The majority of the measures we studied are amenable for QTL mapping, as showed by high genetic effect sizes. Among these measures, 64.61% have genetic effects accounting for greater than or equal to 30% of the phenotypic variation. For those that are under weaker genetic influences, a genetic reference population offers a remedial measure to increase power by increasing the sample size within strain, especially when heritability is low. This improved phenotypic precision allows mapping of QTL for traits with heritability as low as 10% in smaller BXD populations (Belknap 1998). For lower heritability traits (h2 = 0.10), use of 58 RI strains with a within-strain sample size of 20 is equivalent to mapping in an F2 population of size 760, without the expense of genotyping individual mice and with the added value of genetic correlation across many traits. Crusio (2004) showed that maximum similarity between additive QTL effects and the correlation between a molecular marker and a strain behavioral or neuronal phenotype is achieved either when heritability approaches unity or within-strain sample sizes are infinite. At a heritability of
