In the large and confused family of the clan, all the members of which were bound together by a strict solidarity of interests and a real or fictitious kinship, the restricted family became gradually established by a reaction of individual interests. On account of the more or less complete confusion of sexual unions, the first to become detached from the consanguine clan was the maternal family, based on uterine filiation, the only filiation capable of sure proof; but the great association of all the members of the clan still existed. By the simple fact of birth in this little ethnic group, the individual had rights to the territory of the clan and his share in the common resources; his clan were bound to give him aid, assistance, and, at need, vengeance also. In proportion as the family assumed more distinct proportions in the clan, it tended to become separate from it, and then, nearly always, it was based not on maternal but on paternal filiation. This did not come to pass in a day; it took a long time to arrive at the point of attributing to such or such a man the ownership of one or more women and their progeny. The ridiculous ceremonial of the couvade was probably invented during this period of transition, when it was no easy matter for a man to obtain the recognition of his paternal title and rights by the other men of the clan. For a long time the maternal family resisted the enthronisation of the paternal family, and here and there it succeeded in maintaining its existence, and in serving as a basis for the transmission of inheritance. For, whether paternal or maternal, the institution of the family, when well consolidated, had for its result the parcelling out of the possessions of the ancient clans, and the creation of familial or individual property on the ruins of the ancient common property. Finally, nothing more remained of the clan, or _gens_, but the sign or _totem_, the name, and a kinship, also nominal, between the various families that had come from it. The system and the vocabulary of kinship were then renewed; to the classificatory mode, grouping the relations by classes, without much care as to consanguinity, has succeeded the descriptive mode, which carefully specifies the degree of consanguinity of each person, and distinguishes a direct line from collateral lines, and in which each individual is the centre of a group of relations.