Biometric method has laid great stress on variability as an index of racial purity, low variability being regarded as an indication of lack of mixture in a population. Genealogic studies on the American Negroes have shown that the American Negroes are more mixed than is generally supposed, only about 20% being unmixed Africans, while the mixture not only comprises Negro and White stock, but American Indian as well. Yet when the variabilities of this highly mixed population for various traits are compared with those for pure African, European, and Indian stocks for the same traits, those of the mixed American Negroes are found to be as low or lower than those for the unmixed populations, rarely higher. The present studies, coupled with other work on populations of known inbred biologic character, show that low variability is an index of the extent to which a population is homogeneous rather than racially pure, and a greatly mixed population which, through social or geographic isolation, has bred within itself, as has the American Negro population, will show this homogeneity after a few generations of this process. This does not argue that an unmixed racial group can not manifest the same phenomenon, for it will if inbred.