Advances in our understanding of the physiology of many quantitative phenotypes combined with better measurement abilities is providing a means for pursuing a measured genotype approach to partitioning the phenotypic variance into the contribution of separate loci. The standard estimate of the contribution of a single locus to the phenotypic variance applied recently in the human genetics literature is a biased statistic. We compare the biased estimates from several published studies with biased corrected estimates to illustrate the general problem.