Wonambi naracoortensis - Digital Library - The University of Adelaide

43 Pages·2005·2.03 MB·English

PUBLISHED VERSION
Scanlon, John D. Cranial morphology of the Plio-Pleistocene giant madtsoiid snake Wonambi
naracoortensis. Acta Palaeontologica Polonica, 2005; 50(1):139-180.
© 1997-2011 Institute of Paleobiology PAS

COPYRIGHT PERMISSIONS

http://www.app.pan.pl/copyright-policy.html
Acta Palaeontologica Polonica is published by Institute of Paleobiology, Polish Academy of
Sciences. APP authors are permitted to deposit the published pdf version of the article in their

Institutional Repository.

Permission to make digital or hard copies of all or part of APP articles for personal or classroom
use is granted without fee provided that copies are not made or distributed for profit or
commercial advantage and that copies bear the notice Institute of Paleobiology, Polish Academy

of Sciences, and full citation on the first page.

22nd March 2011

http://digital.library.adelaide.edu.au/dspace/handle/2440/16371
Cranial morphology of the Plio−Pleistocene giant
madtsoiid snake Wonambi naracoortensis
JOHN D. SCANLON
Scanlon,J.D.2005.CranialmorphologyofthePlio−PleistocenegiantmadtsoiidsnakeWonambinaracoortensis.Acta
Palaeontologica Polonica50 (1): 139–180.
Newmaterialandadditionalmorphologicaldetailsofarareandphylogeneticallysignificantlargefossilsnake,Wonambi
naracoortensisSmith,1976,aredescribedfromPleistoceneandPliocenecavedepositsinsouthernSouthAustralia.The
newdatarefutesomepreviousinterpretationsofthemorphologyofthisspecies,andhaveimplicationsforthephylogen−
eticpositionofWonambirelativetoextantsnakesandotherfossils,includingotherMadtsoiidae.Thenatureofcontacts
amongpalatal,braincase,snout,andmandibularelementsimplysimilarfunctionalattributestothoseofextantanilioid
snakes:maxillaefrommultipleindividualsshowcorrugatedcontactsurfacesfortheprefrontal,implyingatightsuture;
structuresontheanteriorandmedialsurfacesofthepalatinechoanalprocessareinterpretedasformingextensivecontacts
withthevomerandparasphenoid;andthedistinctlyboundedfacetsonthebasipterygoidprocessesandpterygoidimply
little or no capacity for anteroposterior sliding of the palatopterygoid arch, hence absence of the macrostomatan
“pterygoidwalk”.Onthefrontal,interolfactorypillarswereeitherabsentorveryslender,andadeep,sculpturedcontact
surfaceforthenasalimpliesaprokineticjointwasalsoabsent.Marginsofthefrontalandparietalindicatebroadentryof
thesphenoidintotheophthalmicfenestra,asinDinilysia.Similarityofelementsandfeaturesofthebraincase(trigeminal
foramen,earregion,andbasipterygoidprocesses)withbothlizardsandextantsnakesshowthatdifferencesbetween
snakesandothersquamateshavesometimesbeenoverstated.ThecaseformacrostomatanaffinitiesofWonambiisnot
supported by new evidence.
Key words: Ophidia,Wonambi,anatomy,braincase,palate,mandible,Plio−Pleistocene,Australia.
JohnD.Scanlon[[email protected]],DepartmentofEnvironmentalBiology,TheUniversityofAdelaide;
and Department of Palaeontology, South Australian Museum, North Terrace, Adelaide, SA 5000. Present address:
Riversleigh Fossil Centre, Outback at Isa, PO Box 1094, Mount Isa, QLD 4825, Australia.
Introduction
Quaternary deposits in Australia. Descriptions of some of
this material have appeared (Barrie 1990; Scanlon 1997;
Snakefossilswithrelativelycompleteskullsarerare,andin ScanlonandLee2000,2002;Scanlon2003)buttheevidence
mostcases occur inaquatic sedimentsthatimposecrushing for a relatively basal phylogenetic position of Madtsoiidae
and other damage, making interpretation of their morphol− among snakes, and unusually good preservation of the
ogydifficult.Thus,despitethespectacularpreservationofa Australian material, justify more detailed documentation.
number of Cretaceous and early Tertiary snakes [such as The description of Wonambi naracoortensis by Smith
Pachyrhachis problematicus Haas, 1979; Eupodophis des− (1976)wasthefirstreportofanextinctsnakespeciesinAus−
couensi(RageandEscuillié,2000);Haasiophisterrasanctus tralia. When first described this species was represented by
Tchernov et al., 2000; Dinilysia patagonica Woodward, eightlargevertebraeandajawfragmentfromtheMainFos−
1901; Archaeophis proavus Massalongo, 1859; A. turkme− silChamber,VictoriaFossilCave,Naracoorte,south−eastern
nicus Tatarinov, 1963; and Rottophis atavus (von Meyer, SouthAustralia;SmithreferredWonambitoBoidae(asdiag−
1860)],theidentityandmorphologyofmanyoftheircranial nosed by Hoffstetter and Gasc 1969) and considered it to
bones are either undocumented or subject to conflicting in− have probable affinities to Madtsoia and Gigantophis, al−
terpretations (Janensch 1906; Auffenberg 1959; Estes et al. though she did not formally include it in Madtsoiinae
1970;Haas1980a, b;Tatarinov1988;SzyndlarandBöhme (erected for the latter two genera by Hoffstetter 1961; spelt
1996; Lee and Caldwell 1998; Zaher 1998; Rage and incorrectlybySmith1976as“Madstoiinae”).Insubsequent
Escuillié2002;LeeandScanlon2002;CaldwellandAlbino popularworksSmith(1983,1985)referredtoWonambiasa
2002; Rieppel et al. 2003). Madtsoiidae, another extinct “giantpython”,butthiswasapparentlyonlymeantasanin−
snake group already widespread in Gondwana during the formal characterisation of this large and presumably con−
Cretaceous, and hence of comparable phylogenetic interest stricting snake, not a change of systematic assignment to a
tothosementionedabove,hasnotyetproducedsuchfullyar− particular modern boid lineage.
ticulatedskeletonsbutisrepresentedbyassociatedandfully Merrilees(1968)andArcher(1972)hadpreviouslymen−
three−dimensionalcranialremainsfromseveralTertiaryand tionedlargesnakevertebraefromPleistocenecavedeposits
Acta Palaeontol. Pol.50 (1): 139–180, 2005 http://app.pan.pl/acta50/app50−139.pdf
140 ACTA PALAEONTOLOGICA POLONICA 50 (1), 2005
inthesouth−westofWesternAustralia(MammothCaveand Barrie’smaterialwasonlypartlypreparedandadditionalele−
KoalaCave,Yanchep);initiallyunidentifiedorthoughttobe mentshavesincebeenidentified.ScanlonandLee(2000)re−
frompythons,thismaterialwassubsequentlyreferredtoW. portedsomeofthisnewmaterialofW.naracoortensisaswell
naracoortensis(Merrilees1979;Molnar1984;Smith1985) as a smaller, Oligo−Miocene species from Riversleigh de−
but has never been fully described. scribedasW.barrieiScanloninScanlonandLee,2000,and
DuringasalvageexcavationbyvolunteersatNaracoortein providedreviseddiagnosesforthegenusandbothspecies.
the early 1980s, many disarticulated elements representing Rieppeletal.(2002)havegivensomeadditionaldescrip−
two different−sized individuals were found mingled together tionoftheHenschke’sQuarrydisplayspecimen,whichthey
inthelowerfissuresofHenschke’sQuarryFossilCave(Barrie interpret as a macrostomatan snake of uncertain affinities.
1990).D.JohnBarrieassembledbonesattributedtothelarger Theydidnotrefertoanyoftheadditionalspecimensfromthe
individualasapartlyrearticulateddisplayspecimenandpre− samelocalityreportedbyBarrie(1990),oranymaterialidenti−
sented it to the South Australian Museum in 1989 (Barrie fiedafter1987,orfromotherdepositsincludingthetypelocal−
1990:fig.2;Vickers−Rich and Rich 1993:fig.292;coverof ity,butstateincorrectlythatofthetwospeciesofWonambi,
Nature,27January2000).Barrie(1990)attemptedtointerpret “P30178...is...theonlyspecimenwithpreservedpartsofthe
theaffinitiesofWonambiwithinthecontextofabroadlyde− skull”(Rieppeletal.2002:824).Duetolimitedrepresentation
finedBoidae(sensuUnderwood1976),andsupportedSmith’s ofmaterialaswellasnumerouserrorsandquestionableinter−
suggestion that it was close to Madtsoia, and thus a Gond− pretations in their work (some mentioned by Scanlon 2003:
wananrelictwithalonghistoryindependentofextantlineages 975–976,andothersbelow),afulldescriptionofthisextinct
includingpythonines.FromBarrie’sworkuntilveryrecently, snakeisnolessnecessarythanbefore.
W. naracoortensis was the most completely known fossil Thereconstructionandinterpretationofthecranialmor−
snakeinthiscontinent. phologyofWonambipresentedbyScanlonandLee(2000)is
Further vertebral remains referable to this species have slightlymodifiedbynewobservationsandmoredetailedde−
sincebeenidentifiedfromadditionalPleistocenecavedepos− scriptioninthiswork.Postcranialmorphologyandcompari−
its at Naracoorte (Grant Hall in Victoria Fossil Cave, Wet sons with W. barriei will be discussed elsewhere. Recently
Cave,andHaystallCave;ReedandBourne2000),Wellington discovered partial skeletons of Yurlunggur from the Oligo−
Caves,NewSouthWales(Scanlon1995),andTightEntrance cene and Miocene of Riversleigh, Qld, currently being pre−
Cave, southern Western Australia (Gavin Prideaux, personal pared and studied, provide additional tests of many of the
communication 2002). Skeletal material referred to W. cf. interpretations in this work.
naracoortensis is also known from the Curramulka Local
Institutional abbreviations.—AMS, Australian Museum,
Fauna, obtained from a cave on the Yorke Peninsula, South
Sydney (R, Herpetology); AR, Michael Archer reference
Australia,andisconsideredtobeearlyPlioceneinage(orper−
collection, University of New South Wales (also used
haps latest Miocene; Pledge 1992). Based on a figure pub−
as temporary registration for some Riversleigh fossils);
lishedbyPledge(1992:fig.3a),Iearliersuggested(Scanlon
FU,FlindersUniversity,DepartmentofZoology, Adelaide;
1995) that the Curramulka fauna might include a species of
QM, Queensland Museum, Brisbane (F, Palaeontology,
YurlunggurScanlon,1992;however,subsequentexamination
J,Herpetology);SAM,SouthAustralianMuseum,Adelaide
ofthematerialindicatesthatWonambiistheonlylargesnake
(P, Palaeontology, R, Herpetology); UNSW, University of
present. On the other hand, a vertebra from the Pleistocene
New South Wales, Sydney.
Wyandotte Local Fauna of northern Queensland (Qld), ini−
tially referred to Wonambi by G. McNamara (1990), is ex− Other abbreviations.—CL, Corra−Lynn Cave, Curramulka,
cludedfromthatgenusandreferredtoYurlunggursp.(Scan− YorkePeninsula;CQ,CurramulkaQuarry,YorkePeninsula;
lon 1995, 1996; Mackness and Scanlon 1999). A vertebra HQ,lowerfissuresofHenschke’sQuarryFossilCave,Nara−
fromthePlioceneKanunkaLocalFauna,SouthAustralia,re− coorte; VF, Main Fossil Chamber, Victoria Fossil Cave,
ferredtoW.naracoortensisbyPledge(1992),alsobelongsto Naracoorte.
aspeciesofYurlunggur(personalobservations).
Although some cranial bones of other madtsoiid species
Materials and methods
areknown,almostalloftheseareelementsalsoknownfrom
W.naracoortensis(Scanlon1996,1997;Rage1998;Scanlon
and Lee 2000;personalobservations). The late Oligocene to Thedescriptionbelowisbasedonallknowncranialmaterial
mid−Miocene deposits at Riversleigh contain abundant and ofWonambinaracoortensis,whichcomesfromfourlocalities
taxonomicallydiverseremainsofthisgroup,butuntilrecently (seeprevioussection),allinSouthAustralia:VF(Pleistocene;
onlyasinglebraincasefragmentwasrecognized(asphenoid Smith1976;ReedandBourne2000),HQ(Pleistocene;Barrie
of Yurlunggur sp. identified in 1990; Scanlon 1993b, 2003). 1990,1997),CL(?earlyPliocene;Pledge1992),andCQ(fis−
For comparison among madtsoiid taxa and resolution of sure RF 95, probably Pleistocene based on associated mam−
phylogeneticrelationshipswithothersquamatelineages,more mals,JamesMcNamara,personalcommunication2002).All
detaileddescriptionofthecranialmorphologyofWonambiis ofthismaterialisregisteredintheSAMPorFUcollections.
required than that given by Barrie (1990), especially as ReferenceisalsomadetoQMFspecimensofothermadtsoiid
SCANLON—CRANIAL MORPHOLOGY OF GIANT SNAKE WONAMBI 141
taxa studied at UNSW as part of a PhD project (Scanlon sistent with crushing or impacts. In attempting to correct
1996),andotherRiversleighmaterialidentifiedsubsequently Rieppel et al.’s (2002) mis−statement of SAM P30178’s
andcurrentlybeingstudied.Recentcomparativematerialhas originas“Victoria, Australia”, Imade asimilarbutsmaller
been consulted in the AMS R, QM R, SAM R, and AR lapse(Scanlon2003:975)inattributingittoVFratherthan
collections. HQ (the two sites are about 15 km apart).
VFmaterialrepresentsatleasttwoskeletons,currentlyvery Jaw remains known from CL (previously unreported) in−
incomplete (as careful excavation of the deposit initiated by clude a maxilla and a pterygoid described below, which are
R.T.Wellsisstillinprogress)butshowingexcellentpreserva− fragmentarybutshowfinepreservationofbonesurfacesand
tion of teeth and bone surfaces. Smith (1976) reported a jaw teeth.Anotherpartialmaxilla,moreseverelyworn,istheonly
fragment(SAMP16170c),andBarrie(1990)anear−complete cranialfragmentofWonambifromCQ.Thefurtherprepara−
parietal(SAMP27777)fromthisdeposit,butdidnotpublish tionoftheHQspecimens,andstudyoflesscompletematerial
detailedillustrationsordescriptionsofeither.ScanlonandLee notpreviouslydescribedorillustrated,thusallowdescription
(2002) described features of the dentition based on a partial ofsurfacedetailshithertounreported,andcomparisonofone
maxilla of a very large individual (FU1762). Another partial cranialelement(maxilla)frommultipleindividuals.
maxillafromasmallerindividualisalsodescribedbelow. Forpurposesofcomparisonsdiscussedhere,Iassumethe
Material from HQ (registered under the single number relationshipsfoundbyLeeandCaldwell(2000)andLeeand
SAM P30178) comprises two partial skeletons, referred to Scanlon(2002).Thus,phylogeneticallyrelevantcomparisons
below with suffixes A and B for the larger and smaller re− aremadewithothermadtsoiids,otherextinctsnaketaxasuch
spectively(HJD2:84WiandHJD1:83WiofBarrie1990).Al− asDinilysia,PachyrhachisandHaasiophis,“anilioids”(basal
though the fissure containing the deposit no longer exists, alethinophidiansAnilius,CylindrophisandAnomochilus,but
processing and sorting of the excavated material has been not the more derived uropeltines), macrostomatans (Xeno−
continued by John Barrie, and he has made additional ele− peltis,pythoninesandothers),and“lizard”outgroups(partic−
mentsofbothskeletonsavailableforstudyandlodgementin ularly, extant varanoids and extinct mosasauroids) based on
the SAM (Scanlon and Lee 2000); more may remain to be examinationofspecimensordescriptionsintheliterature.For
identified. Most of the cranial elements are attributed to allcomparisonsinvolvingextinctspecies,thecriterionofpri−
SAM P30178A; Barrie reported a complete right maxilla mary homology is topographic similarity of ossified struc−
fromthesmallerSAMP30178B,andsomeotherfragments tures,anddoesnotdependonhypotheticalsoft−tissuerelation−
may be attributed to the same individual. At the time of re− shipsorontogenetictrajectories(thiscorrespondstotheprin−
covery the surfaces of most of the cranial elements were cipleespoused,thoughnotalwayspracticed,byRieppeland
hardened with a PVA wash (Barrie 1990), which consoli− Zaher2000).Perceptually“similar”states,i.e.thoseamenable
datedalayerof“terrarossa”cave−earth,inplacesoverlyinga to identical partial decriptions in different species, are re−
calcite patina, so that much surface detail remained ob− garded provisionally as historically identical effects of com−
scured.Thiswastheconditionofthematerialwhenfirstde− monancestry(Hennig1966;Kluge2004),andtestingofsuch
scribed(Barrie1990),butIcleanedpartsofthebraincasein homologies by character congruence takes place in (and
1991, revealing some additional features (Scanlon 1993b, subsequentto)simultaneousanalysesofmanycharactersand
1996; Scanlon and Lee 2000; see also Rieppel et al. 2002: taxa(e.g.,LeeandScanlon2002).
814).Furtherpreparationofallcranialelementswascarried Terminology for cranial structures follows Oelrich
out either in 1999 (on specimens that were held by D.J. (1956) unless stated otherwise; other valuable sources for
Barrie since collection, then studied at the University of terminology and comparisons include Estes et al. (1970),
Queensland before lodgement in the SAM) or July 2001 Rieppel (1979, 1985), Bellairs and Kamal (1981), Kluge
(subsequent to examination of some SAM material by O. (1993), Cundall and Rossman (1993), Norell and Gao
RieppelandA.G.Kluge):Iremovedmostofthesuperficial (1997), andRieppelandZaher(2000). Iuse“sphenoid” for
calcite with the aid of dilute acetic acid, and also separated the element representing indistinguishably fused basi−
someofthebraincaseelementsandcleanedtheircontactsur− sphenoidandparasphenoid(“parabasisphenoid”ofmanyau−
faces, but was unable to safely separate the sphenoid, thors);“exoccipital”mayincludeopisthoticwherereference
basioccipital and left opisthotic−exoccipital, and some fo− is made to the fused (dorsal) parts of these bones.
ramina and canals remain blocked. The right basipterygoid
processandposteriorpartofthesphenoidwereunfortunately
Systematic palaeontology
damaged during handling (present condition shown in Fig.
8). Parts of the calcitic encrustation also remain on the left
maxilla and right dentary; some of these resistant areas ap− Squamata Oppel, 1811
peartoforminternalmouldsoftubesseveralmillimetresin
Ophidia Brongniart, 1800
diameter, consistent with Barrie’s (1990) inference of ter−
Madtsoiidae Hoffstetter, 1961
mitedamageto(andpartialdestructionof)somepostcranial
elements. None of the other identified cranial elements Diagnosis.—Snakes of small to very large size; hypapo−
showssimilardamage,butsomehavecracksorbreakscon− physes present only in anterior trunk; middle and posterior
http://app.pan.pl/acta50/app50−139.pdf
142 ACTA PALAEONTOLOGICA POLONICA 50 (1), 2005
trunk vertebrae with moderately or well developed haemal knownonlyfromEarlyMioceneandpossiblyLateOligocene
keel(exceptafewnearthecloacalregion),oftenwithshort depositsatRiversleigh,Qld.
laterallypairedprojectionsontheposteriorpartofthekeel. Postcranialvariationwillnotbedescribedfullyhere,but
Alltrunkandcaudalvertebraewithaparazygantralforamen Iaddsomeremarksontwopoints:(A)therelativeindistinct−
(or sometimes several foramina) in a more or less distinct nessofparazygantralforaminainW.naracoortensis,and(B)
fossa lateralto each zygantral facet; prezygapophyseal pro− thenatureandpreservationofhaemalarchesoncaudalverte−
cessesabsent;paracotylarforaminapresent;diapophysesrel− brae in madtsoiids.
atively wide, exceeding width across prezygapophyses at (A)WhenWonambiwasrepresentedbyjustahandfulof
leastinposteriortrunkvertebrae(Scanloninpress,modified vertebrae, it was referred to Madtsoiidae provisionally or
from Rage 1998). doubtfully,becauseitsparazygantralforamina(notinallver−
Remarks.—Polarityofsomecharacterslistedaboveremains tebrae, but in some including the holotype; Smith 1976: fig.
uncertain. Additional apomorphies have been identified in 2E)weresmallerorlessdistinctthaninthefewgiantspecies
phylogeneticanalysisbutmaynotapplytoallincludedtaxa ofMadtsoiaandGigantophisthenknown(Smith1976;Rage
(LeeandScanlon2002:table2;seeremarksondiagnosisin 1984).AccordingtoRieppeletal.(2002:812),“theassocia−
Scanlon in press). Monophyly of Madtsoiidae is accepted tionofWonambiwithmadtsoiidswasconsideredtentativeby
provisionally while there is insufficient contradictory evi− Rage(1984;seealsoSmith1976),anditstillis.”Thisambigu−
dence,butremainsweaklysupportedaslongasanyofthein− ous phrasing either attributes to J.C. Rage a view which he
cludedtaxaarepoorlyknown(Rage1998;RageandWerner does not actually hold (e.g., Rage 1998), or asserts that the
1999;Scanlon2003).Duetotheinferredphylogeneticposi− groundsforearlierdoubtarestillvalid.Infactwenowknow
tion of this group outside the clade containing all extant rathermoreaboutvariationofparazygantralforaminainAus−
snakes(Scanlon1996;ScanlonandLee2000;LeeandScan− tralianmembersofthisgroup,andthreeindependentcompo−
lon2002),itisreferredtoOphidia(definedasinLee1998) nentsofvariationcanbeidentified.ManyW.naracoortensis
butnottoSerpentesLinnaeus,1758;thelatterisrestrictedto vertebrae have one or more of the following features: (1)
“crownclade”snakes,i.e.,theleastinclusivecladecontain− subdivsionofthelarge,singleforamenoneachsideusualin
ing all extant snake taxa (Scolecophidia+Alethinophidia, other madtsoiids; (2) shallowness of the depression (“fos−
collectively referred to as “modern snakes”). sette”,Hoffstetter1961)containingtheforamina;and(3)pres−
ence of numerous smaller additional foramina or pits in the
WonambiSmith, 1976 parazygantral area, beyond the depression. These features
(1–3) have a restricted distribution within Australian Madt−
Type species:Wonambi naracoortensisSmith, 1976.
soiidaeandarepotentiallyusefulastaxonomiccharacters,as
Referredspecies:WonambibarrieiScanloninScanlonandLee,2000.
they vary independently within and between some species
Diagnosis.—Neuralspinesofvertebraehigh,slopingpostero− (particularlythoseofWonambiandYurlunggur),whileothers
dorsally, with sharp−edged anterior lamina extending to near (includingspeciesofAlamitophis,Patagoniophis,andNano−
anterior edge of zygosphene; transverse processes extending wana) have bilaterally single, relatively large, usually
laterally beyond zygaphophyses in most trunk vertebrae, distinctly recessed foramina and no additional pits (Scanlon
diapophyses with concave dorsaledge in lateralview;zygo− 1992,1993a,1995,1996,1997,inpress).
sphenerelativelynarrow,withsteepfacets(20–30°fromverti−
(B) Extant snakes have the caudal haemal arches
cal); zygapophyses inclined 20° or more above horizontal;
(haemapophyses, corresponding to chevron bones in most
haemalkeelinmiddleandposteriortrunkregionnarrowand
other amniotes), if present, fused to the centra proximally,
weaklydefinedlaterally,butoftendistinctlybifidortrifidon
and forming laterally paired projections which are unfused
the posterior third of the centrum. Pterygoid tooth row near
and usually separated distally (Hoffstetter and Gasc 1969;
middle of bone, away from medial edge, and basipterygoid
Lee and Scanlon 2002, characters 204–206). Paired ventral
facet narrow and facing medially as much as dorsally; ecto−
projectionsofcaudalcentrainmadtsoiidswerethusinitially
pterygoid process of pterygoid triangular in palatal view.
identified as “haemapophyses”, despite having flat or con−
Maxilla and dentary relatively elongate and depressed;
cave facets rather than finished bone distally (e.g., Scanlon
maxilla with deep, anterolaterally directed trough on sub−
1993a; Rage 1998). However, one caudal vertebra of W.
orbitalsurface(ScanlonandLee2000).
naracoortensisisknown(partofSAMP30178,HQ)thatre−
Remarks.—All characters in the generic diagnosis are ob− tains a complete chevron (V−shaped element, open proxi−
served directly in both species except for the shape of the mallyandfuseddistally)partiallyfusedtothedistalfacetsof
ectopterygoidprocessinW.naracoortensis,whichisinferred the “haemapophyses”, indicating that the latter are actually
belowtobesimilartothatofW.barriei(andunliketheother pedicelsforthechevronlikethoseinvaranoidlizards(Scan−
madtsoiidswherethisstructureisknown).Diagnosesofboth lon and Lee 2000: fig. 2g). This lizard−like condition in
species are given in Scanlon and Lee (2000). All Wonambi Wonambi unambiguously supports a phylogenetic position
materialfromthePlioceneandPleistocene(southernWestern outsidethecladeofmodernsnakescharacterized bytypical
AustraliaandSouthAustralia,andcentralNewSouthWales) haemapophyses.Rieppeletal.(2002:813,824)raiseseveral
is referred to W. naracoortensis; the smaller W. barriei is objectionstothisinterpretation:(1)Barrie(1990:fig.11)il−
SCANLON—CRANIAL MORPHOLOGY OF GIANT SNAKE WONAMBI 143
Fig.1.Wonambinaracoortensis.A.LeftmaxillaofSAMP30178Aindorsalview.B.RightmaxillaofSAMP30178Binlateral(B ),ventral(B ),medial
1 2
(B), and dorsal (B ) views.
3 4
lustrated a considerably larger caudal vertebra of P30178 smaller; but it is also possible that the caudal with chevron
which lacks a chevron; (2)the vertebra withattached chev− comes from the smaller of the two skeletons comprising
ron was not among the material examined by Rieppel and SAMP30178.(2,3)Thevertebrainquestionwasnotinthe
Kluge;(3)thesuggestionwasmade(attributedtoM.Hutch− display cabinet (donated in 1989) that Rieppel et al. (2002)
inson, SAM Herpetology) that the specimen illustrated by examined; it remained with the collector and donor (D.J.
Scanlon and Lee was from a different locality and age; and Barrie) until positively identified as Wonambi in 1998. The
(4) “it seems that chevrons are a plesiomorphic trait at the finalobjection(4,repeatedbyRieppeletal.2003:538)rep−
levelofSerpentes,andhencecannotbeusedinanargument resentsabasicmisunderstanding oftheprinciplesofphylo−
supporting the basal position of Wonambi.” These are an− genetic inference: synapomorphy is evidence for mono−
sweredbrieflybelow.(1)ThelargecaudalfiguredbyBarrie phyly, while symplesiomorphy is not (Hennig 1966). All
isfrom immediately posterior to the cloacal region (itisal− modern snakes either have fixed, paired haemapophyses in
mostthesamesizeasthecloacalvertebrafiguredadjacentto place of chevrons (Macrostomata, with few exceptions),
it)andlacksanysignofanossifiedhaemalarchorpedicels; fixedmedian “caudal hypapophyses” (Uropeltinae), orlack
thepresenceofatleastonesuch“pygal”vertebraistypicalof any such structures (Scolecophidia, Anilius, Cylindrophis)
squamates and has also been demonstrated in W. barriei (Hoffstetterand Gasc 1969; Leeand Scanlon 2002, charac−
(Scanlon 1996; Scanlon and Lee 2000). More posterior ters204–206;SzyndlarandRage2003).Whethertheserep−
vertebrae of the same caudal series are, not surprisingly, resent separate derivations from the plesiomorphic state or
http://app.pan.pl/acta50/app50−139.pdf
144 ACTA PALAEONTOLOGICA POLONICA 50 (1), 2005
Fig.2.MaxillaeofthreeindividualsofWonambinara−
coortensisinmedioventralview.A.SAMP30178A,
left(A )andright(A )sides.B.P30178B.C.P16172.
1 2
(as it may be more parsimonious to assume) true chevrons Hoffstetter 1939 and Szyndlar 1984 applied the term “pre−
were lost only once in an ancestor of modern snakes, these frontalprocess” to the anteriormedialprocessin colubroids,
synapomorphies are indeed evidence for grouping modern here considered equivalent to the palatine process of the
snakes to the exclusion of Wonambi, and thus for the maxilla).Inoverallshape,includingrelativelygreatlengthof
relatively basal phylogenetic position of the latter (and of the bone and correspondingly high tooth number, and also
Eupodophis, Rage and Escuillié 2000). shapeofthepalatineprocess,themaxillaissimilartothoseof
some macrostomatan snakes, particularly Loxocemus and
somepythons(speciesofLiasissensulato);buttheprefrontal
Description of cranial elements of process is quite different from these extant taxa and more
closelyresemblesthatofanilioidsnakes(e.g.,Cylindrophis).
Wonambi naracoortensis
Comparison with maxillae of other madtsoiids (complete in
Nanowana godthelpi Scanlon, 1997, partially known in N.
Maxilla(Figs.1,2;formeasurementsseeTable1).—Partial schrenki Scanlon, 1997, Yurlunggur sp. or spp. [Scanlon
tocompletemaxillaefromVF,HQ,andCLprovideanindi− 1996], Madtsoia camposi Rage, 1998, and Wonambi barriei
cation of the number and size range of individual skeletons ScanloninScanlonandLee,2000)revealsvariationinpropor−
represented;thesinglemaxillafromCQ(SAMP40158)lies tionsanddetails,butallpossessahigh(lizard−likeoranilioid−
withinthesizerangeoftheothermaterialbutisomittedfrom like)prefrontalprocess;itisrelativelylowerinWonambithan
Table1because itistooworn forcomparably precisemea− theothermadtsoiids,butunlikeanymacrostomatan.
surements. SAM P16172 was not mentioned by Smith Inlateralview,theboneislongandlow,thealveolarmar−
(1976), despite the registration number (very close to those ginnearlystraight.Theapexoftheprefrontalprocess(above
ofthetypematerial)suggestingitwascollectedataboutthe 9–10) and a lower rise above 17–18 define three concave
same time; the “anterior left maxilla” fragment reported by sectionsofthedorsalmargincorrespondingtothepreorbital,
Smith(P16170c)ishereidentifiedaspartofarightdentary suborbitalandpostorbitalregions.Thesuborbitalconcavity
(seebelow).Theunidentified“maxillaryfragment”shownin is the deepest, formed by a broad and deep trough passing
lateralviewbyRieppeletal.(2002:fig.2)isnotconsistent obliquely across the dorsal surface. Anterolaterally this
withanymaterialthattheyreportedexamining;comparison trough is bounded fairly distinctly by a ridge parallel to the
with all SAM Wonambi material indicates it is actually the alveolar margin, extending anteriorly to the level of the 5th
dentary fragment P16170c (Fig. 12B). alveolus(Fig.1B;notpreviouslydescribed,andunknownin
Thereare22–23alveoliincompletemaxillae,andalveoli othersnakes).Theprefrontalprocessisroundeddorsally,but
(numberedfromanterior)areusedbelowtoindicatetherela− fallsoffsharplyposteriorlyatthe10th,thencontinuesasan−
tivepositionsofotherfeatures.Prefrontal(=ascending,nasal, other dorsally convex ridge until obscured by the postero−
or facial) and palatine (medial) processes are level with the lateralrise.Sevensmalllateralforamina(alveolarforamina,
7th–10thand10th–12thalveolirespectively(Fig.1;notethat for branches of the trigeminal nerve) are present in the
SCANLON—CRANIAL MORPHOLOGY OF GIANT SNAKE WONAMBI 145
Table1.Measurements(mm)ofjawelementsofWonambinaracoortensis.HQ1,HQ2etc.=singleindividualsfromHenschke’sQuarryFossilCave
(HQ),VictoriaFossilCave(VF),Corra−LynnCave,Curramulka(CL);L=left,R=right.Wherepossible,measurementsaredefinedinthesameway
asforNanowanagodthelpi(Scanlon1997:table1);valuesinbracketsareminimaformeasurementsaffectedbydamage.Noprecisemeasurements
couldbeobtainedfromtheotherknownmaxilla,SAMP40158.Maxilla:maxalv=totalnumberofalveoli;mtl=totallength;map=lengthfroman−
teriortiptoposteromedialangleofpalatineprocess;m7−12=lengthfromanterioredgeof7thtoanterioredgeof13thalveolus;mpw=totalwidth
acrosspalatineprocess;mph=totaldepthatprefrontalprocess(excludingteeth,fromventrolateralmarginofalveolitocrestofprocess,notpro−
jectedinparasagittalplane);m12−p=lengthfromanterioredgeof12−lastalveolustoposteriortip.Palatine(ventralview):palalv=totalnumberof
alveoli;ptl=lengthofpalatinefromanteriortipofdentigerousprocesstoposteriortipoftoothrowspineorchoanalprocess;pcl=baselengthof
choanalprocessfromintersectionofanterioredgewithdentigerousprocesstoapexofposteriornotch;pl−11=lengthfromanteriortiptoanterior
edgeof11thalveolus;ptw=widthacrosschoanalandmaxillaryprocesses;pcw=width(insamelineaspreceding)ofchoanalprocess;prw=width
(insamelineaspreceding)oftoothrowbar;pmw=width(insamelineaspreceding)ofmaxillaryprocess.Pterygoid(ventralview):ptealv=total
numberofalveoli;tl5=lengthacrossmostposteriorfivealveoli(5–9);lbp=lengthofbasipterygoidfacet;wbp=maximumwidthoffacet;taw=
widthbetweennear−paralleledgesanteriortoectopterygoidprocess.Dentary:denalv=totalnumberofalveoli;mff=numberofmentalforamina;
dtl=totalstraight−linelength;dl−15=lengthtoanterioredgeof15thalveolus;dlf=lengthtolateralfossa;d4t=posterioredgeof4thalveolustopos−
teriorextremity;d4−15=posterioredgeof4thtoanterioredgeof15th;d4f=posterioredgeof4thtolateralfossa;dl−7=anteriortiptoanterioredge
of 7th alveolus; dmd = depth from dorsolateral to ventromedial edge in middle part of bone; dpp = depth of upper posterior process.
Mus. cat. no. SAMP30178A SAMP30178B FU1762 SAM P16172 SAM P31801 SAM P31785
Specimen HQ1L HQ1R HQ2R VF1L VF2R CL1R CL?L
max alv 22 (12) 23 (13) (14) – (7)
mtl 81.0 – 69.8 – – – –
map 38.9 40.5 34.85 – 31.3 – –
m7−12 22.2 24.0 21.7 – 17.1 – –
mpw 14.0 12.9 10.9 – 9.5 – (10.5)
mph 15.0 14.5 (9.2) – 10.0 – 12.4
m12−p 46.0 – 35.5 47.9 – – –
pal alv – (12) – – – – –
ptl – (33.45) – – – – –
pcl – 15.4 – – – – –
pl−11 – (26.2) – – – – –
ptw – (17.4) – – – – –
pcw – 8.0 – – – – –
prw – 6.5 – – – – –
pmw – (2.1) – – – – –
pte alv – – (7) – – (5) –
tl5 – – 9.1 – – 12.0 –
lbp – – 8.0 – – – –
wbp – – 2.4 – – – –
taw – – (6.7) – – 7.1 –
den alv 25 25 – – – – –
mff 1 1 – – – – –
dtl 75.2 74.5 – – – – –
dl−15 41.2 40.6 – – – – –
dlf 43.3 43.1 – – – – –
d4t 65.1 64.6 – – – – –
d4−15 30.0 30.0 – – – – –
d4f 34.6 33.9 – – – – –
dl−7 16.8 16.8 – – – – –
dmd 14.6 15.0 – – – – –
dpp 4.9 5.1 – – – – –
http://app.pan.pl/acta50/app50−139.pdf
146 ACTA PALAEONTOLOGICA POLONICA 50 (1), 2005
dorsolateralconcavity;mostoftheseformasinglelongitudi− tooth;thishasbeeninterpretedastheattachmentsurfacefor
nal row (almost in one−to−one association with the alveoli), a robust premaxillary ligament (Scanlon 1996). This flat−
withtheremaindermoredorsal.Thetipsofthefirst13teeth tenedareaismoredistinctlydemarcatedandrelativelylarger
liealmostinasingleplaneparalleltothatofthealveoli(Fig. insuccessivelylargermaxillae(Fig.2);thattheligamentwas
2C);thefewmostanteriorteetharesomewhatmoreerectbut veryshort(aswellasthick)issuggested bythesquared−off
notmuchlonger,whileposteriorteethareshorter,andmore shape of the maxillary tip (cf. Cylindrophis) suggesting
strongly inclined posteriorly and medially. The maxillary near−contactwiththepremaxilla,althoughthelatterelement
dentition is thus not really proterodont (as described by remainsunknown.Theconcavemedioventralsurfaceofthe
Scanlon 1997) but nearly isodont. anterior part of the maxilla is bounded ventrally by a blunt
In dorsal view the general outline is approximately as ridgeadjacenttothealveoli,anddorsallybythesharpercrest
shown by Barrie (1990: fig. 5), but additional features are leading up to the prefrontal process; the crest is interrupted
nowvisible:alongitudinalcrestoverhangingasinusandfo− abovethe4thalveolusbyanarrowobliquenotchresembling
raminaonthepalatineprocess;theposterolateral(aswellas thatinpythonines(Scanlon2001),butnotcontainingafora−
anteromedial)boundaryofthedorsaltrough;andlateraland men.Twoforaminaoccurabovethe7th(asmallfollowedby
medialgroovesfortheectopterygoid(abovethelastfoural− a larger foramen); immediately posterior to the foramina,
veoli)convergingatashortdorsoposteriorcrest.Theanterior above the 8th to 10th alveoli, the prefrontal process is
tip is not smoothly rounded (as shown by Barrie 1990; stronglyconcave,formingadeepsinuswithintheprefrontal
Rieppel et al. 2002: 814) but somewhat square in all speci− processsimilartothatinlizards,butapparentlyunmatchedin
mens,withanglesdefiningadistinct,slightlyconvexantero− modernsnakes(anilioidsandleptotyphlopidsbeingtheonly
medialmargin;inthisitresemblesanilioidsandxenopeltids oneswithanequivalent medialsurfaceoftheprocess).The
ratherthan“core”macrostomatans.Theposteriorendisalso dorsomedial“lip”oftheprocess(abovethesinus)facesme−
not rounded (as previously described) but bluntly angular, diallyandsomewhatposteriorly,andbearsverticalgrooves
withridgesandgroovespresumablyforligamentousorten− and ridges implying an interdigitating suture with the pre−
dinous attachments (to the ectopterygoid, and apparently frontal; this sutural surface is similar in all maxillae but in−
also the m. pterygoideus). creasesinrelativedepth,andextentofdigitation,insucces−
Inventralviewthealveoliandtoothbasesaresomewhat sivelylargerspecimens(threeindividualsrepresentedinFig.
rectangular in shape, with an anterolateral ridge on each 2),refutingthesuggestionofRieppeletal.(2002:814)that
tooth forming a bluntly acute angle; distinct anterior and thesculpturewasartefactualinSAMP30178.Theprefrontal
posterolateral cutting ridges are present on the more com− sinusisflooredbytheanteriorpartofthepalatineprocess.A
pletecrowns.Aconcavemedialfaceoftheboneisvisibleup separate,large,posteromediallydirectedsinusonthedorsal
to about the 6th tooth, defined above by the dorsomedial face of the process encloses a smallmedially−directed fora−
ridge. Between the 7th and 8th alveoli this ridge bifurcates menandalarger,deeperonefacingposteriorly(twosimilar
around a medial sinus, forming the edges of the (dorsal) foraminaareseparatedbymoreorlessflatboneinthebooids
prefrontalprocessand(medial)palatineprocess.Themedial examined, and separated by a prominent, oblique dorsal
ridge remains distinctly dorsal to the alveolar row from the ridge in anilioids). The lateral margin of this sinus forms a
8thto10th,convergingwithitmoreposteriorly.Thepalatine posteriorextensionoftheprefrontalprocess,thetwodorsally
process has a relatively symmetrical, bluntly square profile convex crests separated by a dorsal concavity or saddle.
inventralview,slightlylongerthanwide(verysimilarinout− These crests form the medial margin of the oblique dorsal
line to some pythonines including Liasis olivacea Gray, trough, which continues posteriorly as a shallow dorso−
1842,butwithoutthestronglydownturnedmedialedgetypi− medial concavity as far as the 18th alveolus.
cal of pythons); its posterior edge converges gradually to−
As reported by Scanlon and Lee (2000, 2002), breaks
wardthetoothrowuntiladistinctinflexionbetweenthe15th
through some teeth reveal a basal zone of curtain−like
and 16th alveoli (indicating the anterior extent of the
infolding affecting both the outer (enamel) and inner
ectopterygoid, see below). The bone then tapers in width
(dentine)surfaces.Thisconditionisconsidered comparable
fromthe16thto21stalveolus,thenmoresteeplyfrominner
to the “plicidentine” of varanoid lizards, but in Wonambi
andouteredges.Thereisashallowlongitudinalgrooveme−
(and Yurlunggur) it is normally hidden by bone of attach−
dialtothetoothrowfromthe14thalveolus totheposterior
mentinankylosed teeth,and hasbeen revealed fortuitously
extremity, the medial edge of the bone projecting as a low
inonlyafewcases(e.g.,FU1762).Hence,failuretoobserve
ventral crest from 17 to 20. The extent of bone posterior to
infolding on teeth with intact bone of attachment does not
thelastalveolusvaries,beingrelativelygreatestinthelargest
provideevidenceofitsabsenceinanysnaketaxon,anditis
specimen(FU1762)wherethereisamedioventraloutgrowth
difficult to exclude the possibility that the condition occurs
with a strongly pitted surface (Scanlon and Lee 2002: fig.
more widely in extant snakes (Scanlon and Lee 2002).
3a),inferredtobeforpartofthetendinousanteriorinsertion
of m. pterygoideus. Palatine (Fig. 3).—The nearly complete right palatine of
In medial view, the blunt anteromedial tip of the pre− SAMP30178Aistheonlyoneknownfromthisspecies,and
maxillary proces is roughened and pitted above the first can be compared with practically complete palatines of
SCANLON—CRANIAL MORPHOLOGY OF GIANT SNAKE WONAMBI 147
orlyasfarasthemiddleofthe7thalveolus.Theanterioredge
ofthechoanalprocesscrossesabovethemedialedgeofthe
dentigerousbarattherearofthe4th,curvinganteriorlyand
laterally to define the “sub−circular perforation” mentioned
byBarrie,andextendsanteriorlyasfarastherearofthe2nd
alveolus;itsnarrowseparationfromthedentigerousprocess
is best seen in anterior view (Fig. 3C). The choanal groove
occupies the anterior half of the ventral surface of the
choanal process, divided from a flatter posterior area by a
curvedridgecontinuouswiththemedialmarginofthedenti−
gerousprocess,theridgeextendingtothemedialedgeatthe
palatine’s broadest point, between the 7th and 8th alveoli.
Theanteriorpartofthechoanalprocessformsdistinctlateral,
dorsomedial and medioventral flanges, the hollow between
thetwomedialridgesfurtherdividedbyaridgeintotwodis−
tinctconcave“sockets”(Fig.3C,D).Thedeepanteromedial
socket was presumably for articulation with the vomer (cf.
Anomochilus, Cundall and Rossman 1993: fig. 5B), while
theshallowposteroventralonemighthavereceivedaninter−
choanal process of the parasphenoid; however, in Anomo−
chilusandCylindrophistheinterchoanalprocessisdorsalto
thevomerinearticulationanddoesnotformadistinctgroove
onthepalatine.Barrie(1990)reportedpresenceofaforamen
within the anterior socket, but this is not confirmed after
cleaningofthespecimen.Therestofthemedialface,dorsal
and posterior to the sockets, is sculptured and pitted rather
Fig.3.RightpalatineofWonambinaracoortensisSAMP30178Ainventral
(A), dorsal (B), anterior (C), medial (D), and lateral (E) views. than smoothly finished bone (similar in appearance to the
facet of the basipterygoid process, or the anterior tip of the
maxilla),suggestingthatitformedanarticulatingsurfaceor
Nanowana godthelpi and N. schrenki, Yurlunggur sp., and tightfibrousconnection.Similarsculptureandpittingofthe
(?)Wonambi barriei (Scanlon 1996, 1997; Scanlon and Lee medial palatine surface is seen in Cylindrophis (SAM
2000), as well as one from Madtsoia camposi (Rage 1998; R36779)wherethepalatineabutstheinterchoanalkeelofthe
alsomentionedbyRage1984)that,iforiginallycomparable parasphenoid.Thetransverseposteriormarginofthechoanal
inmorphology,mustbeveryincompleteposteriorlyandme− process is unusually blunt and almost square, but close in−
dially. The following points can be noted in addition to the spectionrevealsthatthisshapeisnatural,withnotransverse
informationgivenbyBarrie(1990:fig.6)andRieppeletal. break(previouslyregardedasprobable;Scanlon1996;Riep−
(2002: fig. 3; note that their figure is reversed and thus ap− peletal.2002), and damage islimitedtotheposteromedial
pears to show a badly encrusted left palatine instead of a margin (shown dashed in Fig. 3A, B).
right). Because the maxillary process isbroken there isno evi−
There are 11 complete alveoli preserved (ventral view dence of a foramen or groove piercing it for the palatine
Fig.3A),andpartofoneanteriortothese,throughwhichthe nerve in W. naracoortensis; other known madtsoiids lack
bone is broken; the complete alveoli are referred to here as suchaforamenbutsometimeshaveashallowgroove(Scan−
2–12, but it is possible that one or even several more are lon 1996, 1997; Rage 1998; multiple losses of the foramen
missinganteriorly.Ankylosedstumpsofteetharepresentin are known among modern snakes; Lee and Scanlon 2002,
2,3,8,and9.Thealveoliareclosetothelateralmarginofthe character98).Cleaningofthespecimenhasrevealedtwofo−
bone, which forms two slightly convex arcs separated by a ramina piercing the choanal process, one close to the tooth
concavity level with the 10th (an apparently minor feature, row level with alveoli 7–8 (its dorsal opening facing medi−
butnotinsignificantassimilarlyinflectedtoothrowsareseen ally), and a larger one closer to the medial edge, level with
in other madtsoiid taxa; Scanlon 1996, 1997). The denti− the 9th alveolus, which opens posteriorly. Similarforamina
gerous process has a sloping dorsomedial surface partly arecommoninlizards,andknowninotherAustralianmadt−
flooringthechoanalpassage(dorsalviewFig.3B),bounded soiids(Scanlon1996,1997);amongextantsnakestheyhave
by a dorsolateral ridge that becomes higher posteriorly, ex− been reported in Acrochordus (Hoffstetter and Gayrard
pandingabovethemiddleofalveolus4intoaraisedsurface 1965; Underwood 1967) and illustrated in Tropidophis
that branches laterally (forming the maxillary process) and (McDowell 1975: fig. 4), and are also present in Cylindro−
medially(formingtheanterioredgeofthechoanalprocess). phis ruffus (Laurenti, 1768) (single foramen piercing each
Themaxillaryprocess,mostlybrokenaway,extendsposteri− palatine of SAM R36779). These are also among the few
http://app.pan.pl/acta50/app50−139.pdf