Flinders Petrie has acutely observed that the only meaning the term "race" now can have is that of a group of human beings, whose type has become unified by their rate of assimilation exceeding the rate of change produced by foreign elements[123]. We are also reminded by Gustavo Tosti that "in the actual state of science the word 'race' is a vague formula, to which nothing definite may be found to correspond. On the one hand, the original races can only be said to belong to palaeontology, while the more limited groups, now called races, are nothing but peoples, or societies of peoples, brethren by civilization more than by blood. The race thus conceived ends by identifying itself with nationality[124]." Hence it has been asked why, on the principle of convergence, a fusion of various races, if isolated long enough in a given area, may not eventually lead to a new racial type, without leaving any trace of its manifold origin[125]. Such new racial types would be normal for the later varietal groups, just as the old types were normal for the earlier groups, and a general application might be given to Topinard's famous dictum that _les peuples seuls sont des realites_[126], that is, peoples alone--groups occupying definite geographical areas--have an objective existence. Thus, the notion of race, as a zoological expression in the sense of a pure breed or strain, falls still more into the background, and, as Virchow aptly remarks, "this term, which always implied something vague, has in recent times become in the highest degree uncertain[127]." Hence Ehrenreich treats the present populations of the earth rather as zoological groups which have been developed in their several geographical domains, and are to be distinguished not so much by their bony structure as by their external characters, such as hair, colour, and expression, and by their habitats and languages. None of these factors can be overlooked, but it would seem that the character of the hair forms the most satisfactory basis for a classification of mankind, and this has therefore been adopted for the new edition of the present work. It has the advantage of simplicity, without involving, or even implying, any particular theory of racial or geographical origins. It has stood the test of time, being proposed by Bory de Saint Vincent in 1827, and adopted by Huxley, Haeckel, Broca, Topinard and many others.