Mr. Frazer, in 1899, begins with numerous totem groups, primarily and originally arranged for mere purposes of co-operative magic, in the social interests of a large friendly tribe, itself no primitive institution, one thinks. Then he supposes that the exogamous bisection occurred (and _why_ did it occur?), and then ‘if the existing totem groups were arranged, as they naturally would be, some in one of the two new classes, and the rest in the other, the exogamy of the totem groups would follow, _ipso facto_.’[13] Mr. Frazer does not here pretend to guess why the bisection occurred. The rest is quite obvious: but it is unavoidably inconsistent with Mr. Frazer’s earlier theory, that a tribe begins (or that the Thlinkets began) with two original totem groups, made them exogamous, and then ‘subdivided’ them up (or did they merely swarm off?) into many totem groups. It is against that almost universal theory, in 1899 abandoned (as I conceive) by Mr. Frazer, that I have so long been arguing. There was not first an exogamous bisection of a tribe, or the addition of the exogamous rule to two ‘original clans,’ or totem groups, and then the subdivision of each of the two sections into a number of totems. This cannot have occurred. Totems, I venture to think, did not come in that way, but pre-existing totem kins, granting the bisection, might fall into one or other phratry, if they had always been exogamous. ADVANTAGES OF THE SYSTEM HERE PROPOSED On my system, as has been already stated, the origin of exogamy may have been sexual jealousy, in small primitive groups, perhaps aided by ‘sexual tabu,’ with the strange superstitions on which it is based, and these causes would be strengthened enormously by the totem superstition, later. The totem name would now be the exogamous limit. The ‘phratries’ might result, quite naturally, and even gradually, now in one region, now in another, from the interlocking and alliance, with _connubium_, of two large friendly local totem groups, an arrangement of which the advantages are so obvious that it might spread by way of imitation and accretion.