Defection on the bounty? Kinship and cooperative exploitation of a rich, essential but dangerous resource

Abstract

We develop a game-theoretic model to explore the question of whether two animals should cooperate in the dangerous activity of obtaining a rich and essential resource. We consider variation in the risks incurred to individuals and in how the activities of the two animals interact to influence the probability of success. We also consider that the animals may be relatives and thus share evolutionary interests. The model is general and can, for instance, be applied to mammalian predators attempting to capture and subdue large and dangerous prey or to female parasitoid wasps that attack and, if successful, paralyse aggressive hosts that then provide the only feeding resource for their offspring. This minimal model of cooperation contains three dimensionless parameters: vulnerability (the ratio between the average time for a lone attacker to subdue the defending resource and the average time for the defender to fatally strike the attacker), the dilution ratio (the extent to which attack by animals acting in tandem reduces a defender's ability to kill its attackers) and the relatedness between the potential attackers. The model predicts that higher values of all three parameters favour cooperation and that for small values cooperation is not evolutionarily stable. Cooperation can arise from an ancestral state of non-cooperation if values of all parameters are sufficiently high but cannot arise among non-relatives, irrespective of other parameter values. Once cooperation has emerged in a population, it can be maintained among non-relatives at modest values of dilution ratio and vulnerability. We discuss these general predictions in particular relation to the parasitoid genus Sclerodermus, in which multiple females may attack unusually large and aggressive hosts and in which host attack behaviour is mediated by kinship.