There were many tribes and many tribe-groups, or, as the latter are usually designated, “stocks,” among the Amerinds. These various stocks differed considerably from each other in manners, customs, possibly in origin, and in languages, the last often being widely different.[14] Yet there was a homogeneity binding them all together as one distinct race while at the same time separating them completely from other races of the world as now constituted. The subdivisions of the Amerind stocks were not always contiguously distributed on the continent, but, as in the case of the Navajo-Apache branch of the Athapascan, sometimes separated from their kindred by wide stretches of territory peopled by other stocks, and also, as in the case of the Navajos, somewhat altered by absorption of people of another stock. Various methods of arranging the distribution and classification of these stocks have been attempted, but the basis of language appears to offer the most advantages and the greatest accuracy. There are some who dispute the correctness of the present analysis of the Amerind languages, and deprecate the classifications obtained by this means,[15] but foremost students, like Brinton, Gatschet, Powell, Steinthal, and others, have pronounced unequivocally in favour of its value when applied with judgment. “Nothing is so indelible as speech,” wrote George Bancroft; “sounds that in ages of unknown antiquity were spoken among the nations of Hindostan still live in their significancy in the language which we daily utter.” And this fact has been the corner-stone of the modern science of linguistics, which maintains accordingly that the possession of similar _language roots_ and grammatical construction by two otherwise distinct tribes proves a relationship or a common descent. In this way, as is well understood, the Indo-Germanic—that is, our European stock—has been traced back toward its origin. The accuracy of this has been questioned, but it doubtless affords the best method attainable.[16] The same principle is applicable to the American languages, which afford an immense field for linguistic study in their great diversity. This diversity is not popularly understood, the majority of our people believing that if a person can speak “Indian” he could converse with every tribe on the continent. Yet within a limited area in Arizona he would find useless in four different tribes the language he had learned, say, in California; and in California itself some twenty or thirty tribes would listen to his words, as well as to those of each other, without a gleam of understanding. And not one of the languages of any of these tribes would serve him in the Mississippi or in the Atlantic region any better than English, for the Iroquois and the Algonquin and other Eastern tongues are as widely different from those of California as they are from each other, while every one contains numerous dialects, or what may be called sub-languages,