Robertson Smith is the first who undertook this work of elaboration. He realized more clearly than any of his predecessors how rich this crude and confused religion is in germs for the future. It is true that MacLennan had already connected it with the great religions of antiquity; but that was merely because he thought he had found here and there the cult of animals or plants. Now if we reduce totemism to a sort of animal or plant worship, we have seen only its most superficial aspect: we have even misunderstood its real nature. Going beyond the mere letter of the totemic beliefs, Smith set himself to find the fundamental principles upon which they depend. In his book upon Kinship and Marriage in Early Arabia,185 he had already pointed out that totemism supposes a likeness in nature, either natural or acquired, of men and animals (or plants). In his The Religion of the Semites,186 he makes this same idea the first origin of the entire sacrificial system: it is to totemism that humanity owes the principle of the communion meal. It is true that the theory of Smith can now be shown one-sided; it is no longer adequate for the facts actually known; but for all that, it contains an ingenious theory and has exercised a most fertile influence upon the science of religions. The Golden Bough187 of Frazer is inspired by these same ideas, for totemism, which MacLennan had attached to the religions of classical antiquity, and Smith to the religions of the Semitic peoples, is here connected to the European folk-lore. The schools of MacLennan and Morgan are thus united to that of Mannhardt.188
During this time, the American tradition continued to develop with an independence which it has kept up until very recent times. Three groups of societies were the special object of the researches which were concerned with totemism. These are, first, certain tribes of the North-west, the Tlinkit, the Haida, the Kwakiutl, the Salish and the Tsimshian; then, the great nation of the Sioux; and finally, the Pueblo Indians in the south-western part of the United States. The first were studied principally by Dall, Krause, Boas, Swanton, Hill Tout; the second by Dorsey; the last by Mindeleff, Mrs. Stevenson and Cushing.189 But however rich the harvest of facts thus gathered in all parts of the country may have been, the documents at our disposal were still fragmentary. Though the American religions contain numerous traces of totemism, they have passed the stage of real totemism. On the other hand, observations in Australia had brought little more than scattered beliefs and isolated rites, initiation rituals and interdictions relative to totemism. It was with facts taken from all these sources that Frazer attempted to draw a picture of totemism in its entirety. Whatever may be the incontestable merit of the reconstruction undertaken in such circumstances, it could not help being incomplete and hypothetical. A totemic religion in complete action had not yet been observed.
It is only in very recent years that this serious deficiency has been repaired. Two observers of remarkable ability, Baldwin Spencer and F. J. Gillen, discovered190 in the interior of the Australian continent a considerable number of tribes whose basis and unity was founded in totemic beliefs. The results of their observations have been published in two works, which have given a new life to the study of totemism. The first of these, The Native Tribes of Central Australia,191 deals with the more central of these tribes, the Arunta, the Luritcha, and a little farther to the south, on the shores of Lake Eyre, the Urabunna. The second, which is entitled The Northern Tribes of Central Australia,192 deals with the societies north of the Urabunna, occupying the territory between MacDonnell's Range and Carpenter Gulf. Among the principal of these we may mention the Unmatjera, the Kaitish, the Warramunga, the Worgaia, the Tjingilli, the Binbinga, the Walpari, the Gnanji and finally, on the very shores of the gulf, the Mara and the Anula.193
More recently, a German missionary, Carl Strehlow, who has also passed long years in these same Central Australian societies,194 has commenced to publish his own observations on two of these tribes, the Aranda and the Loritja (the Arunta and Luritcha of Spencer and Gillen).195 Having well mastered the language spoken by these peoples,196 Strehlow has been able to bring us a large number of totemic myths and religious songs, which are given us, for the most part, in the original text. In spite of some differences of detail which are easily explained and whose importance has been greatly exaggerated,197 we shall see that the observations of Strehlow, though completing, making more precise and sometimes even rectifying those of Spencer and Gillen, confirm them in all that is essential.
These discoveries have given rise to an abundant literature to which we shall have occasion to return. The works of Spencer and Gillen especially have exercised a considerable influence, not only because they were the oldest, but also because the facts were there presented in a systematic form, which was of a nature to give a direction to later studies,198 and to stimulate speculation. Their results were commented upon, discussed and interpreted in all possible manners. At this same time, Howitt, whose fragmentary studies were scattered in a number of different publications,199 undertook to do for the southern tribes what Spencer and Gillen had done for those of the centre. In his Native Tribes of South-East Australia,200 he gives us a view of the social organization of the peoples who occupy Southern Australia, New South Wales, and a good part of Queensland. The progress thus realized suggested to Frazer the idea of completing his Totemism by a sort of compendium201 where would be brought together all the important documents which are concerned either with the totemic religion or the family and matrimonial organization which, rightly or wrongly, is believed to be connected with this religion. The purpose of this book is not to give us a general and systematic view of totemism, but rather to put the materials necessary for a construction of this sort at the disposition of scholars.202 The facts are here arranged in a strictly ethnographical and geographical order: each continent, and within the continent, each tribe or ethnic group is studied separately. Though so extended a study, where so many diverse peoples are successively passed in review, could hardly be equally thorough in all its parts, still it is a useful hand-book to consult, and one which can aid greatly in facilitating researches.
II
From this historical résumé it is clear that Australia is the most favourable field for the study of totemism, and therefore we shall make it the principal area of our observations.
In his Totemism, Frazer sought especially to collect all the traces of totemism which could be found in history or ethnography. He was thus led to include in his study societies the nature and degree of whose culture differs most widely: ancient Egypt,203 Arabia and Greece,204 and the southern Slavs205 are found there, side by side with the tribes of Australia and America. This manner of procedure is not at all surprising for a disciple of the anthropological school. For this school does not seek to locate religions in the social environments of which they are a part,206