An 'ameridelphian' marsupial from the early Eocene of Australia supports a complex model of Southern Hemisphere marsupial biogeography.

Recent molecular data strongly support the monophyly of all extant Australian and New Guinean marsupials (Eomarsupialia) to the exclusion of extant South American marsupials. This, together with available geological and fossil evidence, has been used to argue that the presence of marsupials in Australia is simply the result of a single dispersal event from South America during the latest Cretaceous or Palaeocene, without subsequent dispersals between the two continents. Here, I describe an isolated ankle bone (calcaneus) of a metatherian from the early Eocene Tingamarra Local Fauna in northeastern Australia. Strikingly, this specimen, QM F30060, lacks the 'continuous lower ankle joint pattern' (CLAJP), presence of which is a highly distinctive apomorphy of the marsupial clade Australidelphia, which includes Eomarsupialia, the living South American microbiotherian Dromiciops and the Tingamarran fossil marsupial Djarthia. Comparisons with a range of marsupials and stem-metatherians strongly suggest that the absence of the CLAJP in QM F30060 is plesiomorphic, and that this specimen represents the first unequivocal non-australidelphian ('ameridelphian') metatherian known from Australia. This interpretation is confirmed by phylogenetic analyses that place QM F30060 within (crown-group) Marsupialia, but outside Australidelphia. Based on these results, the distribution of marsupials within Gondwana cannot be explained by simply a single dispersal event from South America and Australia. Either there were multiple dispersals by marsupials (and possibly also stem-metatherians) between South America and Australia, in one or both directions, or, alternatively, there was a broadly similar metatherian fauna stretching across southern South America, Antarctica and Australia during the Late Cretaceous-early Palaeogene.