Data from: Social network analysis of mating patterns in American black bears (Ursus americanus)

Non-random mating can structure populations and has important implications for population-level processes. Investigating how and why mating deviates from random is important for understanding evolutionary processes as well as informing conservation and management. Prior to the implementation of parentage or pedigree analyses, understanding mating patterns in solitary, elusive species like bears was virtually impossible. Here we capitalize on a long-term genetic dataset collected from black bears (Ursus americanus) (N=2,422) in the Northern Lower Peninsula (NLP) of Michigan, USA. We identified mated pairs using parentage analysis and applied social network (selection) models to quantify the effects of individual characteristics, and spatial and population demographic factors on mating dynamics. Social network analysis is particularly useful in this context as it allows for analyses of attributes at the level of the individual and pair, while taking into account effects of the overall social environment. Logistic regression models, which controlled for features of the social network, revealed that black bear mating was associated with spatial proximity of mates, male age, local population density, and relatedness. Mated pairs were more likely to contain older males. On average, bears tended to mate with nearby individuals to whom they were related, which does not support the existence of kin-recognition in black bears. Pairwise relatedness was especially high for mated pairs containing young males. Restricted dispersal and high male turnover from intensive harvest mortality of NLP black bears are likely the underlying factors associated with younger male bears mating more often with female relatives. Our findings illustrate how harvest has the potential to disrupt the social structure of game species, which warrants further attention for conservation and management.