New skeletal material sheds light on the palaeobiology of the Pleistocene marsupial carnivore, Thylacoleo carnifex

Who was the real sabertooth predator: Thylacosmilus or Thylacoleo?

Sabertoothed mammalian predators, all now extinct, were almost exclusively feloid carnivorans (Eutheria, Placentalia): here a couple of extinct metatherian predators are considered in comparison with the placental sabertooths. Thylacosmilus (the "marsupial sabertooth") and Thylacoleo (the "marsupial lion") were both relatively large (puma-sized) carnivores of the Plio-Pleistocene in the Southern Hemisphere (Argentina and Australia, respectively). Both carnivores have captured the public imagination, especially as predators that were somehow analogous to northern placental forms. But a more detailed consideration of their morphology shows that neither can be simply analogized with its supposed placental counterpart. While Thylacosmilus did indeed have saber-like canines, many aspects of its anatomy show that it could not have killed prey in the manner proposed for the sabertoothed felids such as Smilodon. Rather than being an active predator, it may have been a specialized scavenger, using the hypertrophied canines to open carcasses, and perhaps deployed a large tongue to extract the innards. Thylacoleo lacked canines, and its supposedly "caniniform" incisors could not have acted like a felid's canines. Nevertheless, while its mode of dispatching its prey remains a subject for debate, it was clearly a powerful predator, likely to be capable of bringing down prey bigger than itself while hunting alone. In that regard, it may have filled the ecomorphological role proposed for placental sabertooths, and so despite the lack of canines can be nominated as the true "marsupial sabertooth" out of the two extinct taxa.

The skeleton of Congruus kitcheneri, a semiarboreal kangaroo from the Pleistocene of southern Australia

The macropodine kangaroo, Wallabia kitcheneri, was first described in 1989 from a Pleistocene deposit within Mammoth Cave, southwestern Australia, on the basis of a few partial dentaries and maxilla fragments. Here, we recognize W. kitcheneri within the Pleistocene assemblages of the Thylacoleo Caves, south-central Australia, where it is represented by several cranial specimens and two near-complete skeletons, a probable male and female. We reallocate this species to the hitherto monotypic genus Congruus. Congruus kitcheneri differs from all other macropodid species by having a highly unusual pocket within the wall of the nasal cavity. It is distinguished from C. congruus by having a longer, narrower rostrum, a taller occiput and a deeper jugal. Congruus is closest to Protemnodon in overall cranial morphology but is smaller and less robust. In most postcranial attributes, Congruus also resembles Protemnodon, including general limb robustness and the atypical ratio of 14 thoracic to five lumbar vertebrae. It is distinguished by the high mobility of its glenohumeral joints, the development of muscle attachment sites for strong adduction and mobility of the forelimb, and large, robust manual and pedal digits with strongly recurved distal phalanges. These adaptations resemble those of tree-kangaroos more than ground-dwelling macropodines. We interpret this to imply that C. kitcheneri was semiarboreal, with a propensity to climb and move slowly through trees. This is the first evidence for the secondary adoption of a climbing habit within crown macropodines.

A new family of diprotodontian marsupials from the latest Oligocene of Australia and the evolution of wombats, koalas, and their relatives (Vombatiformes)

We describe the partial cranium and skeleton of a new diprotodontian marsupial from the late Oligocene (~26–25 Ma) Namba Formation of South Australia. This is one of the oldest Australian marsupial fossils known from an associated skeleton and it reveals previously unsuspected morphological diversity within Vombatiformes, the clade that includes wombats (Vombatidae), koalas (Phascolarctidae) and several extinct families. Several aspects of the skull and teeth of the new taxon, which we refer to a new family, are intermediate between members of the fossil family Wynyardiidae and wombats. Its postcranial skeleton exhibits features associated with scratch-digging, but it is unlikely to have been a true burrower. Body mass estimates based on postcranial dimensions range between 143 and 171 kg, suggesting that it was ~5 times larger than living wombats. Phylogenetic analysis based on 79 craniodental and 20 postcranial characters places the new taxon as sister to vombatids, with which it forms the superfamily Vombatoidea as defined here. It suggests that the highly derived vombatids evolved from wynyardiid-like ancestors, and that scratch-digging adaptations evolved in vombatoids prior to the appearance of the ever-growing (hypselodont) molars that are a characteristic feature of all post-Miocene vombatids. Ancestral state reconstructions on our preferred phylogeny suggest that bunolophodont molars are plesiomorphic for vombatiforms, with full lophodonty (characteristic of diprotodontoids) evolving from a selenodont morphology that was retained by phascolarctids and ilariids, and wynyardiids and vombatoids retaining an intermediate selenolophodont condition. There appear to have been at least six independent acquisitions of very large (>100 kg) body size within Vombatiformes, several having already occurred by the late Oligocene.

The extraordinary osteology and functional morphology of the limbs in Palorchestidae, a family of strange extinct marsupial giants

The Palorchestidae are a family of marsupial megafauna occurring across the eastern Australian continent from the late Oligocene through to their extinction in the Late Pleistocene. The group is known for their odd 'tapir-like' crania and distinctive clawed forelimbs, but their appendicular anatomy has never been formally described. We provide the first descriptions of the appendicular skeleton and body mass estimates for three palorchestid species, presenting newly-identified, and in some cases associated, material of mid-Miocene Propalorchestes, Plio-Pleistocene Palorchestes parvus and Pleistocene Palorchestes azael alongside detailed comparisons with extant and fossil vombatiform marsupials. We propose postcranial diagnostic characters at the family, genus and species level. Specialisation in the palorchestid appendicular skeleton evidently occurred much later than in the cranium and instead correlates with increasing body size within the lineage. We conclude that palorchestid forelimbs were highly specialised for the manipulation of their environment in the acquisition of browse, and that they may have adopted bipedal postures to feed. Our results indicate palorchestids were bigger than previously thought, with the largest species likely weighing over 1000 kg. Additionally, we show that P. azael exhibits some of the most unusual forelimb morphology of any mammal, with a uniquely fixed humeroulnar joint unlike any of their marsupial kin, living or extinct.