How should we explain variation in the genetic variance of traits?

Recent work has called attention to large differences among traits in the amount of standardized genetic variance they possess. There are four general factors which could play a role in causing this variation: mutation, elimination of deleterious variation, selection of favorable alleles, and balancing selection. Three factors could directly influence the mutational variability of traits: canalization, the mutational target size, and the timing of trait expression. Here I carry out simple tests of the importance of some of these factors using data from Drosophila melanogaster. I compiled information from the literature on the mutational and standing genetic variances in outbred populations, inferred the relative mutational target size of each trait, its a timing of expression, and used models of life history to calculate fitness sensitivities for each trait. Mutational variation seems to play an important role, as it is highly correlated with standing variance. The target size hypothesis was supported by a significant correlation between mutational variance and inferred target size. There was also a significant relationship between the timing of trait expression and mutational variance. These hypotheses are confounded by a correlation between timing and target size. The elimination and canalization hypotheses were not supported by these data, suggesting that they play a quantitatively less important role in determining overall variances. Additional information concerning the pleiotropic consequences of mutations would help to validate the fitness sensitivities used to test the elimination and canalization hypotheses.