THE ESTIMATION OF GENETIC CORRELATIONS FROM PHENOTYPIC CORRELATIONS -A TEST OF CHEVERUDS CONJECTURE

The estimation of genetic correlations is central to the study of evol utionary change in populations. However, sample sizes required to achi eve a small standard error are typically enormous. This precludes larg e-scale comparative analyses. Cheverud has conjectured that in some ci rcumstances the phenotypic correlation can be substituted for the gene tic correlation. This suggestion is examined using a large set of morp hological traits in the sand cricket, Gryllus firmus. In this case the difference between the two estimates is very small. Further, by simul ation it is shown that the phenotypic correlations are as good as, or better than, the estimated genetic correlations as estimates of the tr ue genetic correlations. Examination of other data sets of morphologic al traits suggests that the phenotypic correlation may, in general, be a suitable substitute for the estimated genetic correlation. However, because the number of such examinations is still small, a protocol is suggested in which two sets of genetic analyses are undertaken to con firm the assumption in a large comparative analysis.