Nimbadon, a I:,Iew Genus and Three New Species Of

Home , Kolopsis, Neohelos, Nototherium

NIMBADON,A I:,IEWGENUS AND THREENEW SPECIESOF TERTIARY ZYGOMATURINES(MARSUPIALIA: DIPROTODONTIDAE) FROM NORTHERN AUSTRALIA,WITH A REASSESSMENTOF N'OI''ZOS S.J. HAND, M. ARC}IER,H. C,ODTHELP.T.H. RICH AND N.S. PLEDGE

Hand, S.J.,Archer, M., Codthelp, H., Rich, T.H. & Pledge,N. S. 199306 3O:Nimbadon,a new genusand three new speciesofTeniary zygomaturines(Marsupialia: Diprotodontidae) from nonhem Australia, with a reassessmentof Neohelos.Memoirs of the Queenslnnd Museum 33(l\: 193-210.Brisbane. ISSN 0079-8835.

Thrce new species of Oligo-Miocene zygomatu.ine dip.otodontids are described from northem Australia. All are small,l, plesiomorphicplesiomorDhicand appearto comprise a distinctive clade of zygomaturines.naEed here Nimbadon.Nimbalon. TlreThe clade is Dartlvpartly defined on the basis of the posleriorly iqclined P" parastyleand blade exrendingfrom the parametaconeto the lingual halfofhalfofthe the P'crown and then to lhe anterolinpualanterolingualcinpulum.cingulum. Two ofthe sDe.iesspe.ies areiue known from Oligo-Miocenelocalfaunas ofRiversleigh Slalion,nonhwestern Queensland. The third is from the middle Miocene Bullock Creek Local Faunaof Camfield Station. norlhwestern Nonhern Territory. Descriptionof additionalfossil materialreferable ro Neohelostrrarensrs Stinon, 1967,a middle Miocene zygomaturinefrom the Kutjamarpu lacal Faunaoi South Australia, enablesthe genus to be distin9tished frcm Nimbadon. The chronoslratigraphic significanceof the ncw zygomaturinesis consideredand found lo approximalelycorrespond to curent undeBtandingofthe relative agesof the dcpositsfrom which the specimenswere obtained. n Diprotodontidae, Z;gomaturinae, Ninbadon, Neohelos, Oligo-Miocene, Riversleigh, Bullock Creek local Fauna, Henk's Hollow lacal Fautv, Fig Tree lncal

Suzonne Hand, M. Archer & H. Godthelp, School of Biologicol Science, UniversiD, of New South Wales, PO Box I, Kensinglon, New South llales 2033, Australia: T.H. Rich, Palaeontology,Museum ofVictoria, Russell St,Melbourne, Victoria 3000, Australia: N.S. Pledge,Palaeontology, South Australian Museutn,NorthTerrace, Adelaide, SouthAustru- lia 5000, Australia;24 November, 1992.

Specimensrepresenting a very small zygoma- (Rich et al., 1982, l99l; Archer & Flannery, turine werc collectedbetween 1984and 1986by 1983; Archer & Hand, 1984; Woodbume et al., Archer. Godtheloand Hand from the Henk's Hol- 1985;Munay, 1990a;Munay&Megirian, 1990; low Site, in an ds yet unnamedfreshwater lime- Archeret al., 1989,1991).Recent reappraisal of stoneon the Gag Plateauof Riversleigh Station the ageofcentral AustralianTeniary mammalian I (Archer, Hand & Godthelp, 1986;Archer & FIan- faunas,based on the study of foraminifera (Lind- nery, 1987;Archer er al., 1989,1991). Refened say, 1987), suggeststhat some of these faunas material includes maxillary and dentary frag- (e.g. the Ditjimanka, Pinpa and Ericmas Local ments and isolated teeth, A second, larger Faunas)are appreciablyolder than once thought Riversleigh species, representedby a maxilla, (e.g. Woodbume et al., 1985; but see Tedford, was collected in 1986 by Plcdge from Fig Tree 1966;.possibly latc Oligocene.By comparison. Site Locality adjacent to Godthelp Hill, the Bullock Crcek Local Faunais estimatedto be Riversleigh Station. ?middle Miocene in age, i.e. irnmediately post- A small Bullock Creek zygomaturineis known Wipajiri (Munay, 1990a; Murray & Megirian, from a palate collected in 1981, by Rich and 1990).On-going studies (Archer et al., l99l ) of colleagues,from HorsesboeWest tocality (WV the Riversleigh fossil faunas suggest that the I 13) in the Camfield Beds of Camfield Station, Henk's Hollow L,ocalFauna is ?middle Miocene northwest€m Northem Ter.itory (Rich et al., in age and the Fig Tree Local Fauna somewhere 1982.l99l ). betweenlate Oligoceneand early Miocene in age. Estimatesof the age of the Henk's Hollow, Fig The new zygomaturines described here are Tree and Bullcrck Creek l-ocal Faunasare based most similar to forms currently referred to the mainly on stage-of-evolutioncomparisons oIpar- genusNeohelos. Their generic distinction, how- ticular diprotodontian marsupialswith those of ever, has beenclarified by discovery at the type other northern and central Australian faunas locality of new material rcferable ro Neohelos 194 MEMOIRSOF THE QUEENSLANDMUSEUM

TABLE l. MeasuremeltsoI Neohelostirarensis t' r ing unchallengedconvention, is called the hypo- (AMF87625) and M (At\4F87626)from the Leaf cone.Ch€ektootb homology is that proposedby I-ocality,Lake Ngapakaldi, South Australia. In mm. Archer (1978). Higher level systematicnomen- Abbreviations:ant, anterior;post, posterior; paramet, clature (1987). parametaconeiprotoc, protocone; parac, pamcone. followsAplin & Archer

Lenqth 15.8 SYSTEMATICS width 13.9 SuperorderMARSUPIALIA Illiger, l8l I c widlh(parametto protoc lips) Order DIPROTODONTIA Owen, 1866 heisht(parameo I t.4 Family DIPROTODONTIDAEcill, 1872 h€ish(parastyle) 6.5 SubfamilyZYGOMATURINAE Stirton, lenqth 18.8 Woodbume& Plane,196? M3 width(anr) t1.2 width(post) r5.8 NeohelosSrinon, 1967 heisht(parac) 1.4 (Fig. I, Table l ) tirarensis, the tyW andonly namedspecies of the Typespecies. Neohelos rirzr"dJri Stinon,1967 genus.The original description(Stirton, 1967)of this Miocene zygomaturinefrom the Kutjamarpu REVTSEDGENERTc DrAcNosls Local Fauna of central Australia was basedon The following combinationof featuresappears five isolated teeth collected in 1962 at the haf to distinguishspecies of Neoftelosfrom all otber Locality, t-ake Ngapakaldi, Etadunna Station, zygomaturines:P'with large,erect parastyle that South Australia. The holotype, a broken P', pre- is well isolatedfrorn the parametaconeiparastyle serves only the posterior portion of the tooth. is conical without tip developedas-a biade; in- Though lesscomplete than the other teeth,it was terolingual basin well defincd in Pe by distinct selectedas the holotype becauseof the impor- basalcingulum extendingbetween parastylar cor- tanceof premolar morphology in diprotodontoid neroflooth lo anterolingualbase of protocone:P' systematics(Stirton, 1967; Stirton, Woodbume parametaconcwith an undividedtip: paranlcta- & Plane,1967). cone tip not developcdinto rnrcrobucall\ori, In 1982 two isolated teeth referable to M ri, ented bladc: linc r-ulrrccn prot()cone. rcrezsis were collectedfrom the type locality by paramclaconcand s idcstbuccal point on cro* n Archer, Hand, T. Flannery,G. Hiclie, J. Cascand is appro\inralcl\r!'itilincdr lnd dirrdcs crown P. Bridge: a well-preserved RPr (AMF8762-5) eitherin hlrlf r)r lca\c\ lnlcnr)r division shorter and RM' (AMF87626). These add considcrably than the F)stcn('rdir rrion: P *ith well-devel- to knowledgeof this species.Additional materi. opedhrg'--^onc or hlFEonc shelf; M2-rwith well als rcferableto Neolelos from the Bullock Creck dcvelol^-dpara\t) lc and metastyle. l.ocal Fauna (Plane & Gatehouse.1968: Planc. l97l; Munay & Megirian, 1992) and various \imbadon n.gen. Oligo-Miocene Riversleigh local faunas(Archcr tFigs2-5, Table 2) et al., 1989,l99l) arepresently under stud1,. 'Irpe lnstitutional abbreviationsuscd hereare as fol- sptcus..\'i badonlarorackorutn n.sp. lows: QMF, fossil collectionof the Queensland Addiiono! sp.cies.Nitnbadon wltitelawi n.sp.. Mrr- Museum; AMF, fossil collection of the Auslra- DAdotrscoIloft ot um n.sp. lianMuseum; SAMP, palaeontolosical col lecrion of the Soulh AustralianMuseuri; NMVP, pa- GENERIcDIAGNOSIS laeontologicalcollection of the Museum of Vic- Specicsof Nirnbadon differ fiom thosc ol all toria; AR, researchcollection of the School of other zygomaturines(with special rcfcrcnce 10 Biological Sciences,University of New South Neohelos,Ihetaxon to which they are othcrwise Wales. Cusp nomenclature follows Archer most similar) in the following combinationof (1984) and Rich, Archer & Tedford (1978). The features:small size (althoughthey are largerthan posterolingualcusp of theupper molars, conven- RaeneotheriutnRich, Archer & Tedford, 1978); lionally called the hypocone(e.g. Stirton, Wood- in havinga molargradienl lhilt does nol apprecia- burne & Plane, 1967), is now called the bly increaseposteriorly (in contrast to all other metaconule followine Tedford & Woodbume zygomaturinesexcept Raeneotlrcriu,r); posteri- (198?) but the postero=lingualcusp of P3,follow- orly inclined parastyleon P' (nther than conical NEWDTPROTODONTIDS FROM NOR-IHII{N AUS ]RALIA r95

l l(' | ,\eoltelostirat?/rJlr, baf l,o,calily,Lake Ngrpakaldi, $outh Austmlia. AMF87625, RP': A-A', occlusrl 't.rcopair:B. buccal!iew; C, liogualview. AMF87626, RM': D D', trclusal stereopirir.Ilar irrdicnlcslOnt l. .rrrJ( recr .rs in othcrz) gnn rrlurine\ e\cept possi zygomaturinesexcepl Neo lrc [ o s aDdA Ibrc rt atlr blt PLaisiodoncerr,'.?l/.r Woodbume, 1967); an erilal Munay, 1990b);an"antcrobuccal hlade on 3nlerolingualblade on the parastyleof P' that the panmelacoDcof P' that sunnounts an surmountsa postcrolinguallyoriented thegotic obliquelyoricntcd anlcrolingual thcgotic frcette lrcctte (in contrastto no apicalblade such as (in contrastto eitherno bladeor a poolly devcl- nrar'[?]charactcrise Neo he Lo s); completelyundi- opcd antcriorlyoricntcd bladc such as may['/] \idcd parametaconeon P' (in contrastto all other characteriseNeohelos); very snlall to absenthy- 196 MEMOIRSOF THE QUEENSLANDMUSEUM

LOWERDENTITION A Species Ninbador lavaracko.un TABLE 2. Measurements of OMF 23t42 23t53 23154 23155* 23158 23t59 Nimbadon lavaroc korum L 12.5 lo.9 PJ n.sp.,1V.writelalri n.sp.and N. 1.O 5.8 scottorrofum n,sp. from L 14.0 Riversleigh and Camfield Sta- 13.3 ll.0 M2 tions, northem Australia. ln 8.4 8.3 mm. Bmckets around figures 9.4 8.6 8.4 indicate estimates.Abbrevia- L 15.4 tions: ant,anterior:posl. poste- M3 l0.l rior: *Gag Site specimens. t0.0 L |5.5 t4.0 Ma l l.0 I0.6 10.2 L 15.9 t4.3 M5 l l.l I1.2 10.0 '72.1 Pr-M5 L Mz.s L 60.9

UPPERDENTTTION

Species Ninbodon lavarackorun qhilclasr r.?toronrr

QMF 2314\ 23143 23t44 2314523].46 23t41 23148 21t49 u:rso::roo- plls\L :;rsr n!hr lell J L t2.7 l4.l 12.4 lt 9 r3.0 14.3 10.3 10.4) 10.9 t3.4 L I1.9 IJ.3 i.r.I ll_l 3.2 1.1 M2 w(ant) 10.2 10.9 I l.l (9.3) 5.0 9.9 10.8 10.5 9.2 5.8 L t2.3 l3.l t4.3 I3.l 7.9 11.3 1.6 10.9 10.1 10.9 t0.4 (9.5) L 13.8 l4.l lt.5 t8_5 M- I 1.5 12.2 I l.l t0.7 10.4 L 13.5 18.r M" w(ant) ll.3 w(Dos0 9.8 Palatewidrh 29.2 P,-M, L (84.4) P,.M' L 52.8 66.3 P3-M2 L z5.3 25.8 48.2 L (71.5) M'24 L 39.5 53.1 NEW DIPROTODONTIDSFROM NORT}IERN AUSTRALIA 197 l l j l

F1G.2. Nimbadonlavarackorum n.gen., n,sp.,.Henk's llollow Sile,niversleigh Slaiion A-A', QMF23l4l, holetype,Ieft maxillaryfragment \.\ith M'. M'and buccalhrlf of P', occlusalstereopirir. B-D, QMF23l4l. RP': B. lingualview: C-C , c'cclusalstereopair; D, buccalview. Bar indicateslOmm poconeon Pr (in contrasttoAll<.wertatherium and developed);poorly developedto absentanterol- most but not ^ll Neohelosand in contrastto all ingual basalcingulumjoining baseofparastyle to other zygomaturineP3s where this cusp is large, baseof protocone(in contrast to most other zy- including Maokopla Flannery 1992); pos- gomaturinesexcept some Neohelos where there terobuccalbasal cingulum on P' (in contrast to may be only a vgry small cingulum); an anterol- most other zygomaturines except ALkwertath- ingual crcston Fpthat extendsftom the pammeta- einnandNeohelos wherethis featurcis variably cone towards lhe protocone but. veering 198 MEMOIRSOF THE QUEENSLANDMUSEUM

anteriorly, bypassesthe buccal crest from the Archer et al., 1991), Microsite (containing the protoconeand extendsto the anterolingualbasal Nooraleebal-ocal Fauna; Sig6, Hand & Archer, cingulum (in contrasttodirectly connectingto the 1982;Hand, 1993)and Gag Site (nearthe baseof buccalcrest from the protoconeas occurs in most System C, containing the Dwornamor hcal if not all other zygomaturines);a line extended Fauna; Flannery & Archer, 1984; Hand, 1985; buccally tbrough the protocone and parameta- Arcber et al. l99l), but is close to (though still cone passesanterior to the position of the widest above)the level ofTwo TreesSite (containingthe point on the buccal baseofthe crown (in contrast Two Trees Local Fauna Flannery & Archer, to intersectingthis point as it does in Neorelos 1987). On the basis of its stratigraphicposition and Alkwertatherium, altbough this feature is and stageof evolution comparisonsof its fossil variableh Nimbadon wlrirelawi n.sp.);this same mammals(work in prepamtion),the Henk's Hol- line divides the crown such that the anteriormoi- low Local Faunais tentativelyinterpreted herc to ety is longer than tbe posteriorone (as it does in be middle Miocene in age (Archer et al., 1989, Alkwertatheriumbut in contrastto eitherdividinc 1991). the crown into approximatelyequal lengthsoi into shorteranterior and longerposterior moieties REFERREDSPECIMENS AND THEIR LOCALITIES asitdoes in Neohelos);poorly developedparasty- QMF23142from the Henk'sHollow Locality,a left les and metastyleson uppermolars (in contrastto dentarycontaining P3-M5, is tentativclyrcfbned to this soeciesof Neohelosbut as in someanterior teeth species,as are the following iso,lated teeth lrom Henk's oTother zygomaturinessuch as P. centralis). tiollow: QMF23l43,a qgit P3;QMF23 | 44, a left Pl; QMF23l45, a left M'; QMF23l46, a right M'; ETYMol-ocY QMF23I47,a left M'; QMF23I48, a left ?lvf; Nimba is a nofihwesten QueenslandAborigi- QMF23l49,a right M'; QMF23150,a right M'; nal word (Wanyi language)meaning 'small'; don QMF23l5l, a right P3;QMF23I52, a left Pr; 'tooth'. is Greekfor The namealludes to the small QMF23l53,a right P3; QMF23l54, a left ?M;. molars found in speciesof this genusand to the QMF23l55,a rightM2, and QMF23l60, a right M'-", fact that the Dosterior molars do not increase werecollected from the Cag Siteof thesame plateau, markedly in si2e from M2 to M4 as they appear fuversleighStation. to do in lhe otherwisesimilar Neohelos tirarensis. REFERRED[nCALITIES AND AGES Nimbadon lavarackorum n.sp. The stratigraphicrelationships of the Henk's (Figs2,3) Hollow l-ocal Fauna and the Dwomamor Lrcal Fauna are indicated above. At present,both are HoLoTYPE interpretedto be middle Miocene in age. Theholotype^is qMF23l4l, a left maxillaryfragment containinsM'. M' andthe buccal half of Fp. SPECIESDIAGNoSIS This speciesis distinguished from Nirnbadon ETYMol-ocY scottorrorutn n.sp. from Riversleigh's Fig Trce The species is named after Sue and Jim locality{,see belowt by its smallersizc. rclatrvel; Lavarack in recognition of their invaluable sup- longerPe. b€tler developed hypocone on P'. less port in the field at Riversleigh and as founding well-developed parastyle on M' and probably members and hard working Councillors of the more elongate posterior upper molars and less Riversleigh Society, a suppon group for Austra- well developedpostmetacristabut more discrete Iian palaeontologicalresearch. metastyleon M". It is distinguished from the similar-sized N. TYPELOCALITY AND AcE whitelawi n.sp. by its almost square(Iather than Henk's Hollow Locality (Archer & Flannery, rectangular)upper molars. 1987) occurs within the sequenceof Teniary limestones outcrcpping on Riversleigh Station, DEScRIPTIoN no hwestemQueensland (Archer & Hand, 1984; F is representedby three specimens:a LP3 Archeret al., 1989,l99l). It is stratigmphically from the holotype,qMF23141iap isotatednl3, nearthe top of the SystemC sequenceas defined QMF23143;and an isolatedLP, QMF23144. by Archeret al. (1989,l99l). This is interpreted The tooth is widest acrossthe protocone,sharply to b€ stratigraphicallyhigher than Site D (System pinching in antedorly and smoothly rounding A containing the Riversleigh l,ocal Fauna; posteriorly. It is subtriangularand is compnsed NEWDIPROTODONTIDS FROM NORI'HEIiN AUSTRALIA 199

I'lo. J. Ninbatlon [av,r/.].tdrrn n.gen.,n. sp., HcDk s Hollow Site.Iliversleigh Slalion. QMF23 1 ,12. lc li dcntnry rrith P:.M:. M:, Mr andMj. occlnsalslcreopair. Bar indicales10nlm. trl lour cusps:the parastyle, paramctaconc, proto toconeis lingual to the plramctlicoDcand thc .one and hypocone(the laller however being hypoconelics postcrobuccrlto thc protr)coDc. \lfilblv dcvclopcd).Thc paranrctaconcis the The hypoconeis moresharply sepafated by deep highcstcusp. lt is bladelike with a strongshear flssurcslrcnr thc buccalcusps lhiin is the proto lns creston lhe poslerioredge and a shorl an cone.Thc parast)ilcis postcriorlyinclined with a lerobucallyoriented apical crest. The protocone shortlnterolinBurlly oricnlcd apicll blilde.This ilnd parastyleare subequalto eachother in sizc. largc conicalcusp forrnsa stronganterobuccal The hypoconcis thc smallestcusp and is cven projectionin theocclusrl outlinc ol thccrorvn. lt .rbsentin onespecimen (QMF23l44). There is no is seprrrtedlioln adjaccntcusps by a chevron .uggcstionof a division in thc paranrclaconcshapedtransverse cleft that is bctlcr-dcvclopcd 'rhich is pyramidshaped having relatively Uat- lingurlly thanbuccally. Postedor to thiscleft, and rcnedanterior. buccal and lingual faccs. Thc pro on the anteriortlanl( of thc paramctacurc.a con- 200 MEMOIRSOF THE QUEENSLANDMUSEUM

spicuousridge extendsanterolingually from the tip of the metaconeto the posteriorcingulum with tip of the pammetaconeto the anterolingualcin- which it merges.A weakly developedpostparac- gulum at a point anterobuccalto the baseof the rista on the oosterior flank of the oaraconeex- protocone. From the tip of the protocone, a tends toward the midvalley but stops short of poorly-definedtransverse crcst runs buccally un- contacting similar premetacdstaethat extend til it almostmeets (but doesnot) theparametacone from the metaconeand from just lingual to the :: ridge, its counter?art,at the commisurethat sepa- metaconetowards the midvalley, From the proto- rutesthe protoconeand parametacone.Anterolin- cone and metaconule,wide but very faint crcsts gually, a very poorly developed (to ?absent) extend posterobuccally across the posterior cinsulum extendsfrom the anterobuccalbase of flanks of the protoloph and metaloph. From the : the-protoconeto the lingual baseof the parastyle Drotocone.this crest extends to the tmnsverse i but doesnot continueup theflank ofthe parastyle. valley nearthe midline; from the metaconulethis A cingulum is also developed posterobuccally crestextends to the posteriorcingulum also near from the buccal flank of the pammetaconeto the the midline of the tooth. most posterior point of the crown (but is better- Wear facetteson thesecrests exhibitnolish and developedin somespecimens than in others)and parallelsrriations {paflicularly in QMF23l45) appearsto be developed posterolingually from suggestingthat their function was maintainedby this point to at least the base of the hypocone. thegosisagainst (respectively)the entoconid of There is no evidenceof an anterobuccalcingu- Mz and the metaconidof M:. The anteriorcingu- lum. From thetip of the parametacone.a promi- lum extendsfrom the parastylearound the lingual nent crest runs to the posterior margin of the baseof the protoconewhere it is interupted and tooth, dividing the posterolingual and pos- thencontinues across the lingual cnd ofthe trans- terobuccalcingula. A small crestruns posteriorly versevallev to the anterolinsualbase ofthe meta- from the tip of tbe protocone to a swelling or conule where it is interruptJdand then continues cuspule.Another thin crest on the posterior face around to lhe metastvlc.Therc is no buccal cin- of the hypocone appearsto converge with the gulum. The M2 may fiavc beenthree-rooted with posterolingualcingulum (at leastin QMF23l44). a cvlindrical root beneaththe pamconeand an- The tooth is two rooted.The anteriorroot is coni- othir beneaththe protoconeand asingle wideroot cal and slightly anteriorly sloping;the largerpos- beneaththe metaloph. Precise thegotic facettes terior root is anteroposteriorly flattened and occur on the anterior trailing edgesof the pro- anteriorly convex. Many of the medial surfaces toloph and metalophblades. The anterior flanks of the four crowns are covered in fine, poorly of thc protoloph and metaloph are "hollow- develoDedcrenulations. There are two roots: one ground" in anticipationofthe thegoticsharpening anterioi oval one beneath the parastyle; and a of the blades. Fine vertical crenulationson the much longer antercposteriorlycompressed root flanks of the protoloph and metaloph produce thatextendstansversely b€neaththe whole width secondary,vertical ?beta thegotic blades as the of the DosteriorDart of the crown. tooth sustainsabrasive and thegotic wear. l. ih" hrrt ,ipp", adult molar is known from M3 5. M3'5are similar to M2 in basic morphol- thg holotype, QMF23l4l, and ?an isolated left ogy but are slightly larger and higher-crowned. Mr, QMF23I45. It is a.low-crowned almost The paraconeis larger and morc buccally- situ- squaretansversely lophodont tooth that is only atedthan the metacone.The protoloph an^d.meta- slightly longer than wide. It is comprisedof an loph increaseslightly in width from M'-. The anterior moiety, the protoloph, and a posterior parastvleand poslrnelacristaare markedlyre- t moiety, the metaloph.Both areanteriorly convex. buced'inMt suchthar in M' theyare represinted The oaraconeand metaconeare similar in size as by terminalswcllings in theanlerior rnd poslerior arethe slightly shorterprotocone and metaconule. cingularespectivelyandarc lbsenl inM' andM'. The metaconuleis slightly higher than the proto- In somespecimcns (e.g. M'in QMF23l60),how- cone and more lingually-situated.The metaloph ever, there is a very tiny cuspuleon the posterior is thus wider than the protoloph. The paracone flank of the metaconeof posterior molars. The and metaconeare slightly closertogether than are postpamcristais less well-developed as are the the protoconeand metaconule.In the anterobuc- crestsfrom the Drotoconeand metaconule.The cal corner of the tooth a weak parastyleis devel- mesostyleappcais to be lacking. The part of the oDed at the buccal end of the short anterior crown antedor to the protolophis shorterand the cingulum. There is no metastylealthough there is swelling in the transversevalley betterdeveloped. a DrominentDostmetacrista that extendsfrom the The metaloph is much more strongly curved in NEWDIPROTODONTTDS FROM NORTHERN AUSTRALIA 201

M3-5and is more lingually offset with respectto apex of the protoconid. Taken together, these the protolopb. flanking cuspules and their associatedvertical cristids comDrisea bandve$e structurethat inter- Dentary. QMF23I42, a dentarycontaining P3, Mz-5,is t€ntativelyreferred to this species.It is of sectsthe crown at its apex at dght anglesto the a size appropriare for some of the referred upper main longitudinal shearingcristid. The posterior molars of this speciesand in its suite of plesio- cingulum is rnuchbetter developd than the ante- morphic zygomaturinefeatures (e.9. its lack of a riorcingulum whicb is presentonlyat the anterol- posteriorly increasing molar size gmdient), is ingual tip of the crown. The posterior cingulum analogousto the phylogeneticstate of the upper extendsfrom the lingual baseoftbe protoconidto molars of N. lavarackorum. However, because the posteriorcuspid where it is met by the buccal the specimen appearsto be approximately 10olo baseofthe protoconidbut is bestdeveloped in the larger than the holotype, there is still someslight region of the posterior cuspid where it is met by doubt aboutthe propdety of referringthis dentary the main longitudinal crest of the crown. In this to N. lavarackorum. arca of inteNection, the posterior cingulum is conspicuouslycrenulated in somespecimens (i.e. The dentary is badly fractured and is missing QMF23l52 and QMF23153) and slightly lessso the incisor, medial symphysis,angle and ascend- in others(i.e. In a wear ing ramus. It is deepest the hypolophid of QMF2315l). QMF23l52, below lacelteon lhe poslerobuccalcingulum. possibly ventral posleriorly upward Ma. The borderslopes producedby lhe P'parametacone,is developed from this point and anteriorly upward to at least which is not seenin other specimens.In this same the hypolophid of Mz. The anterior edge of the specimen,justanterior to the crenulatedposterior process posterodorsally coronoid leads from the cinsulum. a transversecrest extends anterolin- postalveolar body of the horizontal ramus. The gually from the tip of the posterior cusp to the process is missing. There is no distinct digastdc posterolingualcingulum. In all specimens,a wear fossaand the areabeneath the leadingedge ofthe facettefor Mz is developedat the tallest point of process coronoid is smooth suggestingthat the the posterior cingulum. The tooth is double- massetericfossa was not deeD.The mental fora- rooted, the anterior root being conical in shape, men is located approximately4mm in front of a the postedor root being anteroposteriorlyflat- vertical line marking the mostanterioredge ofthe teneo. P' crown. It is round in lateml view and opens anterodorsallyonto the buccal surfaceof the di- Sp€cimensQMF23l52 and QMF23l53, al- astemal region. Although the mandibular fora- though similar in morphology, are considerably men is not Dreseryed.the mandibular canal is smaller than the Pi preservedin the dentafy. The representedby a sulcus at about the leyel of tbe Pl in this speciesmight be sexuallydimorphic. tooth row. M2. Tlne Mz, known from QMF23142 and (andthe Gag Site is an The P3 is preservedin QMF23142 as well as QMF23154? QMF23l55), elongale.subrectangular loolh that is narrower QMF23l52, an isolatedLP3, QMF23l53, an iso- anteriorly than posteriorly.The trigonid of Mz is lated RP3 and QMF2315l, an isolatedRPt. The tooth is longer than it is wide and more narrow rcpresentedby a transvene protolophid ard an anteriorly than posteriorly.It bearsa centralprin- arcuate anterclingually directed paracristid that cipal cusp, the protoconid, and a much smaller extendsto the anteriormargin of the tooth. From medially positioned postedor cingular cusp. the protoconida short indistinct crest(?protocris- These are joined by a prominent longitudinal tid) appearsto extend posterobuccallyto termi- shearingcrest that also extends anteriorly from nate in a slight swelling which is perhaps the the protoconidto terminateas a cuspuleor cingu- protostylid. From the metaconid,a crcst extends lar swelling at the anterior edge of the crown. anteriorly a short distance.From the entoconid,a Lingual to the principle centml cusp there is a short entocristidextends anterolingually. variably distinguishedcuspule (stylid or vertical In occlusalview. the oaracristidis convexbuc- cristid). This structure is the apex of a vertical cally, the protolophid is transverse or vcry flanking crestthatextends ventrally to a pointjust slightly posteriorlyconvex and the hypolopbid is posterior to the lingual base of the protoconid. markedly posteriorly convex.The protolophid is This crestis most conspicuousin posteriorview. narrowerthan the hypolophid. An anteriorcingu- A posterolingualcingulum leadsposteriorly from lum is present both buccal and lingual to the the baseof this crest to the baseof the Dosterior anterior end of the paracristid.The cingulum on cusp. A similar but much less well-developed the buccal side is not preseryedin QMF23142, is vertical cristid extendsventrally from the buccal distinct (between the protolophid and hypolo- 2fr2 MEMOIRSOF THE QUEENSLANDMUSEUM

phid)in QMF23l55but not in QMF23l54.There TypeLocA]-rry AND AGE is no lingual cingulum or cristid obliqua in The type locality is fossil vertebratelocality QMF23I42 but there is a lingual cingulum in WV I l3 in Bultitude(1973) (17"7'S, 131"32'E), QMF23155 (again betweenthe protolophid and tbe HorseshoeWest locality of the Camfield Beds hypolophid). The posterior cingulum connects at Bullock Creek.Camfield Station.northwestern the posteriorbase of the hypoconid to that ofthe Northem Teffitory. The Bullock Creek lncal entoconid. A wear facette is develoDedat the Fauna (Plane & Gatehouse, 1968; Archer & midpoint of theposterior cingulum. Diltincl wear Bartholomai, 1978; Rich et al., 1991; Archer & facettesextend along the posteriorlengths of the Hand, 1984; Murray & Megirian, 1990, 1992)is Fotolopbid and hypolopbid. ln lingual view, the currently interpretedto be ?middle Miocene, i.e. floor of the tuansversevalley is U-shaped.The irnmediatelypost-Wipajid, in age(Woodbume et protoconid is taller than the subequalmetaconid, al., 1985; Munay, 1990a; Munay & Megirian, bypoconid and entoconid. The tooth is double- 1990,1992). rooted but these are not well Dreservedin the isolatedreeth QMF23l54 and QMF23l55. SPECIESDIAGNoSIS QMF23154 (Mz?) has a paracristid and pre- This speciesis distinguishedfrom M lavnrack- metacristidthat arc much lessdistinct than those orutn and N. scoltorrorumby its markedly more preservedin thedentary and GagSile specimen. elongate(i,e. lesssquare) upper molars and rela- Only the paracristidextends to the antedorcingu- liyely moreanteriorly silurted buccal swelling on lum of the tooth. ln this soecimen.the enamelon P'. ft is alsodistinsuished from M scottorrorutn the posterior face of the protolophid has promi- by its relatively riuch longer P3 and markedly nent vertical crenulations.The cdstid oblioua is convex anterior'andposterior molar crown mar- alsodistinct althoush low andextends as a minor gins such that theseteeth have a much reduced verticalcrenulation on to the Dosteriorflank ofthe arcaof interdentalcontact. protolophid. Mt Unlike M2.lrigonids of theposlcrior mo- s. DEscRtPTroN lars lack a prominent paracristid and arc more The speciesis described rectangularin appearancewith the prctolophid insolar as it differs from Nitnbadon lavarqckorwn. b€ing slightly wider than the hypolophid in M3.5. The palate prcservesmost of the right and a In M3, the protolophid and hypolophid are ap- largepart ofthe left maxillae and small fragments proximately the same width. The paracristid ofthe left andright palatines.Anteroventrally, the forms a poorly-definedvertical crest that extends palate is fractured at or behind the premaxillo- only partly down the anteriorface of the protolo- maxillary suture. Posteroventrally,it preserves phid. The premetacristidis also poorly defined the maxillo-oalatine suture. The latter extends and extends approximately the same distance. anteri-orlyto-a point medial to the posterior part Protolophidsand hypolophidsbecome more pos- of M'. Palatal ridges extend longitudinally and teriorly convex from Mz to M:. In Mr s, the pro- anleriorlyalong the lengthof thepalate from the toloDhids are markedlv wider and the level of M'. The right diastemalcrest runs an- hypolophids are lingually displaced.Ma and M5 terolingually. The looth rows are parallel arc larger than Mt. to slightly convex buccally. Anteriorly, doming of the median region of the palate is marked. IJ Nimbadon whitelawi n.su. extendsposteriorly to the level of the rear of M' (Fig.4) and, though shallowing perhaps,anteriorly to at Ieastjust behindthe premaxillo-maxillarybound- HoLoTYPE ary. There is a oair of nutrient foramina on the The holotype and only known spec^imgnis right sideofthe ialate medial to P3and on the left NMVPI86506.It Dreservesthe oalate.RPj-M) and side a single foramen. Medial to the hypoloph of LP3-M3.AII teethexcept LP3 are damaged and LM3 M', is anothertiny foramen on each side of the andRM'are missingparls of theirposterior halves. Dalate.The infraorbital foramen canal is about 30mm in length.The infraorbitalcanal opensontg the face l4mm above the anterior edgeof the P' ETYMoLocY alveolus.Sutural relationships of the palatine. This soeciesis namedafter Michael Whitelaw lacrimal,jugal and maxillary bones are unclear. who assistedin the collection and processingof There is no evidence for maxillarv or oalatal sD€cimensfrom Bullock Creek. vacuiliesallhough lhe anteriorand posreriorex- NEW DIPROTODONTIDSFROM NORTHERN AUSTRALIA 203

FIC.4. \inbatlon wl,i/eldwi n.sp.,Bullock Creek, Norlhern Territory. NMVP)86506. palate with RPj-M) and LP--M". occlusal stereoDair.Bar indicatesl0mnl. tremities ofthe palateare missing.The maxillary badoninLteinglongerrelative to C andin appear- processnear the roof of the zygomatic arch, al ing to be^rectangularmther than squarc. though damaged,is down{urned and conspicu The M' appearsto differ from the Riversleigh ous but doesnot extendventmlly below the level speciesin a simil3rway as M'. bul it appeatslo of the palate.There appearsto have been a very be relativclyc\ en moreelongare. M' and\4' are reduced orbital wing of the maxilla, although approximatclycqual in length. bone bounda es in this region are unclear. The Mo differs less frcm that tooth in the other Upper dentition. The Pp differs from most speciesof Nir?D.rdo,r. specimens(all exceptQMF23l44 AR5513)ofM All of the cheekteethofN. )rriteldrri appearto laverackorum in the greaterdegree of develop- differ from those of M scottorrorun in having ment of thc hypocone. ln N. $,hitelq$,i the anteriorly and posteriorlymarkedly convextooth parastyleappears to be less posteriorly inclined marginsso that the crowns abut with limited contact. and the anterolingually-directedcrest that extendsfrom thc parametaconeto the anterolingual cingulumis lessdistinct although it is possibly Nimbadon scottorrorum n.sp. (Fig. diminished by wear. It neverthelessapproaches s) the buccal crest from the protocone at a steep anglerather than perpendicular. HoLoTYPD The M' on both sidesof the holotypc is dam- Theholotype and only known specimen is QMF23l57 agedbuccrlly and lingually. However. it appc:rrs {Iornrcrl\SAMP278l5i. J flcht nr.rxillarvIrasnlent to differ from both fuversleigh speciesof Ni,r- conrainingP'. M . Mr.fM":rni M' whileP'r;d M2 204 MEMOIRSOF THE QUEENSLANDMUSEIM

areintact, M3-5 are rnissing the buccalmargins ofthe and is subroundedin shaoe,Unlike N. lavarack- ctown. orzm, in which C is equaiin lengthorlongerthan any adult upper molar, in M scottorrorum thrs ETYMoLocY tooth is the sbortestin the tooth row. The hypo- The species is named after Sue and Don Scott- cone appearsto have been very poo_rlydeveloped. Orr, in recognitionof their long-term supportfor The anterobuccalcomer of M' has a mucb the Riversleigh Palaeontological Research Pro- squarerappea.rance due to greaterparastylar de- ject. Witb the lnvaracks, they are also founding velopmentin the anterior cingulum. This is true membersand Councillors of the Riversleisb So- also of the posterobuccal corner of the tooth ciety. where metastylar developmentoccurs rnid-way along the postmetacrista. TYPE l-ocAl-rrY AND AcE Pamstylar development in M'is relatively The type locality, Fig Tree Site, is adjacentto marked compared with its condition in M Godthelp Hill, Riversleigh Station,northwestem lau^alackoruti.'I"be teeth are buccally fractured in It is laterally adjacentto units rc- M" making it impossible to detemine the de- Queensland. 5. gardedby Arcber et al. (1989, 1991)to be part of greeof deve-lopmentof the buccal cuspsin Ma System B although it may actually be from the basal part of System B or even upper part of DISCUSSION SystemA, hencestratigraphically below Henk's Hollow (upperpart ofSystem C; Archer, Hand & The threediprotodontids described here appear Godthelp, 1986) and Gag Site (lower part of to compdse a new clade of zygomatudnes.As SystemC;Flannery & Archer, 1984;Hand, 1985) speciesof the new genusNitnbidont , they may and possibly above or equivalent to Microsite be distinguishedfrom other zygomaturinesby, (?SystemA; Sig6, Hand & Archer, 1982;Archer among otber features, a combination of upper et al., 1989, 1991)and SiteD (SystemA; Tedford, premolar attributes. Premolar morphology was 1967;Archer et al., 1989,l99l ). On this basisand extensivelyused by Stifton (1967) and Stirton et stageof evolution compadsonsof the marsupials al. (1967) to help resolye distinctions between in tbese faunas, the Fig Tree Local Fauna is diprotodontidlineages and has subsequently been interpretedto be ?late Oligocene-earlyMiocene used by most workers. Few other charactersys- ln age. tems have been found to be as useful ir distin- guishing probable inter-relationships among SPEcrEsDrAGNosrs Tertiary diprotodontids. This speciesis distinguished fuom Nimbadon SinceStirton et al.'s (1967)review of Tertiary laverackorwn and N. whitelawiby itslaryer size, diprotodontids,diverse diprotodontoid materials anteroposteriorlycompressed P (which is much have been collected from Oliso-Miocene fossil shorter than any upper molar), its more robust sitesat Riversleigh,Bullock Cieek, Alcoota and cingulaonall cheekleethand its greaterparastylar Beaumaris and from Pleistocenesites in New developmentin M'-' (contributingto the squared Guinea (e.g. Flannery, 1992) and many areasof appearanceofthese teeth).It also hasless convex Australia.This material includes specimensrep- anterior and posterior molar crown margins resentingnew generaand probably new diproto- wbich thereforehave much wider interdentalcon- dontoid subfarnilies.Munay (1986, 1990a,b)has tact with each other. namedseveral new Tertiary taxa but the bulk of It is further distinguishedfrom most (but not the material, particularly from Riversleigh sites, all) N. lavarackorun by its rnuch smaller hypo- is yet to be described.As a result of the new cone on P'. discoveries,the superfamilyDiprotodontoidea is rn neeoor major revrslon. DEscRIFftoN At present(e.g. the reviews of Archer, 1984; Nimbadon scottorrorum differs from M Aplin & Archer, 1987;Marshall, Case& Wood- lavarackorun and N. whitelawi as follows. bume, 1989;Murray, 1990b)two diprotodontoid The P" is markedly anteroposteriorly com- families are recognised:the Palorchestidae;and pressedsuch that it is almost as wide as it is long the Diprotodontidae with two subfamilies, the I llre generic narneNimbadon was used by Munay (1990b) in a cladogram of hypothetical phylogenetic inter-rclationshipsamong Miocene diprotodontids.However Murray did not designatea srlF-ciesof Ninba.lon in the cladogram or in the ensuing discussion,and, as outlined in Article 13B of the Code of Z-oological Nomeneclature,'Nirrbadon' was thus usedas a nomemnudum- NEW DIPROTODONTIDSFROM NORT}iERN AUSTRALTA 205

F"IC.5. Nlnba!.onscottorrorum n.sp., Flg^free Site, Riversleigh Station. QMF23l57, dght maxillaryfragment withRP1 M" . A A'. occlusalstereoDair: B. buccal view. Bar indicates IOmm.

Diprotodontinae and the Zygomatudnae. Al- siderable variability in cranial morphology has though zygomaturines have traditionally (e.g. been observedin some zygomaturines(e.9. in a Stirton et al., 1967)been considered to be among Neohelos samplefrom the Bullock Crcek I-ocal the most pdmitive of diprotodontoids,it is by no Fauna; P. Murray, pers.comm.),and the molar meanscertain that this is so (Murray, 1990b).It dentitionsof the zygomaturinesPlaisiodon cen- is not clear what might be usedas an appropriate rraliJ Woodburne, 1967 and Alhrertatheriutn outgroup in a phylogeneticanalysis of zygoma- webbi Munay, 1990b and the diprotodontine turines nor, consequently,are polaritiesof chal- Pyranios alcootense Woodburne, 1967 have acter state morphoclines within the been found to overlap in size and morphology Zygomaturinaeconfidentlydetermined. (Munay,l990b). Munay (1990b)hasdiscussed other difficulties It is perhapsa measureofthe difficulties en- in analysingdiprotodontid phylogenetics. He counleredin determiningspecies and- generic notes tliat despiie the fact tiai miny Teniary boundariesfor diprotodontidsthat NimbQdonis diprotodontids are representedby almost-com- oneof few non-monotypicTertiary zygomaturine plete skull and dentary material, discontinuities genera, the others being Kolopsis Woodbume, in some charactercomplexes and tbe continu- 1967 and ZygotnaturusOwen, 1859. ously varying nature of otherstend to obfuscate Recognition of the genusNitnbedon has been clarification of relationshipsbetween taxa. Con- facilitated by re-diagnosisof Neohelos Stirton, m6 MEMOIRSOF THE QUEENSLANDMUSEUM

1967based on new material obtainedby Archer is more closely relatedto Alkwertatherium webbi andcolleagues 20 yearsafter the original material than to any other zygomaturine. was collectedby Stirton and colleaguesfrom the A phylogenetichypothesis of diprotodontidin- lraf Locality of l-ake Ngapakaldi,South Austra- ter-relationships,including the three new Nim- lia. This orovides novel information about the badon speciesand broadly based on Munay's anterior morphology of F, which exhibits, cladisticanalysis of denlalcharacters. is given in among other distinctive featurcs, a large, ercct Fig. 6. The polarity of somecharacters has been parastyle.This fearure is also prcbentin speci- interyreteddifferently from Munay (1990b) but mens refened to Neohelos from the Bullock basic intergeneric relationships remarn un- Creek and Riversleigh Tertiary faunal assem- changed except for the position of Plaisiodon blages(Plane & Gatehouse,1968; Plane, 1971; centralts. Rich et al., 1982; Munay & Megirian, 1990, ln Fig. 6 and Munay (1990b),the Zygom tlt',- 1992;Archer et al., 1989,l99l ). naearc clusteredon the basisof a large parastyle ln his review of Oligo-Miocene diprotodontid on F which is separatedfrom the parametacone taxa,Munay (1990b)postulates a numb€rofphy- by a deep cleft. The Kolopsoides-Plaisiodon- logenetic hypotheses,only one of which is ex- N i mb adon- N e o he I os - Ko Iop sis -Zy go tnat ur u s pressedin his cladogram (hg. l4). The latter is clade sharesas g synapomorphydevelopment of basedprimarily on analysisof dental characters a hypoconein P'. The basicproportions ofP are (in particular, P), using speciesof the Oligo- apomorphically sharedby species of the Ni,r- Miocene palorchestid genera Ngapakaldia and badon,Plaisiodotr, Neohelos and Kolopsis clade Pitikantia as outgroups.ln it, Murray identifies Out seeMurray, 1990b),with a division in the speciesof KolopJiJ a\d Zygotnaturusas the mosl Dammetaconeintemrcted to havg occurred sub- derived diprotodontids and Neohcft)stirare si.\ iequently in thc K;topsis lineage.Retraction of as their closest relative. Species of Nimbadon themesostyl€ towards the cingulum in P clusters folm the sister-grouplo the Neohelos-Kolopsis- spccies of Naohelos anci Kolopsis (Murray, Zygotrraturus clade, with Plaisiodon centalis 1990b). being the sister-groupof a Nitnbadon-N eo lte lo s- ln both cladograms,the postcriorly inclined (or KolopsisZygomaturas clade, Kolopsoides cul - hooked)paraslyle of P'in specicsot Nitnbadon tridens Plane, 1967 the sister-group of a is regardcd to be apomorphic. Only one other P lais i o don-N im b adon-N eo heI os-Ko I opsi s- zygomaturinecxhibits this fcaturc- P/cisiodon Zygomaturus clade and Alwertatheriun webbi cerrtral/.r. AJthough species of Nintbado,l are the most plesiomorphic of zygomaturines.Spe- readily distinguishedfrom P. cerlralis by aspccts ciesof Pyrainios,Euryzygotna, Benatheriwn and of upper premolar morphology (including in P. Diprotodon form a separate (diprotodontine) centralis lhe vcry large hypoconcand the lingual clade, Raemeotheriumyatkolai Rich, Archer & crcst from thc paramclaconelinking to the proto- Tedford, 1978 from the ?lateOligocene Narnba conerather than thc anterolingualcingulum) as Formation oi Lakc Pinpa, South Australia, is well as by their smallsizc, lack of a posteriorly interpretedto be the most plesiomorphicmember increasingmolar gradient and absence ol melalo- of the family Diprotodontidae. phids on lower molars,sharing of the distinc- On the basisof his broaderstudy of dental and tively shapedparastyle may indicatean albeit cranial characters,however, Munay (ibid,) con- distantphylogenetic relatioDship. cludesthat thereale probablytwo minor zygoma- Although the threc taxa describedhere as spe- turine lineages: one representedby species of ciesof Nimbadortmay, when bettcrknown, plove Nhnbedon, Neohelos and Kolopsis, a group he not to bc monophylctic.all exhibitdc dislinclive suspectsmay consist of taxa related largely by creston P'running lingurlly from the paramcta. symplesiomorphies:the other lineage possibly conc to the protoconc then anteriorly to the an- containing speciesof A1l'werrctheriutn, Plaisio- tcrolingual cingulum, a fcature not previously don and Kolopsoides,although for this cladc he notcd in any other diprotodontoidtaxa. Other can find even lessconcrete evidence. He suggests fealureslhat charactcriscthe new taxa include that thesetwo lineagesmight be relatedthrough theirsmall size and an Mz:M4length ratio which common ancestryin Nirabadon.He also suggests approachcs1.0. Thc Iatterare interpreted herc and in the text, but not the cladogram,thatNinbadon generally (see below) to be plesiomorphic fea- speciescould representbasal zygomaturines with tures among diprotodontids,but they might somesoecific affinity to Plaisiodoncentralis ^nd equally be interpretcd to reprcsentapomorphic Neoheiostiraren s is,-andthat Plaisioclon centralis reversalswithin cc ain zygomaturinelineagcs. NEW DIPROTODONTIDSFROM NORTHERN AUSTRALIA 201

Ngopakaldia & Pitikantio sPP.

Rae oreo Iher i um lat koIa i Diprotodonlines

Alkwertatherium webbi

K o Io p so ide s cuI t r ide ns Plaisiodon centralis

..-...... _ Nimhadon -- s....._ scolrcrrorum 7---) I fto- N lavarackorum L-9 I L11- N ,*,hitelah,i

Ncohclosttarensis ''-L-,.-- Kolopsrrspp.

FIG.IG. 6. A phylogenetichypothesis of zygomrturines, broadly bascdon Murray's (1990b) cladistic analysisof dental charactersin lhe Diprotodontidae.Apomorphies are as follows: l, loss,incorporation or supprcssionof slylar cuspsC ilmdD with respectlo the lophson upper molars; 2, reductionof paralophidcrest on I&; 3, largc parastyleon P' separatedqom parumetaconeby a degpclcli; .1.devek'pmcnt o[ ir hypoconegn P'l 5. brsic propgrlionalsimilarity ofP-;6. horrkedpareslylc on P"; 7, rlicrl bl c irnddiagq|l;rl cresl on P'l 8. shorlening of P"; 9, elongationoflmsteriorof F)sterior molars (i,e(i.e- irt lc{stlclsr-Ml.-4)i M3,4); I9.10, hypertrophyof P' hynoconq; I l,I , elongarionelongirtionof anledormolars, i.e. M2: 12,mesostyle on P'lr retractedIowards lypelmnhl.ofcingulunr; 13, division l'.hylocong;in P' parrmctecone.

We can, however,find no pressingevidence to yalliol.ii but is tcntatively considercdby Tedford closely afign individual species of Nirnbaclon et al. (1975) to be a palorchestid.Apart from its with any previouslyknown zygomaturine taxa. slightly largcr size,Nitnbodott I avarac korwn, the Most zygomaturincs- including specicsof only speciesof the genus so far representedby Zygonalurus, Kolopsis, Kolopsoides and lowcr molars,like all other zygomaturinesdiffers Plaisiodon - ex\ibit a markedincrease in molar from R. yatkolai in its rcduction of antedor en- sizefrom M2 to M4 (with an M2-Malength ratio tocristids.Nirlrd./on sDcciesfufiher differ fron'l of 0.85or lesstwith M5 Aenemll)dccreasing in Rutttcotherirrnluttolni in beinglargcr; in hav- size (see also Rich, Archer & Tedford, 1978). ing a moremassive dentary which, in transverse Whether this is also true of topotypicalNeorclos width,is approximatclytwicc the width of M:; in tiraret$isis not known but M,dlclos sDecimcn\ havingless wcll developcdcristids obliqua;rela- from Bullock Crcck and most from Riversleish tivcly wider trigonidson all molars;tfigonid ol' exhibilthis feature. M: nearlythc samewidth as the talonid(in con- A molar gradientof M2:M4 lcngthapproaching lrast to R. )ntkolai wherc it is mrrkedly nar- 1.0is regardedby Rich, Archer & Tedford ( 1978) rowcr);metaconid ofMz ashigh as thc protoconid to be plcsiomor?hic in zygomaturines. Rde- andassociatcd with a prominent,steeply inclined, ncother iutn \atko lai. commoIllv considered1o be swollenanterior buttress; metacristid of M? tmns- the most primitive of known zygomaturines(but vcrsely orienlcd (in contrastto R. I?/tolci wherc seeMunay, 1990b, and below) and rcpresented it is posterobuccallyoricnted); "arcuate" (rathcr by a dentary, isolatcd lower teeth and an uppcr lhan anleroposteriorlyrcctilinear) paracristidon incisor from the Oligo-Miocene Namba Forma- M:; and lophidsof lowcr nolars lessocclusally tion of Lake Pinpa,also hasuniformly sized mo- concave. lars (with a ratio of Mz:M.r lcngth of 0.96). It is possiblcthat molar sizc gradientsin zy- Palo.chestidsand diprotodontincsvariably ex- gomaturinesare allometricallyrclatcd to body- hibit this featurcand one of thc oldestknown sizc, with largcr animalsexhibiting relativcly diprotodontoids, reprcscntcdin the early Mio largerpostcrior nrolars. Frorn Oligo-Miocene de- cene Geilston Bay ^t cal Fauna by a maxilla positson RivcrsleighStation, small diprotodon- preservingpart of M'and M', alsohas uniformly tincs have bcen found with M2:M4 leneth sizedmolars. The latter is similar in sizc to R. gradicntsapproaching 1.0. Equally. ho$ cver.ihc 208 MEMOIRSOF THE QUEENSLANDMUSEUM

larger of the Riversleigh speciesof Nimbadon all membersof other lineageswith ubiquitously describedhere (N. scottorrorum, aDqearsto have small members(e.g. burramyids and acrobatids). an beensimilar in overallsize lo al li;st one unde- Becauseso few dental charactersystems have scribed Riversleigh Neohelos ihat shows a been found to be useful in diprotodontid sys- marked progressiveincrease in molar size from tematics,polarity swings of this kind greatly in- th M2 to M4. It seems,therefore, that the featureis fluence the interDretedrelationshios of taxa and N not always dependenton absolutesize although their biostratigrabhicsignifi cance. If elongation la it may vary allometrically within lineages.ln Fig. is plesiomorphicamong diprotodontoids, then M a 6 it has been interpreted to b€ autapomorphic whitelawi wo!ld,be the mostplesiomo{phic ofthe S within various clades. threeNimbadon species.Its presencein the mid- r Intragenericrelationships within ly'irnDcdon dle Miocene Bullock Creek l-ocal Faunaand the arc 2 not much easierto intercrcLNimbodon prescnceof what would then be interpretedto be whitelawi t differs fiom both N. lavarackorun the most apomorphicin the older Fig Tree lrcal and N. scot- l torrorutn in its markedly more elongate upper Fauna, would not be support for the currently molars (particularly with respect Io N. scotbr- understoodstratigraphic relationships. If, on the rorum) and relativelymore anteriorlysitualcd other hand, elongation is interpretedto bc apo- buccaf swellins on P. Nitnbadon lavarackorutn morphic, then the most derived species (N. differs fiom Nl scotorrorwn in its s4aller size, whitelawi) occurs in one of the younger faunas less well-developed parastyleon Mr and prob- and the most plesiomorphic(N. :tcotorrorum) in ably more elongate posterior upper molars and the oldest of the faunas(Fig Trce). lesswell developedpostmetacrista but more dis- The stratigraphicrclationships of the fossil as, crcte metastyfeon M'. Ninbadon sco orrorunt semblagcshave becn discussedin the Syslemat- is larger than tnth N. lavarackorum and N. ics section above. The Fis Tree Site occurs whitelawi, has more robust cingula on all cheel: adjacent to units regarded to be part of teelhand greaterDarastvlar develooment in M" fuversleigh's System B sequenceas defined by (contributing to the squlred appeaianceof rhcsc Archeret al. (1989,l99l ), thoughit mayactually teeth). Within N nbadon, features such as the be basalSystem B or upper SystemA. and is hypenrophyof the Pr hypoconein somespeci- interprelcdlo be latcOligocene to carly Mioccne mens of N. laverackorutn and the shorteningof in age.The Henk's Hollow Site occurs near the Pr in N. scottorrorumappcar to beautapomorfhic top of Riversleigh's SystemC sequenceand the Iealuresunique to thosespecies. The moslstrik- Gag Site nearits base.Both lie stratigraphically ing differencesbetween the taxa lie in the degree above thc Fig Trcc level. On the basis of their of squaring(or converselyelongation) of the up- positions and stageof evolution comparisonsof Per morars. their fossil mammals,they are interyretedto be middle Miocene in age and are probably similar There are two quite different phylogeneticin- in age to the Bullock Creek assemblage.The terpretationsof Ni,nbadon species,depending on Bullock Creek Local Faunahas been intemreted whetherelongation {versus squaring) of lheupper by Woodburneet al. ( 1985). Murrry ( 1990i)and mola.rsis consideredto be plesiomorphicor apo- Murray& Megirian(1990) tob€ middleMiocene, morphic. Commonality would indicate squared i.e.immcdialely post-Wipajiri. in age. uppermolars to b€ plesiomorphicamong dipro- todonlianmarsupials including phalangeridans, On balance,wc considerthe phylogenetic rcla- tionships of Nitnbadon species (Fig. vombatimorphiansand even plesiomorphickan- 6) to ap- proximatelyconespond to current garoos.However, within the Diprotodontidaethe understanding 'plesiomorphic' ofthc relativeages ofthe deposits otherwise (i.e. simple) Rae- fromwhich the specimenswere obtained meotherium yatkolai has relatively elongalc but concludethat for the momentat leastcaution is appropriatein attempts lower molarsand, althoughno uppersmolars are to use zygomaturinesto correlatenorthern yet known, they too would probably have been Aus- tralian Tcniary relatively elongate,using as a guide the relative mammal-bearing faunas. proponions of lhe upper and lower molars of N. lavarackorum. It should be noted that the Dre- ACKNOWLEDGEMENTS sumptionlhat R. yatkolaiisthe mostplesiomor- phic diprotodontidis basedon its simplicity. Grants(o P.V. andT.H. Rich from the Austra- Simplicity, however, often characlerises the lian ResearchCouncil and the National Geo- smaller members of some diDrotodontianline- graphic Society madepossible the collection and ages(e.g. pseudocheiridsand macropodids)and preparationof Nimbadon whitelawi. Ttre South \TA DCROTOOO]VTIDSFROM NORTHERNAUSTRALIA 209

Ausu-alian}ltEan ar$lcd N. Pledseto collect Mineral Resources, Geology and Geophysics and prepar rhc holcrlpc of li. sittorrorurn. Australia, sheetSB52-8. Work ar R.rrrrslagh has been supponed by the FLANNERY, T.F., 1992.New Pleistocenemarsupials Australian Rcs.acl| Conncil, the Departmentof (Macropodidae, Diprotodontidae) from subal- the Ans. Spofl. lhc Envimnment and Terrilories, pine habitatsin Irian Jaya,Indonesia.Alcheringa National EJrarc Programme Grants (Queens- 16:321-331. land).the AusrralianGeographic Society,ICl, the FLANNERY. T.F. & ARCHER. M. 1984.Revision of QueenslardMuseum and the University of New lhe extinc! gigantic rat kangaroos (Potoroidae: South Wales. During the courseof tbis work, S. Marsupialia),with descriptionofa new Miocene Hand and H. Godthelp were supportedby a Uni- genusand species. and a newPleistocene species versity of New South specialr€search grant. The of Propleopus. Journal of Palaeontology 89: authorsthank Dr PeterMurray for his construc- I l3r-l149. tive criticism of the manuscriot.and Dr David Ride who providedhelpful nomenclaturaladvice. 1987a. Strigocuscusreidi andTrichosurusdicksoni, new phalangerids(Marsupialia: The photos were taken by Ross Arnett of the two fbssil Pha- University of New South Wales. langeridae) from the Miocene of nonhwestern Queensland.h. 527-536. In Archer (1987). 1987b.Bettongia noyesi, a new and plesiomorphic LITERATURE CITED kangaroo (Marsupialia: Potoroidae) from Mio- cenesediments of northwesternQueensland. Pp. APLIN, K. & ARCHER,M. 1987.Recent advances in 759-767.In Archer(1987). marsupialsyslematics with a newsyncretic clas- HAND, S.J. 1985.New Miocene megadermatids(Chi- sification.Pp. xvlxxii. In Archer(198?). roptera: Megadermatidae)from Australia with ARCHER,M. 1978.The nature ofthe molar-prcnrolar commentson megadermatidphylogenetics. Aus- boundaryin marsupialsand a re-interpretationof ralian Mammalogy8: 5-43. thehomology of marsupialcheekteeth. Memoirs 1993. First of the Museuml8: 157-164. skull of a species of Hipposideros Queensland ( Brachippos i.lercs (M icrochiroptera : Hip- 1984.The Australian marsupial radiation. Pp.633- ) posideridae), from Australian Miocene sedi 808.In Archer,M. & Clayton,C. (eds),'Verte ments.Memoirs of the Museum 33: brate zoogeographyand evolution in Queensland (Hesperian t'l9-r92. Aus!-alasia'. Press:Penh). MARSHALL,L.G., CASE, J.A. & WOODBURNE, 1987.'Possums and opossums:studies in evolu- M.O. 1989. Phylogenetic relationships of the lion'. (SurreyBearry and Sonsand the Royal families of marsupials.Current Mammalogy 2: ZoologicalSociety of New SouthWales: Syd- 433-502. reY)' ARCHER.Nt. & BARTHOLOMAI,A. 1978.Teniary MURRAY, P. 1990a. Primitive marsupial tapirs manmalsof Australia:a synoplicreview. AI- (Propal o rc heste s novac u Iacephal us Munay and cheringa::l- 19. P. ponticulus sp. nor.) from the mid-Miocene of ARCHER.TT.. GODTHELP, H., HAND, S,J.& norlh Australis (Marsupialia: Palorchestidae). MEGIRIAN.D. 1989.Fossil mamnals of The Beagle, Records of the Nonhern Tenitory Museum of Ans and Sciences7: 39-51. Riversleigh,Donhwestem Queensland: prelimi naryor en ieu of biostratigraphy,correlation and l990b. Alkwertatheriwn webbi, a new zygomatur- environmentalchange. The Austaalian Zoologist ine genus and species from the late Miocene 25: 35-69- Alcoota lrcal Fauna,Northen Territory (Mar- ARCHER,M. & HAND.S.J. 1984. Background to the supialia:Diprotodontidae). The Beagle,Re.ords s€archfor.{uslinlia s oldeslmammals. Pp. 517- of the Northem Ter.itory Museum of Ans and 565.In A-rcher.I\,|. & Clayton,G. (eds),'Vene- Sciences7:53-80. brate zoogeographyand evolutionin MURRAY. P. & MECIRIAN. D. 1990.Furtherobser' Australasia'.(Hesperian Press: Perth). vations oD the norphology of Wakaleo vander- ARCHER,M.. H.A,}.'D.S,J, & GODTHELP,H, 1986. /ert (Marsupialia: Thylacoleonidae) from the 'UncoveringAustralia's dreamtime'. (Surrey mid-Miocene Camfield Beds, Norlhem Teni Be.tty andSons: Sydney) tory. The Beagle,Records ofthe NorthernTeri- ARCHER,M.. HAND,S.J. & GODTHELP,H, I99I. tory Museunrof Ans and Sciences7: 9l-102 'Riversleigh . {ReedBooks: Sydney). 1992.Continuity and conlrastin the middle andlate BULTITUDE.R.J. 1973.Wave Hill, NonhernTerri- Miocene verlebratecommunities ftom the North- tory - l:150.000Geological Series. Bureau of ern Territory. The Beagle, Recordsofthe North- 210 MEMOIRSOF THE QUEENSLANDMUSEUM

ern Tenitory Museum of Arts and Sciences SIG6, B., HAND, S.J. & ARCHER, M. 1982, AN 9:195-218. Australian Miocene Brachipposideros (Mam- OWEN,R. 1859.On someoutline-drawings and pho- malio. Chiroptera)related to Miocene repre- tographsof the skull of theZyBomaturus trilobus senlatives from France. Palaeovertebrata 12: Macl-eay(Nototherium Owen?). Proceedings of 149-t72. theGeological Society of london l5: 168-176. STIRTON, R.A. 1967.A diprolodontid from the Mio- PLANE,M. 1967.Two new diprotodontidsfrom the ceneKuljamarpu fauna,Soulh Australia.Bulletin PlioceneOtibanda Formation, New Guinea. Bul- oflhe BureauofMineral Resources,Geology and letinof the Bureauof MineralResources, Geol- Geophysics,Ausralia 85: 45-51, ogy andGeophysics, Australia 85: 105-128. 1971.The genusNeoielos (Marsupialia:Diprolo- STIRTON, R.A,, WOODBURNE, M.O. & PLANE, phylogeny dontidae).ANZAAS, Zoology,43rd Congress, M.D. 1967. A of the Teniary Dipro- Sydney.Abstract. todontidae ond its significance in correlation. PLANE. M. & CATEHOUSE.C.G. 1968.A new Bulletin of the Bureau of Mineml Resources, venebratefauna from the Tertiarvof Northern Geologyand Geophysics, Australia 85: 149-160. Australiu.Auslralian Jounral of ScienceJ0: 272' TEDFORD, R.H. 1967.Fossil mammal remainsfrom 2'73. the Carl Creek Linrestone,northwestem Queens- RICH, T.H., ARCHER, M., HAND, S.J., land. Bulletin of the Bureau of Mineral Re- GODTHELP,H., MUIRHEAD, J., PLEDCE, sources,Geology and Ceophysics,Australia 92: l 21'1,23't. N.S., , FLANNERY, T,F., WOODBURNE 1 M.O.W.,CASE, J.A., TEDFORD, R.H., TURN. TEDFORD, R.H., BANKS, M.R., KEMP, N.R., I BULL, W.D., LUNDELIUS, E.L. JR, RICH, MCDOUCALL, I. & SUTHERLAND, F.L, L.S,V,,WHITELAW, M.J., KEMP, A. & RICH, 1975. Rccogrrition of the oldest known fossil P.V. 1991.Australia's mammal rccord: Austr.- marsupialsfrom Auslralia.Nature 225: 141-142. lian Mesozoicand Tertiaryterreslrial mamnral TEDFORD, R.H. & WOODBURNE, M.O. 1987.The localities.Pp. 1005-1070In Vickers-Rich,P., Illariidae, a new family ofvombatiform marsupi- Monaghan,J.M., Baird, R.F. & Rich,T.H. (eds), als from Miocene st.ata of South Australia and 'Venebratepalaeontology of Australasia'.(Pio- an evalu tion of lhe homology of molar cuspsin neerDesign Studio and Monash University Pub- the Diprolodonlidae.Pp.40l-418. In Archcr licalionsCommitteet Melbourne). ( 1987). zuCH, T,H,, ARCHER,M,, PLANE, M.D,, FI-AN- NERY,T,F., PLEDGE, N,S., HAND, S.J.& WOODBURNE, M.O. 1967.Three new diprodontids RICH, P.V. 1982.Australian Tertiary mammal from lhc Tertiary ofthe Norlhern Teritory, Aus- localities.Pp. 525-72 1n Rich, P.V. & Thompson, lralia. Bullctin of the Bureau of Mineral Re- 'The E.M.(eds), fossilvenebrate record of Au$- sources,Geology and Geophysics,Australia 85: tralasia'.(Monash University Offset Prinring 53-103. Unit: Clayton,Victoda) WOODBURNE,M.O., TEDFORD, R.H., ARCHER, RICH,T,H,, ARCHER, M. & TEDFORD,R.H. I978. M.,TURNBULL, W.D., PLANE, M.D. & LUN- Roemeotherium)akolai gen. et sp. nov.. a primr- DELIUS,E.L. 1985. Biochronology of thecon- tive diprotodontidfrom the medialMiocene of tinentalmamnral record of Australiaand New Soulh Australia.Memoirs of the NationalMu- Guinea.Special Publication, South Australian seumof Victoria39: 85-91. Deparlmentof Minesand Energy 5: 347-363.