The population variance of the proportion of genetic admixture in human intergroup hybrids

For each individual in a human hybrid population there is a proportion mu(i), whose value is usually unknown, that expresses the fraction of his genes deriving from a specified parental population. The distribution of these individual proportions about the mean proportion mu is not known for any large hybrid population in man. It is of interest to know whether the population variance of individual proportions (mu(i)) can be estimated from the variation between different, independent estimates of the mean proportion (mu). This possibility was tested with data on Negroes of the Oakland, California area, by the use of some of the principles of analysis of variance. Even with a large sample and the useful Duffy blood-group system to indicate admixture, almost no information about the population variance of individual proportions is provided by between-sample variation in estimates of mu. It is concluded that group data on admixture proportions usually do not give useful information about the population variance. It is further concluded that a recent estimate of this variance by Shockley is in error.