Contrasting context dependence of familiarity and kinship in animal social networks

The social structure of a population is a crucial element of an individualʹs environment, fundamentally influencing the transfer of genes, information and diseases. A central question in social network analysis is how different traits affect associations within populations. However, previous studies of animal social networks have typically focused on a single predictor or stage in the life cycle whereas social interactions within populations are known to be dynamic and not fixed through time and/or context. Relatively few animal network studies have explored how individual traits affect decisions across different ecologically relevant contexts. We collected detailed behavioural data (personality, dominance, familiarity) and high-resolution genetic data from a flock of 43 captive barnacle geese, Branta leucopsis, to understand how these traits affect association patterns in two different evolutionary and ecologically highly relevant contexts: foraging and mate choice. Using a novel analytical framework for node label permutations, we found that barnacle geese preferentially associated with close kin and other individuals familiar from earlier in life when foraging, but selected unfamiliar partners during mate choice. We found no effect of either personality or dominance on foraging associations or mate choice. Our study shows how using social network analysis can increase our understanding of the drivers behind population structure (in our case kin selection and inbreeding avoidance). Moreover, our study demonstrates that social networks can be largely determined by long-term processes, in particular early life familiarity.