Canalization Versus Developmental Conversion in a Spatially Variable Environment

A model is presented for the evolution of developmental control in a spatially variable environment. Individuals are assumed to disperse at random into one of two patches (one harsh and the other benign) and use one of three developmental strategies for the production of one of two discrete morphological types. Two of the strategies are unconditional (develop as either the stress-tolerant or the nontolerant morph), and the third strategy depends on the environment. Invasion criteria are used to determine the conditions under which each of the three pure strategies are evolutionarily stable and the conditions under which the population is expected to contain some mixture of these strategies at equilibrium. The results demonstrate that environmental control of development requires a cost to the stress-tolerant morphology, and that the average probability of making the right choice is greater than 50%. The range of patch frequencies over which environmental control is stable increases with increases in these two parameters. The results also suggest that the two morphs may be maintained by a mixture of unconditional strategies (i.e., genetically determined polymorphism) for a narrow range of patch frequencies. Unlike the environmentally induced dimorphism, however, the genetically determined dimorphism requires for its maintenance that the morphs compete within the patches for resources. Finally, the results suggest that a mixture of genetic and environmental control is an evolutionarily stable state under some conditions. These conditions are the same as those required for the maintenance of genetically determined dimorphism, except that the average probability of making the right choice must be greater than 50%. Under mixed control, some fraction of the population shows canalized development and the remaining fraction shows developmental conversion when exposed to the appropriate cue. That such a mixture of strategies could persist in evolutionary time might explain the results of some studies that indicate partial genetic and partial environmental control of phenotypes.