40 
Records of the Australian Museum (2009) Vol. 61 
Koalas are generally rare components of the Australian 
fossil record, presumably reflecting their arboreal habits 
(Murray, 1991), and fossil koala material is commonly 
fragmentary and/or poorly preserved. Thus, determination 
of intra- and interfamilial phascolarctid relationships has 
proven difficult. Pre-Holocene koalas are known from Oligo- 
Miocene deposits of Riversleigh, Frome Basin, and Tirari 
Desert (Stirton et al, 1967; Woodburne et al, 1987; Black 
& Archer, 1997; Myers et al, 2001; Louys et al., 2007); Late 
Miocene-Pliocene deposits of Corra Lynn Cave and Waikerie 
(Pledge, 1987); and Pleistocene deposits of Koala Cave, 
Mammoth Cave, Devil’s Lair, Tight Entrance Cave, Lake 
Eyre region, Madura Cave, Lake Menindee, Lake Victoria, 
Nelson Bay, Naracoorte region, Wellington Caves, and Gore 
(Bartholomai, 1968; Merrilees, 1968; Archer, 1972; Balme 
et al., 1978; Lundelius & Turnbull, 1982; Tedford & Wells, 
1990; Archer et al. , 1997; Dawson & Augee, 1997; Moriarty 
et al, 2000; Reed & Bourne, 2000; Piper, 2005; Ayliffe et 
al, 2008; Price, 2008a) (Fig. 1). The koala fossil record 
from the central to north eastern margin of the Australian 
continent is particularly depauperate, with only one specimen 
known (type specimen of Phascolarctos stirtoni from Gore, 
southeast Queensland). Koobor, a koala-like vombatiform 
marsupial is also known from isolated specimens from the 
Bluff Downs (central eastern Queensland) and Chinchilla 
Local Faunas (southeastern Queensland (Archer & Wade, 
1976; Archer, 1977a). Although originally considered to be 
a koala, more recent morphology-based phylogenic analyses 
suggest that Koobor sits outside the Phascolarctidae (Black 
& Archer, 1997) and may actually be a primitive sister-group 
of wynyardiids and ilariids (Myers & Archer, 1997). 
Here we report new specimens of fossil koalas that were 
recovered during recent systematic excavations from several 
Plio-Pleistocene deposits of eastern Queensland, including 
the regions of Chinchilla, Marmor and Mt. Etna (Fig. 1). 
Although the new specimens are fragmentary, the paucity 
of information about koalas in the Plio-Pleistocene makes 
the new eastern Australian material particularly noteworthy. 
Thus, the aim of this paper is to describe the new material 
and discuss the taxonomic and palaeoecological implications 
within a reliable geochronological framework. 
Materials and methods 
Dating. Samples of bone and post-depositional calcite 
growth within long-bone hollows from the Marmor and Mt. 
Etna cave fossil deposits were targeted for thermal ionization 
mass spectrometry (TIMS) U/Th dating. Each sample 
was pre-treated and processed at the Radiogenic Isotope 
Facility, The University of Queensland, following techniques 
described in Zhao et al (2001) and Yu et al. (2006). 
Speleothem calcites and aragonites are secondary 
mineral deposits that form in cave environments. Uranium 
is commonly leached from downward percolating meteoric 
waters and becomes co-precipitated within speleothem 
calcite (or aragonite) during genesis. At the time of 
speleothem formation, some U, but little or no Th, is 
incorporated into the calcite (or aragonite) lattices, and 
disequilibrium in the U-series decay chain occurs. The U/ 
Th age is calculated by determining the amount of 230 Th 
that was produced by the decay of 238 U (via intermediate 
isotope 234 U). Thus, dating of speleothem material provides 
the true age of initial calcite crystallization. However, in 
this dating study, the calcite precipitated within long bone 
hollows at some stage after deposition, thus, the calculated 
U/Th ages will represent minimum ages for the associated 
faunal assemblages. 
Fresh bone and teeth contains little or no U. However, 
after burial, U is taken up from the environment by bone 
apatites that scavenge U, but exclude Th, during diagenesis. 
Unlike speleothem, bones and teeth are open systems for 
U (Grim et al, 2008), therefore, the U/Th dates commonly 
represent the mean age of U-uptake history. Thus, a 
calculated age most likely represents a minimum age for the 
dated bone or tooth. This has previously been demonstrated 
for eastern Australian cave deposits where U/Th dating 
of deposit-capping speleothem (thus, also representing 
minimum ages) return dates that are always older than U/ 
Th dated bone and teeth recovered from within the deposit 
itself (Hocknull et al, 2007). 
Unfortunately, U/Th datable material that could 
potentially produce maximum ages of deposition was not 
recovered from the Marmor and Mt. Etna koala deposits. 
No dateable samples were obtained from the Chinchilla 
fossil deposits. 
Terminology. Dental nomenclature follows Fuckett (1993) 
where the adult unreduced cheek tooth formula of marsupials 
is PI-3 and Ml-4 in both upper and lower dentitions. Dental 
cusp terminology follows Archer (1978) except for what 
was then interpreted to be the hypocone, is now regarded 
to be the metaconule, based on its linkage through the 
postprotocrista with the protocone and metacone (Tedford 
& Woodburne, 1987; Tedford & Woodburne, 1998). Higher- 
level systematic nomenclature follows Aplin & Archer 
(1987). All measurements were made using callipers and 
are given in millimetres (mm). 
Institutional abbreviations. QMF, Queensland Museum 
Fossil specimen, Queensland Museum, Brisbane, Australia; 
QMJ, Queensland Museum modem specimen, Queensland 
Museum, Brisbane, Australia; QMF, Queensland Museum 
fossil Focality, Queensland Museum, Brisbane, Australia; 
SAMP, South Australian Museum Palaeontological 
specimen, South Australian Museum, Adelaide, Australia. 
Geographic and geological settings 
Chinchilla. Site QMF7 is located in Chinchilla, south¬ 
eastern Queensland (Fig. 1). Vertebrate fossils, constituting 
the Chinchilla Focal Fauna, are predominantly derived 
from the Chinchilla Sand, a lithostratigraphic sequence 
of fluviatile sediments exposed in the Condamine River 
between Nangram Fagoon and Warra (Woods, 1960). 
The Chinchilla Sand includes well-lithified calcareous 
sandstones grading into siltstone and conglomerate (quartz 
and ferruginous sandstone), and weakly consolidated 
sands that grade into silt and sandy clay. Such sediments 
were derived from erosion of the Orallo Formation and its 
lateritized profiles (Bartholomai & Woods, 1976). There 
are no analytical dates associated with the Chinchilla Sand. 
However, biochronological correlation based on fossil 
vertebrates to other important radiometrically-dated faunas 
from elsewhere in Australia (e.g., Kanunka Focal Fauna, 
