Fitness Set Analysis of Mimetic Adaptive Strategies

Although the occurrence of mimicry is increasingly well documented and the objections to its evolution have largely been answered (see Sheppard 1959; Wickler 1968; Rettennieyer 1970), the adaptive strategies underlying various types of mimicry are not clearly understood. The ecological concepts of environmental grain or patchiness and of fitness sets as combined by Levins (1962, 1968) permit a consideration of optimal adaptive strategies. Here, we apply this system of analysis to mimicry, and particularly to diverse mimetic strategies exhibited by butterflies (Lepidoptera: Papilionidea). Although Levins (1968) stresses the power of this style of analysis in understanding adaptive polymorphism, his emphasis is on theory. Cody (1966) and McNaughton (1970) apply fitness sets to biological problems, but none of their specific examples combines a consideration of the effects of differences in grain with a consideration of the effects of differences in the factors which change the general shape of the fitness set. We have generally adopted Levins (1968) notation to facilitate comparisons. Examination of optimal adaptive strategies in mimicry using the fitness set approach requires a specification of the shape of the fitness set and the forml of the adaptive function; both are determined by the biological features of the particular case. We have restricted our treatment to mimicry among butterflies, but suggest that our arguments can be applied, with appropriate adjustments for biological realism, to other cases of mimicry. Levins's (1968) parameters can be specified as follows: yi is the mimetic phenotype, Y2 is the "ancestral" nonmimetic phenotype, S1 is a predator encounter state of the environment where yi is the most fit phenotype, S2 is a mating encounter state of the environment where Y2 is the most fit phenotype (alternatively, the environmental axis can be the ratio of predators to potential mates), WV1 is fitness relative to predator avoidance through mimetic resemblance to a relatively unpalatable model, and W2 is fitness relative to visual attractiveness as a potential mate (fig. 1), p is the proportion