidsphenethrnoidmoid areare badlybadiy crushed,crushed, and thethe suturalsutural sur­sur- facesfaces adjacentadjacent toto thethe pterygoid are entirelyentirely exposed.exposed. TheThe posterior papartrt of thethe brain cascasee isis broken awayaway fromfrom thethe restrest of theththee brbrainain case, and onlyonly thethe lowerlolowerwer part of thethe exoccipitals, and thethe basbasibasioccipitalioccipitaloccipital and ththee foramenfoforamenr amen magnum can be readily seen. TheThe dorsaldorsaL part of the exoccipitals and theththee otic are missingmissing.. TheThe fenestrafenestra ovalis isis present on either side of the ventraventrall surface of thethe otic, butno trace of thethe Nvii foramenforamen ccana n be seen anterior toto the fenestrae.fenestrae. Nx and Nxii foraminaforarnlna are visible on the llefteft exocciexoccipitalexoccipi-tal,pi-tal,, but thethe rreste st of that eelementlementl ement has been broken away posterioposteriorly.r ly. Anteriorly the sphenethmoid region and ththee anterior part of the paraspaparasphenoidrasphenoidphenoid are badly crushed, exposing the vomeronasal nerve canalca.nal or first cranial nerve canal.can:rl. ThThee preorbprpreorbiialeorbitaitall flare and anterior end of the sphenethmoidsphenethrnoid region are poorly preservedpreserved.. The prevomerinprevomerinee toothtooth craters and posterolateral elevations are well displayed thouthoughgh disorienteddisdiscrientedo riented by crushicrushing.ng. An unusual feature is the presence of a double ectopterygoidectopterygoid ttoothooth andandpit pit on the right side. The left ectopterygoid crater is developed normally. The premaxillaprernaxillariesries areate !!otr:.otnot well weil expoexposeds e d ventventrally,rally, althoughalihough their commoncornrnon suture can be seen along the mamarginrgin of the left premaxil­prernaxil- lary. The numbernurnber of prprernaxill.aryemaxillary teeth and pitspii:s cannot bbee count­count- ed precisely; it is estimated,estirnated, on the leftllefieft side, at 12,!2, but mayrnay haveh ave been 13, as inir E. megacephalus.g."q"phrlor. MaxillaryM~illary teethteethr are estimatedestirnated ata t 37 on the right and llefteft sides, with about two-thirds of the pits filled. fil1ed. Teeth are largestl argest in the "canine"Itcaninerrregion regioregionn of the antanteriorerior part of the maxillary,rnaxillary, some of these at­at- taining the size of the larger l arg er of the prernaxillarypremaxillary teeth. PostPosieriorly,ereriiorlyorly,, the maxillaryrnaxillary teethteech decreased e crease rather gradually in size. The right pterygoidpte rygoid has been-been crushed posteriorly into the adad-ad­­ ducductorductortor fenestra. TheThe left adductor fenestrafenestra is well preserved, but the right is crushed and obscured by the right mandirnandible.mandible.ble.

PLATEPI,ATE 1I - Eryops cf.d. E. avinoffi (Romer) from thethe MorgantownMorgantown Sandstone Member,-nearMernber, near Sutton, Braxton County, West Virginia. Cleveland Museum of NaturalNatural Hi.story,Hi.story, no. 11025. ll0?5, 1.I. Left laterallatlateraleral view of skull, X 0.4. 2. Dorsal view of skull, X 0.4. 3,3. Ventral view of skull, with leftleft rnandiblemandible removed,removed, XX 0.4,O. 4. 4. Right , XX 0.9.O. 9.

268

------Southeastern Geology,Geology,, Vol. XIIIXIII,,, NoNo.. 44-Murphy- Murphy Plate 1I

,fs.1 t:ffidh*-;,: ffi

I

« Mandibles

(Plate 1, figure 3; Plate 2, figures 1,l, 2)

The right mandible is prepreserved served nearly in its natural position, though forced upward under the maxilla (Plate 1, figure 3). Though it is vivirtuallyrtually complete, none of the teeth are expexposed.osed. The left mandible has been folded over under the skull, its base lying along the innerianer margin of the right pterygoid, cutting across the prevomerine lateral ridge, and jutting out beneath the maxillary-premaxillarymaxillary-prernaxillary suture. The anterior 45-5045-50 mmrnm of the left mandible is missing. Both mandiblesrnandibles differ only in minor respects from those of E. megacephalus, as described and illustrated by Sawin (1941).(194 l). There areaTe the obvious differences ofof. smaller size andandfiner finer surface ornamentation. AlsAlso,o, the mandibular foramen occurs relatively morernore posteriorly on the present specimen,specimerS posterior to ththee angular posterior extension of the precoronoid. In E. megacephalus the foramenforarnen occurs beneath the precoronoid extension, well anterior to the acutely pointed posterior extremity; at least such is the case withwitb- the specimen dedescribed sc ribed by Sawin (1941). Size a.ndaad position of the mental foramen, dental foramen and inframecklian fossa agree with those of E. megacephalus.lg"g.""ph"1*.. CoronoidCoroncid suturesguttrres are difficultdifficultto to recognizerecognizebecause because of thethebattery battery of fine coronoid denticles,deaticles, a feature notedooted by Romer in the holotype of _E.E. avinoffi and by LangstonI-angston (1953) in E. grandis.

Pectoral Girdle and Limb Elements

(Plate 1, figure 4; PlatePIate 2, figures 3, 4, 6, 7)7l

The right (Plate 2, figures 3, 4) is complete, while the distal half of the left humerus is also preserved. WhenWtren compared withwit.l the humerihurneri of E. megacephalus, a numbernurnber of differences are ob­ob- servable. PerhapsthePerhapJth. most*"st conspicuous is the less oblique trend of the supinatorsupioator process in the Conemaugh specimen.specirnea. In posterior view, this process extends only slightly above the ectepicondylar process; the ectepicondylar notch is thus confined to a rather narrow groove best seen in a view of the outer side of the humeruhumerus.s. The latissimi dorsi process and the deltoid crest are somewhat less developed than typical of EE. megacephalusryg""ephalgs though the former has been damagedciamaged and is still partially"f cocovered,vered, making observation difficult. ThThee articulatory sur­sur- face for the radius is quite pronounced in the ConemaughConernaugh speCimen,specirnen, distinctly bulbobulbousus in side view (Plate 2,Z, figure 3). The entepicondylar processprocess does not extend downward much beyond the ectepicondylar pro-pro­ cecess ss and the radial articulatory surface, makingrnaking the lowelower r part of the humerhurnerusus decidely morernore transverse than in E. megacephalus. As a conseconsequencequence of the development of the radial articulatory surface, the lower outline of the humerus, in antp.rioranterior and posterior views, is

269 distinctlydistinctly convexconvex ratherrather thanthan concaveconcave asas inin E.E. rnegacephalus.megacephalus. OnOn thethe posterior side, thethe articulatoryarticulatory surface f""for tft"if""l.the ulna is confined toto aa posterior "ottfined small patchpatch onon thethe peripheryperiphery ofof thethe distaldistal edgeedge ofof thethe hurnerus.humerus. Neither Cope's (1888) illustrations of thethe humerushumerus of g.E. mega­ Neither Copets (1888) illustrations of "{ T"g"-- @_norcephaluscepha1us nor MoodietsMoodieMoodie's' s (1910)(1910) drawingdrawing ofof thethe hurnerushumerus ofof E.E. -;'i1~i~il~iwillistoni [Ftpermit a.taileddetailed comparison.comparison. AA fragmentary,fragmentary, poorlypoorly ptut"fr"dpreserved hffi;;humerus ofof E.E. grandisgrandis isis notednoted byby LangstonLangston (1953)(1953) butbut isis tootoo poorlypoorly preservedpreserved forfor cornpariJon.comparison. MinertsMiner's (1925)(1925) study ofof thethe pectoralpectoral girdlegirdle ofof E.E. megacephalusmegacepha1us has beenbeen reliedrelied upon heavilyheavily both here andand inin thethe follow-follo-;­ inging description. The rightright scapulocoracoid (Plate 2, figure 6, 7) is very well preserved inin thethe 'W'estWest VirginiaVirginia specirnen,specimen, sirnilarsimilar innearlyininnear1y nearly all respects toto thatthat of E. megacephaLusmegacephalusmegacepha1us and E. grandis, except for thethe much srnalL-small­ er size. It is also relatively shorter thanthan illustrated scapulocoracoidsscapu10coracoids of thosethose twotwo species. No tracetrace of the cleithrum has been recognized in the collection from Sutton.gutton. The rightclavicleright clavicle is present(Platepresentpre sent (Plate 1, figure 4), slightly crushed ventrally, so that thethe ventrolateralventrolateral aagulationanguangulationl ation is 0 nearly 90o.90 •. The rather coarse sculpture of the exterior surface is well shown. A considerablconsiderabl8considerab1~ quantity of isolatedisolated rib and vertebral elements are in the collectioncollection fromfrom Sutton. Neither the axis nor atlas has been recognized,recognized, and mostrnost of the vevertebraertebrae appearappeal to represent the cervical and perhaps part of the dorsal section of the spinal column.colurnn. There are about seven relatively complete neural archearches s (Plate 2, figurefigtre 5) and five readilyreadity identifiable intercentra.intercentra. NumerousNurnerous small fragmentsfragrnents mayrnay reEreBrepresent.,resent pieces of pleurocentra, though somesorne of these fragments are definitely pieces of nneurneuraleuralal arch. None of the vertevertebralbral elementselernents were !t'ere found articulatedarticulated..

TAXONOMIC ASSIGNMENT

The Sutton specimenspecirnen isis so well preserved and relativelyrelatively com­corn- plete thatthat all but twotwo previprevi9us1yprevigusly9usly describeddescribed rhachitomerhachitome genera are im­im- mediatelymediately removedremoved fromfrom consconsideration.ideration. InIn soso farfar asas discerniblediscernible,, thisthis ConemaughConemaugh specimenspecirnen agagreesagleesree s in in eeveryeve11.very particular withwith thethe wellwell knownknown PermianPermian genusgenus EryopEryopss Cope.Cope. AsAs discusseddiscussed above,above' therethere is is somesorne

- PLATEPLATE 22 - - Eryopsftyop" d. E.E-. avinoffiavinoffi CRomer)(Romer)(Romer) fromfrom thethe MorgantownMorgantcwn SandstoneSandstone Member,-nearMember,"f. near Sutton,Sutton, BraxtonBraxton County,County, WestWest Virginia.Virginia. ClevelandCleveland MuseumMuseurn ofof NaturalNatural History,History, no.no. 11025.11025. 11,1,, 2.Z. MesialMesial andand laterallateral viewview ofof leftleft mandible,rndndible, XX 0.5.0.5. 3,3, 4.4. OuterOuter andand antelioranterioranterior viewsviews ofof thethe rightright humerus,hurnerus' XX 0.8.0.8. 5.5. Posterior Posterior viewview ofof cervicalcervical neuralneural arch,arch, XX 1.1. 1.1. 6,6, 7.7, InnerInner andand outerouter laterallateralviews views ofof rightright scapulocoracoid,scapulocoracoid, Xx 0.7.0.7.

270270 Southeastern Geology,Geology, Vol. XIII,XIII, No.No.4 4-Murphy4-- Murphy PlarePlate 2

1 B#g /,

6 7 question about the nature of the bones of the skull, specifically the prpre-e­e ­ sence of an interfrontal element. Uncertainty on this point is especial­especial- ly critical for, while all known Eryops specimens possess an inter­inter- frfrontal,ontal, RomerRorner (1952) has erected the genus Glaukerpeton, distinguish­distinguish- ed from Eryops primarily by its smallersrnaller size, finer ornamentationornarnentation and the absence of an interfrontal. The holotype of Glaukerpeton, 0..G. avavinoffi,ininoffi,offi, is a fragmentary skull from a stratigraphic position some­sorne- wherewvrhereh ere in the Pittsburgh LimestoneLirnestone Member, at the very toptopof of the Cone­Cone- maugh Group, found within the cityciiy of Pittsburgh. A variety of small,srna1l, isolated bonee.bones fromfrom the Round Knob Shale MembMemberMernberater at Pitcairn, Pennsyl­Pennsyl- vanvania,ia, originally refreferrede rred to Eryops by Case (1908) were transferred to GlaukerpetonGiaukerpeton by Romer.Rorner. This latter material comes from strata slighslightlytly lower sstratigraphically tratigraphically than the Sutton occurrence. Romer surmssurrnisedurmisedised fromfrorn ththe e dearth of unquestionable Eryops materialrnaterial in the ConemaughConernaugh Group that the genus does notnotoccur occur ininrocks rocks as old asasCone- Cone­ maugh.rnaugh. This appears to have been his major reason in referring muchrnuch ofof the ConemConemaughConernaughaugh eryopsid materialrnaterial to his neneww genus. It is regrettable that only the holotypholotypee of Gla1JkerpetonGla1(kerpetonGlaUkerpeton shows the mostrnost diagnostidiagnosticc ffeatureeature of the genus-- the lack of aan n interfrontal-- and even this has been seriously questioned. VVaughanaughan (1958)(i958) has clearly shown that the small size and fine reticulation cited by Romer in his diagnosis of Glaukerpeton are of doubtful value at the generic level. Vaughan also restores the GlaukGlaukerpetone rpeton skull fragments fragrnents in such a mannerrnanner that an interfrontal seemsseerns to be present. He places GlaukerpGlaukerpetone ton in synonymysynonyrny with Eryops and concludes that Eryops does indeed occur in the Conemaugh GrouGroup Group.p.. In light ofoi ththG-GEertaintye unce rrtaintytainty regardregardinging the prepresence sence or absence of an interfrontal eelementlement in the SuSuttontton specimen, there mustrnust rremainemain some quesquestioquestiontionn aaboutbout the genericg e neric identity of the specimen,specimen, just as there mumustst rremainrernainemain a modicumm.odicurn of uncertainty about the synonymysynonyrny of GlaukerpetonGlaukerpeion with ErEryops.yops. NonNonethelNonetheless,e the l ess,es s , thethe close degree in which the Sutton rharhachi­rhachi-chi­ tome matches the manyrrrany known features of Eryops preponderates so greatly over the possibility that it differs from Eryops in a single character (absence of thethe interfrontal)interf rontal) that the mostrnost suitable assignment of the Sutton amphibian iiss to Eryops. In view of ththee factfact thatbat t thithiss one distingudistinguishingishing character is not cercertainly tainly knoknownwn to exist in even the genoholotype of GlaukGlaukerpeton,erpeton, this assignment seemsseems to bbee the bbestesestt solution for thethe prepresent. sent .. Romer may well be correct in his beliefb e lief that there is a Cone­Cone- mamaughmarlghugh rhachitome idideniicalentical with rvith Eryops in evereveryy rree sspectpect exceexceptpt ththee prprepresenceesencsence ofoi anaa::n interfrontal;interfronta l; it i iss even*u'En-!6-"sible possible thatthat both genera occur iniir-rn the ConConemaugh,emaugh. Such parallelism isi-s pperhapsererhapshap s not unkr'.Own,unkp-unknown,.own, butbui it wouwrvouldoull d be vveryery difficultdiffiCifiicultc ult ttoo proveprove on ththee basbasisi s of only two orol three speci­speci- mensmeirr.ens.ns. ThThee very rarity ranty of relativerelativelyly wewellll prpreservedeserveeservedd eryopsideryop sid remains in the ConemaConemauConernaughughgh Group of the AppalacAppalachianhiaann BasinBasin makesmakes it inadinadvisablevisableisable to erecerectt new ttaxaaxa on thethe basis of uniqueunique specimens s p ecim e n s w whichhich differd i ffer fromf rom previously describdescribeded matematermaterialrialial in only a singssingleinglel e chacharacter,racter, especespeciallyially wwhenhen the presence or aabsencebsenceb sence of eveevenn that one character is in doubdoubt.dcubt.t.

271 At the species level, assignment of the Sutton specimen is less difficult. The small size of the individual and perhaps the relative pro­ portions of the scapulocoracoid and the humerus suggest that the speci­ men is immature; but it is remarkable that of the half dozen or so Eryops specimens thus far recovered from the upper and Dunkard strata of the Appalachian Basin all are considerably b elow the average size of Eryops megacephalus. It is difficult to believe that all of these are immature specimens, and it is much more probable that a distinct species is represented, one characterized in part by a considerably smaller size and finer ornamentation. In view of the stratigraphic and geographic proximity of the Sutton occurrence to the holotype of Eryops avinoffi, the two may well be conspecific. The poorly preserved nature of the holotype of E. avinoffi precludes the removal of all doubt on this point, as it has also injected a certain element of doubt at the generic level. Nevertheless, assignment or comparison to E. avinoffi seems the best course in this preliminary study of the Sutton amphibian. Eryops avinoffi, as interpreted here, differs from other des­ cribed species of Eryops by virtue of its small size and finer surface ornamentation. The sole exception is E. grandis, which Langston (1953) gives reason to believe is distinctly smaller than typical E. megacepha­ Ius. Other minor charact eristics noted by Langston include a "dense s hagreen" of coronoid denticles, quite like that of E. avinoffi; a man­ dible relatively more slender than that of~. megacephalus; a thinner, less robust scapulocorac oid. .The Sutton s pecimen herein compared to E. avinoffi differs from~. grandis in being even smaller-- only one­ half to two-thirds as large-- with a mandible more like that ofE. mega­ cephalus in proportions, and a scapulocoracoid that is stouter-;- wider, than that of either of the other two species. As Langston notes, evaluations of such criteria will not be pos­ sible until a thorough re study of all known Eryop s mate rial is unde r ­ taken. But for the present there is no reason to believe that E. grandis and E. avinoffi are conspecific.

C ONC LUSIONS

Preliminary study of a rhachitomous amphibian from the Cone­ maugh Group near Sutton, Braxton County, West Virginia, suggests that the specimen represents Eryops avinoffi (Romer). The specimen is the best preserved example of Eryops known from the Appalachian Basin, but a critical taxonor.lic character-- the presence or absence of an interfrontal-- remains uncertain. It is believed, following Vaughan (1958), that the genus Glauker­ peton Romer is a junior synonym of Eryops. The Sutton specimen thus confirms the presence of Eryops in rocks as old as mid-Conemaugh in age.

272 REFERENCES CITED

Case, E. C., 1908, Description of vertebrate from the vicinity of Pittsburgh, Pennsylvania: Ann. Carnegie Mus., v. 4, p. 234-241, 8 text-figs., 1 pI. Cope, E. D., 1888, On the and extremities of Eryops: Am. Phil. Soc. Trans., v. 16, p. 362-367, 1 pI. Hennen, R. V., 1917, Braxton and Clay Counties: W. Va. Geol. Sur­ vey, Co. Rept., 883 p., 2 maps, 16 figs., 29 pIs. Langston, W., Jr., 1953, from New Mexico: Univ. California Publ. in Geol. Sci., v . 29, p. 349-416, 24 text-figs. Miner, R. W., 1925, The Pectoral limb of Eryops and other primitive tetrapods: Am. Mus. Nat. Hist., Bull., v. 51, p. 145-312, 104 figs. Moodie, R. L., 1910, The Temnospondy1us amphibia and a new species of Eryops from the Permian of Oklahoma: Kansas Univ. Sci. Bull., v. 5, no. 3, p. 235-253, pIs. 49-54 . Romer, A. S., 1952, Late Pennsylvanian and early Permian vertebrates of the Pitt~ burgh- We st Virginia region: Ann. Carnegie Mu s., v. 33, p. 47-113, 13 text-figs., 2 pIs. Sawin, H. J., 1941, The Cranial anatomy of Eryops megacephalus: Bull. Mus. Compo Zool., v. 88, p. 407-463, 6 text-figs. , 12 pIs. Vaughan, P. P., 1958, On the geologic range of the labyrinthodont am­ phibian Eryops: J. Paleontology, v. 32, p. 918-922.

273