Mem. Qd Mus. 19(3): 299-307. pis. 2. [1979] w ABULAROO NAUGHTONI GEN. ET SPo NOVo, AN ENIGMA TIC KANGAROO (MARSUPIALIA) FROM THE MIDDLE TERTIARY CARL CREEK LIMESTONE OF NORTHWESTERN QUEENSLANDo RESULTS OF THE RAY Eo LEMLEY EXPEDITIONS, PART 4 MICHAEL ARCHER* Queensland Museum ABSTRACT Wabularoo naughtoni gen. et sp. nov. is a middle to late Miocene kangaroo from the Carl Creek Limestone of Riversleigh Station, northwestern Queensland. The dentary morphology is potoroid-like but the molars are lophodont and therefore macropodid-like; the very large plagiaulacoid premolar shares characters of both groups. Its systematic and phylogenetic position are obscure. Although it may be structurally ancestral to either the potoroids or macropodids it post dates the appearance of both families in the fossil record. Tedford (1967) summarizes the Riversleigh Wabularoo gen. nov. fauna from the Miocene Carl Creek Limestone noting three diprotodontid genera including TyPE SPECIF.5:Wabularoo naughtoni gen. et sp. Bematherium angulum Tedford 1967, and an nov. undeterJ:Ilined genus of kangaroos. In 1976, the GENERIC DIAGNOSiS: It differs from all genera author and museum assistants Messrs H. of the Potoroidae (Hypsiprymnodon. Propleopus. Godthelp and R. Kohout made further collections Potorous. Caloprymnus. Bettongia and from the Carl Creek limestone at the locality Aepyprymnus) in having lophodont molars. It referred to by Tedford ( 1967, figs. 1-2) as '0'. differs from all genera of the Macropodidae in These collections included two additional genera having a combination of a short, shallow, swollen of kangaroos, as well as crocodiles, birds, and dentary with a greatly enlarged masseteric canal diprotodontids. and a trenchant but wide and tall plagiaulacoid The new kangaroo described here was found in PJ. Of all known macropodid genera, it most an isolated block of the very hard Carl Creek closely resembles the monotypic Hadronomas but Limestone. Using an electric jack-hammer, it was it also differs from the single species of this genus collected as part of a smaller chunk of limestone. (H. puckridgei) as follows: the molars have Fine preparation was carried out in the laboratory narrower anterior cingula; the PJ is proportionate- using a compressed-air vibrotool. ly much wider and taller-crowned, obliquely oriented in the tooth row, narrowed posteriorly, Terminology of individual teeth follows Archer with recurved more numerous and finer serrations, (1978a) and that of crown morphology follows and no buccal or lingual cingula. Archer (1976a, b) or Bensley (1903). Registration Origin of the generic name: Wabula (wa'bula} numbers prefixed with Fare in the palaeon- means 'long-time-ago' in the Waanyi language as tological collections of the Queensland Museum. spoken by Ms Ivy George of Riversleigh Station; roo is a common non-specific Australian term for SYSTEMATICS Superfamily: MACROPODOIDEA *Present address: School of Zoology, University of New Family: Incertae sedis South Wales. 300 MEMOIRS OF THE QUEENSLAND MUSEUM a kangaroo. The generic name is regarded to be ORIGIN OF THE SPECIESNAME" In honour of Mr and Mrs E. Naughton, owners of Riversleigh Station, who masculine. graciously allowed us to work on the property as well as extended many kindnesses to us during our stay. Wabularoo naughtoni sp. nov. (Fig. I; PIs., I, 2) DIAGNOSIS: That of the genus until additionai species are known. HOLOTYPE-Queensland Museum F9177, broken right dentary with P3, M2_3' and part of M4. DESCRIPTION: The dentary is potoroine-like TYPE LOCALITY- From an isolated boulder of the resembling for example Bettongia and upper clastic arenaceous limestone member of the Carl Aepyprymnus in being short, heavy-bodied, with a Creek Limestone, site '0' (of Tedford 1967), Riversleigh marked inflection of the ventral border below the Station, northwestern Queensland. molar row, a prominent swelling of the lateral wall AGE- The absolute age is unknown but, based on faunal comparisons, the Carl Creek Limestone is below PJ, and a large and laterally swollen interpreted to be mid to late Miocene in age (Tedford masseteric canal. The point of inflection of the 1967, Archer and Bartholomai 1978). ventral border is below M4 as in some 30.8 5.1 4.8 4.9 4.5 \ \ ' , 9.8 16.1- 22.8. mJ \ ;m2\ / PJ" ,m4\8.0. 6.7-' 6.3-\.L-10.7 ~ r3.31 r3.2l r2.~ ./5~4 22.0 FiG I. Occlusal and buccal views of F9177, the holotype of Wabularoo naughtoni, showing measurements (in mm) and tooth nomenclature. In the buccal view, the intercusp distances 3.2,2.9 and 3.3 were measured between the metaconid (lingual end of the protolophid) and entoconid (lingual end of the hypolophid) of each lower molar. In measurements involving length of the broken M4, the posterior point is taken from the posterior edge of the posterior root. In measurements involving length of the PJ, the anterior edge is taken from the antero-most basal edge of the enamel. Molar widths were measured along axes passing transversely through the major cusps. Molar cusp heights involve only the protoconid (buccal end of the protolophid) and hypoconid (buccal end of the hypolophid). Abbreviations: m.c., masseteric canal in section; sym, posterior end of the symphysis. 302 MEMOIRS OF THE QUEENSLAND MUSEUM MJ: The morphology of MJ is as in M2 except as mid-Miocene Etadunna Formation (M. 0. follows: The whole tooth is larger; the protolophid Woodburne, pers. comm.) and at least one is subequal in length to the hypolophid; the high-crowned macropodid tooth is known from the protolophid is symmetrically concave anteriorly mid-Miocene Namba formation from the Frome and occlusally; the crest extending anteriorly from Embayment of South Australia (Archer and Rich, the metaconid to the anterior cingulid does not in preparation). These probably older occurrences have an inflection immediately anterior to the discount the first two of the above possibilities. metaconid; the anterior cingulid is lower such that There is at present insufficient information to its anterior edge is no higher than the midvalley of decide between the remaining three. the tooth; the trigonid is noticeably wider than the Speculations about kangaroo evolution and talonid and is the widest part of the tooth; the classification have been going on continuously anterior cingular shelf is longer; the swelling since Bensley's ( 1903) comparative study of the anterior to the entoconid is not as well-developed; teeth of marsupials. More recent speculation and there is a poorly-developed but distinct lingual concerning the relative primitiveness of the known vertical crest on the posterior flank of the subfamilies (e.g. Pearson 1950, Ride 1971, metaconid (this may also have been present in M2, Bartholomai 1972) has highlighted the fact that but the tooth is very worn in this area). there is still no agreement about whether potoroids were ancestral to macropodids, macropodids were M4: The hypoconid and posterior face of the ancestral to potoroids, or both groups were derived hypolophid are missing. The morphology of M4 is independently from a common ancestor referable as in MJ except as follows: There is a point of to neither group. There has even been renewed inflection between the anterolingual corner and interest and uncertainty about the composition of the end of the anterior cingulid and the base of the the subfamilial groups (e.g. Woodburne 1967, metaconid; the anterior flank of the entoconid is Kirsch 1968, Archer 1978b). Elsewhere (Archer not extended by a swelling, but rather projects 1978b) I have proposed a speculative rearran- gement of all kangaroo genera into two families: anteriorly. The Potoroidae containing the Hypsiprim- Meristic gradients along the tooth row: The nodontinae (Hypsiprimnodon and Propleopus) protolophid and trigonid increase in width from and the Potoroinae (all other potoroid genera); M2 to M4; the hypolophid and talonid also and the Macropodidae containing the Sthenurinae increase in width at least from M2 to MJ; the (possibly including the genera Sthenurus, preprotocristid increases in length from M2 to M4; Procoptodon. Setonix. Dorcopsoides. Dorcop- the anterior cingulid decreases in height from M2 sulus, Dendrolagus and Hadronomas) and the to M4; and the protoconid increases in mass from Macropodinae (containing all other previously M2 to MJ, but is subequal in MJ and M4° described genera). The Sthenurinae, in the expanded form used by Archer (1978b), is DISCUSSION essentially a plesiomorphic group and probably not monophyletic. WabuJaroo naughtoni is in most characters There are other different and as yet unnamed except molar morphology, a decidedly potoroid- middle Miocene Kangaroos, from central South like kangaroo. Its lophodont molars which also Australia (M. 0. Woodburne, pers. comm.) that lack posterior cingula are however decidedly cannot be referred to any of the previously non-potoroid-like characters. Its systematic and described subfamilies without significantly phylogenetic position within the MacroPo<ioidea altering the current concepts of these subfamilies. are therefore very much in doubt. There is a possibility that Wabularoo naughtoni WabuJaroo naughtoni could theoretically and these other aberrant taxa are referable to a represent anyone of at least five evolutionary fifth subfamily of kangaroos. I have not proposed stages: (I) a potoroid developing into a anew subfamily for W. naughtoni because of the macropodid; (2) a macropodid developing into a limited information provided by the single potoroid; (3) a specialized macropodid par;llleling dentary. It would be desirable to consider incisor potoroids; (4) a specialized potoroid paralleling and cranial structure as well, aspects of which are macropodids; (5) a representative of an as yet preserved in the Miocene material under study by unrecognized group equivalent in rank to known M. 0. Woodburne. kangaroo subfamilies but a derivative or ancestor The diprotodontids of the Riversleigh local of none of the other known groups. However. at fauna were concluded by Tedford (1967) to be least one undoubted potoroid occurs in the contemporaneous with or slightly younger than ARCHER: WABULAROONAUGHTONI.