the power of driving all cobras out of the place. Should rain drive the house snake out of his hole, he is worshipped. No image of a cobra or other venomous snake is ever made for purposes of worship. Ant-hills are believed to be the homes of snakes, and there the people offer sugar, rice, and millet for forty days.' These correspond to the benevolent domestic snakes, of whom Aubrey says that "the Bramens have them in great veneration; they keep their corne. think it is Tavernier mentions it." 2 I They are, in fact, as we have already seen, the representatives of the benevolent ancestral ghosts. Hence the deeprooted prejudice against killing the snake, which is both guardian and god. "If," says Mr. Lang,3 "the serpent were the deity of an earlier race, we could understand the prejudice against killing it, as shown in the Apollo legend." The evidence accumulated in this chapter will perhaps go some way to settle this question, as far as India is concerned. 1 "Panjab Notes and Queries," iii. 92, 59. 2 "Remaines," 39. Edition), i. 42; ii. 249. He perhaps refers to Tavernier, "Travels' Ball's 3 "Custom and Myth," ii. 197. VOL. II. L CHAPTER III. TOTEMISM AND FETISHISM. Olim truncus eram ficulnus, inutile lignum, Cum faber incertus scamnum faceretne Priapum, Horace, Sat. I. viii. 1-3. "A TOTEM is a class of material objects, which a savage regards with superstitious respect, believing that there exists between them and every member of the class an intimate and altogether special relation."1 As distinguished from a fetish, a totem is never an isolated individual, but always a class of objects, generally a class of animals or plants, rarely a class of inanimate objects, very rarely a class of artificial objects. ORIGIN OF TOTEMISM. As regards the origin of totemism great diversity of opinion exists. Mr. Herbert Spencer considers that "it arose from a misinterpretation of nicknames; savages first took their names from natural objects, and then confusing these objects with their ancestors of the same name, paid the same respect to the material totem as they were in the habit of doing to their own ancestors." The objection to this is, as Mr. Frazer shows, that it attributes to verbal misunderstandings far more influence than, in spite of the comparative mythologists, they ever seem to have exercised. Sir J. Lubbock derives the idea from the practice of naming persons and families after animals, but "in dropping 1 Frazer, "Totemism," I; and his article on "Totemism," in "Encyclopædia Britannica," 9th Edition Principles of Sociology," i. 367. the intermediate links of ancestor-worship and verbal misunderstanding, he has stripped the theory of all that lent it even an air of plausibility."' 1 Recent inquiries in the course of the Ethnographical Survey of Bengal and the North-Western Provinces enable us perhaps to approach to a solution of the problem. To begin with, at a certain stage of culture the idea of the connection between men and animals is exceedingly vivid, and reacts powerfully on current beliefs. The animal or plant is supposed to have a soul or spirit, like that of a human being, and this soul or spirit is capable of transfer to the man or animal and vice versa. This feeling comes out strongly in popular folk-lore, much of which is made up of instances of metamorphosis such as these. The witch or sorcerer is always changing into a tiger, a monkey, or a fish; the princess is always appearing out of the aubergine or pomegranate. We have, again, the familiar theory to which reference has already been made, that the demon or magician has an external soul, which he keeps occasionally in the Life Index, which is often a bird, a tree, and an animal. If this life. index can be seized and destroyed, the life of the monster is lost with it. These principles, which are thoroughly congenial to the beliefs of all primitive races, naturally suggest a much closer union between man and other forms of animal or vegetable life than people of a higher stage of development either accept or admit. With people, then, at this stage of culture, the theory that the ancestor of the clan may have been a bear or a tortoise would present no features of improbability. This theory accounts, as Mr. Frazer shows, for many of the obscure rites of initiation which prevail among most savage tribes and in a modified form among the Brâhmanized Hindus. The basis of such rites is probably to extract the soul of the youth and temporarily transfer it to the totem, from which in turn fresh life is infused into him. 1 "Origin of Civilization," 260, and Mr. Frazer's criticism, loc. cit. Lastly, the result of the Indian evidence is that it is only in connection with the rules of exogamy that totemism at the present day displays any considerable degree of vitality. The real basis of exogamy in Northern India seems to be the totem sept, which, however, flourishes at the present day only among the Drâvidian tribes and those allied to them. But it would, it is almost certain, be incorrect to say that while totemism is at present most active among the Drâvidians, in connection with marriage, it was peculiar to them. It is more reasonable to infer that it continues to flourish among these races, because of their isolation from Brahmanical influence. As among the inferior races of the Gangetic valley, the primitive family customs connected with marriage, birth, and death have undergone a process of denudation from their connection with the more advanced Hindu races which surround them, so to a large degree in Northern India, the totemistic sept names have been gradually shed off, and replaced by an eponymous, local, or territorial nomenclature. In short, under the pressure of higher culture, the kindred of the swan, turtle, or parrot have preferred to call themselves Kanaujiya or "men of Kanauj," Sarwariya or "residents of the land beyond the Sarju river," and Raghuvansa or Bhriguvansa, "descendants of the sages Raghu or Bhrigu." We find, then, among such races, as might have been expected, that at the present day the totemistic sept system exists only in obscure and not easily recognizable forms. Folk etymology has also exercised considerable influence, and a sept ashamed of its totemistic title readily adopts some title of the eponymous type, or a local cognomen sounding something like the name of the primitive totem. It is perhaps too much to expect that a careful exploration of the sept titles or tribal customs of Northern India will lead to extensive discoveries of the primitive totemistic organization. The process of trituration which has affected the caste nomenclature for such a lengthened period, and the obscuration of primitive belief by association with more cultured tribes, have been so continuous as to leave only a |