Competition Avoidance in a Dragonfly Mating System

Mature males of the dragonfly species Aeschna cyanea regularly visit ponds or rivers where they patrol along the shoreline in search for females. Whenever two males encounter another during their visits at the water, a peculiar interaction develops which quite often results in one or even both participants leaving the mating place. In the present paper we shall analyse the question whether the tendency to leave after an intermale encounter can be interpreted as an adaptive response to local circumstances. In particular, the leaving tendency will be viewed as a strategic decision to avoid local high density situations on the basis of private information which is gained by encountering other males. It is easy to see that an adaptive decision to leave a mating place should depend on the presumed reaction of the opponent. Accordingly, an intermale encounter will be modelled as an evolutionary normal form game where the payoffs are given in terms of mating chances which are to be expected on the basis of a male’s information situation. Optimal updating of a male’s private information strongly depends on the behavioural norms which are established in the population. This implies that there is a strategy-payoff feedback: On the one hand, selection induced by payoff differences leads to a change in the strategic structure of the population; on the other hand, the payoffs are not externally given and fixed, but they themselves evolve in reaction to the evolution of the population strategy. The concept of evolutionary stability will be extended in order to cope with such strategy-payoff feedbacks which arise quite naturally in biological applications. On the basis of this concept, several evolutionary games will be analysed. In all our models, we get a unique evolutionarily stable leaving tendency that corresponds to a completely mixed strategy. Even for realistic parameter constellations, however, our models generally overestimate the leaving tendency when compared to empirical data. In the last section, we shall outline a refined model which is in better agreement with field observations.