Etienne Sirot
Université de Bretagne Sud
Competition and negotiation over antipredatory vigilance in groups of prey
[When and where]
The topic of vigilance in groups of prey has been attracting the attention of behavioural ecologists for decades. Game theoretical models, in particular, have been used to study the proportion of time each prey should allocate to vigilance, as a function of the number of its companions, with whom it shares public information about the possible arrival of a predator.
Field studies, however, reveal that interactions between group members do not only concern the overall proportion of time dedicated to vigilance, but run of a fast time scale, because predatory attacks are sudden, and what counts for the prey is to react immediately. As a consequence, even when no predator is detected, the instantaneous probability of raising one's head to scan for predators depends on the level of vigilance of neighbours, which makes collective vigilance a responding game between flock members.
To study these responding rules and their effects on collective patterns of vigilance, I derived a game-theoretical model considering a pair of interacting foragers, and concerning competition over the task of collective vigilance. In this model, the strategy of each individual is described by the instantaneous probability of raising or lowering its head, as a function of the other's attitude. Each individual then responds to the attitude of its companion, and influences this companion through its own behaviour. The ultimate pattern of vigilance for the pair is then the outcome of a negotiation process, whereby each individual constantly tries to influence the behaviour of its companion to selfish ends.
The game played by the animals and the resulting strategies take radically different forms, depending on the predators’ targeting strategy. If predators choose their target at random, the prey respond by displaying moderate vigilance and taking turns scanning. Competition over vigilance takes then the shape of a war of attrition, each individual waiting for the other to be the sentinel. By contrast, if predators tend to preferentially target individuals that do not react instantaneously, vigilance increases and there is always at least one individual scanning. Lastly, when predators’ preference for stragglers becomes extreme, vigilance decreases again and the pair scan simultaneously. In this last situation, the evolutionarily stable responses perfectly correspond to the behaviour expected under a fully cooperative scenario.
Invited talk Mini-symposium 2
Updated May 13, 2015, by Minus