The question arises whether any single rule can as yet be formulated for identifying the original seats of existing peoples. By some ethnologists and historians such homes have been sought where the people are distributed in the largest area, as the Athapascan and Algonquin Indians are assigned to a northern source, because their territories attained their greatest continuous extent in Canada, but were intermittent or attenuated farther south. The fact that colonial peoples often multiply inordinately in new lands, and there occupy a territory vastly greater than that of the mother country, points to the danger in such a generalization. Of the ten millions of Jews in the world, only a handful remain in the ancient center of dispersion in Palestine, while about eight millions are found in Poland and the contiguous territories of western Russia, Roumania, Austria-Hungary and eastern Germany. Moreover, history and the German element in the "Yiddish" speech of the Russian Jews point to a secondary center of dispersion in the Rhine cities and Franconia, whither the Jews were drawn by the trade route up the Rhone Valley in the third century.[238] A more scientific procedure is to look for the early home of a race in the locality around which its people or family of peoples centers in modern times. Therefore we place the cradle of the negro race in Africa, rather than Melanesia. Density often supplies a test, because colonial lands are generally more sparsely inhabited than the mother country. But even this conclusion fails always to apply, as in the case of Samos, which has a population vastly more dense than any section of the Grecian mainland. The largest compact area including at once the greatest density of population and the greatest purity of race would more nearly indicate the center of dispersion; because purity of race is incompatible with long migrations, as we have seen, though in the native seat it may be affected by intrusive elements. When this purity of race is combined with archaic forms of language and culture, as among the Lithuanians of Aryan speech among the Baltic swamps, it may indicate that the locality formed a segregated corner of the early center of dispersion. It seems essential to such an original seat that, whether large or small, it should be marked by some degree of isolation, as the condition for the development of specific racial characteristics. The complexity of this question of ethnic origins is typical of anthropo-geographic problems, typical also in the warning which it gives against any rigidly systematic method of solution. The whole science of anthropo-geography is as yet too young for hard-and-fast rules, and its subject matter too complex for formulas. NOTES TO CHAPTER IV [126] H.J. Mackinder, Britain and the British Seas, pp. 179-187. London, 1904. W.Z. Ripley, The Races of Europe, pp. 306-310, 319-326. New York, 1899.