Students of languages have shown that the most enduring characteristics of man are to be found in his speech. The fundamental principles of a language outlive not only political and social changes, but even physiological distinctions, and are inherited from a primitive stock by all its branches. We might reasonably expect, therefore, that a study of the languages spoken in America in pre-Columbian days would be a sure index as to the primitive stock from which the various tribes came, and show to which of the many other branches of the linguistic tree they are most closely related. Turning with this question to those who have made a critical study of the languages of the American aborigines, and no one is better qualified to bear testimony in this connection than J. W. Powell, the honoured director of the American Bureau of Ethnology, we find a definite answer. He says: "The North American Indian tribes, instead of speaking related dialects, originating in a single parent language, in reality speak many languages, belonging to distinct families, which have no apparent unity of origin." To the north of Mexico (Plate VI) the aborigines are divided into 58 linguistic families. In a large portion of these languages there are tribal dialects not understood by members of other tribes of the same family. Thus the Algonquin linguistic family contains some 30 or 40 distinct languages. In the Athapascan the diversity is nearly as great. The smaller families present similar conditions in proportionate degree, although there are stocks which speak but one language. Four of the linguistic families referred to extend into Mexico, but to the south of the territory occupied by them other languages and dialects are spoken. Ethnologists who have studied the tribes of Mexico report 19 linguistic stocks, containing 108 distinct languages, among which there are upward of 60 dialects. In Central America a similar diversity in the native tongues exists. Reclus, in his great work The Earth and its Inhabitants, states that in the New World 450 native languages are spoken--a number greater than that of all the languages in use in the rest of the earth. Not only are the American linguistic stocks different from each other, and fail to furnish evidence of having been derived from a single parent tongue, but, as philologists assure us, no one of them is analogous to any language spoken in other lands. As is well known, a language is not created _de novo_, but by a slow process of development. Since the first acquirement of articulate speech by man a succession of languages has appeared owing to the growth, differentiation, etc., of pre-existing forms of speech. It is a warrantable inference, therefore, that the marvellous diversity in speech present in America could only have arisen by a process of evolution involving a very long period of time.