https://doi.org/10.24199/j.mmv.1978.39.04 7 July 1978

PROPLEOPUS CHILLAGOENSIS, A NEW NORTH QUEENSLAND SPECIES OF EXTINCT GIANT RAT-KANGAROO (MACROPODIDAE: POTOROINAE)

By Michael Archer,* Alan Bartholomai* and Larry G. Marshall!

* Queensland Museum t Princeton University

Abstract

Propleopus chillagoensis is ^ described as a second species in the genus. The genus is rediagnosed. The new species differs markedly in details of premolar and molar morphology from the Pleistocene P. oscillans (De Vis), but reveals no additional information about inter- generic relationships. In the unique characters which differentiate it from P. oscillans, it is less like structurally ancestral forms such as Hypsiprymnodon than is P. oscillans, and could be a descendent of the lineage leading to P. oscillans. A specimen of Propleopus from the Pleistocene Wellington Caves is similar to P. chillagoensis, and suggests the possibility that the two species, or lineages leading to them, were contemporaneous during part of the Pleistocene, and probably represent end members of a late Tertiary radiation.

Introduction nomenclature for individual premolar teeth, as Propleopus oscillans was described on the used elsewhere (e.g. by Bartholomai, 1972), is basis of a single dentary fragment by De Vis not used here, but not because it has been (1888). Since then a lower incisor and two shown to be any less appropriate than the sys- additional dentary fragments have been re- tem used by Archer (1974) and some other ported (Woods, 1960 and Tedford, 1955), as authors (e.g. Mahoney and Ride, 1975). well as a maxillary fragment (Bartholomai, 1972). All of this Pleistocene material has Rediagnosis of Propleopus Longman, 1924 been referred to P. oscillans with little hesita- Recognition of a second species broadens tion. the concept of Propleopus as follows: Poto- Gill (1953, 1957) described a tooth from the roine macropodids differing from other poto- Pliocene Grange Burn local fauna which Ride roine genera in unique extent of subhorizontal (1964) and Turnbull and Lundelius (1970) wear facet on Ia ; also differ from Hypsypri- regard as similar to but smaller than P. oscil- mnodon in having enamel on Ij covering only lans. lower half of labial and lingual faces; posterior

The present material, comprising a maxillary half of P4 shorter crowned; M x trigonid wide; fragment and two isolated premolars repre- differing from Caloprymnus and Potorous in sents a second very distinct species. The exis- distribution of enamel of Ii, larger more finely tence of at least two distinct forms of Propleo- serrated and arcuate plagiaulacoid premolars, pus in the Pleistocene indicates that there was in having a vestigial tooth immediately pos- probably a late Tertiary radiation of the genus. terior to Ii, and in having lingual cingula Registration numbers are those of the (sometimes as rudiments) on upper molars; Queensland Museum (e.g. F6675, fossil col- differing from Aepyprymnus in having more lection; e.g. J6824, modern mammal collec- arcuate plagiaulacoid premolars, poorly- tion), the National Museum of Victoria (e.g. developed lingual cusp or crest develop- 4 P15917), the American Museum of Natural ment on P , wider molars, poorly-developed History (e.g. AMNH 65279) and Museum of transverse lophs on molars; vestigial Palaeontology, University of California (e.g. tooth immediately posterior to l u obliquely set UCMP 51697). plagiaulacoid premolars, and lingual cingula Dental terminology follows that used else- (sometimes rudiments) on upper molars; also where by Archer (1974, 1976). An alternative differing from Bettongia in having more arcu-

55 56 MICHAEL ARCHER, ALAN BARTHOLOMAI, and LARRY G. MARSHALL ate, mcsially swollen plagiaulacoid premolars confluent with occlusal cutting crest of tooth. with narrow posterior ends on permanent pre- Posterior to this anterior crest are eight pairs molars, and a vestigial tooth immediately pos- (lingual and labial) of prominent vertical terior to Ix. ridges which reach the occlusal edge of tooth. Ninth pair fails to extend as far as occlusal Propleopus chillagoensis sp. nov. edge. Longitudinal spacing between ridges de- (Plate 11, fig. 1) creases posteriorly. Bases of second pair of Holotype: PI 59 17, right maxillary fragment ridges bifurcated. Bases of eighth and ninth 4 with erupting P and erupting and broken M\ buccal ridges confluent. Ninth ridges more fragment of 1 and 2- 3 M , M , donated by Mr massive than others and form shallow vertical Hugh C. Mainwaring in 1910. fossettes between themselves and eighth Referred Material: PI 59 isolated 18, LP4 ; ridges. Relative ridge lengths as follows 3 sub- P15919, isolated and broken 1 RP ; both from equal to 4, 2 subequal to 5, 6, 7, 8, 9. Eighth type locality. ridge interrupted by gap on each side. Base of Type Locality: Fissure deposit in Chillagoe crown lacks cingula. Judging by position of Formation, Chillagoe District, N.E. Queens- erupting crown, P 4 probably oriented antero- land. According to a letter (National Museum labially, with respect to long axis of molar row. of Victoria) from Mr Hugh C. Mainwaring to 4 P4 : Identical to P except as follows: Pos- Professor Baldwin Spencer, 21 July 1910, the terior edge of crown keeled with non-serrated specimens came from \ . . a bluff 800 ft. long, vertical continuation of occlusal cutting crest; 300 ft. wide and 50 ft. above the plain of the widest part of crown closer to middle of tooth; surrounding country. This bluff is being worked ninth ridges less massive than corresponding [1908-11] as a quarry to supply the smelter 4 ridge in P ; base of second ridge not bifurcated; with flux. In breaking down the stone a number base of second and third lingual ridges con- of small caves were found. These were fitted fluent. with material resembling silicified clay in 1 M : Crown represented by only posterior one- which the fossils were imbedded'. Mr William third. Roots indicate anterior one-half of Morrow (pers. comm.) notes that the locality tooth much wider than posterior one-half, and was 'Smelter's Junction, one mile from Chilla- that lingual one-half of tooth longer than goe Railway Station, along the Mungana line, labial one-half, result of evidently very large on the left hand side on the way to Mungana.' protocone. Remnant of posthypocrista and Origin of Name: Reference to type locality. postmetacrista linked in arc, middle of which Diagnosis: Differs from the only other species functions as a posterior cingulum. Base of in the genus, Propleopus oscillans, in having a 4 crown very swollen, such that transverse dia- markedly larger P and P4 , each with nine ver- meter of crown base below hypocene and meta- tical ribs; anterior ribs on P 4 are rectilinear cone estimated at three times distance between rather than curved; P 4 crown base more steeply those cusps. inclined posteriorly; upper molars shorter, with 2 more swollen bases; lingual cingulum around M : Anterior one-half of crown wider than pos- 2~ 3 terior one-half. Lingual and labial sides of base of protocone on at least M ; preproto- crista connects directly with preparacrista; pos- crown subequal in length. Protocone, hypo- terior half of upper molar crowns markedly cone and paracone subequal in height, and narrower than anterior half. taller than metacone. Preprotocrista and pre- 4 paracrista connect as anteriorly arc. Description: P : Two-rooted tooth; occlusal an convex outline oval, widest at anterior one-half; an- Postimetacrista and posthypocrista joined as in 1 terior end gently convex, posterior end blunt. M . Prehypocrista and postprotocrista linked in Occlusal cutting edge gently convex, ventrally barely obtuse angle. Postparacrista and pre- with maximum convexity occurring towards metacrista become irregular towards middle of posterior one-half of edge. Anterior edge of crown, and are only just linked by low lon- tooth has curved, labially convex, vertical crest gitudinal crest. Prominent, transverse lingual EXTINCT GIANT RAT-KANGAROO PROPLEOPUS 51 ridges descend flanks of paracone and meta- flank of hypocone to posthypocrista near its cone from longitudinal midline of crown. labial end. M 3 less worn than M 2 and reveals Poorly-defined transverse labial ridge descends fine crenulations over much of crown surface flank of hypocone and meets lingual rib of enclosed by principal cusps. Also evidence of metacone in longitudinal midvalley. Very accessory vertical crenulations parallel to la- poorly-defined anterolingually directed labial bial ridges of protocone and hypocone, and ridge descends protocone, but intersects the connecting posthypocrista with longitudinal preprotocrista anterior to the protocone. La- midvalley of crown. biad to point of intersection of postparacrista 4 M : Only anterior one-half of crown remains. and premetacrista are two small adjacent cusps From remnant of base of posterior one-half, along labial end of transverse midline. Prom- evident that anterior one-half wider than pos- inent lingual anterolingually inclined ridge terior one-half. Protocone taller than paracone. ascends flank of hypocone to base of proto- Preprotocrista connected to preparacrista as cone, adjacent to small basal lingual cuspule, 3 in M . Postparacrista not connected to pre- which in turn is adjacent to lingual end of an- metacrista. Lingual ridge of paracone promin- terolingual basal cingulum which terminates ent and extends transversely to longitudinal 1 2 labially at point where abuts against . M M midvalley to meet labial ridge of protocone. Very small cingular bulge occurs on postero- Some evidence for lingual protocone ridge as lingual base of hypocone. 3 in M . Tooth erupting so anterior cingulum not 3 M : Anterior one-half of tooth much wider visible, if present. Edge of enamel near pos- than posterior one-half. Lingual one-half terior base of protocone suggests lingual hy- subequal in length to labial one-half. Proto- 3 pocone ridge present as in M . Very poorly- cone subequal in height to hypocone, aller developed vertical ridge occurs on anterolabial than paracone, and much taller than re- flank of paracone linking preparacrista with duced Prehypocrista, postproto- 3 metacone. base of crown. As in M , clear indications of crista, preprotocrista and preparacrista linked accessory crenulations occur in crown valleys 2 in M . Posthypocrista does not contact ves- and on flanks of cusps. tigial postmetacrista, being separated by Meristic changes along tooth row : Pos- posterolabial depression. Poorly developed teriorly, teeth become relatively narrower and premetacrista and postparacrista only just longer; metacone markedly reduces in size; fail in transverse midvalley of to contact paracone reduces in size but at a slower rate; para- tooth. Prominent lingual rib descends crests linking metacone to other crests become cone in posterolabial direction from longitu- smaller; transverse midvalley of crown be- poorly de- dinal midline of crown. Much more comes more open at labial end; bases of pro- labial ridges flanks veloped linqual and descend tocone and hypocone become less swollen; of metacone and hypocone but just may not distance between hypocone and metacone meet each other in longitudinal midvalley. diminishes. Tooth eruption sequence indicated proto- 4 4 Poorly-developed labial ridge descends suggests P and M erupt at approximately the 2 cone in same manner as in M . Prominent same time. lingual rib ascends flank of hypocone to base Discussion: The type specimen of Propleopus of protocone. Vertical lingual ridge, also des- oscillans is a dentary while that of P. chillago- cends from base of protocone to point halfway ensis is a maxillary fragment. A possibility down its flank. Base of this ridge vaguely con- therefore exists that both specimens represent tacts oblique lingual hypocone ridge and an- P. oscillans. We have considered this possi- terolingual basal cingulum on anterior flank of bility and discounted it for the following rea- protocone. Anterolingual cingulum terminates sons. First, Bartholomai (1972) has described labially just short of point at which M2 abuts F6675 from the Darling Downs, a maxillary 2 4 2 . M3 Poorly-developed (but better than M ) fragment with LP -M , and referred it to P. posterior cingulum descends from posterior oscillans. It is specifically distinct from the 58 MICHAEL ARCHER, ALAN BARTHOLOMAI, and LARRY G. MARSHALL

Chillagoc maxillary fragment, for reasons noted in the diagnosis. In addition, although 1 P of the type of P, chillagoensis is larger than

( P of F6675, its molars are proportionately smaller. For this reason, the size differences in the two maxilla cannot be the result of allo- metry.

Additional evidence for reference of F6675 to Propleopus oscillans comes from the num- ber of ridges on upper and lower adult pre- 0.96 1.10 molars of living potoroines. In modern Aepy- prymnus rufescens, of twenty-three 1.25 specimens examined with relatively unworn premolars, nine have the same number of ridges (seven or eight) on the upper and lower premolars, while the remainder differ by only one ridge. Varia- tion in ridge 4 number is as follows: P : one 1.50 has seven ridges; thirteen have eight; and eight have nine; P.,: three have seven ridges; thirteen have eight; and three have nine. In Hypsiprimnodon moschatus, of all ten speci- mens with relatively unworn adult premolars, 4 P has six and P4 seven ridges. Both upper and 1.47 the single lower premolars of P. chillagoensis have nine ridges, while the P 4 of the holotype B of P. oscillans and P 4 of the referred specimen both appear to have seven ridges. The con- stancy of ridge number in Hypsiprimnodon, the consistently different ridge number in P. chilla- goensis from the holotype of P. oscillans, and 0.97 the apparently close relationship of Propleopus to Hypsiprimnodon (see below) which has a species constant number of ridges, convince us of the correctness of Bartholomai's (1972) re- ference of F6675 to P. oscillans. Added to the other morphological 0.81 characters noted in the 0.69 diagnosis, it also convinces us of the distinction between the two species of Propleopus.

Further, but less convincing evidence for the association suggested by Bartholomai (1972), is the fact that both the holotype of P. oscillans and F6675 come from the Darling Downs of southeastern Queensland, while the specimens of P. chillagoensis come from far northeastern Queensland.

These data, as well as the close approxima- Figure 1—Measurements of specimens of Propleo- tion of upper and lower molar sizes, support pus chillagoensis sp. nov. A-C, P15917, holotype. D-E, Bartholomai's (1972) reference of F6675 P15918, . to LP4 F, P159I9, 4 RP . Measurements are in centimetres. Propleopus oscillans and convince us that the EXTINCT GIANT RAT-KANGAROO PROPLEOPUS 59

Chillagoe material with its nine vertical ridges, of one is likely to be reflected by reduction of and relatively short and wide premolars can- the other. Its taxonomic position is considered not be referred to P. oscillans. as doubtful until better material becomes avail- Most authors who have ^ considered the affini- able. ties of Propleopus oscillans regard it to be Tedford (1967) concludes that the late Pleis- closely related to Hypsiprymnodon, a conclu- tocene, Lake Menindee, New South Wales sion with which we generally agree, although specimen (UCMP51697) of Propleopus is refer- some characters more closely resemble Betton- able to P. oscillans, a conclusion with which gia than Hypsiprymnodon. P. chillagoensis re- we agree. A second tooth fragment from Men- veals no characters that more closely ally it indee (UCMP51698) was compared by Ted- with Hypsiprymnodon or Bettongia than with ford (1967) with the Chillagoe material. It is P. oscillans. The characters which differentiate likely that this specimen is also referable to it from P. oscillans are unique derived charac- P. oscillans. ters and are of neutral value in assessing inter- These occurrences suggest the possibility generic relationships. Because P. oscillans re- that the lineages resulting in Propleopus oscil- tains more structurally pr vitimiecharac lans and P. chillagoensis may have been con- tains more structurally primitive character temporaneous during the Pleistocene and rep- states (i.e. fewer ridges on P\ relatively un- resent parts of a late Tertiary radiation of the swollen molars, poorly-developed lingual cin- genus. gula, and unreduced metacones on the upper molars), and has no derived characters of its Acknowledgements own which are not also present in P. chilla- We thank Dr T. Rich (National Museum of goensis and hence diagnostic of the genus as a Victoria) for making the Chillagoe specimens whole, it can be regarded as structurally, if not available to us. The late Professor R. A. Stir- actually ancestral to P. chillagoensis. ton provided the Queensland Museum with a The precise age of the eastern Darling cast (F3287) of the Wellington Caves speci- Downs specimens within the Pleistocene is in men. Mr A. Easton and Mr R. Burgess doubt. Similarly, the age of the Chillagoe speci- (Queensland Museum) produced the photo- mens is in doubt but the presence of Sarco- graphs for the plate. Ms J. Utz and Ms R. philus, suggests that the Chillagoe deposit is Owens (Queensland Museum) typed drafts of no older than Pleistocene (Bartholomai and the paper. Mr S. Van Dyke (Queensland Mu- Marshall, 1973). Tedford (1967) mentions seum) assisted in study of premolar ridge num- that the material described here as P. chilla- bers in modern potoroines. goensis is associated with Sarcophilus and small macropodines apparently related to liv- References ing species. Archer, M., 1974. The development of the cheek- The Propleopus specimen (UCMP 45171) teeth in Antechinus flavipes (Marsupialia, Dasyu- ridae). /. R. Soc. West. figured by Tedford (1967, fig. 5b) as P. oscil- Aust. 51: 54-63. , 1976. Phascolarctid origins and the po- lans from Pleistocene deposits in Wellington tential of the selenodont molar in the evolution Caves, New South Wales, although only a frag- of diprotodont marsupials. Mem. Qcl Mus. 17: mentary dentary lacking P4 , shows in some re- 367-71. spects, the kinds of characters which might Bartholomal, A., 1972. Some upper cheek teeth in Propleopus be expected in lower molars of P. chillagoen- oscillans (Dc Vis). Mem. Qd Mus. 16: 211-213. sis. Compared with P. oscillans, the anterior , in press. The fossil vertebrate fauna from half of Mi is relatively wider, the molars are Pleistocene deposits at Cement Mills, Gore, somewhat shorter, and most importantly, the south-eastern Queensland. Mem. Qd Mus. 18. Bartholomai, A., and L. G. Marshall, 1973. talonid of M 4 is narrower as noted by Ted- The identity of the supposed dasyurid marsupial, ford (1967) and its entoconid is more reduced. Sarcophilus prior De Vis, 1883, with comments The entoconid is the occlusal counterpart of on other reported 'Pliocene' occurrences of Sarco- the metacone in the upper molars. Reduction philus. Mem. Qd Mus. 16: 369-374. 60 MICHAEL ARCHER, ALAN BARTHOLOMAI, and LARRY G. MARSHALL

De Vis, C. W., 1888. On an extinct genus of the , 1967. The fossil Macropodidae from marsupials allied to Hypsiprymnodon. Proc. Lake Menindee, New South Wales. Univ. Calij. Linn. Soc.N.S.W.I: 5-8. Publ. Geol. Sci., 64: 1-156. Lundelius, Jnr., 1970. Gill, E., 1953. Australian Tertiary marsupials. Aust. Turnbull, W. D., and E. Pliocene mamma- J.Sci. 16: 106-108. The Hamilton fauna—a late lian fauna from the Grange Burn, Victoria, , 1957. The stratigraphical occurrence and Australia. Fieldiana (Geology) 19: 5-84. palaeoecology of some Australian Tertiary mar- Woods, J. T., 1960. The genera Propleopus and supials. Mem. Nat. Mus. Vic. 21: 135-203. Hypsiprymnodon and their position in the Macro- Mahoney, J. A. and W. D. L. Ride, 1975. Index podidae. Mem. QdMus. 13: 199-212. to the genera and species of fossil Mammalia described from Australia and New Guinea be- Explanation of Plate tween 1838 and 1968. Spec. Pubis. West. Aust. PLATE 11 Mus. 6: 1-250. Propleopus chillagoensis sp. nov. A, stereopair, oc- Ride, L., Australian fossil 4 4 W. D. 1964. A review of clusal view, P15917, holotype, RP -M . B, labial marsupials. /. R. Soc. West. Aust. 47: 97-131. view, P15917, holotype. C, stereopair, occlusal view,

. E, stereopair, Tedford, R. H., 1955. Report on the extinct mam- P15918, LP4 D, labial view, P15918, 4 malian remains at Lake Menindee, New South occlusal view, P15919, RP . F, labial view, P15919. Wales. Rec. S. Aust. Mus., 11: 299-305. White line is one cm in length. MEM. NAT. MUS. VICT. 39 PLATE 11