Familiarity affects network structure and information flow in guppy (Poecilia reticulata) shoals

How individuals respond toward one another can depend on the level of familiarity between them. Variation in the proportion of familiar individuals comprising a group can shape group-level outcomes and group members’ fitness, but less is known about how this variation shapes the emergence and structure of social networks or the resulting consequences for social processes. We formed guppy (Poecilia reticulata) groups in which individuals were: 1) all familiar with one another, 2) all unfamiliar, or 3) part of a mixed group of familiar and unfamiliar individuals. We then examined the fission–fusion dynamics of these shoals, their social network structure, and the speed and pattern by which foraging information diffused through them. Although fission–fusion dynamics were not driven by group composition, differences in social network structure were observed. Both familiar and unfamiliar groups exhibited nonrandom network structure, whereas mixed groups expressed more homogeneous social organization. How quickly—and in what order—individuals discovered foraging sites was socially influenced, but there was little evidence for social transmission. More likely, closely associated individuals discovered foraging sites at similar times to one another as a result of traveling together. Group composition affected the speed of information diffusion, with knowledge of the foraging site spreading most rapidly through mixed groups, potentially due to either their less structured networks or the dense clustering of highly social individuals observed in those groups. Our study joins a growing body of literature pointing to the importance of group composition in driving group-level patterns and outcomes.