Any and every attempt to trace the evolution of religion seems at first necessarily to involve the assumption that from the beginning religion was there to be evolved.  That was the position assumed by Robertson Smith in The Religion of the Semites, which appeared in 1889.  At that date the aborigines of Australia were supposed to represent the human race in its lowest and its earliest stage of development.  In them, therefore, if anywhere, we might expect to find what would be religion in its lowest and earliest stage indeed but still religion.  Reduced to its lowest terms, religion, it was felt at first, must imply at least belief in a god and communion with him.  If, therefore, religion was to be found amongst the representatives of the lowest and earliest stage in the evolution of humanity, belief in a god and communion with him must there be found.  He who seeks finds.  Robertson Smith found amongst the Australians totem-gods and sacramental rites.  Indeed, it was at that time the belief universally held by students of the science of religion that in Australia a totem was a god and a god might be a totem.  It was conjectured by Robertson Smith that in Australia the totem animal or plant was eaten sacramentally.  Since, then, the totem in Australia was held to be both the god and the animal or plant in which the god manifested himself, it followed that in Australia we had, preserved to this day, the earliest form of sacrifice—­that in which the totem animal was itself the totem god to whom it was offered as a sacrifice, and was itself—­or rather himself—­the sacramental meal furnished to his worshippers.  The totem was eaten, it was conjectured, with the object of acquiring the qualities of the divine creature, or of absorbing them into the person of the worshipper.  That the totem was eaten sacramentally rested, as has just been said, on the conjecture made by Robertson Smith in 1889; but in 1899 Dr. (now Sir James) Frazer declared that, thanks to the investigations of Messrs. Spencer and Gillen, ’here, in the heart of Australia, among the most primitive savages known to us, we find the actual observance of that totem sacrament which Robertson Smith, with the intuition of genius, divined years ago, but of which positive examples have hitherto been wanting.’

On the foundation thus laid by the intuition of Robertson Smith and approved in 1899 by Sir James Frazer, a simple and complete theory of the evolution of religion was possible.  In any one tribe there were several totem-kins.  The totem of each kin was divine, but the personality of a totem was so undeveloped in conception that, though it might, and on the theory did, develop into a deity, it was originally more of the nature of a spirit than a god, and totemism proper might easily pass into polydaemonism, that is a system in which the beings worshipped were conceived to possess a personality more clearly defined than that attributed to totems but less developed than that