of these vanished tribes—their form and stature, their arts, their mental capacity, their state of civilization; but the races, ethnologically speaking, to which they belonged, can not be ascertained, except by other data than those which we now possess. Whether they were Aryan, Iberian, Uralian, Semitic, Eskimo, Algonkin, Dakotan, Zuñi, Navajo, or whether they belonged, as Prof. Boyd Dawkins supposes of the river-drift men, to a race now utterly extinct, will never be known, unless, as in the case of the Assyrian mounds, some relics are discovered from which their speech can be ascertained. Until this is learned, their affiliations of race will be merely matter of conjecture; and conjecture is not science. As soon as the language is determined, the race will be known. The instant assent which every ethnologist will give to this assertion proves at once, without need of further argument, the truth of the proposition that language is the sole test of race. As the proper deductions from the foregoing facts and arguments, the following propositions are presented for the consideration of anthropologists:
1. The only sure and scientific method of grouping the tribes of men, to show their descent and affiliations, is by the evidence of language. The grouping of men by their languages constitutes the science of ethnology.
2. Ethnologically speaking, the terms "race" and "linguistic stock" are synonymous. The people of each linguistic stock, in their original and unmixed condition, are distinguished from those of other stocks by various peculiarities of physical traits and character, as well as of religion, customs, and arts. The physical differences may, in certain cases, be comparatively slight, as among the American aborigines, and the stocks of Central Africa; but to a practiced eye they are always apparent. When the differences in this respect between two stocks are slight, the inference is simply that, since those stocks originated, the climatic and other influences which affect the physical type have been nearly the same for both.
3. Whenever a mixture of races is indicated in any community by peculiarities of physical traits which can not be ascribed to climate or other natural causes, the language, on a careful analysis, will always show traces of a corresponding mixture; and, on the other hand, a mixture of languages belonging to different linguistic stocks is an invariable indication of a mixture of blood.
To sum up briefly our conclusions, a scientific treatise on ethnology will commence, like a treatise on chemistry, with the primary elements, which, as has been said, are the linguistic stocks. It will determine, as far as possible, the mother-tongue and the original geographical center of each stock. It will describe the moral and intellectual traits and the physical characteristics of the people. It will ascertain their mythology, their social system, their industries, and arts. It will trace their migrations, their interminglings with other septs, and the moral