The linguistic customs of these tribes, as well as of most totem races, reveals a peculiarity which undoubtedly is pertinent in this connection. For the designations of relationship of which they make use do not take into consideration the relationship between two individuals, but between an individual and his group; they belong, according to the expression of L. H. Morgan, to the ‘classifying’ system. That means that a man calls not only his begetter ‘father’ but also every other man who, according to the tribal regulations, might have married his mother and thus become his father; he calls ‘mother’ not only the woman who bore him but also every other woman who might have become his mother without violation of the tribal laws; he calls ‘brothers’ and ‘sisters’ not only the children of his real parents, but also the children of all the persons named who stand in the parental group relation with him, and so on. The kinship names which two Australians give each other do not, therefore, necessarily point to a blood relationship between them, as they would have to according to the custom of our language; they signify much more the social than the physical relations. An approach to this classifying system is perhaps to be found in our nursery, when the child is induced to greet every male and female friend of the parents as ‘uncle’ and ‘aunt’, or it may be found in a transferred sense when we speak of ‘Brothers in Apollo’, or ‘Sisters in Christ’. The explanation of this linguistic custom, which seems so strange to us, is simple if looked upon as a remnant and indication of those marriage institutions which the Rev. L. Fison has called ‘group marriage’, characterized by a number of men exercising conjugal rights over a number of women. The children of this group marriage would then rightly look upon each other as brothers and sisters although not born of the same mother, and would take all the men of the group for their fathers. Although a number of authors, as, for instance, B. Westermarck in his _History of Human Marriage_[13], oppose the conclusions which others have drawn from the existence of group-relationship names, the best authorities on the Australian savages are agreed that the classificatory relationship names must be considered as survivals from the period of group marriages. And, according to Spencer and Gillen[14], a certain form of group marriage can be established as still existing to-day among the tribes of the Urabunna and the Dieri. Group marriage therefore preceded individual marriage among these races, and did not disappear without leaving distinct traces in language and custom.