Similar conditions must have prevailed among the Athapascans, one of the great linguistic families of North America. The great body of people speaking languages belonging to this linguistic stock live in the northwestern part of America, while other dialects are spoken by small tribes in California, and still others by a large body of people in Arizona and New Mexico.[3] The relationship between all these dialects is so close that they must be considered as branches of one large group, and it must be assumed that all of them have sprung from a language once spoken over a continuous area. At the present time the people speaking these languages differ fundamentally in type, the inhabitants of the Mackenzie River region being quite different from the tribes of California, and these, again, differing from the tribes of New Mexico (Boas). The forms of culture in these different regions are also quite distinct: the culture of the California Athapascans resembles that of other Californian tribes, while the culture of the Athapascans of New Mexico and Arizona is influenced by that of other peoples of that area (Goddard). It seems most plausible to assume in this case that branches of this stock migrated from one part of this large area to another, where they intermingled with the neighboring people, and thus changed their physical characteristics, while at the same time they retained their speech. Without historical evidence, this process cannot, of course, be proved. These two phenomena,—retention of type with change of language, and retention of language with change of type,—apparently opposed to each other, are still very closely related, and in many cases go hand in hand. An example of this is, for instance, the distribution of the Arabs along the north coast of Africa. On the whole, the Arab element has retained its language; but at the same time intermarriages with the native races were common, so that the descendants of the Arabs have often retained their old language, and have changed their type. On the other hand, the natives have to a certain extent given up their own languages, but have continued to intermarry among themselves, and have thus preserved their type. So far as any change of this kind is connected with intermixture, both types of changes must always occur at the same time, and will be classed as a change of type or a change of language, as our attention is directed to the one people or the other, or, in some cases, as the one or the other change is more pronounced. Cases of complete assimilation without any mixture of the people involved seem to be rare, if not entirely absent.