Predator psychology and mimicry evolution - a theoretical analysis

Abstract: The evolution of Müllerian mimicry depends on many factors, among which predator psychology is the most important one. Predator avoidance learning, generalization and discrimination are primary selective agents during the evolutionary process. This thesis investigates an issue that has been a matter of debate since the 19th century; the question of how a mutant of a potential mimic population can escape the apostatic selection caused by predators that neither recognize it as a member of its own population, nor as being similar to its future model. In paper I, we investigate one of the ideas concerning this issue. Fisher stated that mimicry can evolve gradually by the displacement of an adaptive peak in a fitness landscape. We find this to be a plausible scenario, under certain conditions regarding predator generalization. In paper II, we further investigate the gradual evolutionary process and the implications of different generalization gradients. We find that the gradual process might be even more general than has previously been assumed and that abundant genetic variation in prey populations is an important factor in combination with predator generalization. In paper III, we examine another suggested solution to the problem of apostatic selection, the so-called two-step process, which can be problematic to extend to multiple prey traits. We find that the two-step process works for multidimensional traits provided that predators use feature-based generalization. The selective landscape of mimicry evolution is also shaped by the cost-benefit relationships for models and mimics. In paper IV we explore this matter, by applying the Rescorla-Wagner theory of learning to mimicry. We find that if variation in prey unpalatability gives rise to surprise, the learning rate is increased. This leads to unexpected kinds of mimicry that are more mutualistic than Müllerian mimicry has previously been assumed to be.