The marsupial lion, Thylacoleo carnifex, was the largest-ever marsupial carnivore, and is one of the most iconic extinct Australian vertebrates. With a highly-specialised dentition, powerful forelimbs and a robust build, its overall...
moreThe marsupial lion, Thylacoleo carnifex, was the largest-ever marsupial carnivore, and is one of the most iconic extinct Australian vertebrates. With a highly-specialised dentition, powerful forelimbs and a robust build, its overall morphology is not approached by any other mammal. However, despite >150 years of attention, fundamental aspects of its biology remain unresolved. Here we analyse an assemblage of claw marks preserved on surfaces in a cave and deduce that they were generated by marsupial lions. The distribution and skewed size range of claw marks within the cave elucidate two key aspects of marsupial lion biology: they were excellent climbers and reared young in caves. Scrutiny of >10,000 co-located Pleistocene bones reveals few if any marsupial lion tooth marks, which dovetails with the morphology-based interpretation of the species as a flesh specialist. When humans first set foot in Australia around 50,000 years ago they entered a unique landscape occupied by large reptiles, birds and mammals seen nowhere else. These included the anatomically-bizarre Thylacoleo car-nifex. The ubiquity of this species and its evocative depiction in Aboriginal rock art 1 suggest an important role in Australian ecosystems, but despite numerous skeletal studies, interpretations have remained controversial. The species was initially described by Richard Owen in 1859 as " one of the fellest and most destructive of predatory beasts " 2 , a view to which he was largely led by its greatly enlarged slicing premolar. It was soon after reinter-preted as a herbivore 3 , because it retained the herbivorous, diprotodontian template of enlarged first incisors, tiny canines, blade-like premolars, and large masseter and pterygoid muscles. Subsequently, the diet and behaviour of T. carnifex have been intensely debated. The species has been speculatively portrayed as a consumer of crocodile eggs 4 , a hyaena-like scavenger 5 , a melon specialist 6 , a leopard-like predator that dragged prey into trees 7 , a slow-to medium-paced runner incapable of climbing 8 , a terrestrial version of a cookie-cutter shark or raider of kangaroo pouches 9 , and a bear-like super-predator 10. The doubts over how to interpret its bizarre combination of features are not due to a lack of fossil bones: T. carnifex is better represented in Pleistocene localities that any other large carnivore 11,12 , and more complete or partial skeletons are known from caves than for any other extinct Pleistocene species 8,12–14. Although the current consensus is that T. carnifex was a carnivore, other lines of evidence are required to generate further insights into its behaviour and ecology. Trace fossils, such as trackways or burrows, can provide insights into locomotory abilities and behaviours unobtainable via functional analyses of the skeleton alone. However, it is seldom possible to associate skeletal and trace fossils 15. This underscores the significance of a claw-mark assemblage in the main chamber of Tight Entrance Cave (TEC), southwestern Australia (Fig. 1), where a now-blocked entrance in the ceiling provided access to the surface for species capable of navigating the steep, convoluted cave terrain. TEC also contains a diverse Pleistocene bone deposit 16 , which allows us to generate a shortlist of two extinct and five still-extant species of candidate claw markers based on known or potential climbing ability or cave utilisation. Past studies have noted cave surfaces scratched by Pleistocene cave bears 17 and humans 18 , and fossil burrows scratched by their rodent 19 and xenarthran 20 makers. In particular the long deep marks of cave bears are thought to be associated with navigation in darker areas, or in locomotion through the complex 3D cave environment 21. However, analyses have been largely qualitative and overlooked as a reliable information source beyond determining their maker. To analyse the TEC traces we pioneered a quantitative analysis involving comparisons of manual claw-mark dimensions with those made by living animals and simulated scratch sets of T. carnifex. We also sought evidence