at least such claims and such ambitions should excite mistrust, even hostility, and the prestige of anthropology may have also suffered at times from the indiscretion of its friends. Still, its work is of wide vogue, its energy exuberant, and its influence in the future assured. In considering, therefore, the aims and methods of the comparative science of religion, it has appeared to me that its relations to anthropology {7} are now one of the main points in the inquiry. And we may seem to have reached a stage where it is desirable to test our position, to take stock as it were, to examine our methods, and to consider whether they are capable of improvement. The task is difficult, and in facing it one must face the imputation of presumption, especially as in a short course of lectures one must be brief, and may therefore appear over-dogmatic. If the comparative study of religion is to examine, as on the ground of its title it must, the various recorded or discoverable religions of every branch of the human family, then a part of anthropology, limited, as it has usually chosen to limit itself, to the study of the savage races, is obviously a sub-department of the whole. And its work, conducted often under great difficulties, has been solid, well-organised, and of high importance. Even those who deny its claim to be called a science, whatever that word may mean, must admit that it is at least an indispensable branch of historic {8} inquiry, and that it has deepened the self-knowledge of mankind. Some of its pioneers may have been overeager in their theorising, premature in their attempts to reveal the origin of all religion in some savage ritual or in the background of savage thought, for instance in ancestor-worship or totemism. Such rash generalisations are inevitable in the opening periods of a new study, and may be discredited or abandoned without discrediting the investigations that gave rise to them. We may have come to be aware of the excesses of the students of totemism: we may have come to the conviction that neither theirs nor any other special and single hypothesis has as yet supplied us with the master-clue by which we can penetrate to the aboriginal source of human religion: we may have found scientific reasons for rejecting the belief that all gods arose as ghosts of departed ancestors. But if we discard such theories of origin, we owe this negative result to the maturer study of anthropology {9} itself; and we may owe to it the positive induction that the religious product at the different stages and in the different branches of mankind was a complex growth from many different germs.