Intraspecific variation in Artiodactyla social organisation: A Bayesian phylogenetic multilevel analysis of detailed population-level data

Understanding inter-specific variation in social systems is a major goal of behavioural ecology. Previous comparative studies of mammalian social organisation produced inconsistent results, possibly because they ignored intra-specific variation in social organisation (IVSO). The Artiodactyla have been the focus of many comparative studies as they occupy a wide diversity of habitats and exhibit large variation in life history patterns as well as other potential correlates of social organisation. Here we present the first systematic data on IVSO among Artiodactyla, infer their ancestral social organisation, and test whether habitat, sexual dimorphism, seasonal breeding, and body size predict inter- and intraspecific variation in social organisation. We found data on social organisation for 110 of 226 artiodactyl species, of which 74.5% showed IVSO. Using Bayesian phylogenetic multilevel models, the ancestral artiodactyl population was predicted to have a variable social organisation with significantly higher probability (0.77, 95% CI 0.29-1.00) than any non-variable form (i.e. solitary, pair-living, group-living). Greater sexual dimorphism and smaller body size both predicted more IVSO; smaller body size also predicted a higher likelihood of pair-living. Our results challenge the long-held assumption that ancestral Artiodactyla were pair-living and strongly imply that taking IVSO into account is crucial for understanding mammalian social evolution.