between a man and his father's sister in the Banks' Islands may be explained on similar lines, and may be the survival in mother-right of a preceding condition of patrilineal descent. Though following so naturally, such a conclusion would, in my opinion, be flying in the face of every probability. All through Melanesia we have at the present time the clearest evidence that the population is in a state of transition from matrilineal to patrilineal descent, the change having been completed in the Western Solomons, and it would seem in the highest degree improbable that this change now going on should be merely a reversal of one in the opposite direction which has left Its mark in the functions I have recorded We cannot, however, afford to dismiss any hypothesis in anthropology merely on the ground of its improbability and it will therefore be well if we keep in mind, as the first working hypothesis suggested by these customs, that they may be survivals of a condition of father-right or at least of patrilineal descent which preceded the present state of mother-right.
A second possibility is that the functions of the father's sister may have been due to the fact that she was at one time also the wife of the mother's brother. It is a frequent feature of systems of relationship,—and tokens of it are found in the Banks' Islands,—that the father's sister and the wife of the mother's brother receive the same name, and it is clear that this is either because they are actually one and the same person or have been so in the past. This is the natural result of either of two different customs, both of which are found in Melanesia, viz., the cross-cousin marriage and the custom of exchange of brother and sister with sister and brother, i.e., the custom through which, when a man marries a woman, it is at the same time arranged that the brother of the woman shall marry the sister of the man. Of these two customs, either of which would have made the fathers sister identical with the mother's brother's wife
