Records of the Western Australial1 Museum Supplement No. 57: 299-306 (1999).

First Australian Pliocene species of Hipposideros(Microchiroptera: Hipposideridae)

Suzanne J. Hand and Henk Godthelp

School of Biological Science, University of New South Wales, Sydney, NSW 2052;email: s.hand@unsw.edu.au

Abstract A new Australian Pliocene hipposiderid is described on the basisof maxillary fragments representing at least four individuals recovered fromRackham's Roost Site, Riversleigh World Heritage property, Lawn HillNational Park, northwestern Queensland. Hipposideros wil1sburyorum sp. novois distinguished from all other Hipposideros species (and all otherhipposiderids) by its very large, round infraorbital foramen. Dentally, itshares most similarities with members of the diverse and widespreadHipposideros bicolor group of Hill (1963) as well as Hipposideros megalotis ofEast Africa and the extinct H. (Syl1desmotis) vetus lineage from the middleMiocene of Morocco and Pliocene of France.

INTRODUCTION

Tertiary deposits in the Riversleigh WorldHeritage property, Lawn Hill National Park,northwestern Queensland are particularly rich inthe remains of members of the bat familyHipposideridae (Sige et al. 1982; Hand 1993, 1997a,1997b, 1997c, 1998a, 1998b). This Old World familycontains 65 living species, referred to the genusHipposideros (53 species) and eight other genera (ofone or two species each): Rhinonicteris, Coelops,Paracoelops, Triaenops, Cloeotis, Anthops, Asellia, andAselliscus (Koopman 1994). Within Hipposideros, anumber of species groups is generally recognized(e.g., Gray 1866; Peters 1871; Tate 1941; Hill 1963).Recent authors (e.g., Koopman 1994) generallyaccept Hill's (1963) seven species groups orsubgroups (i.e., the H. megalotis, bicolor, cyclops,speoris, pratti, armiger, and diadema subgroups),with some also recognizing the subgenusSyndesmotis for H. megalotis (e.g. Legendre 1982).Hill (1963) further grouped Hipposideros species intothree primary divisions (i.e., the bicolor, cyclops anddiadema groups) interpreted to represent threedistinct evolutionary trends within the genus.

Recent Australian hipposiderids include fivespecies of Hipposideros (representing the bicolor,cyclops and diadema groups) and the endemicRhinonicteris aurantius, all of which inhabit warmhumid caves in tropical to subtropical latitudes. Inthe Australian (i.e. Riversleigh) Tertiary recordseveral hipposiderid genera or subgenera arerepresented: Brachipposideros, Hipposideros,Rhinonicteris, Miophyllorhina, Xenorhinos andRiversleigha in Oligo-Miocene deposits (Sige et al.1982; Hand 1997a, 1997b, 1997c, 1998a, 1998b), andBrachipposideros and Rhinonicteris in Riversleigh's

only Pliocene deposit, Rackham's Roost Site(Archer et al. 1994). A number of other distinctiveOligo-Miocene hipposiderid taxa awaitsdescription.

In this paper, the first species of Hipposideros fromthe Australian Pliocene is described, based onmaxillary fragments from Rackham's Roost Site.Skull terminology follows Hill (1963) and Hand(1997a). Dental terminology follows Sige et al.(1982). The prefix QM F refers to specimens held inthe fossil collections of the Queensland Museum,Brisbane.

SYSTEMATIC PALAEONTOLOGY

Authorities for all scientific names not specificallyreferenced in this paper can be found in Koopman(1994).

Suborder Microchiroptera Dobson, 1875

Superfamily Rhinolophoidea Weber, 1928

Family Hipposideridae Miller, 1907

Hipposideros Gray, 1831

Hipposideros winsburyorum sp. novoFigure 1; Table 1

Material Examined

HolotypeQM F30571, a right maxilla fragment with P4_M2

and alveoli for Cl and M3.

Pa ra typesQM F30572, a right maxilla fragment with P4_M2

300 S.J. Hand, H. Godthelp

Figure 1 Hipposideros winsburyorum sp. nov., from the Pliocene Rackham's Roost Site, Riversleigh World Heritageproperty, Lawn Hill National Park, northwestern Queensland. QM F30571, holotype, right maxilla contain­ing p4, MI-2. A-A', oblique-occlusal view, stereopair; H, lateral view. Scale = 2 mm.

and alveoli for Cl and M3; QM F30573, a left maxillafragment with MI and alveoli for p4 and M2-3; QMF30574, a right maxilla fragment with p4-MI andalveolus for Cl and M2; QM F30575, a left maxillafragment with P4_MI; QM F30576, a right maxillafragment with P4_M2.

Type Locality, Age and Associated FaunaThe type locality, Rackham's Roost Site

(Godthelp 1987, 1997; Archer et al. 1989, 1994;Hand 1995, 1996), occurs at 19°02.09'5.,138°41.60'E. (Global Positioning Satellite device) inthe Riversleigh World Heritage property, LawnHill National Park, northwestern Queensland. Itappears to represent the indurated floor of a long,narrow cave developed in Cambrian ThorntoniaLimestone. The sediment is a breccia of tiny,mostly fragmented, bones and teeth set in a fine­grained, pink-coloured (presumably iron-stained)limestone. The deposit covers an area ofapproximately 200 m 2, with a maximum depth of0.5 to 1.0 m.

The deposit is interpreted to be Pliocene in age

because it contains a macropodid similar toProtemnodon snewini Bartholomai, 1978 from theearly Pliocene Bluff Downs Local Fauna ofnortheastern Queensland (Archer and Wade 1976)and abundant plesiomorphic murids, includingZyzomys rackhami Godthelp, 1997, several species ofPseudomys and a number of new genera. TheRackham's Roost deposit also contains the remainsof crustaceans, fish, frogs, lizards, small crocodiles,snakes, birds, dasyurids, peramelids, apseudocheirid, an extinct potoroid, twoemballonurids (Taphozous spp.), at least four otherhipposiderids, four vespertilionids, the smallmegadermatid Megaderma richardsi and the large,living Ghost Bat Macroderma gigas (Hand 1995, 1996;Boles 1998). The very finely broken remains of thesmall vertebrates and depressed fractures andimpressions in the bones characteristic of RecentMacroderma gigas canines suggest that the depositlargely represents the remains of prey accumulatedby an early population of that species. Boids, theemballonurids, hipposiderids and vespertilionidsprobably cohabited the megadermatid roost; the

Australian Pliocene Hipposideros

Table 1 Measurements (in mm) of types ofnov. from the

Pliocene Rackham's Roost Rlversleigh,northwestern Queensland. I maXinlUmlength; W wIdth; tooth absent.

Holotype Paratypes

QMF 30571 30572 30573 30574 30575 30576

P' 0.91 086 0.80 0.81 082W 1.25 128 120 120 1.10

M1 1.36 1.34 1.39 1.30 122 128W 1.34 1.30 1.38 1.35 1.29 1.37

M2 1.25 121 1.18w 1.39 1.39 1.34

macropodid remains were recovered from whatwas possibly an entrance to the cave.

DiagnosisThis Hipposideros species differs from all others in

the following combination of features:exceptionally large, round infraorbital foramen,low on the face and dorsal to P'-MI; anteriorlyconvergent toothrows; anterior margin of palate U­shaped; loss of p2; p4 wider than long, with shallowlingual cingulum and poorly developedanterolingual cingular cusp; MI with small,posterolingually developed heel and posteriorlyopening protofossa; M2 with little heeldevelopment; M3 reduced in length and width.Assignment of the new species to the genusHipposideros is discussed in Remarks (below).

EtymologyThe species is named for Keith and Janet

Winsbury, long-term supporters of the Riversleighfossil research.

DescriptionThe new taxon is known from six maxillary

fragments: the holotype QM F30571 and the fiveparatypes listed above. The material represents atleast four individuals. All specimens preserve theinfraorbital foramen and all clearly lack an alveolusfor pl. The lower dentition is not yet known.Several dentary fragments of appropriate size, andpossibly dental morphology, have been recoveredfrom among hundreds of Rackham's Roosthipposiderid specimens, but they are notsufficiently complete to confidently assign toHipposideros willsbllryorum (see also Remarksbelow).

The infarorbital foramen is extremely large andround, with a minimum diameter of 0.5 mm. It issituated low on the face and dorsal to P4_M l . It isenclosed by a narrow bar of bone (the antorbital

301

bar) that is straight, almost perpendicular to thezygoma and of equal thickness throughout (i.e. itlacks a wing). A foramen Ulacrimal) opens directlyabove the antenor attachment point of theantorbital bar at the edge of the circumorbital rim.The lateroventral fossa, a deep and very broadfossa in the ventral part of the anterior recess ofthe orbital fossa, exposes the maxillary coveringthe roots of the posterior cheek teeth; in this region,the maxillary is perforated by many foramina. Inlateral view, the ventral margin of the orbit extendsto a point level with the posterior face of MI, itsanterolateral rim being low and rounded (ratherthan raised and sharp).

The midline of the palate is preserved posteriorlybut not anteriorly; it appears, however, to havebeen quite short. The posterior extension of themidline of the palate is level with the posterior faceof M2; it does not extend posteriorly beyond thetoothrow. The palation (Hill 1963) appears to havebeen shallow but round with no postpalatal spine.The posterolateral indentation (incisura) extendsanteriorly to at least the level of the posterior faceof M2. The palate's anterior border, for the junctionwith the premaxillae, extends posteriorly at least tothe level of the metacone of MI, has a roundedlateral margin and was evidently a broad U-shape.Assuming that the anterolateral margins of thesphenorbital bridge were subparallel (as in mosthipposiderids), the toothrows would haveconverged anteriorly.

The alveolus for Cl is broken anteriorly in QMF30571 but is complete in QM F30572 and QMF30574. These alveoli indicate that it was not aparticularly long tooth, with a relatively roundedrather than flattened root. p2 is completely lacking,there being no sign of an alveolus for this tootheither within or extruded (lingually or buccally)from the toothrow. p3 is also absent, as in allhipposiderids. p4 is wider than long, and only justnarrower than MI and M2. The lingual cingulum isshallow but distinct and virtually complete. Theanterobuccal cingular cusp is poorly developedand the anterolingual cingular cusp lacking.

MI-3 all have three roots. In MI2 the heels arepoorly developed such that the teeth areconspicuously shorter lingually than buccally (i.e.,they are not squared). MI has a small but distinctposterolingually-directed heel, while M2 virtuallylacks a heel. In these teeth the postprotocristaapproaches but does not meet the base of themetacone so that the protofossa is open posteriorly,noticeably more so in MI than in M2. There is nocrest issuing from the end point of thepostprotocrista in either MI or M 2 • The lingual andposterior cingula are continuous in MI2 and theposterolingual cingulum surrounding the heel onlyslightly thickened. TI1e three alveoli for M3 indicateit was reduced in both length and width. The

302

paracone root alveolus occurs on the buccal marginof the toothrow, but that for the metacone islingually displaced such that it occurs on theposterior border of the maxilla. The alveolus forthe metacone root is damaged posteriorly but M3appears to have been between one-half and two­thirds the length of MI or M2. The alveolus for theprotocone root (also damaged) occurs level withthe protofossa of MI-2, indicating that M3 was alsosignficantly narrower than the other cheekteeth.

RemarksIn lacking p2, the Rackham's Roost hipposiderid

differs from extinct species of Palaeophyllophora,Pseudorhinolophus, Brachipposideros and Vaylatsia,and extant species of Coelops, Paracoelops, Triaenops,Rhinonicteris, Anthops and Aselliscus. (Paracoelopsmegalotis, known only from the type specimen, wasnot examined in this study; its description suggestsit is very similar to species of Coelops.) In thisfeature H. winsburyorum resembles living andextinct species of Asellia, Cloeotis and Miophyllorhinaas well as some species of Hipposideros, Le. H.megalotis, H. sabanus of the bicolor group, H. stenotisof the cyclops group and the extinct H. (Syndesmotis)vetus lineage of the Miocene of North Africa andPliocene of southern France (Lavocat 1961; Sige1976; Legendre 1982).

The new hipposiderid further shares with speciesof Asellia and some species of Hipposideros asignificantly reduced MJ. In A. tridens, H. megalotisand H. vetus, M3 is significantly reduced such thatthe premetacrista is lost and the tooth is muchshorter Gust more than one-half the length of MI orM2) and much narrower than P"-M2. In H. sabanusand the French Miocene Asellia mariaetheresae Mein,1958, M3 is slightly less reduced, in this way perhapsmore closely resembling H. winsburyorum. TheRackham's Roost hipposiderid differs fromRiversleigh's Miocene Miophyllorhina riversleighensisHand, 1997c (the only other Riversleigh hipposideridlacking P2) in its broad p4 with indistinctanteroIingual cingular cusp, and poorly developedheel on M2.

Of hipposiderids lacking p2 and with a reducedM3, only H. megalotis and H. vetus further sharewith H. winsburyorum a P" that is wider than longand lacks an anterolingual cingular cusp.

Hipposideros winsburyorum resembles manyspecies of the bicolor group, as well as H. megalotisand H. vetus, in its small molar heels, posteriorlyopen protofossae and gracile nature of MI-2. Theseteeth differ markedly from the squared MI-2typically found in species of Brachipposideros,Rhinonicteris, Triaenops, Xenorhinos, Riversleigha,Asellia, Aselliscus and many (but not all) diademagroup taxa. Coelops species differ from H.winsburyorum in their MI-2 with pronounced heelsand very tall posterolingual cingular cusps.

S.J. Hand, H. Godthelp

Further, in Coelops species the postprotocristacontinues posterobuccally as a metacingulumthereby closing the protofossa in MI-3; in species ofCloeotis the protofossa is also closed in MI-3, buthere the postprotocrista meets the base of themetacone (as it does also in species of the cyclopsgroup and Palaeophyllophora).

The anterior convergence of the toothrowsobserved in H. winsburyonlm is similar to that seen inH. megalotis and members of the bicolor group as wellas species of Coelops and Cloeotis. In Cloeotis speciesthe palate extends posteriorly only to the posteriorface of M2, as in H. winsburyonlm; in Coelops, H.megalotis and in most H. bicolor species it extendsfurther posteriorly. In species of Asellia, Aselliscus,Triaenops, Anthops, Brachipposideros, Rhinonicteris andthe H. cyclops and diadema groups the toothrows aresubparallel reflecting the typically broad rostrum inthese groups. The anterior margin of the palate is U­shaped in species of Asellia, Coelops, H. megalotis andsome members of the cyclops and diadema groups, asit is in H. winsburyorum, but is more typically V­shaped in species of the bicolor group and otherhipposiderid genera.

Cl morphology distinguishes varioushipposiderid groups. Hipposideros megalotis, H.vetus, and species of Rhinonicteris, Triaenops,Coelops, Cloeotis, Anthops, Brachipposideros,Xenorhinos and Riversleigha possess a well­developed secondary posterior cusp on Cl. Speciesof Aselliscus, Palaeophyllophora and members of thecyclops group lack this cusp. In Pseudorhinolophus,Asellia and the H. diadema group, the cusp, ifpresent, is generally less well developed. In thebicolor group, a tall posterior cusp is variablypresent and/or developed. The morphology of theCl of H. winsburyorum is not yet known. Mostisolated hipposiderid Cls from the Rackham'sRoost deposit have a discrete posterior secondarycusp but some completely lack the secondary cusp.In hipposiderids, the relative length of the Clalveolus does not necessarily reflect secondarycusp development (e.g. Coelops and Cloeotis spp.).

No dentary fragments have been referred to H.winsburyorum. The ascending ramus istaxonomically important in hipposiderids but ismissing in all specimens of appropriate size anddental morphology (e.g., P2 and M3 reduced) topotentially represent this species. This part of thedentary is diagnostic for H. megalotis and H. vetus(Le. low coronoid process and long and outwardlydeflected angular process; Legendre 1982) andappears to be characteristic for some speciesgroups of Hipposideros (e.g., tall coronoid processin species of the diadema and cyclops groups). Itsheight and shape are also diagnostic for species ofthe Brachipposideros-Rhinonicteris lineage (i.e. lowcoronoid process and broad angular process; Sigeet al. 1982; Legendre 1982).

Australian Pliocene Hipposideros

Hipposidcros winsburyorulll is unique amonghipposiderids in its exceptionally large, roundedinfraorbital foramen. At least four individuals, allof similar morphology, are represented in theRackham's Roost sample, indicating that this is notan aberration. The foramen is unlike the typicallyelongate infraorbital foramen characterizing mosthipposiderids, including Il. mcgalotis and membersof the bicolor group (notwithstanding the fact thatthe fora men was described by Hill (1963) assomewhat rounded in H. mcgalotis, H. pygmacus, H.cur/us and H. bentus among others). The foramen isalso relatively much larger and rounder than in thespcoris, pratti, annigcr and diadcma groups.Hipposideros colllmersoni of the diadema group is anotable exception but otherwise differs strikinglyfrom H. winsburyorum in its very large size, P4_M2morphology, and in the position of the infraorbitalforamen (high on the face and dorsal to M 23 ratherthan very low and dorsal to P4_Ml in H.winsbllryonml). The foramen is perhaps moresimilar to the foramen characterizing manymembers of the cyclops group, which otherwisediffer from H. winsbllryonml in, among otherfeatures, their distinctive molar morphology andsubparallel toothrows, as discussed above.

Based on comparisons of tooth size (e.g., Hill1963, table 2), Hipposideros winsbllryorum would besimilar in size to Hipposideros a/er and H. mcgalo/isboth of which are small hipposiderids withforearm lengths of 34-38 mm and 33-43 mmrespectively (e.g., Koopman 1994). Other bicolorspecies of similar size include H. cineracells (32-37mm), H. pygmaeus (36-40 mm), H. sabanlls (37-48mm), H. pomona (37-44 mm), and H. hallophyl/lIs(35-39 mm).

DISCUSSION

On the basis of its dental formula and dental andpalate morphology, the Rackham's Roosthipposiderid appears to be most similar to EastAfrica's Hipposideros megalotis, the Tertiary H. vc/uslineage from North Africa and France, andmembers of the H. bicolor group, 33 species of whichare distributed widely from Africa to northernAustralia and Vanuatu. The relationship of megalotisto other Hipposideros species is far from clear. Hill(1963) interpreted mcgalotis to be an early butspecialized offshoot of the Hipposideros lineage andthe sole living member of the megalotis group, acombined mcgalotis-bicolor group comprising hisDivision 1 of Hipposideros. Legendre (1interpreted megalotis to be a derived offshoot of theBrachipposideros radiation, resurrecting the subgenusSyndesmotis for megalotis and the Miocene H. ve/usfrom North Africa and the Pliocene H. sp. cf. H.vetlls from southern France (Lavocat 1961; Sige1976; Legendre 1982). Hand and Kirsch (1998), in

303

their phylogenetic analysis of fossil and livinghipposiderids, found no clear division between thebicolor and megalo/is groups; in some analysesmegalo/is appeared to be among the mostplesiomorphic of Hipposidcros taxa (figure Ib) whilein others it nested among bicolor taxa (figures laand 2b). Additionally, they found no specialrelationship between H. mcgalotis and theAustralian Brachipposideros- I<.h inonic/cris lineage.Bogdanowicz and Owen (1998) also failed toresolve the relationships of megalotis in theirphylogenetic analysis of 57 living hipposiderid taxa,although in some trees (e.g. figure 3) megalotisclustered with H. sabanlls (among other bicolor taxa).In no case, however, did it group with Rhinonicterisallrantills.

Hill (1963: 15) described the megalotis-bicolorgroup as a wide variety of loosely-allied species,few of them widespread, and many apparentlyrepresenting independent lineages. He interpretedthem to be the most primitive group in the genusand the group from which the other morespecialized sections had been derived. Members ofthis group are generally smalL with broad, usuallyrounded, ears, typically a simple noseleaf, anelongate skull with moderately inflated braincase, anarrow rostrum and an unspecialized auditoryregion. Hand and Kirsch (1998) concluded that thegenus Hipposideros is probably paraphyletic (seealso Hugueney 1965; Sige 1968; Legendre 1982;Bogdanowicz and Owen 1998) with species ofAsel/ia, Palaeophyliophora and Pseudorhinolophllsnesting within it (these three commonly beingassociated with cyclops taxa). They agreed, however,that Hill's (1963) Division 1 taxa (the megalotis­bicolor group) was probably the most plesiomorphicwithin Hipposideros.

The new Pliocene hipposiderid from Rackham'sRoost appears to be a member of the megalotis­bicolor group. However, although H. winsbllryorumis perhaps closer in overall dental morphology toH. megalotis and H. vetus than bicolor taxa such asH. sabanlls, the lack of complete dentaries and/orassociated upper canines precludes a more preciseplacement of this Riversleigh hipposiderid withinthe megalotis-bicolor group at this stage. Furthercomplicating placement is the fact that this speciesexhibits a complex mix of plesiomorphic andapomorphic features. For example, the small heelsize of M12 is probably a plesiomorphic feature,while the lack of a prominent anterolingualcingular cusp on the broad p 4 is possibly derivedamong hipposiderids (see Hand 1997a; Hand andKirsch 1998). The loss of p2 is a derived feature forhipposiderids, evidently independently acquired ina number of separate lineages (e.g., in species ofAsel/ia and Cloeotis) including perhaps thewinsbllryonml, megalotis and sabanlls lineages. Thereduced M 3 is probably derived among

304

hipposiderids (e.g., in Hipposideros species) but ispossibly plesiomorphic in some subgroups (forexample, species of the H. cyclops group; see Hand1997a). The short palate with U-shaped anteriormargin and convergent toothrows are probablyalso derived features among hipposiderids butplesiomorphic among Division 1 taxa. The verylarge infraorbital foramen appears to be anautapomorphy of this species, although the roundshape could be plesiomorphic (being found inrhinolophid and megaderrnatid sister-groups).

The unique combination of features exhibited byH. winsburyorum precludes it from direct ancestryof any living Australasian members of the bicolorgroup (or indeed any hipposiderid living in theAustralian Region) including H. ater, a commoncave-dwelling species in the Riversleigh regiontoday. It is likely that skull material will berequired to more precisely determine therelationships of H. winsburyorum. Hand and Kirsch(1998) found in their phylogenetic analyses of 37hipposiderid taxa that although dental features (20characters) alone were insufficient to resolverelationships among hipposiderid genera andspecies groups, a combination of cranial and dentalfeatures (56 characters) provided greaterresolution. Ongoing taxonomic revisions ofIndonesian Hipposideros by Kitchener and co­authors (e.g., Kitchener et al. 1992; Kitchener andMaryanto 1993; Kitchener and Maharadatunkamsi1995; Kitchener et al. 1996) have highlighted thetaxonomic problems still to be resolved in thisgenus.

Hipposideros winsburyorum is the second species ofHipposideros described from the Australian Tertiaryand one of few recorded from the Tertiaryworldwide. Hipposideros bernardsigei was describedfrom Riversleigh's early Miocene Neville's GardenSite on the basis of near-complete skull materialand is the first and only known fossil member ofthe cyclops group (Hand 1997a). Hipposideros felixhas been described on the basis of an upper thirdmolar and lower first molar from early Miocenedeposits at Li Mae Long, Thailand (Mein andGinsburg 1997). From East African Miocenedeposits, Butler (1969, 1984) recorded threeunnamed hipposiderids, one of which wasreported to be very similar to the EuropeanPseudorhinolophus bouziguensis. The living H.commersoni and Triaenops persicus, and a smaller,now locally extinct, unidentified hipposiderid,have been reported from the Pliocene ofMadagascar (Sabatier and Legendre 1985). NorthAfrica's H. (5yndesmotis) vetus and France's H. (5.)sp. cf. H. vetus (Lavocat 1961; Sige 1976; Legendre1982) are the only other records of Hipposiderosfrom the Tertiary.

Compared with the early Eocene appearance ofEuropean species of Pseudorhinolophus and

S.}. Hand, H. Godthelp

Palaeophyllophora, Hipposideros species appearrelatively recently in the world hipposideridrecord. However, this is possibly an artefact of theTertiary fossil bat record. Hand and Kirsch(1998) found Eocene Pseudorhinolophus andPalaeophyllophora to be relatively derivedhipposiderids nesting within Hipposideros. Basal tothis Hipposideros-dominated clade was a groupcontaining widely distributed species ofBrachipposideros, Rhinonicteris, Coelops, Cloeotis andTriaenops among others. If the phylogeneticinterpretation by Hand and Kirsch is correct, thenrepresentatives of almost all hipposiderid lineages,including the bicolor group, might be expected tobe found in the early Tertiary record, with theirearliest members possibly in the Asian orAustralasian region. Sige (1977) and Sige andLegendre (1983) have suggested that the origin ofthe family was outside Europe, probably in theSouthern Hemisphere (Sige 1991). Unfortunately,the Asian fossil bat record is poor and there is aprofound gap in the Australian fossil mammalrecord from the earliest Eocene to late Oligocene.Nevertheless, the appearance of a member of thehighly specialized H. cyclops group in theAustralian early Miocene (Hand 1997a) indicatesthat species of Hipposideros had already radiated bythat time.

Hipposideros almost certainly had its primaryradiation in Asia or Australasia. The bicolor anddiadema groups are interpreted to bepredominantly Indo-Australian-centred (e.g. Hill1963), while most members of the cyclops group areendemics of Australo-Papua (although the roots ofthe latter specialized group possibly lie in Africa:Hill 1963; Flannery and Colgan 1993; Hand 1997a;Hand and Kirsch 1998). The genus appears to haveradiated rapidly and widely with many closelyrelated species pairs living in Asia and Africa (Hill1963). There appear to have been few barriers todispersal for hipposiderids in the middle and laterTertiary as indicated by the wide distribution ofspecies of Brachipposideros and related taxa (Sige etal. 1982, 1994; Hand 1997b, 1998a, 1998b). The longhistory and wide distribution of the H. megalotislineage, from contemporary East Africa to middleMiocene North Africa and Pliocene Europe (andperhaps even Australia), is further evidence of thisdispersal capacity.

By Riversleigh standards, Hipposideroswinsburyorum, known from only six maxillaryfragments from a single deposit, is poorlyrepresented. Most hipposiderids recovered fromRiversleigh's Oligo-Miocene deposits arerepresented by several if not dozens of specimens,typically skulls. The remains of these older, better­represented hipposiderids are interpreted to havegradually accumulated within or adjacent to thelimestone caves in which they roosted. Hipposideros

Australian Pliocene Hipposideros

willsbllryorllm, however, appears to have sufferedthe same fate as most Rackham's Roost vertebrates:to become the finely comminuted prey of the GhostBat Macroderma gigas.

ACKNOWLEDGEMENTS

Work at Riversleigh has been supported by theAustralian Research Council, the Department ofthe Environment, Sport and Territories, NationalEstate Programme Grants (Queensland),Queensland National Parks and Wildlife Service,the Australian Geographic Society, the LinneanSociety of New South Wales, lCI, the QueenslandMuseum and the University of New South Wales.Study of the Riversleigh bat material would nothave been possible without the support andencouragement of Professor Michael Archer of theUniversity of New South Wales. The followingpeople kindly provided access to comparativespecimens in their institutions: B. Engesser, H.Felten, T.F. Flannery, W. Fuchs, L. Gibson, J.E. Hill,M. Hugueney, P. Jenkins, D.J. Kitchener, K.F.Koopman, P. Mein, R. Rachl, B. Sige, N.B.Simmons, G. Storch, M. Sutermeister, and S. VanDyck. Constructive criticism of an earlier draft ofthe paper was provided by B. Sige, N.K. Cooperand A. Baynes. We thank Jenni Brammall for theSEM photographs which were produced in theUniversity of New South Wales ElectronMicroscopy Unit.

REFERENCESArcher, M., Godthelp, H., Hand, 5.). and Megirian, D.

(1989). Fossil mammals of Riversleigh, northwesternQueensland: preliminary overview ofbiostratigraphy, correlation and environmentalchange. The Australian Zoologist 25: 29-65.

Archer, M., Hand, 5.). and Godthelp, H. (1994).Riversleigh. The story of animals in ancient rainforests ofinland Australia: 1-264, reprinting with revisions,Reed Books, Sydney, New South Wales.

Archer, M. and Wade, M. (1976). Results of the Ray E.Lemley expeditions, part 1. The AllinghamFormation and a new Pliocene vertebrate fauna fromnorthern Queensland. Memoirs of the QueenslandMuseum 17: 379-397.

Bartholomai, A. (1978). The Macropodidae (Marsupialia)from the Allingham Formation, northernQueensland; results of the Ray E. Lemleyexpeditions, part 2. Memoirs of the QueenslandMuseum 18: 127--143.

Bogdanowicz, Wand Owen, RD. (1998). In theMmotaur's labyrinth: a phylogeny for theHipposideridae. In Kunz, T. (ed.), Bats: phylogeny,morphology, echolocation, and conservation biology: 27­42, Smithsonian Institution, Washington, D.C., USA.

Boles, W.E. (1998). A Budgerigar Melopslttacus undulatusfrom the Pliocene of Riversleigh, north-westernQueensland. Emu 98: 32-35.

305

Butler, P.M. (1969). Insectivores and bats from theMiocene of East Africa; new material. In Leakey,LSB. (ed.), Fossil vertebrates of Africa, 1: 1-37,Academic Press, New York, N.Y., USA.

Butler, P.M. (1984). Macroscelidea, Insectivora andChiroptera from the Miocene of East Africa.Palaeovertebrata 14: 117-200.

Flannery, T.F. and Colgan, D.). (1993). A new speciesand two new subspecies of Hipposideros (Chiroptera)from western Papua New Guinea. Records of theAustralian Museum 45: 43-57.

Godthelp, H. (1987). Riversleigh scene 4: Rackham's Roost- the beginnings of the modem world. In Hand, 5.).and Archer M. (eds), TIle Antipodean Ark: 81-83, Angusand Robertson, Sydney, New South Wales.

Godthelp, H. (1997). Zyzomys rackhami sp. novo (Rodentia,Muridae), a rockrat from Pliocene Rackham's RoostSite, Riversleigh, northwestern Queensland. Memoirsof the Queensland Museum 41: 329-333.

Gray, ).E. (1866). A revision of the genera ofRhinolophidae, or horseshoe bats. Proceedings of theZoological Society, London 1866: 81-83.

Hand, 5.). (1993). First skull of a species of Hipposideros(Brachipposideros) (Microchiroptera: Hipposideridae),from Australian Miocene sediments. Memoirs of theQueensland Museum 31: 179-192.

Hand, S.j. (1995). First record of the genus MegadermaGeoffroy 1810 (Microchiroptera: Megadermatidae)from Australia. Palaeovertebrata 24: 47-66.

Hand, 5.). (1996). New Miocene and Pliocenemegadermatids (Mammalia, Microchiroptera) fromAustralia, with comments on broader aspects ofmegadermatid evolution. Geobios 29: 365-377.

Hand, 5.). (1997a). Hipposideros bernardsigei, a newhipposiderid (Microchiroptera) from the Miocene ofAustralia and a reconsideration of the monophyly ofrelated species groups. Miinchner GeowissenschaftlicheAbhandhmgen A 34: 73-92.

Hand, 5.). (1997b). New Miocene leaf-nosed bats(Microchiroptera: Hipposideridae) from Riversleigh,northwestern Queensland. Memoirs of the QueenslandMuseum 41: 335-349.

Hand, 5.). (1997c). Miophyllorhina riversleighensis gen. etsp. nov., a Miocene leaf-nosed bat (Microchiroptera:Hipposideridae) from Riversleigh, Queensland.Memoirs of the Queensland Museum 41: 351-354.

Hand, 5.). (1998a). Xenorhinos, a new genus of Old Worldleaf-nosed bats (Microchiroptera: Hipposideridae)from the Australian Miocene. Journal of VertebratePaleontology 18: 43G-439.

Hand, 5.). (1998b). Riversleigha williamsi gen. et sp. nov., alarge Miocene hipposiderid (Microchiroptera) fromRiversleigh, Queensland. Alcheringa 22: 259-276.

Hand, S]. and Kirsch, ).A.W. (1998). A southern originfor the Hipposideridae (Microchiroptera)? Evidencefrom the Australian fossil record. In Kunz, T. (ed),Bats: phylogeny, morphology, echolocation, andconservation biology: 72-90, Smithsonian Institution,Washington, D.C., USA.

Hill, I.E. (1963). A revision of the genus Hipposideros.Bulletin of the British Museum (Natural History), Zoology11: 1-129.

306

Hugueney, M. (1965). Les chiropteres du Stampiensuperieur de Coderet-Branssat. Documents duLaboratoire geologique de la Faculte des Sciences de Lyon9: 97-127.

Kitchener, D.J., How, R.A., Cooper, N.K. and Suyanto, A.(1992). Hipposideros diadema (ChiropteraHipposideridae) in the Lesser Sunda Islands,Indonesia: taxonomy and geographic morphologicalvariation. Records of the Western Australian Museum16: 1-60.

Kitchener, D.J., Konishi, Y. and Suyanto, A. (1996).Morphological variation among eastern Indonesianisland populations of Hipposideros bicolor (Chiroptera:Hipposideridae), with descriptions of three newsubspecies. Records of the Western Australian Museum18: 179-192.

Kitchener, D.J. and Maharadatunkamsi (1995). TheHipposideros bicolor group (Chiroptera,Hipposideridae) from Sumbawa Island, NusaTenggara, Indonesia. Records of the Western AustralianMuseum 17: 309-314.

Kitchener, D.J. and Maryanto, 1. (1993). Taxonomicreappraisal of the Hipposideros larvatus speciescomplex (Chiroptera: Hipposideridae) in the Greaterand Lesser Sunda Islands, Indonesia. Records of theWestern Australian Museum 16: 119-173.

Koopman, K.F. (1994). Chiroptera: systematics. Handbookof Zoology, VIII, 60, Mammalia: 1-217, Waiter deGruyter, Berlin, Germany.

Lavocat, R. (1961). Le gisement de Vertebres miocenes deBeni Mellal (Maroc). Etude systematique de la faunede Mammiferes et conclusions generales. Notes etMemoires de Geologie Maroc 155: 1-121.

Legendre, S. (1982). Hipposideridae (Mammalia:Chiroptera) from the Mediterranean Middle and LateNeogene and evolution of the genera Hipposiderosand Asellia. Journal of Vertebrate Paleontology 2: 386­399.

Mein, P. (1958). Les mammiferes de la faunesiderolithique de Vieux-Collonges. Nouvelles Archivesdu Museum d'Histoire naturelle de Lyon 5: 1-122.

Mein, P. and Ginsburg, L. (1997). Les mammiferes dugisement miocene inferieur de Li Mae Long,Thai:land: systematique, biostratigraphie etpaleoenvironnement. Geodiversitas 19: 783-844.

Peters, W. (1871). Uber die Gattungen und Arten der

S.J. Hand, H. Godthelp

Hufeisennasen. Rhinolophi. Monatsberichte derKonigliche Preussische Akadamie der Wissenschaften zuBerlin 1871: 301-332.

Sabatier, M. and Legendre, S. (1985). Une faune arongeurs et chiropteres plio-pleistocenes deMadagascar. 11(1 Congres national des Societes savantes,Sciences 4: 21-28.

Sige, B. (1968). Les chiropteres du Miocene inferieur deBouzigues. 1. Etude systematique. Palaeovertebrata 1:65-133.

Sige, B. (1976). Les Megadermatidae (Chiroptera,Mammalia) miocenes de Beni Mellal, Maroc. Geologiemediterraneenne 3: 71-86.

Sige, B. (1977). Les insectivores et chiropteres d uPaleogene moyen d'Europe dans l'histoire des faunesde mammiferes sur ce continent. Jurij A. OrlovMemorial, Journal of the Palaeontological Society of India20: 178-190.

Sige, B. (1991). Rhinolophoidea et Vespertilionoidea(Chiroptera) du Chambi (Eocene inferieur deTunisie). Aspects biostratigraphique, biogeo­graphique et paleoecologique de l'origine deschiropteres modernes. Neues Jahrbuc11 flir Geologie undPaliiontologie Abhandlungen 182: 355-376.

Sige, 8., Hand, S.J. and Archer, M. (1982). An AustralianMiocene Brachipposideros (Mammalia, Chiroptera)related to Miocene representatives from France.Palaeovertebrata 12: 149-171.

Sige, B. and Legendre, S. (1983). L'histoire despeuplements de chiropteres du bassinMediterraneen: l'apport compare des remplissageskarstiques et des depots fluvio-lacustres. Memoires deBiospeologie 10: 209-225.

Sige, 8., Thomas, H., Sen, S., Gheerbrandt, E., Roger J.and Al-Sulaimani, Z. (1994). Les chiropteres deTaqah (Oligocene inferieur, Sultanat d'Oman).Premier inventaire systematique. MunchnerGeowissenschaftliche Abhandlungen A 26: 35-48.

Tate, G.H.H. (1941). A review of the genus Hipposideroswith special reference to Indo-Australian species.Bulletin of the American Museum of Natural History 78:353-393.

Manuscript received 6 February 1998; accepted 30 September1998.