Peer Pressure: Do the Fitness Benefits of Animal Behavioral Traits Vary With Conspecific Density?

ABSTRACT Environmental fluctuations may maintain the substantial variation exhibited in behavioral phenotypes if selection on behavior varies spatiotemporally. Spatial or temporal variation in conspecific density is experienced by most organisms and has been hypothesized to maintain variation in a variety of phenotypes, as fluctuations in density often alter patterns of selection. We tested the hypothesis that the fitness consequences of two behavioral traits (activity and aggression) in juvenile North American red squirrels ( Tamiasciurus hudsonicus ) are modified by the conspecific density they experience during the spring of their year of birth. Red squirrels in the Yukon, Canada, experience yearly changes in conspecific density due to substantial variation in their primary food source (conifer seeds), and these fluctuations in conspecific density can alter the association between juvenile traits and overwinter survival. Juveniles generally must acquire a territory to survive their first winter, and they experience heightened competition for vacant territories when adult densities are elevated due to fewer vacant territories being available, which may favor different behavioral traits. We measured activity and aggression of juvenile red squirrels using standardized behavioral assays (open field trials and mirror image stimulation, respectively), along with survival through summer to autumn or from birth through their first winter. Activity, but not aggression, was associated with overwinter survival, but not survival to autumn, in a density‐dependent manner: more active individuals had higher survival under high density, whereas less active individuals had higher survival under low density. Aggression did not associate with survival in a density‐dependent fashion, but individuals who were both more aggressive and active were more likely to survive regardless of density, suggesting an adaptive advantage to the positive covariance between these two behaviors. Our results offer partial support for the hypothesis that environmental fluctuations maintain behavioral variation, while also highlighting how specific combinations of behavioral traits may be favored by correlational selection regardless of environmental context.