Considering, now, the laws of inheritance as outlined before, it would appear that in a community of this type, the members of which number little more than two hundred individuals, the ancestor of every single individual from the eighth generation back must have been the same, because the eighth generation would require theoretically two hundred and fifty-six individuals,—a greater number than are actually found in the community; and the occurrence of any individuals who have not a good many near and remote ancestors in common with the whole rest of the community is highly improbable, if not impossible. It follows from this at once that the variability of the whole series around its typical average must be rather small, because all the members of the group will have a certain amount of family resemblance. This uniformity of type will, of course, be the greater, the more uniform the ancestral group. On the other hand, these conditions bring about another peculiar characteristic of the isolated group. Owing to the constant recurrence of the same ancestral types among the whole group, the type of the whole people becomes very similar to the characteristic traits of the small ancestral group; and the smaller this group, the stronger will be the probability of the type of the local group being quite distinct from the type of the whole people to which they belong. It seems to my mind that these conditions explain to a considerable extent the occurrence of distinct local types in primitive races. When we find, for instance, that in North America a very marked type belongs to the Arctic coast of the continent, that a quite distinct type is found in the Mackenzie basin, still others in well-defined localities on the Pacific coast, again others in the Mississippi basin, one in the southeast, and one along the Rio Grande and in Mexico,—it seems plausible to ascribe their origin to the increase of small isolated groups, which, as we have seen, must necessarily lead to differentiation of type. This view of the origin of local races is quite in accord with the remarkable results obtained by Johannsen in his studies of heredity. He repeated artificially the conditions which prevail in a small community, and exaggerated them by selecting types of equal form, and by propagating them by self-fertilization. He raised in this way bean-plants from beans of equal weight, and was able to restrict variation of the type, so that practically any kind of bean of particular weight and particular form could be bred true to the type of its ancestor, and without perpetuating the accidental variations of the ancestors. In cases where the ancestry goes back to a limited number of individuals, as among our isolated tribes, the variation will, of course, not be restricted to the same extent; but the characteristic development of a stable type is quite analogous to the experiment made by Johannsen.