The evolution of imperfect mimicry

Examples of imperfect resemblance between Batesian mimics andtheir models appear widespread in the natural world, but sofar few quantitative models have been proposed to explain thephenomenon. I used a simple signal detection model to showthat the relationship between model—mimic similarity andmimic effectiveness is typically nonlinear. In particular, Ifound that there will be little or no further selection toimprove model—mimic resemblance beyond a certain levelif the model species is costly to attack, if the mimic speciesis not particularly profitable (e.g., hard to catch), or ifthe mimic is relatively rare. When there are two different sympatricmodel species, then mimics should usually evolve a phenotypicsimilarity to one or the other model species, but not to both.In contrast, when several model species occur in differentareas (or emerge at different times) and individual mimicsuse each of these areas, then the optimal phenotype should bea "jack-of-all-trades" intermediate phenotype that does notclosely resemble any particular model species. Somewhat surprisingly,the theory predicts that if mimics spend an equal amount oftime with each model species, then the optimal intermediatephenotype should more closely resemble the least numerous andleast noxious model. This phenomenon arises because a vaguesimilarity to an extremely noxious species is usually sufficientto guarantee significant protection, whereas a much closerresemblance to a mildly noxious model species is necessaryto afford a similar level of benefit.