ANALYSIS OF DIVERSIFICATION HISTORIES IN EXTINCT CARNIVOROUS MARSUPIALS (SPARASSODONTA, METATHERIA) USING A BAYESIAN FRAMEWORK SERGIO D. TARQUINI1*, SANDRINE LADEVÈZE2, AND FRANCISCO J. PREVOSTI1, 3 1 Centro Regional de Investigaciones Científicas y Transferencia Tecnológica de La Rioja (CRILAR), Provincia de La Rioja, UNLaR, UNCa, SEGEMAR, CONICET, Entre Ríos y Mendoza s.n., CP 5301, Anillaco, La Rioja, Argentina. 2 Centre de Recherche en Paléontologie, Paris (CR2P, UMR 7207), MNHN CNRS Sorbonne Université, Muséum national d'Histoire naturelle, 57 rue Cuvier CP 38, F-75005 Paris, France. 3 Departamento de Ciencias Exactas, Físicas y Naturales, Universidad Nacional de la Rioja (UNLaR). Av. Luis M. de la Fuente s/n (5300), La Rioja, Argentina. Corresponding autor: [email protected]

South America was isolated during a great time of the Cenozoic, which allowed an endemic terrestrial biota to evolve in that continent. Within mammals, the carnivore guild was occupied by the Sparassodonta. The oldest records of these carnivorous metatherians (Metatheria being the clade including all mammals more closely related to marsupials than to placentals) would be from the early Palaeocene of Bolivia (just after the Cretaceous/Tertiary crisis), while the last come from the Pliocene of Argentina. Moreover, from the late Miocene - Pliocene, carnivorous placentals (Carnivora) began to colonize the continent from North America, following the formation of the Isthmus of Panama (ca. 3 Myrs). Given the evidence, some authors justified the extinction of the sabertooth sparassodont (Thylacosmilus) by the arrival of the sabertooth placental (Smilodon), while other authors are opposed to the “competitive displacement” hypothesis and propose an “opportunistic replacement”. We compiled a data set of fossil occurrences for Sparassodonta collected from the bibliography and museums databases. We also compiled fossil data for the six families of Carnivora that subsequently dispersed into South America and have been potentially competing for resources against sparassodonts, as suggested by previous hypotheses. We carried out analyses of the fossil datasets using a Bayesian framework implemented in the PyRate software. First, we estimated the preservation rate, the times of speciation and extinction for all species, and the speciation and extinction rates through a reversible jump Markov Chain Monte Carlo (RJMCMC). We ran 20,000,000 RJMCMC iterations and sampled once every 5,000 to obtain posterior estimates of the parameters. Additionally we tested whether the diversification dynamics of Sparassodonta may be linked with changes in body mass (using the Covar birth–death model) and with changes in global temperature (using a birth–death model with time-varying rates). Information of body mass and temperatures were obtained from the literature. Finally we assessed the effect of competition on the diversification of Sparassodonta with a Multiple Clade Diversity Dependence model. Our results show temporal changes in both speciation and extinction rates for Sparassodonta. Therefore, our results support the idea that the demise of a clade is controlled by the two factors. With respect to body size there is a trend where the larger body mass in Sparassodonta appears later in the evolutionary history of the group. Although the changes in body mass are not related to the speciation rate, body mass being related to the extinction rate. That is to say that as the size increased, the extinction rate decreased (the opposite pattern was previously discovered in large mammals). No significant correlations emerged between the global temperature curve and changes in diversification rates. Finally, clade competition did not affect the diversification dynamics of Sparassodonta; speciation rate fell before the dispersion of Carnivora in South America. In conclusion, these new techniques can improve our knowledge of the evolution of the taxa but more studies are required to elucidate the demise of Sparassodonta.