But this affords no argument for holding that magic preceded religion. Rites are not necessarily magical because they are not addressed to defined personalities. They may be yet inchoate. Not until reflection has begun to clarify in some degree man’s relation to his environment (a slow and tedious process, slowest and most tedious of all in the early stages) can we reckon them satisfactorily under the one head or the other. If I am right in contending that magic and religion flow from a common source, rites may remain for generations in an indecisive condition which is neither, but may crystallize in either shape according to the specific occasion, the environment, or the dominant mental and institutional tendency of the social group. Such a transformation will be gradual and piecemeal, and in large part, if not entirely, unconscious. Many things done “for luck,” even in the higher civilizations, are {120} still in this indeterminate state. The intellectual atmosphere is unfavourable; their development is arrested, probably for ever. I suspect that an accurate appreciation of the Intichiuma rites practised by the Arunta and their neighbours would show that they too are not finally to be assigned to either category. The part played by society in the generation of religion demands some further observations. From whatever type of anthropoid ape man has been evolved, it is safe to believe that he has from the first lived in communities. But for this he could have made no progress, if even he could have existed as man. The condition of the solitary apes is incapable of improvement. It is incredible that if rudimentary human beings had lived like them in a group consisting at the utmost of a male, female, and still dependent young, they would ever have emerged into humanity, or that if they had emerged they would have been able to hold their own against the foes that surrounded them. The lowest human beings are never found solitary. If they wander on the food-quest, or are driven away from higher and more powerful societies, they do not fail to come together at certain times to enjoy the companionship of their fellows, to exchange experiences, to plan hunts or raids, to perform rites in common and partake of common pleasures. This implies organization. In fact, such communities, when they meet and live the communal life, are not found to be a mere incoherent congeries of individuals. They are true societies, organized, some more, some less closely, on a definite plan, in which every individual has his place. The Australian natives have evolved social institutions of proverbial complexity. The Bushmen of South Africa, persecuted and broken by intrusive races, have left us on the walls of the caverns they haunted representations {121} marvellous in their skill of ceremonies apparently totemic. And if this interpretation of the drawings be doubtful, such remains as have been preserved of their traditions afford evidence of an organization by no