The Arunta still marry, at pleasure, in or out of the totem, merely because their totems are now scattered about among their exogamous divisions. This is not the ‘natural arrangement’ (as Mr. Frazer assures us), is not the inevitable original arrangement, and is not the case with their neighbours, the Urabunna, who are confessedly ‘less developed than the Arunta.’ The Urabunna system, therefore, is more archaic, _ex hypothesi_ than that of the Arunta, which must be less archaic. It is, I repeat, peculiar, isolated, needs explanation, and the Arunta traditions give the explanation. The ancestors took women in or out of the totem, as at present the Arunta do; exogamy by _classes_ was later imposed, says the myth. Dr. Durkheim appears here to hold the more logical position. There was, I conceive, with Dr. Durkheim, and have stated, though Messrs Spencer and Gillen and others deny it, ‘a primary relationship between the totemic system and exogamy.’[41] [Footnote 1: _Ethnological Bureau, Annual Report_, 1893-1894, pp. 200, 201.] [Footnote 2: _Studies in Ancient History_, p. 221.] [Footnote 3: Suppose we take a group ranging in a given locality, and known to its neighbours as the Emu group. Let us also take a similar and similarly situated Kangaroo group. Let us suppose that each such group has raided for its wives among Opossum, Grub, Cat, and Dingo groups. By female descent, both the Emu and Kangaroo groups will contain persons of the Opossum, Grub, Cat, and Dingo groups. This being so a man of the Emu local group, named Grub by totem, might marry a woman of the Emu local group, by totem of descent an Opossum; and similarly in the Kangaroo group. But, as Dr. Durkheim remarks in another case, ‘the old prohibition’, deeply rooted in manners and customs, survives (_L’Année Sociologique_, v. 107, note). Now ‘the old prohibition’ was that a man of the Emu group was not to marry a woman of the Emu group. That rule endures, though the Emu group now contains men and women of several distinct totem kins. To escape from the difficulty, by my theory, Emu local totem group makes connubium with Kangaroo local totem group. Any Emu man may marry any Kangaroo woman not of his own totem by descent. But this does not, automatically, throw Opossum and Grub into one, Cat and Dingo into another, of the two local totem groups, Emu and Kangaroo, now become phratries, with loss of their local character. For if a man, by phratry Emu, and by totem of descent Cat, marries a woman, by phratry Kangaroo, and by totem of descent Grub, their children, by female descent, are Kangaroo Grubs. Meanwhile, if a man, by phratry Kangaroo, and by totem Cat, marries a woman, by phratry Emu, and by totem Grub, their children are Emu Grubs. There are thus Grubs in both phratries, a thing that never occurs (except among the Arunta). Therefore the division of the totem kins, some into one phratry, others into the other, is not automatic.