Plant defence signals and Batesian mimicry

In a game theory context, we investigated conditions for an evolutionarily stable equilibrium of defended, signalling plants, and plants mimicking these signals - that is, conditions for a stable mimicry complex. We modelled this in three steps. First, we analysed conditions for selection for defended, signalling plants, in a population of undefended plants. Second, we analysed conditions for when mimicking plants can invade a population of defended, signalling plants, leading to a stable equilibrium between the two strategies. Third, we analysed how sampling of signalling plants by herbivores affects the equilibrium between the strategies. The predictions show that mimicry of plant defence signals may be common, and even imperfect mimics could invade a population of defended, signalling plants. Whether the latter prediction holds or not depends on how herbivores generalize over signals, and on the length of their 'avoidance sequence'. The length of the avoidance sequence is the number of signalling plants that a herbivore avoids to attack, after attacking a defended plant. If herbivores always sample signalling plants, then mimicry cannot evolve, whereas if herbivores have a long avoidance sequence, this may allow selection even for imperfect mimics.