The Role of Environmental Grain in the Maintenance of Genetic Variation

The graininess of an environment is a measure of how rapidly the environment changes within the lifetime of an individual (Levins 1968). In coarse-grained environments, each individual experiences only a single environment, while in fine-grained environments each individual experiences a rapid succession of environments. This succession may reflect actual temporal changes in the environment or movement of the individual through a sequence of spatially differentiated environments. Levins (1968) has argued, using his fitness-set echnique, that environmental graininess will affect levels of genetic variation occurring in a population in such a way that variation becomes less likely in a fine-grained environment. Recently, Selander and Kaufman (1973) used this argument to account for the lower levels of enzyme variation observed in vertebrate populations. As yet no analytic treatment of the relationship of environmental grain and genetic polymorphism has appeared. Since conditions for polymorphism in random environments are now known (Gillespie 1973), it seems appropriate to examine this relationship in detail. Here I do this at two levels of generality. First, I consider the following general questions about graininess and polymorphism. 1. Is it possible to maintain a stable polymorphism in a fine-grained environment due to the effects of a changing environment? 2. Under what conditions is it easier to maintain variation in a coarsegrained rather than in a fine-grained environment? 3. What is the role of sensitive periods in the development of an organism on the interpretation of the effects of graininess? Second, I consider the relevance of the answers to the above questions to the explanation for the occurrence of enzyme variation in natural populations as recently proposed by Gillespie and Langley (1974). This model assumes that heterozygotes are intermediate at the level of gene function and that this intermediacy is sufficient for polymorphism in temporarily and/or spatially fluctuating environments. Assuming that the GillespieLangley model is appropriate for electrophoretically detectable variation, I show here that polymorphism is much more likely in coarse-grained environments and, consequently, that the Selander and Kaufman result supports the Gillespie-Langley hypothesis.