Additive genetic variance for fitness maintained by balancing selection

Levels of additive genetic variation for fitness (${{V}_A}( w )$) and its components (e.g., viability and fertility) appear to be high in nature, substantially exceeding levels attributable to deleterious mutations at mutation-selection balance. Balancing selection can stably maintain genetic polymorphisms and potentially explain this "excess" fitness variation. Yet, according to classical population genetics theory, balancing selection contributes nothing to ${{V}_A}( w )$ in a population at equilibrium, which has cast doubt on the potential role of balancing selection in maintaining ${{V}_A}( w )$. However, populations are only expected to be at equilibrium when there is no genetic drift and the fitness parameters of balancing selection are completely invariant over time. We explore how violations of these conditions affect the amount of additive genetic variation for fitness maintained by balancing selection. We show that drift and modest temporal fluctuations of balanced polymorphic equilibria each result in substantial ${{V}_A}( w )$, with each balanced polymorphic locus contributing as much to ${{V}_A}( w )$ as thousands of loci at mutation-selection balance. We discuss our results in reference to the surprisingly high levels of ${{V}_A}$ for fitness and life-history traits reported from lab populations of Drosophila and intensively monitored vertebrate populations in the field.