A SPECIALISED THYLACINID, THYLACINUS MACKNESSI; (MARSUPIALIA: THYLACINIDAE) FROM MIOCENE DEPOSITS OF RIVERSLEIGH, NORTHWESTERN QUEENSLAND

JEANElTE MUIRHEAD

M uirhead, J ., 1992. A specialised thylacinid, Thylacinus macknessi, (Marsupialia: Thylacinidae) from Miocene deposits of Riversleigh, northwestern Queensland. Australian Mammalogy 15: 67-76.

Thylacinus macknessi is described from Miocene sediments of Riversleigh, northwestern Queensland. Comparisons with other thylacinids and dasyurids reveal it to be a new species of Thy/acinus. In most features it is as specialised as T. cynocepha/us and it is not considered to be ancestral to any other taxon. The presence of such a specialised thylacine in the Riversleigh deposits argues for a pre-Late Oligocene divergence of this group from the Dasyuridae.

Key words: Thylacine, 1h)'lacinus macknessi, Thylacinidae, Riversleigh, Tertiary, Queensland, Marsupialia.

I. Muirhead. Schoo/ of Bi%gica/ Sciences, University of New South Wa/es, PO Box I Kensington New South Wales 2033. Manuscript received /4 September 1991.

THE Thylacinidae is a small family consisting of a abbreviations used are: QMF, Queensland Museum recently extinct form Thy/acinus cynocepha/us Harris, palaeontological collection; AR, temporary catalogue and two Tertiary taxa. Although thylacinid premolars number in School of Biological Science, U niversity of have been recovered from the Miocene Wipajiri New South Wales. Measurements of tooth dimensions Formation of South Australia and the late Pliocene of 7: macknessi are presented in Table I. The regions Otibanda Formation of New Guinea (Campbell 1976, measured are presented in Fig. 1. Plane 1976,Archer 1982),only two distinct thylacinids have been described from the fossil record: the late Miocene Thy/acinus potens from the Alcoota Local SYSTEMATICS Fauna (Woodbume 1967) and the late Oligocene to middle Miocene Nimbacinus dicksoni from the Henk's Thylacinidae Bonaparte, 1838 Hollow Local Fauna, the Riversleigh Local Fauna Thylacinus Temminck, 1827 and the Bullock Creek Local Fauna (Muirhead and Thylacinus macknessi n. sp. Archer 1990). Thy/acinus potens is larger than T. HOLOTYPE cynocepha/us and slightly less specialised. It was considered by Archer ( 1982) to represent a sister QMFI6848a (formerly ARI6491), right dentary taxon to the modem form. Nimbacinus dicksoni is a portion with broken M3, complete M4 and M5. smail and very plesiomorphic thylacinid, evidently not far removed from the group's dasyuroid origins REFERREDSPECIMENS and a possible ancestor to Thy/acinus. Material QMFI6848b(formerlypartof ARI6491), right canine described here as Thy/acinus macknessi is of com- closely associated but not attached to holotype, parable age to the plesiomorphic Nimbacinus dicksoni but, in contrast, represents a specialised thylacine QMFI6849formerly (ARI1046) isolated left M4, and QMFI6850 formerly (AR4794), isolated right M2. whose closest relative appears to be the modem Thy/acinus cynocepha/us. TYPE LOCALITY, LocAL FAUNA AND AGE

Dental and taxonomic nomenclature follows that QMFI6848 Neville's Garden Local Fauna, Neville's of Muirhead and Archer (1990). Dental terminology Garden Site, Riversleigh Station, northwestern is also presented in Fig. I. Catalogue number Queensland; early t{) middle Miocene. QMFI6849, 68 AUSTRALIAN MAMMALOGY

Fig. J. Diagram showing terminology of morphology on an upper and lower molar of 7: macknessi. Regions used for measurements in Table I are also presented using the measurement for two teeth as examples taken from Table I. A = QMFI6850. B = QMFI6849. Arrow indicates anterior of molar. Star indicates region of lost metaconid. Abbreviations: a.c., anterior cingulum; co, cristid obliqua; end, entoconid; hyd, hypoconid; hyld, hypoconulid; mc, metaconid; mcl, metaconule; med, metac paracone; pacd, paracristid; pad, paraconid; p.c., posterior cingulum; phd, posthypocristid; pmc, postmetacrista; ppc, preparacrista; prd, protoconid. MUIRHEAD: THYLACINUS MACKNESSI FROM MIOCENE DEPOSITS IN QUEENSLAND 69

LOWER MOLARS MJ M4 M5 QMF number 16848 16848 16849 16848 Tooth length 9.10 9.00 8.75 Tooth width 4.38 4.92 4.30 4.66 Proto-hypo 5.08 5.49 6.02 Hypo-hypoconulid 2.09 2.27 2.25 1.01 Hypoconulid-ento 1.10 0.85 0.95 Proto-para 3.72 3.75 4.17 Proto-para (horizontally) 4.87 5.25 5.04

UPPER MOLARS M2 QMF number 16850 Anterior 5.50 Tooth length 6.55 Posterolingual dimension 6.65 Para-meta 2.70 Meta-proto 4.35 Proto-para 2.50

Table I. Tooth measurements of Thylacinus macknessi in mm. All measurements are actual distance between cusps except 'Proto-para (horizontally)' in which measurements are taken from a horizontal plane above these cusps.

Mike's Menagery Local Fauna, Mike's Menagerie DESCRIPTION Site, Godthelp Hill, Riversleigh Station, northwestern The dentary is preserved posteriorly from the region Queensland. QMFl6850 Dwornamor Local Fauna, of the M3. The coronoid process departs from the Gag Site, Gag Plateau, Riversleigh Station, north- ramus at approximately 120°. The anterior margin of western Queensland. Neville's Garden Site and Mike's the coronoid process is very straight. The articular Menagerie Site are part of System B of the Riversleigh condyle and angular process are absent. The articular Sequence and are estimated to be approximately condylar appears to have departed from the remaining Early to Middle Miocene in age (Archer, Godthelp, bone at a level no higher than the ramus. The inferior Hand and Megirian, 1989).Gag Site is part of System dental foramen is low lying and directly below the C estimated at Middle to early? Late Miocene (Archer highest dorsal point of the coronoid process. et al 1989). Precise locality details have been lodged with the Queensland Museum. Lower canine represented by Q MF 16848b(Fig. 2). Crown broken at tip. Root laterally compressed and SPECIFICNAME striated. Crown thick at base. Thegosed facet on Named in honour of Brian Mackness in recognition anterior surface (sensu Every 1975). M3 represented of his support for, and long term involvement in by QMFI6848a (Fig. 2). Crown broken anteriorly at Australian vertebrate palaeontology . protoconid. Hypoconid tallest cusp on talonid follow- ed in decreasingheight byentoconid and hypoconulid. Well developed posterior cingulum extends from DIAGNOSIS hypoconulid to posterobuccal comer of crown. Cristid Differs from all other species of Thy/acinus in the obliqua orientated almost perpendicular to antero- following combination of features: I, distinct entoco- posterior dimension of tooth. Both entoconid and nid; 2, vestigial metaconid; 3, long cristid obliqua hypoconulid high and distinct. M4 is represented by orientated centrally on the crown and parallel to the QMF 16849a (Figs 3A, 3B) and QMFI6848 (Figs long axis of the dentary; 4, no stylar shelf; 5, well- I B, 2). Protoconid largest cusp followed in decreasing developed and unnotched anterior cingulum that is height by paraconid, hypoconid, entoconid and continuous with the preparacrista; 6, centrocfista hypoconulid. Vestigial metaconid represented by .(postparacrista plus premetacrista) straight; 7, small slight thickening of enamel in position usually metaconule but no paraconule; 8, relatively unreduced occupied by metaconid on dasyurids and in the paracone; and 9, M5 antero-posteriorly shorter in plesiomorphic Nimbacinus. Anterior cingulum length than preceding molar. continues to anterobuccal base of protoconid. In 70 AUSTRALIAN MAMMALOGY

Fig. 2. ThylacinusmacknessiholotypeQMFI6848a. Bar= I cm. A, occlusal view stereo pair, anterior at top. B, lingual view, anterior to left.

occlusal view the protocristid is the longest crest. It is retained St E). Although postmetacrista broken, 'V'-shaped and almost vertical in the valley between posteriorly it appears to have been the longest crest on the paraconid and protoconid. Cristid obliqua slightly crown, followed (in decreasing order) by prepara- shorter, terminating on flank of protoconid. Remain- crista, postprotocrista, preprotocrista, premetacrista, ing crests (in decreasing length): metacristid, post- postparacrista. Preprotocrista connects with anterior hypocristid, pre-entocristid. Metacristid terminates at cingulum and extends continually up anterior face of vestigial metaconid. M5 represented by QMFI6848a crown. Angle formed at junction of crestsat protocone (Fig. 2). Morphology follows that of M4 except as approximately 90°. Postparacrista and premetacrista follows. Talonid reduced in size by lingual shift in connect in straight line as centrocrista and forms hypoconid and reduction of height, loss of entoconid sharp valley between cusps. Postmetacrista departs and reduction in posterior cingulum. from metacone almost parallel to premetacrista and then bends. Preparacrista forms sharper angle with M2 represented by QMFI6850 (Figs IA, 3C, 3D). postparacrista. Anterior width of crown lessthan buccal length which is less than posterior width. Major cusps present (in COMPARISON decreasing height): metacone, paracone, protocone. Minuscule metaconule present. No cusps or crests Thy/acinus macknessi is a very specialised thylacinid present on stylar shelf which is represented by steep sharing many derived features with Tocynocephalus. regular slope to buccal flank of crown. Posterobuccal It is in some respects more plesiomorphic than To region of crown missing (mayor may not have cynocepha/us and, in most features, far more special- 70 AUSTRALIAN MAMMALOGY

Fig. 2. 1hylacinusmacknessiholotypeQMFI6848a. Bar= 1 cm. A, occlusal view stereo pair, anterior at top. B, lingual view, anterior to left.

occlusal view the protocristid is the longest crest. It is retained St E). Although postmetacrista broken, 'V'-shaped and almost vertical in the valley between posteriorly it appears to have been the longest crest on the paraconid and protoconid. Cristid obliqua slightly crown, followed (in decreasing order) by prepara- shorter, terminating on flank of protoconid. Remain- crista, postprotocrista, preprotocrista, premetacrista, ing crests (in decreasing length): metacristid, post- postparacrista. Preprotocrista connects with anterior hypocristid, pre-entocristid. Metacristid terminates at cingulum and extends continually up anterior face of vestigial metaconid. M5 represented by QMFI6848a crown. Angle formed at junction of crestsat protocone (Fig. 2). Morphology follows that of M4 except as approximately 90°. Postparacrista and premetacrista follows. Talonid reduced in size by lingual shift in connect in straight line as centrocrista and forms hypoconid and reduction of height, loss of entoconid sharp valley between cusps. Postmetacrista departs and reduction in posterior cingulum. from metacone almost parallel to premetacrista and then bends. Preparacrista forms sharper angle with M2 represented by QMFI6850 (Figs IA, 3C, 3D). postparacrista. Anterior width of crown lessthan buccal length which is less than posterior width. Major cusps present (in COMPARISON decreasing height): metacone, paracone, protocone. Minuscule metaconule present. No cusps or crests Thylacinus macknessi is a very specialised thylacinid present on stylar shelf which is represented by steep sharing many derived features with 7: cynocephalus. regular slope to buccal flank of crown. Posterobuccal It is in some respects more plesiomorphic than T. region of crown missing (mayor may not have cynocephalus and, in most features, far more special- MUIRHEAD: THYLACINUS MACKNESSI FROM MIOCENE DEPOSITS IN QUEENSLAND 71

Fig.3. Thy/acinus macknessiparatypes QMFI6849 and QMF16850. Arrows point anteriorly. Bar = I mm. A, QMFI6849 occlusal view stereo pair. B, QMFI6849 lingual view showing lack of metaconid. C, QMFI6850 occlusal view stereo pair. D, QMFI6850 lingual view.

ised than N. dicksani. Comparison to 7: patens is M4 of T. macknessi is taller than the hypoconulid. It restricted because of the poor preservation of known also has a well defined pre-entocristid which defines teeth in T. patens. The features that set T. macknessi the lingual side of the talonid basin. In this respect 7: apart from other speciesof Thy/acinus are discussed macknessi is more plesiomorphic than 7: cyno- below. cephalUY.

Metaconid: Thylacinus macknessi has a slight Length of M5: Thy/acinus macknessi differs from thickening of the enamel on the posterolingual flank both 7: cynocephalus and 7: potens in the lack of of the protoconid at the position where the metacristid elongation of the M5. On T. cynocephalus and T. terminates. This appears to represent a remnant potens, the length of the lower molars increase metaconid (Fig. 38). A distinct metacristid continues posteriorly such that each molar is antero-posteriorly from this region to the tip of the protoconid. Marked longer than the preceding tooth. This may be related reduction and posterior shift in position of the to facial elongation, a feature that is characteristic of metaconid is a synapomorphic feature of the Thyla- these species (D. Ride pers. comm.) and as such is a cinidae while the retention of the metaconid and apomorphic specialisation away from the shorter metacristid is the plesiomorphic condition. Nimba- faced and reduced M5S of dasyurids. Thy/acinus cinus dicksoni displays the extreme plesiomorphic macknessi is therefore more plesiomorphic than these condition for thylacinids with a distinct metaconid on two species of Thy/acinus in retaining the short M5. M 2 while T. cynocepha/us has lost all traces of the Unfortunately, the premolar region of this speciesis metaconid and metacristid. , unknown and the anterior elongation of the snout also typical to speciesof Thy/acinus remains unknown Entoconid: The entoconids on the molars of 7: for this species. cynocephalus occur only on M 3 and M4 and increase in size posteriorly. They are the smallest cusps on the Stylar shelf: Reduction of the stylar shelf is a typical crowns and lack any linking crests. No entoconid is thylacinid feature. Stylar shelf reduction occurs in present on the M2 of N. dicksoni and the talonid derived thylacinids and dasyurids (for example, T. region of this species is not preserved for any other cynocephalus, T. potens, Sarcophilus harrisi,). How- molar. The entoconid is typically smaller on the M2 ever, this specialisation is achieved in different ways in compared to all other molars in dasyuroids and thylacinid and dasyurid lineages. The stylar shelf therefore the size of this cusp on a tooth comparable region of M4 in specialised thylacinids is distinctly with other speciesis not known. The entoconid on the different from that on M2-3. Thylacinus macknessi is 72 AUSTRALIAN MAMMALOGY

represented only by an M2 which limits comparisons and no wall is formed. Nimbac;nus d;ckson; and S. and T. potens cannot be compared to 7: macknessi harris;; do not have straight centrocristae. A sharper because of the poor preservation of this tooth. angle is formed, and in both the centrocristae are high Thylacinus macknessi compares closely with the M2 and form a lingual wall. Sarcoph;lus harris;; differs of T. cynocephalus in lacking stylar shelf development. from N. d;ckson; by having a sharp groove running Stylar cusps D and B are present on the M2 of T. up the flank of the centrocrista. The relative position- cynocephalus but no crests are developed. Stylar cusp ing of the paracone and metacone together with the B is present on T. macknessi but becauseof damage it short, straight centrocrista is an apomorphic special- is not known if St D was present. One clear difference isation culminating in the condition in ~ cyno- between these two speciesis that 7: cynocephalus has cephalus and ~ mackness;. a distinct bulge of enamel in the area of the stylar shelf while in T. macknessi the surface is a steep constant Elongation of postmetacrista: Thylacinus mack- slope from the tips of the paracone and metacone to nessi appears to have the thylacinid feature of an the buccal edge of the tooth. Reduction of the stylar elongated postmetacrista to a degree similar to that of shelf is an apomorphic feature that contrasts with the T. cynocephalus. Absolute length of this crest, how- plesiomorphic state in most dasyurids where stylar ever, can only be approximated becauseof damage to cusps and crests are often prominent and well the posterobuccal corner of the crown. Elongation of developed. Thus it appears that T. macknessi is the postmetacrista is an apomorphic carnivorous slightly more specialised in this feature than T. specialisation resulting in a relatively longer primary cynocephalus. cutting blade. This elongation occurs to a similar degree in species of Sarcophilus. The stylar shelf region of the M2 of S. harris;; is unlike that of speciesof Thylac;nus. In S. harris;; St B and D are always prominent and connected to the Metaconule and protoconule: The presence of a paracone and metacone (respectively) by short metaconule and protoconule is considered to be the transverse crests. A crest is also present connecting St plesiomorphic condition among dasyurids (Archer D to the posterobuccal corner of the crown. These 1976). Neither is present in T. cynocephalus while the stylar cusps have moved linguallyand are very close to condition of these cusps is unknown for the M2 and the paracone and metacone. This shift in position has MJ of T. potens. Thylacinus macknessi has a vestigial the same function as loss of these cusps in that metaconule but no trace of a protoconule. Thylacinus incorporation of these cusps into the major cones macknessi would therefore be more plesiomorphic produces a single large cutting blade with an un- than T. cynocephalus in this character. Nimbacinus obstructed buccal face. This carnivorous blade has dicksoni is more plesiomorphic in having a well- therefore been achieved in two different ways -by developed metaconule and protoconule on all upper loss of stylar cusps in thylacinids and by the lingual molars. shift in stylar cusps in dasyurids. It is possible that a further lingual shift of stylar cusps to a point of Reduction of the paracone: Reduction of the entirely combining with the paracones and metacones paracone is an apomorphic feature displayed in the could produce a condition similar to that seen in T. greatest degree in T. cynocephalus, T. potens and S. mackness;. It is therefore possible that T. mackness; is harris;;. N;mbac;nus d;ckson; shows no such reduc- a product of this Sarcoph;lus-type specialisation tion. The condition of the paracone in 7: mackness; is rather than that of the thylacinid-type. similarly unreduced and thus plesiomorphic among thylacinids. Postparacrista and Premetacrista: Many marsupial carnivores (for example, species of Sarcophilus, Upper anterior cingulum: A notched anterior 1hylacinus, Borhyaena) have the paracone and cingulum is a feature common to all dasyurids. metacone close together with the paracone lying Nimbacinus dicksoni has a notched anterior cingulum directly anterior to the metacone. As a result, the on all upper molars. In these the cingulum does not crests linking these cusps (the postparacrista and continue lingually past the paracone. Thylacinus premetacrista) are shortened and connect at a more cynocephalus has no anterior cingulum on any upper lingual point, producing a linear centrocrista. molars. Thylacinus potens has a small notched 1hylacinus cynocephalus has these crests connecting anterior cingulum on M4. Thylacinus macknessi has a almost in a straight line, a condition very similar to prominent but unnotched anterior cingulum which is that in T. macknessi. The difference between these unusual because it continues lingually past the two thylacinids is that in T. cynocephalus, these crests paracone to connect with the preprotocrista. It is connect well above the crown basin, thereby fonning considered here that the conditions in T. cynocephalus a lingual wall. In 7: macknessi the connection is low and T macknessi are autapomorohic for each. MUIRHEAD: THYLACINUS MACKNESSI FROM MIOCENE DEPOSITS IN QUEENSLAND 73

DISCUSSION entoconid and the presenceof remnant conules. Some of these characters, however, cannot be compared to The family Thylacinidae has very low diversity with T. potens and therefore the typical condition for the two genera, Thylacinus (containing T. cynocephalus, genus in these characters is unknown. Thylacinus 7: potens and T. macknessl) and Nimbacinus macknessi is referred to the genus Thylacinus because dicksoni. These two genera represent morphological it shares many apomorphic states unique only to extremes within the family, the speciesof Thylacinus other species of this genus. In these features T. being highly specialised while N. dicksoni exhibits macknessi is clearly more similar to T. cynocephalus many plesiomorphic features that provide a structural and T. potens than it is to N. dicksoni. Due to these link to plesiomorphic dasyurids. Nimbacinus dicksoni synapomorphies with species of Thylacinus, T. was considered to be a possible ancestor to speciesof macknessi is conservatively placed within this genus. Thylacinus (Muirhead and Archer 1990). The presence of these 'Thylacinus' apomorphies in combination with the plesiomorphic retention of Thy/acinus macknessi is currently known from some features is, at present, inadequate justification early to middle Miocene sediments. Although sharing for erection of anew monotypic genus. apomorphic features with some dasyurids, these are considered here to be convergent carnivorous Despite the consequent expansion of the concept of specialisations. Among dasyurids, T. macknessi is Thy/acinus, morphological variation within this genus most similar to the Late Pleistocene-Recent Sarco- is no greater than it is within many dasyurid genera, phi/us harrisii, however, in some characters it shows a such as Dasyurus or Antechinus. Thy/acinus degree of specialisation unattained by any dasyurid, macknessi is less apomorphic in a number of respects including the very specialised S. harrisii. It is unlikely than 7: cynocepha/us and is therefore positioned as to be related to speciesof Sarcophi/us becauseit does the sister taxon of a combined T. cynocepha/us/ T. not show specialisations typical to this dasyurid potens clade. It cannot itself be considered ancestral lineage (for example, reduction in size of the talonid to 7: cynocepha/us because of the derived character basin with loss of the hypoconulid, substitution of states of its anterior cingulum, cristid obliqua orien- entoconid by metaconid on posterior molars) while tation and specialised stylar shelf. showing apomorphies unknown to Sarcophi/us. Thy/acinus macknessi is further unlikely to represent Archer et al (1989) have summarised current a dasyurid of the Sarcophi/us grade, becauseit shows understanding about the biostratigraphic relationships a much greater degree of specialisation in some of Riversleigh's local faunas. Neville's Garden Local characters while occurring prior to first know Fauna and Mike's Menagerie Local Fauna (the type occurrence of the the G/aucodon-Sarcophi/us lineage and one paratype locality of T. macknessl) is part of (Stirton 1957, Ride 1964) which makes its earliest System B and thus interpreted to be early to middle known appearance in the Late Pliocene-Early Miocene in age. The Gag Local Fauna ( other paratype Pleistocene (Gill 1957, Bartholomai and Marshall locality of T. macknessl) and Henk's Hollow Local 1973). If considered a dasyurid, this species must Fauna (type locality of N. dicksom) are System C represent an otherwise unknown lineage with a local faunas currently interpreted to be middle to early combination of plesiomorphies (for example, relatively Late Miocene in age. Presuming the stratigraphic large talonid basin, large hypoconulid) and apo- relationships to be correctly interpreted, T. macknessi morphies (for example, lost metaconid, lost stylar may span early to middle Miocene time while the time shelf) unknown to any other species but convergent range for N. dicksoni remains unclear. Although a on both Sarcophi/us and Thy/acinus. Although single tooth from the Site D Local Fauna (a System A extreme convergence to thylacinids has been demon- local fauna interpreted to be late Oligocene to early strated in other forms such as borhyaenids, 7: Miocene in age) was statistically unable to be dis- macknessi shows a combination of characters mostly tinguished from N. dicksoni from the Henk's Hollow considered to be morphologically intermediate on a Local Fauna (Muirhead and Archer 1990), it was in lineage between the plesiomorphic N. dicksoni and fact smaller in size and only doubtfully referred to N. the apomorphic T. cynocepha/us. Thus phylogenetic dicksoni. Nevertheless, such a time range for a placement within the Thylacinidae is the most Tertiary thylacinid probably would not be extra- parsimonious placement, evoking fewer character ordinary considering that T. cynocephalus appears to conflicts than if this taxon is considered to be a span at least early Pliocene to Holocene time (Archer dasyurid. 1982, Dawson 1982), an interval of 5 million years.

Thylacinus macknessi falls outside of the currently The presence of a variety of thylacinids in late known speciesof Thylacinus in the lack of elongation Oligocene to middle Miocene sediments (Archer of the M5, unreduced paracone, lesserreduction of the 1982, Muirhead and Archer 1990) rejects the 74 AUSTRALIAN MAMMALOGY

molecular clock date of 7 million years given by David Ride and for help with material and suggestions Sarich, Lowenstein and Richardson ( 1982) and of by Henk Godthelp, Sue Hand and in particular Mike 6-10 million years estimated by Lowenstein, Sarich Archer who also critically read drafts of this and Richardson (1981) for the divergence ofthylacines manuscript. from dasyurids. It now also seemsprobable that this dichotomy occurred well prior to 10-20 million years ago date suggested by Thomas, Schaffner, Wilson REFERENCES and Paabo ( 1989). An early separation date is supported here by the demonstration that highly ARCHER, M., 1976. The dasyurid dentition and its relation- derived thylacinids (T. macknessl) were present in ship to that of didelphids, thylacinids, borhyaenids middle Miocene sediments as contemporaries ofvery (Marsupicarnivora) and peramelids (Peramelina: plesiomorphic taxa (N. dicksom). Clearly a common Marsupialia). Australian Journal of Zoology, Supple- ancestral stock for both must significantly predate the ment Series 39: 1-34. presently oldest known occurrence of either. ARCHER, M., 1982. A review of Miocene thylacinids (Thylacinidae, Marsupialia), the phylogenetic position Presuming that the Nimbacinus lineage (not of the Thylacinidae and the problem of apriorisms in specifically N. dicksom) was ancestral to that of character analysis. Pp. 445-76 in Carnivorous Thylacinus, its occurrence with the specialised T. marsupia/sed by M. Archer. Royal Zoological Society macknessirepresents persistence of this archaic lineage of New South Wales: Sydney. beyond the time of the Nimbacinus/ Thylacinus dichotomy. ARCHER, M., GODTHELP,H., HAND, S. J. AND MEGIRIAN, D. 1989. Fossil mammals of Riversleigh, northwestern Queensland: Preliminary overview ofbiostratigraphy, correlation and environmental change. Australian ACKNOWLEDGEMENTS Zoology 25: 29--{;5.

The study was undertaken with the support from the BARTHOLOMAI,A. AND MARSHALL, G., 1973. The identity of the supposed dasyurid marsupial, Sarcophilus prior Queen Elizabeth II Silver Jubilee Trust For Young De Vis, 1883, with comments on other reported Australians and Australian Commonwealth Depart- 'Pliocene' occurrences of Sarcophilus. Memoirs of the ment of Employment, Education and Training. Queensland Museum 16: 369-374. Material on which this study is based was obtained through financial support the Riversleigh Project has CAMPBELL, C. R., 1976. Tertiary Dasyuridae and Pera- had from: the Australian Research Grant Scheme melidae (Marsupialia) from the Tirari Desert, South (Grant PG A3851506P); the National Estate Grants Australia. Dissertation Abstracts International B. Scheme(Queensland); the Department of Arts, Sport, Science and Engineering 37: 4375. the Environment, Tourism and Territories; Wang Australia Pty Ltd; ICI Australia Pty Ltd; the Queens- DAWSON, L., 1982. Taxonomic status of fossil thylacines land Museum; the Australian Museum; the University ( 1hylacinus, Thylacinidae, Marsupialia) from late Quaternary deposits in eastern Australia. Pp. 527-536 of New South Wales; the Australian Geographic in Carnivorous marsupials ed by M. Archer. Royal Society; Mount Isa Mines Pty Ltd; and Ansett Zoological Society of New South Wales: Sydney. Wridgways Pty Ltd. Critical logistical support in the field and laboratory has been received from organ- EVERY, R. G., 1975. Significance of tooth sharpness for isations such as the Riversleigh Society Inc, the mammalian, especially primate, evolution. pp. 293-325 Friends of Riversleigh Inc, the Royal Australian Air in Approaches to primate biology ed by F. S. Szalay. Force, the Australian Defence Force, the Queensland Contributions to primatology 5. Karger: Basel. National Parks and Wildlife Service, the Riversleigh GILL, E. D., 1957. The stratigraphic occurrence and palaeo- Consortium (Riversleigh being a privately owned ecology of some Australian Tertiary marsupials. station), the Mount Isa City Council, the Northwest Memoirs of the National Museum of Victoria 21: Queensland Tourism and Development Board, the 135-203. Gulf Local Development Association, PROBE, Elaine Clarke and many Riversleigh volunteer workers who LOWENSTEIN,l. M., SARICH,V. M. AND RICHARDSON,B.l., helped to collect the material on which this study was 1981. Albumin systematics of the extinct mammoth based. Light photography was produced by Ross and Tasmanian wolf. Nature 291: 409-411. Amett and Robyn Murphy. SEM photography was produced with the assistanceof Coral Gilkeson from MUIRHEAD, J. AND ARCHER, M., 1990. Nimbacinus dicksoni, a plesiomorphic thylacine ( Marsupialia: Westmead Dental Hospital. I am also grateful for the Thylacinidae) from Tertiary deposits of Queensland assistance given by the Australian Museum (in and the Northern Territory. Memoirs of the Queens- particular Linda Gibson), for suggestions from Dr land Museum 28: 201-221. MUIRHEAD: THYLACINUS MACKNESSI FROM MIOCENE DEPOSITS IN QUEENSLAND 75

PlANE, M. D., 1976. The occurrence of Thylacinus in STIRTON, R. A., 1957. Tertiary marsupials from Victoria, Tertiary rocks from Papua New Guinea. Journalof Australia. Memoirs of the National Museum of Australian Geology and Geophysics I: 78-79. Victoria 21: 121-134.

RIDE, W. D. L., 1964. A review of Australian fossil THOMAS, R. H., SCHAFFNER,W ., WILSON, A. C. AND marsupials. Journal of the Royal Society of Western PAABO, S., 1989. DNA phylogeny of the extinct Australia 47: 97-131. marsupial wolf. Nature 340: 465-467.

SARICH, V., LoWENSTEIN,J. M. AND RICHARDSON,B. J., WOODBURNE,M. 0., 1967. The Alcoota fauna, central 1982. Phylogenetic relationships of Thylacinus cyno- Australia. An integrated palaeontological and geo- cephalus. Marsupialia as reflected in comparative logical study. Bulletin of the Bureau of Mineral serology. Pp. 707-709 in Carnivorous marsupials ed. Resourcesfor Geology and Geophysics. Awtralia 87: by M. Archer. Royal Zoological Society of New South 1-187. Wales: Sydney.