Dr. Rivers defines: "Kin and Kinship.--These terms should be limited to the relationship ... which can be demonstrated genealogically." This is quite a formalistic definition and does not at all meet the full facts of the case. Moreover it seems that in this way we define the unknown by what is still more indeterminate. For to draw up a genealogy we must first know who are the individuals between whom the line of descent is to be drawn; in other words we must know how fatherhood is defined in a given society. Among the Todas, Dr. Rivers had to ascertain in what way the father of a given child is determined, before he could proceed to draw up the genealogies.[574] In any case the problem of kinship requires in the actual state of things not only a purely formal definition, but a detailed analysis. Much more important as regards the present problem is the way in which Dr. Rivers has described the kinship of the Torres Straits Islanders.[575] In introducing the study of the functions of kin he points to a series of important facts which determine some social aspect of kinship and afford an insight into some of the collective ideas concerning this relation. It must be borne in mind, however, that the set of functions described by Dr. Rivers gives us only a partial knowledge of the social aspect of kinship. The every-day functions corresponding to treatment, behaviour, feeding and so forth, which characterize the intimate or home aspect of the kinship relation, ought not to be omitted. They correspond, according to our analysis, to feelings which make an essential part of the relation in question. The social functions of kin collected by Dr. Rivers, expressing certain duties and privileges of the kinsmen involved, correspond to certain customary norms. A complete collection of all legal norms and all moral rules would be an essential addition. That such moral rules do exist among the Torres Straits Islanders appears certain from the precepts given at initiation to youths.[576] [574] Of course I insist here only upon the logical and methodological priority of the psychological determination of kinship over the genealogical. In reality, wherever individual paternal kinship exists, the genealogies may be drawn first, and they possess an independent value, even if we did not know what is the content of the aboriginal idea of kinship. There is a series of highly valuable sociological conclusions that may be drawn from a system of genealogies (compare Dr. Rivers' article on this subject in _Sociological Review_, 1910, pp. 1 _sqq._).