They may not marry within the two primary divisions (which Dr. Durkheim styles ‘phratries’). Each phratry contains two (sometimes four) other ‘classes’ (exogamous), and phratries arose in the combination of ‘two elementary exogamous totem groups’--as I have already suggested. Now phratries, we have agreed with Mr. Howitt and Mr. Fison, were, in all probability, themselves originally totemic. Mr. Frazer also says, ‘We should infer that the objects from which the Australian phratries take their names were originally totems. But there seems to be direct evidence that both the phratries and subphratries actually retain, in some tribes, their totems.’[32] If the opinion be correct, the phratries of the Arunta, which regulate their marriages, were originally local totem groups. On my system, then, namely, that totem kins were originally, or very early became, exogamous, were exogamous before ‘phratries’ arose, and before the so-called ‘bisection’ was made, then the Arunta organisation was originally that of exogamous Totemism. At first, though not now, totems regulated Arunta marriages. Dr. Durkheim, in the passage cited, says that the two exogamous phratries are composed of ‘two elementary totem groups, _également exogames_.’[33] Dr. Durkheim, who here is of my opinion, writes, ‘It is not true that, among the Arunta, the totem has always been’ (as it is now) ‘without influence on marriages, nor, above all, is it true that Totemism, generally, implied endogamy.’ Yet, according to Arunta myth, the ancestors of the ‘dream-time’ (_Alcheringa_) were endogamous, as a general rule, and, as a general rule, were endophagous, ate their totem animals or plants. The ancestors of their traditions fed on their own totems, ‘as if by a functional necessity,’ say Messrs. Spencer and Gillen. But this simply cannot be true, for each totem is not in season, (plums, for instance), or accessible, all the year through, and, if it were, it would be exterminated by endophagy. The traditions, again, do not represent the men of the totem groups as really and religiously endogamous. They exercised marital privileges, not only over the women of their totem group, but over any other woman they could come across. Certain totem groups are represented in the legends as wandering across the land, the men living with women of their totem group, while ‘there is nothing to show definitely that marital relations were prohibited between individuals of different totems.’ The men accepted the caresses of such women of other totems as they encountered; but their habitual mates were the women of their own totem.[34] In the alleged state of perpetual _trek_, the wives were naturally, in the opinion of the myth makers, of the group. At present an Arunta marries in or out of his totem; as he pleases. THE RELATIONS OF TOTEMS AND ‘PHRATRIES’ AMONG THE ARUNTA