Investigating variation in third-party intervention behavior during a fallow deer (Dama dama) rut

Abstract

Pairwise contests are frequently disrupted by the intervention of third-party group members. However, little is known about whether an individual’s engagement in intervention behavior varies over time, or what factors might be associated with such variation. Using a hierarchical “hurdle” model with 2 levels, we investigated the conditions under which focal males: 1) would or would not engage in an intervention, and 2) varied the number of interventions per day they engaged in. The lower level of the model showed that the proportion of unique opponents per day (estimated from the overall number of mature males in the herd) that focal males competed with, and the number of interventions suffered by a focal male were associated with an increased probability that this individual would itself engage in third-party intervention behavior. At the upper level of the model, there was no association between these 2 variables and the rate at which individuals engaged in intervention behavior. The number of matings observed per day and aggression rate within the herd failed to contribute meaningfully to either level of the model. We also show that, although inconsistent over days and between years, some individuals displayed a greater propensity to intervene than others. The data from our study show that intervention behavior is more likely to occur as a result of individual directly experiencing aggressive behavior at a sufficiently high level, and not as a result of individuals monitoring aggressive or sexual activity in the wider social group.