Correspondence. 83
certain ceremonies that the existence of a mutual relationship, consequent upon the possession of a common totemic name, stands out at all prominently " {Native Tribes, pp. 34, 544).
In his criticism of other writers' theories concerning the origin of totemism Dr. Frazer is not always convincing. Against the American theory, according to which the totems of clans are merely the guardian spirits of ancestors transmitted by inheritance to their descendants, he argues that it encounters a serious difficulty in the comparative insignificance of the guardian spirits of women, which is hard to reconcile with descent of the clan totem in the female line. But why could not a person inherit the guardian spirit of his maternal uncle, as he in many cases inherits his property? Dr. Frazer, moreover, maintains that the theory in question is not borne out by the evidence of totemic tribes outside America, with whom personal guardian spirits appear for the most part to be wanting. This argument seems a little curious when we consider that Dr. Frazer's own theory rests on a belief which he has found to prevail in a single group of islands, among people who have no totemism in the proper sense of the word, and which he neverthe- less presumes t.o be at the bottom of totemism wherever it occurs. Nor has Dr. Frazer, in my opinion, been quite successful in his criticism of Wilken's theory that totemism originated in the doctrine of metempsychosis, — a theory which is supported by the idea held by many Bantu tribes that the souls of their dead are incarnate in their totems. Dr. Frazer argues that, if a belief in the transmigration of souls had been the origin of totemism, that belief would have been found lingering among the Australian aborigines, "the most primitive totemic race with which we are acquainted " ; hence he is inclined to regard the Bantu belief as a later development rather than as the source of totemism. But in this, as in some other cases. Dr. Frazer's reasoning is vitiated by the presumption that totemism has everywhere had a similar origin, and that therefore a theory which satisfactorily explains the origin of this institution in any one race will probably explain its origin in all races. Considering the important differences which totemism presents among different peoples, this presumption seems to me very doubtful; and, even if it were correct, we should not be justified in concluding that the form of totemism which prevails
