an assertion of his paternity, thus bringing her functions into line with those which, according to one view, belong to the Couvade. There is little doubt that the latter institution is based primarily on the belief in a sympathetic relation between father and child, but in its more developed forms it is possible that the assertion of paternity may have played a part, and Mr. Hodson's suggestion should be borne in mind as affording a fourth working hypothesis by means of which to seek the explanation of the functions of the father's sister. According to this view it would be expected that the cord and nail-parings would be given to the father's sister to give her a hold over the child, a means of compulsion in disputes between the father and his wife's people.
Of these four hypotheses the first has been advanced chiefly as a matter of form, and I am inclined to attach most importance, so far as concerns the original basis of the customs, to the third, while the conditions assumed in the second and fourth hypotheses have probably been also in action. According to this view the origin of the special functions of the father's sister was in her position as the member of a different social group who stood nearest to the child, whether the actual motive was the fear of magic which I have suggested or some other. Later this special position of the father's sister was strengthened by other relationships to her nephew or niece which came into existence, perhaps as the wife of the mother's brother, but probably still more as the potential mother-in-law, while it is also possible that the desire of the father to assert more definitely the paternity already implied in the functions of his sister may have added to her importance. According to this view we should have in the development of the functions of the father's sister one of those cases of complex causation which I believe to be the rule in sociology.
The foregoing hypotheses are directed towards the
