There is no need, therefore, to reform our conception of totemism from top to bottom. Instead, we must submit to look upon the Arunta as in a special stage of civilisation. According to M. Durkheim's phrase, we find among them at one and the same time the preparation of the future and the perpetuation of the most distant past. It is a mistake to see in them belated representatives of humanity at its first appearance; it would be no less a mistake not to recognise all the relics that subsist among them of the most primitive social forms it has been given us to know. The very complexity of their civilisation constitutes its scientific importance, for it is a rare case and one exceptionally fertile in information and suggestion. A transformation usually wrought only at a much later epoch has, among them, passed over totemic institutions at a moment when they were still strongly organised. By virtue of the reform some of these institutions have been consolidated and have thus become able to maintain themselves more easily and to survive. But they have only survived after changes which give us a glimpse of how totemism is connected with the religious systems which have succeeded it elsewhere.
The length of this analysis of M. Durkheim's luminous Mémoire must be excused by the importance of the essay. The Mémoire itself should be read in order to appreciate the full force of the argument. If I might venture to express an opinion, it would be on the whole one of concurrence, at least in the main lines of his exposition. But it must be remembered that when the results of Messrs. Spencer and Gillen's new journey are published, all these conclusions may have to be reconsidered.
M. Durkheim has in the last paragraph alluded to the survival of archaic conditions in the midst of the advance in civilisation which he claims for the Arunta. I think that this survival might have been insisted on with advantage to his explanation of some of the peculiarities of their organisation. For example, one of the most archaic features of the Arunta traditions is the want of recognition of the physical relation between father and child. The father is merely the master of the mother and provider of food for her and for her relatives. He is not the cause of the child's existence; or if a cause, he is a very indirect cause. The persistence of this view of birth (which I had been led some nine or ten years ago to infer as a primitive condition of thought, but hardly hoped to find surviving in any savages of the present day), in
