Effects of predator associative learning and innate aversion on mimicry complexes

Undefended or weakly defended prey species can evolve to resemble better-defended prey (models) in a potentially parasitic relationship called Batesian mimicry. However, some highly defended prey have lethal defenses that might prevent predators from learning to avoid them, which raises questions as to how and why warning signals evolve in these species. One solution is that the warning signals of lethal species have evolved to resemble those of less defended species, where avoidance learning is possible. To examine the general feasibility of this hypothesis, we modeled associative learning by predators foraging on prey species that were either weakly or highly defended. The highly defended prey had a fixed probability of killing an attacking predator. We found that the weakly defended species was more likely to be a parasitic Batesian mimic when its defenses were weaker. Weakly defended prey were more parasitic when the weakly defended prey was relatively common, and when highly defended prey were less likely to be lethal. Generally, mimicry was more mutualistic (i.e. Mullerian) as the highly defended prey increased in lethality. However, for a relatively lethal mimetic mutant that resembled a weakly defended species to invade a non-mimetic population of highly defended prey, lethality needed to be high, as benefits of mimicry did not accrue at low frequency. Moreover, when we created predators that had innate aversions to dangerous prey, weakly defended mimics were parasitic upon highly defended prey. Innate aversions also evolved in an individual based-simulation. When our analyses of prey lethality and innate aversions are taken together, it is likely that highly defended prey are most often models in Batesian mimicry systems.