Measuring marginal predation in animal groups

Abstract

Predation is a major pressure that shapes animal sociality, but predation risk is not homogenous within groups. Animals located on the group edge typically face an increased threat of predation, although different patterns have been reported. We created a simulation model to determine how changes in predator attack distance and prey density influence predation in relation to within-group spatial position. At large attack distances, peripheral animals were attacked far more than central animals. At relatively short attack distances, central individuals were attacked almost as often as peripheral animals. We used 6 different methods to classify within-group spatial position in our simulations and tested which methods were the best predictors of predation risk at different parameter values. The minimum convex polygon and angle of vulnerability methods were the best predictors of predation risk at large and medium attack distances, respectively. At relatively short attack distances, the nearest neighbor distance and neighbor density methods were the best predictors of predation risk. These patterns demonstrate that the threat of marginal predation is dependent on the behavior of predators and that for some predator–prey systems, marginal predation is predicted to be insignificant. We predict that social prey animals should change antipredatory behavior, such as vigilance, within-group spacing, and within-group spatial choice based on the relative distances at which their predators attack. These results demonstrate the importance of incorporating the behavior of predators in empirical studies and predator–prey models.