New Genus and Species of Extinct Miocene Ringtail Possums

PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY

CENTRAL PARK WEST A T 79TH STREET, NEW YOR K, NY 10024

Number 3560, 15 pp., 3 figures, 1 table March 8, 2007

New Genus and Species of Extinct Miocene RingtailPossums (Marsupialia: Pseudocheiridae)

KAREN K. ROBERTS,1 MICHAEL ARCHER,2 SUZANNE J. HAND,3 AND

HENK GODTHELP4

ABSTRACT

The first unique genus of ringtail possums from the Riversleigh World Heritage Area inAustralia is diagnosed and described. Gawinga aranaea is the type and only species of the newgenus and is known from nine isolated lower molars. It has been recovered from three Riversleighdeposits: two are of early to mid-Miocene age, while the age of the other has yet to be determined.The new possum is larger than Oligo-Miocene species of Paljara, Pildra, and Marlu, but smallerthan most extant taxa. It is characterised by a distinctive lower molar cusp morphology of parallelridges extending primarily from the cristid obliqua, filling the occlusal basins. Additionalautapomorphies include: extended, shelflike protostylid cristids and a bisected posthypocristid onm1, and posterior molars that have a metaconid ridge posterobuccal to the metaconid and ananteriorly positioned protoconid relative to the metaconid. It also possesses an enlargedprotostylid on m1, a feature otherwise known only in extant genera. The precise phylogeneticposition of Gawinga within the pseudocheirid radiation has yet to be determined, but it istentatively identified here to be a highly derived pseudocheirid, apomorphic with respect to Paljaraand Pildra species, and with no known descendants.

Copyright E American Museum of Natural History 2007 ISSN 0003-0082

1 School of Biological, Earth and Environmental Science, University of New South Wales, New South Wales 2052,Sydney, Australia ([email protected]).

2 School of Biological, Earth and Environmental Science, University of New South Wales, New South Wales 2052,Sydney, Australia; Research Associate, Division of Vertebrate Zoology (Mammalogy), American Museum of NaturalHistory ([email protected]).

3 School of Biological, Earth and Environmental Science, University of New South Wales, New South Wales 2052,Sydney, Australia; Research Associate, Division of Vertebrate Zoology (Mammalogy), American Museum of NaturalHistory ([email protected]).

4 School of Biological, Earth and Environmental Science, University of New South Wales, New South Wales 2052,Sydney, Australia ([email protected]).

INTRODUCTION

The ringtail possums and greater gliderfamily (Pseudocheiridae) comprise a group ofreclusive, nocturnal, arboreal marsupial foli-vores. They form one of the most diversegroups within the broad radiation of diproto-dontians (possums, kangaroos, koalas, andwombats). The breadth of diversity withinthe group was formally recognized in 1987with its elevation from subfamilial to familialstatus (Aplin and Archer, 1987). Thisfollowed increasing evidence from molecularstudies that pseudocheirids form a muchmore distinct lineage than previously thought(e.g., Kirsch, 1977; Archer, 1984; Baverstock,1984).

Extant pseudocheirids belong to six generathroughout Australia and the New Guinearegion (Flannery, 1994; Strahan, 1995). Thesegenera had largely been classified as subgeneraof Pseudocheirus (e.g., Tate, 1945). Mostspecies are limited in distribution, inhabitinghigh-altitude tropical rainforest in northeast-ern Queensland or New Guinea (Flannery,1994; Strahan, 1995).

In addition to extinct species of living genera,fossil taxa have been referred to four extinctgenera (Bassarova and Archer, 1999; Crosby etal., 2004). The oldest pseudocheirid fossils, ofthe extinct genera Paljara, Pildra, and Marlu,have been recovered from late Oligocene toearly Miocene sites in the Tirari Desert andFrome Basin in northern South Australia(Woodburne et al., 1987; Bassarova andArcher, 1999). Pseudokoala is the fourth extinctgenus, known from Plio-Pleistocene deposits inSouth Australia, Victoria, and southeastQueensland (Turnbull and Lundelius, 1970;Archer et al., 1997a; Crosby et al., 2004).

Paljara, Pildra, and Marlu have also beenrecorded from the Oligo-Miocene freshwaterlimestone deposits of the Riversleigh WorldHeritage Area in northwestern Queensland(Archer et al., 1989; Archer, 1992; Bassarovaand Archer, 1999). However, the abundantpseudocheirid material from these mid-Tertiary sites at Riversleigh is largely unde-scribed and unpublished, with the exceptionsof Paljara nancyhawardae and P. maxbourkei(Bassarova et al., 2001). Nevertheless, pre-liminary studies reveal a great diversity oftaxa, largely unrecognized beyond Riversleigh

(Archer, 1984, 1992; Archer et al., 1989).These include new species of known Oligo-Miocene genera, the earliest representatives ofextant Pseudochirops and the new genus andspecies described in this paper.

METHODS AND MATERIALS

All described fossil material belongs to theQueensland Museum palaeontology collec-tion, prefix QMF. Lower dental terminologyfollows Archer (1984; see fig. 1) and homologyof molars and posterior premolar followsLuckett (1993). Higher-level systematic termi-nology follows Aplin and Archer (1987).Riversleigh site and system nomenclaturefollows Archer et al. (1989, 1995, 1997b).

Measurements were made using a WildMMS235 Digital Length Measuring Set,attached to a Wild 3MB stereomicroscope.Length (L) represents maximum longitudinaldistance; anterior width (AW), and posteriorwidth (PW) respectively represent width of thetrigonid and talonid, perpendicular to thelengthwise axis. Scanning electron micro-graphs were taken on an FEI Quanta 200ESEM at the Electron Microscope Unit(EMU), University of New South Wales.

SYSTEMATIC PALEONTOLOGY

SUPERORDER MARSUPIALIA ILLIGER, 1811

ORDER DIPROTODONTIA OWEN, 1866

SUBORDER PHALANGERIDA APLIN ANDARCHER, 1987

FAMILY PSEUDOCHEIRIDAE WINGE, 1893

Gawinga, new genusTYPE SPECIES: Gawinga aranaea n.sp.DIAGNOSIS: As for the type species until

additional species are known.ETYMOLOGY: Gawinga is a word meaning

‘‘possum’’, as spoken by the late Ivy George ofRiversleigh Station, northwestern Queensland.Gender is considered to be feminine.

DISTRIBUTION AND AGE: As for the typespecies.

Gawinga aranaea, new speciesHOLOTYPE: QMF52173, isolated right m1

(figs. 2A, 3A).

2 AMERICAN MUSEUM NOVITATES NO. 3560

PARATYPE: QMF52174, isolated left m1(fig. 3B).

TYPE LOCALITY: Arachnea Ridge Site,Riversleigh World Heritage area, Lawn HillNational Park, northwestern Queensland,Australia.

REFERRED SPECIMENS: QMF52175 (fig.3C) right m2; QMF52176 (fig. 2D, 3I) leftm4 from Arachnea Ridge Site. QMF52177(fig. 2B, 3D) right m2; QMF52178 (fig. 3E)left m2; QMF52179 (fig. 3F) right m2;QMF52180 (fig. 2C, 3G) right m3 fromUpper Site. QMF52181 (fig. 3H) right m3from Wayne’s Wok Site.

ETYMOLOGY: From the Latin aranea mean-ing ‘‘spider’’ or ‘‘spider’s web’’. In reference tothe Riversleigh site, Arachnea Ridge.

DISTRIBUTION AND AGE: Riversleigh WorldHeritage Area, northwestern Queensland.Upper Site and Wayne’s Wok Site are inter-preted to be part of Riversleigh’s System Bstrata (Archer et al., 1995, 1997b; Creaser,1997). Biocorrelation of the Riversleigh fossilassemblages, and of Upper Site in particular,suggests an early Miocene age for System Bsites. No age has been determined for

Arachnea Ridge Site, but an early Mioceneage is also possible given the presence of thisnew species.

DIAGNOSIS: Gawinga aranaea is character-ized by a distinct, lower molar cusp morphol-ogy as follows: a reticulated occlusal surface ofrobust ridges, predominantly normal andlingual to the cristid obliqua; a hypoconid basethat is enlarged posterobuccally; an anteriorlyplaced protoconid with respect to the metaco-nid on posterior molars; and a bisected post-hypocristid on m1. The lower molars of G.aranaea are narrower and longer than in mostextinct pseudocheirids, with the exception ofMarlu kutjamarpensis Woodburne et al., 1987,Pildra magnus Pledge, 1987, and all species ofPseudokoala. It differs from known species ofthe other extinct genera by having a shorter,weaker, and more posteriorly oriented m1buccal protocristid; by having a large cuspateprotostylid; and a larger entostylid and ridge(entostylid completely absent in species ofMarlu and Pseudokoala). Posterior molars alsoexhibit a prominent metaconid ridge and anelongate preentocristid that is transverselybuttressed at its anterior end.

Fig. 1. Right m1 diagram of Gawinga aranaea. Abbreviations as follows: abcm anterobuccal cingulum;bprcd buccal protocristid; bprsdr buccal protostylid rib; cora cristid obliqua ridge A; corb cristid obliquaridge B; corc cristid obliqua ridge C; end entoconid; ensdr entostylid ridge; hyd hypoconid; hyld hypoconulid;med metaconid; pacd paracristid; pad paraconid; pcdprsd precristid of the protostylid; pencd preentocristid;pmcd premetacristid; pocdprsd postcristid of the protostylid; poencd postentocristid; pohycd posthypocristid;pomcd postmetacristid; poprcd postprotocristid; prd protoconid; prsd protostylid. Scale bar is 1 mm.

2007 ROBERTS ET AL.: MIOCENE POSSUMS 3

G. aranaea differs further from Paljara spp.in the following features: m1 anterobuccalcingulum reduced; m1 paraconid smaller;protoconid on posterior molars buccally

enlarged; metastylid progressively more com-plex along tooth row; narrower notch betweenpreentocristid and metastylid; entostylid ridgeforms a continuous crest with hypoconulid;larger hypoconulid on posterior molars; morelengthwise orientation of lingual cusps, par-ticularly entoconid; wider buccal cusp angles;buccal cusp apices less lingually curved;shallower trigonid and talonid basins.Features shared with species of Paljara in-clude: m2 longer than m1; reduced or absentanterior margins between paraconid andparacristid; paraconid on posterior molarspositioned on lengthwise axis; elongate, blad-ed metacristids; equidistant positioning of theentostylid ridge between posthypocristid andentoconid; and lack of buccal protostylid orcingulum on posterior molars.

G. aranaea differs further from Pildra spp.by the following features: stronger and broad-er anterior cingula (except P. magnus); ante-rior margins on posterior molars reduced orabsent; m1 paraconid extended further ante-riorly; paraconid on posterior molars posi-tioned much closer to longitudinal axis; m1metaconid relatively lower cusped; widerdistance between apices of m1 metaconidand protoconid; narrower notch betweenmetastylid and preentocristid; preentocristidless steep and more bladed; entoconid lessconical; entostylid ridge forms a continuouscrest with the hypoconulid which is separatefrom the postentocristid; larger hypoconulid.Features shared with species of Pildra include:m2 longer than m1; presence of a lingual ribextending from the m1 paraconid; lingualenlargement of the protoconid on posteriormolars; progressively more complex metasty-lids along the tooth row; similar orientation ofthe posthypocristids on m1–2.

G. aranaea differs further from Marlu andPseudokoala spp. by the following features:stronger anterobuccal cingula; larger m1paraconid; m1 protoconid more linguallypositioned; m1 metaconid relatively largerand separated from protoconid by a deepernotch; greater distance between apices of m1metaconid and protoconid; metastylids pro-gressively more complex along the tooth row;larger, more distinct hypoconulid. G. aranaeadiffers further from all species of Marluand Pseudokoala, except M. praecursor

Fig. 2. Stereo occlusal view of m1–4 of Gawingaaranaea. A. QMF52173 (holotype), right m1; B.QMF52177, right m2; C. QMF52180, right m3; D.QMF52176, left m4. Scale bars 5 1 mm.


Fig. 3. Buccal views of Gawinga aranaea. A. QMF52173 (holotype), right m1; B. QMF52174 (paratype),right m1; C. QMF52175, right m2; D. QMF52177 right m2; E. QMF52178, left m2; F. QMF52179, right m2;G. QMF52180, right m3; H. QMF52181 right m3; I. QMF52176, left m4. Scale bars 5 1 mm.


Woodburne et al., 1987, in the absence of anoverlapping or joined metastylid and preen-tocristid. G. aranaea also differs from speciesof Pseudokoala in having relatively conicallingual cusps, with steeper cristids; shorterpostprotocristid on posterior molars; posthy-pocristid extending further lingually andabsence of a buccal protostylid or cingulumon posterior molars (except P. curramulkensis,originally described as Corracheirus curra-mulkensis Pledge, 1992). Features shared withMarlu and Pseudokoala species include: dis-tinct anterolingual cingula; and broad buccalcusp angles. G. aranaea also shares with Marlua longer m2 than m1.

In comparison with extant taxa, G. aranaeais smaller than all known species exceptPseudochirulus mayeri Rothschild andDollman, 1933, and P. canescens Water-house, 1846, and is very similar in size toPetauroides ayamaruensis Aplin et al., 1999. Italso differs from extant taxa (species ofPseudocheirus, Pseudochirulus, Petropseudes,Pseudochirops, Petauroides, and Hemibeli-deus) as follows: m1 shorter than m2 (exceptPetropseudes); m1 precristid of the protostylidjoins anterior end of the anterobuccal cingu-lum; m1 postcristid of the protostylid well-defined and shelflike, terminating at the cristidobliqua; distinct and strong metaconid ridgepresent on posterior molars; metastylid pro-gressively more complex along the tooth row;metastylid and preentocristid do not overlapor join; posthypocristid more obliquely ori-ented.

G. aranaea differs further from species ofPseudochirops and Petropseudes as follows:stronger anterior cingula (anterior cingulaabsent in Pseudochirops cupreus Thomas,1897, and Petropseudes dahli Collett, 1895);anterobuccal cingulum present on posteriormolars; m1 buccal protocristid shorter andweaker; flat lingual face of metastylid; hypo-conulid relatively larger and more distinctfrom postentocristid and posthypocristid;buccal cusp apices less lingually curved; buccalprotostylid or cingulum on posterior molarsabsent. G. aranaea differs further fromPseudochirops spp. as follows: flat lingual faceon m1 protoconid; greater distance betweenapices of m1 metaconid and protoconid(except P. corinnae Thomas, 1897); m1 pre-

metacristid not linear with very weak post-protocristid; anterior protocristid on posteriormolars reduced or absent; larger entostylidand ridge, forming a continuous crest to thehypoconulid apex; entostylid ridge equidistantfrom entoconid and posthypocristid (except P.archeri Collett, 1884, and P. winteri Macknessand Archer, 2001); wider hypoconid cuspangle. G. aranaea differs further fromPetropseudes dahli as follows: m1 metaconidmuch larger and separated from protoconidby a deeper notch; m1 cristid obliqua connectswith metaconid; much shorter postprotocris-tid on posterior molars; entoconid moreconical; entostylid present; less compressedlingual cristids. Features shared withPseudochirops and Petropseudes species in-clude: lingual rib extending from m1 para-conid and a cuspate protostylid on m1. It alsoshares with species of Pseudochirops an m1protostylid basin.

With the exception of Pseudocheirus stirtoniand P. marshalli (Turnbull and Lundelius,1970), which are compared alongside speciesof Petauroides and Hemibelideus, G. aranaeadiffers further from Pseudocheirus spp. asfollows: stronger and broader anterior cingu-la; anterior margins of posterior molars re-duced or absent; paraconid on posteriormolars positioned much closer to longitudinalaxis; greater distance between apices of m1protostylid and protoconid; m1 metaconidrelatively larger and separated from proto-conid by a deeper notch; premetacristid on m1does not align linearly with the very weakpostprotocristid; wider protoconid cusp angleson posterior molars; protoconid on posteriormolars more lingually swollen, forming a slightanterior protocristid in some specimens; muchshorter postprotocristid on posterior molarsthat does not unite with the metastylid;entostylid present; hypoconulid on m1–2 isnot ribbed posterolingually, and is perpendic-ular to hypoconulid cristids; cristid obliquaterminates at same height as postprotocristid;lingual cristids on posterior molars less steep;buccal cusps less lingually curved; shallowertrigonid and talonid basins. Features sharedwith species of Pseudocheirus include: lingualrib extending from m1 paraconid; a cuspateprotostylid on m1; and absence of a buccalprotostylid or cingulum on posterior molars.


G. aranaea differs further from Pseudo-chirulus spp. as follows: anterior margins ofposterior molars reduced or absent; m1 para-conid smaller and less conical; paraconid onposterior molars positioned much closer tolongitudinal axis; m1 metaconid larger andseparated from protoconid by a deeper notch;shorter, weaker m1 buccal protocristid; ante-rior protocristid on posterior molars absent orvery reduced; entostylid and ridge present;hypoconulid on m1–2 is not ribbed or curvedon the posterolingual surface, and cristids areoriented lengthwise; posthypocristid lesscurved at lingual termination; lingual cuspsless conical (except m1 protoconid); buccalcusp cristids higher crested (except anteriorprotocristid); cristids are less steep; shallowertrigonid and talonid basins. Features sharedwith species of Pseudochirulus include: broad,distinct anterior cingula and a cuspate proto-stylid on m1.

G. aranaea differs further from Pseudo-cheirus stirtoni, P. marshalli and Petauroidesand Hemibelideus spp. as follows: anteriormargins of posterior molars reduced orabsent; paraconid on posterior molars posi-tioned much closer to longitudinal axis; m1metaconid is a more distinct, independentcuspid (except P. marshalli and P. ayamar-uensis); shorter, weaker m1 postprotocristid;much shorter postprotocristid on posteriormolars; entoconid buccally compressed; largerentostylid and much stronger, continuousentostylid ridge on all molars (H. lemuroidesCollett, 1884, has no entostylid or ridge);hypoconulid more peaked and distinct andlacks a ribbed or curved posterolingualsurface; wider hypoconid angles; cristid ob-liqua terminates at same height as postproto-cristid (except P. ayamaruensis); buccal cuspapices less curved lingually; cristid obliquaand paracristid higher crested; buccal proto-stylid or cingulum on posterior molars absent;shallower trigonid and talonid basins. Itdiffers further from species of Hemibelideusas follows: buccally enlarged m1 protostylid;m1 paraconid less conical, and independent ofparacristid (P. ayamaruensis also differs in thisfeature). Features shared with P. stirtoni, P.marshalli and species of Petauroides andHemibelideus include: distinct anterolingualcingula; an enlarged cuspate protostylid on

m1; and flat lingual surfaces of metaconid/protoconid and entoconid on m1–2.

DESCRIPTION

All lower molars are relatively narrow andlow cusped with shallow basins. Their occlusalsurfaces appear reticulated due to the presenceof enlarged ridges off the cristid obliqua andin posterior molars, also off the paracristid.Tooth homology of posterior molars wasdetermined by differences in morphology andmorphometrics (table 1).

m1: The holotype QMF52173 and para-type QMF52174 are virtually identical(figs. 3A, 3B) although the latter is slightlydamaged. Lingual margin of the tooth isalmost flat, and paraconid is low cusped andpositioned slightly buccally to the lingualmargin. A lingual rib runs halfway down thecrown from the paraconid, ending in a short,horizontal shelf. Paracristid is weakly arcuateand extends down the anterior flank of theparaconid. Protoconid is subconical (missingin QMF52174), with steep, moderately bladedcristids that both curve lingually, and a steep,very weak, buccal protocristid. Protostylid ispyramidal and located midway down theprotoconid; a flat, sloping lingual surfaceforms the protostylid basin. Protostylid cris-tids are long and form a shelf around thebuccal protoconid face. Precristid of theprotostylid descends concavely, curving backup slightly to form a small peak, a shortdistance before connecting with the anterior of

TABLE 1

Length (L), Anterior Width (AW), and PosteriorWidth (PW) of Gawinga aranaea

Measurements in millimeters.

QMF L AW PW

m1 52173 3.18 1.51 1.7

52174 3.16 1.48 1.72

m2 52177 3.51 1.82 1.81

52178 3.29 1.78 1.84

52179 3.23 1.57 1.63

52175 3.38 1.77 1.73

m3 52180 3.13 1.68 1.72

52181 3.13 1.57 1.67

m4 52176 3.31 1.73 1.69


the anterobuccal cingulum. Postcristid of theprotostylid travels posterolingually toward themetaconid, following the contour of theprotoconid and hypoflexid, and terminates atthe cristid obliqua. Buccal protostylid rib iskinked midway; basal half of the rib isenlarged and crested, almost reaching the baseof the hypoconid. Metaconid is pyramidalwith a vertical, lingual surface, defined bystraight, moderately sloped cristids.Premetacristid joins the posterolingually ori-ented postprotocristid in a notch. The notchcontinues posterobuccally, forming the hypo-flexid, which is traversed by the postcristid ofthe protostylid. Metastylid is virtually absent,comprising a very small cusp posterior to, andindependent of, the postmetacristid; the tri-angular lingual surface of the metastylid isobliquely oriented and vertical, and its apexcontinues anterobuccally, merging with theanterior crest of ridge A.

Entoconid is subconical; preentocristid isanterobuccally oriented and swings trans-versely lingual at the end. Postentocristid isstraight, running posterolingually to the lin-gual margin at a constant slope. Hypoconulidis also on the lingual margin and has a vertical,triangular lingual face and forms a peakdirectly posterior to the end of the postento-cristid. Hypoconulid merges with the large,anterobuccally directed entostylid ridge,flanked by two V-shaped valleys. The ridgecrest turns anteriorly following the entostylid.Posthypocristid is bisected and runs parallel toentostylid ridge; posterolingual terminus brief-ly deflects lingually to join postcristid of thehypoconulid, forming the posterolingual cor-ner. Hypoconid has a flat, anterolingualsurface, and is slightly shorter than theanteriorly positioned entoconid.

Viewed posterolingually, the cristid obliquaappears zigzagged. Three broad, triangularridges run in a posterolingual direction,perpendicular to the cristid obliqua. Theanteriorly directed buccal section of the cristidobliqua, before the ridges, runs from thehypoconid apex and is broadly bladed andslightly concave, like the buccal half of theposthypocristid, but shorter. Lingual sectionof the cristid obliqua is straight and of equallength to buccal section, and joins metaconidapex perpendicular to the metacristids.

Posterior end of protostylid postcristid joinscristid obliqua at base of this lingual section.The midsection of the cristid obliqua, con-necting the buccal and lingual sections, islinear and supports three ridges, separated byequally wide notches. Ridge B is lowest inheight; ridge C is broadest. Anterior end ofeach ridge is slightly behind the anterior end ofeach adjacent notch; all are joined by thecristid obliqua in a zigzag shape, so that theanterior face of each ridge is flat and leansposterolingually. Crest of ridge A curveslingually as it descends, terminating at themetastylid. Ridge C is linearly aligned with theentostylid ridge. Talonid basin is reduced toa narrow, obliquely oriented valley, parallel tothe cristid obliqua.

m2: Paraconid and anterior point of toothis extended on the lengthwise axis so thetrigonid is pointed. Paraconid is distinct(damaged in QMF52175). Paracristid sup-ports one or two posterolingual, triangularridges, similar to the cristid obliqua ridges. InQMF52178 and QMF52177 (reduced inQMF52179) a small ridge extends posteriorlyfrom the paraconid. A buccal rib descendsfrom the premetacristid in all exceptQMF52177, typically joining a paracristidridge. An anterobuccal cingulum descendslow on the buccal enamel, either proceedingfrom the paraconid or a short distance below;a buccal parastylid is present, except onQMF52178. A slightly shorter anterolingualcingulum, sometimes featuring a lingual para-stylid, extends horizontally from the lingualside of the paraconid, often in conjunctionwith anterior end of the premetacristid.Protoconid is shorter and positioned anteriorto metaconid; it is a conical, lingually enlargedcusp, taller than the hypoconid. A small, shortposterobuccal shelf is positioned midwaydown the cusp (on QMF52175), and a thinposterobuccal rib runs to the protoconid base(QMF52175 and QMF52178), possibly asa vestigial protostylid. Lingual surface ofprotoconid is enlarged, particularly inQMF52179 and QMF52175. The poorlybladed postprotocristid weakly curves poster-olingually, toward the end of the postmeta-cristid. Termination of the postprotocristidvaries: in QMF52178 it joins the cristidobliqua in a continuous crest; in QMF52177


it buccally joins the metaconid ridge; inQMF52175 and QMF52179, cristid obliquaand metaconid ridge are joined, postproto-cristid in the former is detached marginallyfrom this junction and connected in the latter.Metaconid has elongate, bladed cristids; pre-metacristid slopes gradually to lingual edge ofthe paraconid or slightly short of it.Postmetacristid is shorter and steeper thanpremetacristid; it has a small lingual deflectionat its end and either joins the metastylid(QMF52179, QMF52178) or is separated bya notch (QMF52175, QMF52177).

Metastylid forms a lengthwise crest (exceptQMF52178, which divides into two pyramidalcusps) with a small buccal rib extending fromits posterior. Immediately posterobuccal tothe metaconid, a triangular ridge extendsanteriorly from the metastylid (not alwaysjoined). Links between the ridge and sur-rounding features are variable: in QMF52179,perpendicular links at anterior end of the ridgeconnect it to the metaconid and postproto-cristid; in QMF52175, it is joined midway bya buccal link to the cristid obliqua–postpro-tocristid junction; in QMF52177, the ridgejoins the anterior end of the metastylid andforms a perpendicular buccal link with thepostprotocristid; the ridge in QMF52178 isnot connected anteriorly and is bifurcatedposteriorly, of which the anterior crest joinsthe anterometastylid while the posterior crestattenuates toward the posterometastylid.

Entoconid is subconical and taller than theposteriorly positioned hypoconid; it has mod-erately sloped, bladed cristids that are orientedparallel to the metacristids. Preentocristid isanterobuccally oriented but bends anterolin-gually at the base, where it either levels or risesslightly before terminating. Just after inflec-tion, the cristid forks widely, forming a steep,flat, triangular anterior surface. The entoconidnarrows progressively away from apex butwidens again from the inflection point, sothere is a small, transverse, pyramidal buttressanterior to the preentocristid. Opposing thispreentocristid buttress is the flat, posteriorface of the metastylid, separated by a narrownotch. In QMF52175 (reduced inQMF52179), a buccal rib in descends fromthe preentocristid, swinging posteriorly at theentoconid base. Postentocristid is concave and

steeper than preentocristid, extending to thelingual margin. Entostylid ridge is similar tom1, gradually ascending anterobuccally to theentostylid before receding anteriorly, termi-nating buccal or slightly anterobuccal to theentoconid. Entostylid ridge in QMF52177 isdamaged anteriorly. Hypoconulid and hypo-conid are similar to that in m1 (hypoconid isdamaged in QMF52175 and QMF52177).Posthypocristid is not bisected and is almoststraight with a slight posterior curve on thelingual half. Cristid obliqua is similar to thatin m1, but the three ridges are lower in heightand in QMF52179 curve in toward the base ofridge B instead of running parallel. Ridge C issmaller than the entostylid ridge; ridge B splitsinto two ridges on QMF52178 andQMF52177; ridge A splits in all exceptQMF52177. Talonid valley in m2 tends towiden posteriorly. Anterior termination of thecristid obliqua is variable, but is usually eitherjoined to or separated by a minor notch fromthe metaconid ridge and postprotocristidjunction. QMF52178 is the only specimen inwhich the cristid obliqua connects solely to thepostprotocristid.

m3: m3 trigonid is slightly more roundedthan in m2. Anterior cingula are similar to m2,except the anterobuccal cingulum is muchnarrower and buccal parastylid is absent.Paracristid is smoothly curved in QMF52181but more like m2 in QMF52180. Two para-cristid ridges are present; anterior ridge isshorter and oriented anteroposteriorly, bothterminate in the trigonid valley, buccally ofpremetacristid blade. Form of the metaconidand protoconid is similar to that in m2, as isrelative height and position of cusps.Postprotocristid is poorly bladed and joinscristid obliqua in a continuous curve on thelongitudinal axis. Premetacristid terminates onthe lingual edge of the paraconid (QMF52181)or fails to join (QMF52180).

Metastylid is bicuspid and anterometastylidis larger. In QMF52181, the anterometastylidcontinues into the metaconid ridge but inQMF52180 is separated by a notch.Posterometastylid has a short buccal crestand is pyramidal. Metaconid ridge ofQMF52180 is joined to the metaconid andpostprotocristid by shallow, perpendicularlinks on either side.


Entoconid resembles that in m2 but isrelatively smaller, and constriction betweenthe entoconid and anterior preentocristidbuttress is reduced. Entoconid on QMF52181is damaged, but there is a buccal rib descendingfrom the preentocristid, as in QMF52175 (m2).Compared to m1–2: hypoconulid is obliquelyoriented so the flat lingual face forms theposterolingual corner; entostylid, entostylidridge, posthypocristid, and hypoconid arerelatively smaller; entostylid is positionedcloser to hypoconulid. Cristid obliqua isoriented slightly less obliquely than m1–2;cristid obliqua ridges are reduced in heightand progressively widen from ridge A to C. Onm3 specimens, only ridge A bifurcates, and thedivergence occurs at the start of the ridge, onthe cristid obliqua.

m4: QMF52176 is here considered an m4.It has a wider and longer trigonid thantalonid, and the margins are curved so thetooth appears anteriorly bulbous. Relativecusp position and heights are as in m2–3.Anterobuccal cingulum is reduced to a veryshort and stout parastylid-like protrusion nearthe enamel base, below the paracristid.Anterolingual cingulum, also short, has a lin-gual parastylid and runs horizontally fromanterior end of premetacristid. The arcuateparacristid is disrupted by two paracristidridges, enlarged in comparison to m2–3.Protoconid is conical with unbladed cristids;a short posterobuccal ledge occurs on theprotoconid, similar to that in QMF52175(m2). Lingual expansion of the protoconidobserved in m2–3, is developed further into anunbladed, anterior protocristid, directedslightly posterolingually. Protoconid angle isvery wide, more similar to m2 than m3.Postprotocristid descends gradually, joiningthe cristid obliqua in a continuous crest.Postmetacristid is slightly steeper than m2–3.

Metastylid is bicuspid and crested length-wise. It features short, transversely extendedblades from each apex and a lower, parallelblade running the same distance from theposterior base of each cusp. The blades areseparated by small notches and the cusps,postmetacristid and preentocristid are respec-tively separated by deeper notches. Metaconidridge is almost straight, extending between theposterometastylid and the posterior end of the

paracristid. The ridge is swollen centrally anda second, buccal blade extends from a rightangle; this second blade runs posterolingually,parallel to the metaconid ridge, and endsopposite the preentocristid buttress.

Entoconid and hypoconid resemble that inm3. Entostylid ridge is as large as in m2, but ismore curved. It ends adjacent to the lingualpreentocristid rib and almost forms a completecrest around the buccal entoconid surface, asin QMF52175 (m2). Hypoconid has a steeperanterolingual surface than m1–3. Both trigo-nid and talonid basins are completely filled byreticulated ridges off the cristid obliqua andparacristid. Cristid obliqua ridges are moresinuous than m1–3. Ridge C is largest andrather swollen while anterior ridges becomeprogressively thinner. No ridge is divergent,but ridge B has a slightly bifurcated crest at itsposterior end.

COMMENTS: QMF52176 is a damagedspecimen and is tentatively identified as m4,because it differs significantly from otherspecimens identified as m2 or m3 (figs. 2, 3).It is possible that QMF52176 is incorrectlyassigned as m4 but will remain thus untilfurther material is recovered.

DISCUSSION

Gawinga aranaea exhibits the followingpseudocheirid synapomorphies: narrow, sele-nodont lower molars; a distinct entostylid andmetastylid; m1 cristid obliqua connecting withthe metastylid, as opposed to the protoconidor protostylid; and a lingually displaced m1protoconid (Archer, 1984; Woodburne et al.,1987). However, the distinct ridges on thelower molars of G. aranaea represent anautapomorphic feature that immediately se-parates it from all other pseudocheirid taxa.

The form of the ridges is such that thecristid obliqua and paracristid on the posteriormolars are raised and the depth of the trigonidand talonid basins reduced, particularly as theridges themselves fill the basins. This createsrelatively shortened cusp height compared toother pseudocheirid genera, but no apparentdecrease in the overall relative height of crownenamel. The ridges show some resemblance tocrenulations present in species of Pildra andPseudochirops, but the phylogenetic signifi-cance of these similarities is uncertain.


Interestingly, the ridges appear morpholog-ically similar to those found in ektopodontids(particularly Darcius Rich, 1986), an enigmat-ic family of extinct possums allied withphalangerids (Pledge, 1986; 1991). The molarmorphology in ektopodontids is characterizedby multiple ridges arranged in two trans-versely oriented lophs. The ridges often bi-furcate and feature smaller, accessory ribs thatanastomose and reticulate at the base of thelophs (Stirton et al., 1967; Archer, 1984;Woodburne and Clemens, 1986; Woodburne,1987). However, the similarity of cusp detailbetween ektopodontids and G. aranaea ismore likely to be the result of convergencethan close ancestry, as there is little besides theenlarged ridges to unite the two groups. Theunusually wide molars and transverse ‘‘lophs’’of the species of Ektopodon are very muchopposite to the characteristics of G. aranaea orany species of pseudocheirid. Comparison ismore appropriate with species of Darcius(although the oldest material is Pliocene inage), due to their relatively narrow, obliquelyslanted molars with reduced number of ridges(Rich, 1986), and to a lesser extent with themore plesiomorphic (ridges are weaker andreticulated) Oligocene species of Chunia(Woodburne and Clemens, 1986). However,the ridges exhibited by ektopodontids forma disjointed crest or loph, independent of thecristids of the main cusps, whereas G. aranaeahas unpeaked ridges that extend away fromthese main cristids, such as the cristid obliqua.Without the ridges, the molar detail of G.aranaea is unquestionably pseudocheirid innature (e.g., the oblique, selenodont cusps, thelingually shifted m1 protoconid and theenlarged m1 protostylid). Furthermore, theaccepted phylogenetic affinities of pseudo-cheirids and ektopodontids is to separatesuperfamilies — Petauroidea (Baverstock,1984; Baverstock et al., 1987; Westerman etal., 1990; Kirsch et al., 1997; Osborne et al.,2002) and Phalangeroidea (Archer, 1984;Woodburne and Clemens, 1986; Woodburne,1987; Pledge, 1991), respectively—with anestimated divergence date of ,45 m.y.(Springer and Kirsch, 1991).

Besides the ridges, G. aranaea exhibitsa number of other autapomorphies. The m1has extended, shelflike cristids of the proto-

stylid, which are not present in any othergenus. The first lower molar also has anunusually long postcristid of the protostylid, itextends across the hypoflexid, obstructing theshearing plane between the trigonid andtalonid and potentially restricting transversemotion of the m1 paracone and protocone.Bisection of the posthypocristid is anotherunique feature of m1. G. aranaea exhibits thefirst evidence of a cuspate m1 protostylid inextinct pseudocheirids, a feature that untilnow had been regarded as a synapomorphy ofextant genera. Further autapomorphies in-clude the metaconid ridge, a cristid locatedposterobuccally to the metaconid on all of theposterior molars. The form and position ofthis ridge is similar to the anterolingualprotocristid found in species of Pseudo-chirops. However, in G. aranaea, there is nostrong link between the protoconid andmetaconid ridge besides occasional small linksto the postprotocristid (there is only anincipient anterior protocristid in most speci-mens). The ridge is variable in its connectionwith adjacent features, particularly at theanterior end and in QMF52176, in which itforms an additional crest. Paljara nancyha-wardae shows a very small stylid or cuspule atthe posterobuccal base of the m2–4 metaco-nid, which could be homologous to themetaconid ridge of G. aranaea.

The anterior placement of the protoconidrelative to the metaconid on posterior molarsis also a unique feature of G. aranaea, as is theswelling of the posterobuccal base of thehypoconid. However, as the only identifiedmaterial of this species is a number of isolatedteeth, often lacking roots, it is difficult todetermine the precise orientation within thedentary.

The many autapomorphic characteristics ofthe lower molars justify the erection of thegenus, although referable premolars andupper molars have yet to be identified.Characteristic extra ridges should be expectedin them, especially the upper molars. Themorphology of the lower molar dentition,particularly the m1, is highly distinctive withinpseudocheirids and provides one of the bestdiagnostic tools in the identification of species.

Currently, G. aranaea is identified as one ofthe most derived species of the extinct


pseudocheirids with an intriguing a number ofplesiomorphies shared with extinct taxa, sy-napomorphies with extant taxa and a suite ofautapomorphies as outlined briefly below.

Woodburne et al. (1987) identified speciesof Paljara and Pildra as the two most plesio-morphic pseudocheirid groups. Symplesio-morphies shared with G. aranaea include: m1shorter than m2 in length; absence of over-lapped or joined preentocristid and metasty-lid; and presence of an entostylid and ridge.The first symplesiomorphy is also shared withspecies of Marlu and the last with species ofPseudochirops. The ancestral form of thetalonid (entostylid is present; the hypoconulidand posthypocristid are independent of eachother; and the posterolingual corner extendsfurther posterior than the end of the post-entocristid) is shared by G. aranaea andspecies of Paljara, but is also found in speciesof Pseudochirops.

Generic-level synapomorphies with otherpseudocheirids include: a progressively complexmetastylid along the tooth row (shared withPildra); an elongated preentocristid (sharedwith Marlu, Pseudokoala, and extant genera);a lingual rib extending from the m1 paracristid(shared with Pildra, Pseudochirops, Petro-pseudes, and Pseudocheirus) and an enlargedm1 protostylid (shared with extant genera).

Based on these features, G. aranaea appearsparticularly apomorphic in contrast to otherOligo-Miocene pseudocheirid taxa, and themany autapomorphic features make referenceto a particular group difficult. At present, italso appears unlikely that any of the moderngenera were derived from species of Gawinga.The relationships of Gawinga to other pseu-docheirid genera are yet to be fully assessedand will be investigated as part of a largerproject on the systematics and phylogeny ofOligo-Miocene pseudocheirids, incorporatinga number of new taxa in preparation (Robertset al., in prep.).

TEMPORAL AND SPATIAL DISTRIBUTION:Gawinga aranaea is a relatively rare species,found to date in only three of the hundreds oflocal faunas in the Riversleigh World Heritagearea. Two of these deposits, Upper Site andWayne’s Wok Site, are considered to be part ofRiversleigh’s System B deposits, which havebeen interpreted to be early Miocene in age, or

approximately 16–23 m.y. (Archer et al., 1989,1995, 1997b). Upper Site in particular hasproduced a very diverse local fauna and hasbeen biocorrelated with the Kutjamarpu LocalFauna of northern South Australia (Archer etal., 1989, 1995, 1997b). Although Woodburneet al. (1993) argued that the Kutjamarpu LFwas late Oligocene in age, Archer et al. (1997b)argued it was early Miocene. Upper Siteincludes species in the three Oligo-Miocenepseudocheirid genera Paljara, Pildra, andMarlu, representatives of the modern genusPseudochirops, an additional undescribed pseu-docheirid group, and G. aranaea. Wayne’s WokSite has also produced taxa of the three Oligo-Miocene genera in addition to Gawinga andshares many taxa with Upper Site Local Faunaand other local faunas attributed to System B(Travouillon et al., MS.). Numerous otherpetauroid and phalangeroid taxa have beenrecognized from both sites (Crosby et al.,2004). Many System B localities have producedlocal faunas with relatively high levels ofdiversity and relative abundance of arborealanimals. It has been suggested by Archer et al.(1989, 1995, 1997b) that such an ecologicalcommunity indicates a rainforest climate sim-ilar to that found in montane New Guinea,Borneo, or the Amazon of South America.

Arachnea Ridge Site is poorly sampled incomparison to Upper and Wayne’s Wok sites.Aside from the pseudocheirid material de-scribed here, only balbarine kangaroo materi-al has been recovered from Arachnea RidgeSite. Therefore, at the present time little can bedetermined with confidence about the relativeage of this site other than a possible earlyMiocene correlation with Upper and Wayne’sWok sites.

Broad changes in faunal composition be-tween the early to mid-Miocene and lateMiocene, with a general reduction in arborealtaxa, including pseudocheirids, are well docu-mented by the Riversleigh fossil deposits(Archer et al., 1989, 1995, 1997b; Myers etal., 2001). This is thought to correlate withchanges in paleoclimate at the time, relating tothe formation of the New Guinea Highlandsand the onset of reduced rainfall on theAustralian continent (Kemp, 1984; Archer etal., 1995; Martin, 1998). This has beensuggested as the cause of a shift from a closed


forest or rainforest vegetation structure toa more open structure that would be less ableto support a high number of sympatricarboreal taxa (Archer et al., 1989, 1995,1997b; Myers et al., 2001). Five genera ofpseudocheirids are now recognized in the earlyMiocene deposits of Riversleigh, lendingfurther support to this interpretation.

DIET: Extant pseudocheirids are one of themost strictly folivorous possum groups inAustralia and New Guinea (Flannery, 1994).The energy benefits of eating leaf matter arelow, requiring the ingestion of large quantities,and therefore an efficient masticatory anddigestive system (McNab, 1978; Hume et al.,1984; Foley and Cork, 1992). Strict folivory ismost sustainable in mammals of 700 g mass orhigher, but there are known exceptions amongthe rodents (Kay and Hylander, 1978; Foleyand Cork, 1992). All living ringtail taxa areprimarily leaf eaters, but they are also known toeat fruits, flowers, and, in the smallest species,moss and lichens (Archer, 1984; Flannery,1994). Based on molar morphology, G. aranaeais also considered to have been primarilyfolivorous. Australian pseudocheirids range inweight from just under 700–2000 g but averagearound 1000 g (Strahan, 1995). A body-massestimate of G. aranaea from the diprotodontianm3 occlusal area prediction equation of Myers(2001) predicts a mass of 323 g. The twosmallest extant pseudocheirids are the NewGuinean species Pseudochirulus mayeri(,150 g) and P. canescens (,350 g)(Flannery, 1994), and like them, G. aranaeamay have supplemented its diet with a range offood types in addition to leaf matter.

The additional ridges on the cristid obliquaand other blades are likely to have served asextra shearing surfaces. G. aranaea also hasnarrower molars than other extinct pseudo-cheirids, more like those of most living ring-tails. These teeth are also more elongate thanthose possessed by primarily insectivorous,gumivorous, or omnivorous possum groups,which have stouter molars with relatively largerocclusal surfaces (Kay and Hylander, 1978).

ACKNOWLEDGMENTS

Support for the Riversleigh project has beenprovided by the Australian Research Council,

Queensland Parks and Wildlife Service, theUniversity of New South Wales, the Queen-sland Museum, the Australian Museum andZinnifex. We thank M. Bassarova, R. Beck,K. Black, P. Brewer, A. Gillespie, J. Louys, K.Travouillon, and V. Weisbecker for discus-sion, technical advice, and assistance. T.Ennis, S. Ingleby, and N. Pledge kindlyallowed access to comparative specimens fromthe Australian Museum and South AustralianMuseum. We also thank an anonymousreviewer and W.D. Turnbull for valuablecomments and suggestions.

REFERENCES

Aplin, K.P., and M. Archer. 1987. Recent advancesin marsupial systematics with a new syncreticclassification. In M. Archer (editor), Possumsand opossums: studies in evolution: xv–lxxii.Sydney: Surrey Beatty & Sons in associationwith the Royal Zoological Society of NewSouth Wales.

Aplin, K.P., J.M. Pasveer, and W.E. Boles. 1999.Late Quaternary vertebrates from the Bird’sHead Peninsula, Irian Jaya, Indonesia, includ-ing descriptions of two previously unknownmarsupial species. Records of the WesternAustralian Museum 57(suppl.): 351–387.

Archer, M. 1984. The Australian marsupial radia-tion. In M. Archer and G. Clayton (editors),Vertebrate zoogeography and evolution inAustralasia: 633–808. Perth: Hesperian Press.

Archer, M. 1992. Ringtail possums (Pseudo-cheiridae, Marsupialia) from the Tertiary de-posits of Riversleigh. The Beagle 9: 257.

Archer, M., K. Black, and K. Nettle. 1997a. Giantringtail possums (Marsupialia, Pseudoc-heiridae) and giant koalas (Phascolarctidae)from the late Cainozoic of Australia.Proceedings of the Linnean Society of NewSouth Wales 117: 3–16.

Archer, M., H. Godthelp, S.J. Hand, and D.Megirian. 1989. Fossil mammals of River-sleigh, northwestern Queensland: preliminaryoverview of biostratigraphy, correlation andenvironmental change. Australian Zoologist 25:29–65.

Archer, M., S.J. Hand, and H. Godthelp. 1995.Tertiary environmental and biotic change inAustralia. In E.S. Vrba, G.H. Denton, T.C.Partridge and L.H. Burckle (editors),Paleoclimate and evolution with emphasis onhuman origins: 77–90. New Haven: YaleUniversity Press.


Archer, M., S.J. Hand, H. Godthelp, and P.Creaser. 1997b. Correlation of the Cainozoicsediments of the Riversleigh World HeritageFossil Property, Queensland, Australia. In J.P.Aguilar, S. Legendre and J. Michaux (editors),Actes du Congres BiochroM’97, Memoires etTravaux, l’Ecole Pratique des Haute Etude,Institut de Montpellier 21: 131–152.

Bassarova, M., and M. Archer. 1999. Living andextinct pseudocheirids (Marsupialia, Pseudo-cheiridae): phylogenetic relationships andchanges in diversity through time. AustralianMammalogy 21: 25–27.

Bassarova, M., M. Archer, and S.J. Hand. 2001.New Oligo-Miocene pseudocheirids (Marsu-pialia) of the genus Paljara from Riversleigh,northwestern Queensland. Memoires of theAssociation of Australasian Palaeontologists25: 61–75.

Baverstock, P.R. 1984. The molecular relationshipsof Australasian possums and gliders. In A.P.Smith and I.D. Hume (editors), Possums andgliders: 1–8. Sydney: Australian MammalSociety.

Baverstock, P.R., J. Birrell, and M. Kreig. 1987.Albumin immunologic relationships of theDiprotodontia. In M. Archer (editor), Pos-sums and opossums: studies in evolution:229–234. Sydney: Surrey Beatty & Sons.

Collett, R. 1884. On some apparently new marsu-pials from Queensland. Proceedings of theZoological Society of London: 381–389.

Collett, R. 1895. On a new Pseudochirus fromnorthwest Australia. Zoologischer Anzeiger18(490): 464–468.

Creaser, P. 1997. Oligocene-Miocene sediments ofRiversleigh: the potential significance of topog-raphy. Memoirs of the Queensland Museum41: 303–314.

Crosby, K., M. Bassarova, M. Archer, and K.Carbery. 2004. Fossil possums in Australasia:discovery, diversity and evolution. In R.L.Goldingay and S.M. Jackson (editors), Thebiology of possums and gliders: 161–176.Chipping Norton: Surrey Beatty & Sons.

Flannery, T.F. 1994. Possums of the world.A monograph of the phalangeroidea. Sydney:Geo Productions.

Foley, W.J., and S.J. Cork. 1992. Use of fibrousdiets by small herbivores: how far can the rulesbe ‘bent’? Trends in Ecology and Evolution7(5): 159–162.

Hume, I.D., W.J. Foley, and M.J. Chilcott. 1984.Physiological mechanisms of foliage digestionin the greater glider and ringtail possum(Marsupialia: Pseudocheiridae). In A.P. Smithand I.D. Hume (editors), Possums and gliders:247–251. Sydney: Australian Mammal Society.

Kay, R.F., and W.L. Hylander. 1978. The dentalstructure of mammalian folivores with specialreference to primates and phalangeroidea(Marsupialia). In G.G. Montgomery (editor),The ecology of arboreal folivores: 173–191.Washington D.C.: Smithsonian InstitutionPress.

Kemp, E. 1984. Tertiary palaeogeography and theevolution of Australian climate. In M. Archerand G. Clayton (editors), Vertebrate zoogeog-raphy and evolution in Australasia: 61–67.Perth: Hesperian Press.

Kirsch, J.A.W. 1977. The comparative serology ofMarsupialia, and a classification of marsupials.Australian Journal of Zoology (suppl. ser.) 52:1–152.

Kirsch, J.A.W., F-J. Lapointe, and M.S. Springer.1997. DNA-hybridisation studies of marsupialsand their implications of metatherian classifi-cation. Australian Journal of Zoology 45:211–280.

Luckett, W.P. 1993. An ontogenetic assessment ofdental homologies in therian mammals. In F.S.Szalay, M.J. Novacek and M.C. McKenna(editors), Mammal phylogeny: mesozoic differ-entiation, multituberculates, monotremes, earlyeutherians and marsupials: 182–204. NewYork: Springer-Verlag.

Mackness, B.S., and M. Archer. 2001. A newpetauroid possum (Marsupialia, Pseudo-cheiridae) from the Pliocene Bluff Downs localfauna, northern Queensland. Alcheringa 25:439–444.

Martin, H.A. 1998. Tertiary climatic evolution andthe development of aridity in Australia.Proceedings of the Linnean Society of NewSouth Wales 119: 115–136.

McNab, B.K. 1978. Energetics of arboreal foli-vores: physiological problems and ecologicalconsequences of feeding on an ubiquitous foodsupply. In G.G. Montgomery (editor), Theecology of arboreal folivores: 153–162.Washington D.C.: Smithsonian InstitutionPress.

Myers, T.J. 2001. Prediction of marsupial bodymass. Australian Journal of Zoology 49: 1–20.

Myers, T.J., K. Crosby, M. Archer, and M. Tyler.2001. The Encore local fauna, a late Mioceneassemblage from Riversleigh, northwesternQueensland. Memoirs of the Association ofAustralasian Palaeontologists 25: 147–154.

Osborne, M.J., L. Christidis, and J.A. Norman.2002. Molecular phylogenetics of theDiprotodontia (kangaroos, wombats, koala,possums, and allies). Molecular Phylogeneticsand Evolution 25: 219–228.

Pledge, N.S. 1986. A new species of Ektopodon(Marsupialia: Phalangeroidea). University of


California Publications in Geological Sciences131: 43–67.

Pledge, N.S. 1987. Pseudocheirids (Marsupialia:Pseudocheiridae) from the middle MioceneNgama local fauna of northern SouthAustralia. In M. Archer (editor), Possums andopossums: studies in evolution: 681–688.Sydney: Surrey Beatty & Sons in associationwith the Royal Zoological Society of NewSouth Wales.

Pledge, N. 1991. Reconstructing the natural historyof extinct animals: Ektopodon as a case history.In P. Vickers-Rich, J.M. Monaghan, R.F.Baird and T.H. Rich (editors), Vertebratepalaeontology of Australasia: 247–266.Lilydale, Victoria: Pioneer Design Studio.

Pledge, N.S. 1992. The Curramulka local fauna:a new late Tertiary fossil assemblage from theYorke Peninsula, South Australia. The Beagle9: 115–142.

Rich, T.H.V. 1986. Darcius duggani, a new ekto-podontid (Marsupialia; Phalangeroidea) fromthe early Pliocene Hamilton local fauna,Australia. University of California Publi-cations in Geological Sciences 131: 68–74.

Rothschild, W., and G. Dollman. 1932. Onmammals collected in Dutch New Guinea byF. Shaw Mayer in 1930. Proceedings of theZoological Society of London: 211–219.

Springer, M.S., and J.A.W. Kirsch. 1991. DNAhybridization, the compression effect and theradiation of Diprotodontian marsupials.Systematic Zoology 40: 131–151.

Stirton, R.A., R.H. Tedford, and M.O.Woodburne. 1967. A new Tertiary formationand fauna from the Tirari Desert, SouthAustralia. Records of the South AustralianMuseum 15: 427–462.

Strahan, R. (editor). 1995. Mammals of Australia,revised ed: 236–256. NSW: Reed Books.

Tate, G.H.H. 1945. Results of the Archboldexpeditions. No. 54. The marsupial genus

Pseudocheirus and its subgenera. AmericanMuseum Novitates 1287: 1–24.

Thomas, M.O. 1897. On some new phalangers ofthe genus Pseudocheirus. Annali Del MuseuCivico Di Storia Naturale Di Genova 18:142–146.

Turnbull, W.D., and E.L. Lundelius. 1970. TheHamilton fauna, a late Pliocene mammalianfauna from the Grange Burn, Victoria,Australia. Fieldiana: Geology 19: 1–163.

Waterhouse, G.R. 1846. A natural history of theMammalia, vol 1: 297–307. London.

Westerman, M., D.N. Janczewski, and S.J. O’Brien.1990. DNA-DNA hybridisation studies andmarsupial phylogeny. Australian Journal ofZoology 37: 315–323.

Woodburne, M.O. 1987. The Ektopodontidae, anunusual family of Neogene phalangeroid mar-supials. In M. Archer (editor), Possums andopossums: studies in evolution: 603–606.Sydney: Surrey Beatty & Sons in associationwith the Royal Zoological Society of NewSouth Wales.

Woodburne, M.O., and W.A. Clemens (editors).1986. Revision of the Ektopodontidae (Mam-malia; Marsupialia; Phalangeroidea) of theAustralian Neogene. University of CaliforniaPublications in Geological Sciences 131: 1–114.

Woodburne, M.O., B.J. MacFadden, J.A. Case,M.S. Springer, N.S. Pledge, J.D. Power, J.M.Woodburne, and K.B. Springer. 1993. Landmammal biostratigraphy and magnetostrati-graphy of the Etadunna Formation (LateOligocene) of South Australia. Journal ofVertebrate Palaeontology 13: 483–515.

Woodburne, M.O., R.H. Tedford, and M. Archer.1987. New Miocene ringtail possums(Marsupialia: Pseudocheiridae) from SouthAustralia. In M. Archer (editor), Possums andopossums: studies in evolution: 639–679.Sydney: Surrey Beatty & Sons in associationwith the Royal Zoological Society of NewSouth Wales.


This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).

Complete lists of all issues of the Novitates and the Bulletin are available at World Wide

Web site http://library.amnh.org/pubs. Inquire about ordering printed copies via e-mail from

[email protected] or via standard mail from: American Museum of Natural History,

Library—Scientific Publications, Central Park West at 79th St., New York, NY 10024. TEL:

(212) 769-5545. FAX: (212) 769-5009.

Documents

New Genus and Species of Extinct Miocene Ringtail Possums