Genetic variation can be assessed at every level of biological organization. At each level one may ask to what extent the variation is maintained by natural selection versus other influences. Among populations or among congeneric species, phenotypic differences can be quite distinct, and few investigators would question the claim that in many instances these differences reflect genetic adaptation to the particular ecological milieu the organisms inhabit. Detectable genetically based phenotypic variation also exists among individuals within populations, but can this variation be evaluated in a similar manner as can variation at higher levels of organization?

Two arguments bear on the existence of phenotypic variation within populations. First is the null hypothesis that alternative forms are selectively neutral. Variants possess no advantage or disadvantage, and phenotypic differences result from neutral forces such as mutation, migration, or genetic drift (Lande 1976b; Lynch and Hill 1986). The alternative hypothesis is that variation is in some way adaptive, with different forms representing specialized adaptations to a heterogeneous environment (Wilson 1989; Smith and Skúlason 1996). In this chapter we address the adaptive significance of phenotypic variation among individuals within populations and argue for the importance of natural selection – specifically, density-dependent, frequency-dependent, and disruptive selection – in sustaining this variation. To do so, we first examine how variation of this type relates to optimality theory and how it may be useful in addressing questions of adaptation and optimality. Next, we briefly review germane theoretical claims for the maintenance of phenotypic variation in populations, and finally we consider several empirical examples that test hypotheses concerning the adaptive significance of polymorphic variation.