Dickey, B.: Prediction and description of interaction patterns in social dominance structures

Though social dominance hierarchies have been demonstrated and studied in a wide variety of vertebrate and invertebrate taxa, no models exist that explain or predict which animals should interact aggressively within a hierarchy. Data from empirical studies of social dominance are often presented in matrices that summarize which animal won the interactions between each pair of individuals. Many of these matrices contain identifiable patterns of interactions; for instance, interactions may be clustered along the diagonal of the matrix suggesting that individuals prefer to interact with others of equivalent rank. I present a game theoretic model that predicts this and other patterns, depending on the differences in benefits between different ranks in the hierarchy, and the distribution of traits that signal resource holding potential between individuals. The predicted patterns are then compared to empirical data using a statistical approach that permits comparison of matrices. This comparison improves our understanding of the link between social structure, benefits of rank, and traits that signal resource holding potential in natural groups.