Perry[788] who has noted the geographical distribution of terraced cultivation and irrigation and finds that it corresponds to a remarkable extent with that of the "heliolithic" culture-complex, and by J. Wilfrid Jackson[789] who has investigated the Aztec Moon-cult and its relation to the Chank cult of India, the money cowry as a sacred object among North American Indians[790], shell trumpets and their distribution in the Old and New World[791] and the geographical distribution of the shell purple industry[792]. He points out that we have ample evidence of the practice of this ancient industry in several places in Central America, and refers to Zelia Nuttall's interesting paper on the subject[793]. Elliot Smith also discusses "Pre-Columbian Representations of the Elephant in America[794]" and remarks "coincidences of so remarkable a nature cannot be due to chance. They not only confirm the identification of the elephant in designs in America, but also incidentally point to the conclusion that the Hindu god Indra was adopted in Central America with practically all the attributes assigned to him in his Asiatic home." Elliot Smith believes that practically every element of the early civilisation of America was derived from the Old World. Small groups of immigrants from time to time brought certain of the beliefs, customs, and inventions of the Mediterranean area, Egypt, Ethiopia, Arabia, Babylonia, Indonesia, Eastern Asia and Oceania, and the confused jumble of practices became assimilated and "Americanised" in the new home across the Pacific as the result of the domination of the great uncultured aboriginal populations by small bands of more cultured foreigners. These highly suggestive studies will force adherents of the theory of the indigenous origin of American culture to reconsider the grounds for their opinions and will lead them to turn once more to the writings of Bancroft[795], Tylor[796], Nuttall[797], Macmillan Brown[798], Enoch[799] and others. There is no satisfactory scheme of classification of the American peoples. Although there is a good deal of scattered information about the physical anthropology of the natives it has not yet been systematised and no classification can at present be based thereon. A linguistic classification is therefore usually adopted, but a geographical or cultural grouping, or a combination of the two, has much practical convenience. As Farrand[800] points out "It must never be forgotten that the limits of physical, linguistic and cultural groups do not correspond; and the overlapping of stocks determined by those criteria is an unavoidable complication."