On the heritability of developmental stability

Mo*ller and Thornhill (1997) assessed the heritability of developmental stability by a meta-analysis of all published (and some unpublished) heritability estimates of fluctuating asymmetry in bilateral traits. From a limited number of these studies they also calculated the additive genetic coefficient of variation (CVA). The usefulness of CVA is that it may correct for covariance between phenotypic variation and trait size (Houle, 1992). In eight out of the thirty-two studies that they reported, there was a statistically significant estimate of heritability. Twentyeight of these studies, representing fifteen species, were entered into a meta-analysis. From this analysis, Mo*ller and Thornhill (1997) concluded that there was a significant additive genetic component to development stability, but this component was relatively small. They also found that the CVA is approximately eight times larger for developmental stability than for trait size. In this manuscript I wish to highlight some of the benefits and limitations of this analysis, its implications and to suggest alternative approaches to the study of heritability of developmental stability that may prove useful.