The Origin of Exogamy and Totcnnsni. 185
are obviously not given with completeness. Some kins marry into three totems. Some into two. Some only into one other totem. " This table is evidently imperfect." ^- Evidently there are intermarrying combinations of totem- kins within the federation of the phratries. The compact is not between one totem-kin and one totem-kin, but none must marry into ^/Z the totems of the phratry not his own.
In writing all this I am incurring the rebuke of Mr. Golden weiser. He writes that we "Britishers" seize upon " ;i feature salient in thetotemic life of some community only to be projected into the life of the remote past, and to be made the starting point of the totemic,"^^ (in my case of the phratriac) "process." This is " methodologically unjustifi- able." In making a hypothesis, I think I may seize on a salient feature of totemic life in three "nations" more "primitive" than any others known to us. I then try how the feature works into my hypothesis of the origin of phratries in totemic communities. Well, it drops in like the keystone into the arch ! There is the bridge. " Walk over, my Lady Lee ! ", into the land of a theory which, at least, shows how phratries containing each a distinct set of totem- kins viigJit come into existence.
On my theory the primal prohibition was not based consciously on consanguinit}', but on locality and ownership. The semi-brutal Sire says, — " No amours except my own in my camp." When the groups got names, — Emu, Lizard, Grub, Iguana, Kangaroo, — the prohibition was " no amours within the name." When two groups first coalesced into connubiinn, the first rule was " no marriage with peace save into one other totem group." The final rule was " marriage into any totem-kin not in your own phratry." As the rise of the phratries instantly and automatically produced classificatory relationships or " classes," people were confined in marriage to one set of such relations in the opposite
^-Howitt, op. ciL, pp. 18S-9.
^^ The loiimal of American Folk-Lore, vol. xxiii. (1910), p. 2S0.