inconsistent beliefs is that a spirit of the husband's totem follows
the wife and enters into her wherever an opportunity offers, whereas
spirits of other totems would not think of doing so. In the example
supposed, a snake spirit is thought to have followed up the wife of the
snake man and entered into her at the tree haunted by goshawk spirits,
while the goshawk spirits would refuse to trespass, so to say, on a
snake preserve by quartering themselves in the wife of a snake man.
This theory clearly marks a transition from local to hereditary
totemism in the paternal line. And precisely the same theory could,
MUTATIS MUTANDIS, be employed to effect a change from local to
hereditary totemism in the maternal line; it would only be necessary to
suppose that a pregnant woman is always followed by a spirit of her own
totem, which sooner or later effects a lodgement in her body. For
example, a pregnant woman of the bee totem would always be followed by
a bee spirit, which would enter into her wherever and whenever she felt
her womb quickened, and so the child would be born of her own bee
totem. Thus the local form of totemism, which obtains among the Arunta
and Kaitish tribes, is older than the hereditary form, which is the
ordinary type of totemism in Australia and elsewhere, first, because it
rests on far more archaic conceptions of society and of life; and,
secondly, because both the hereditary kinds of totemism, the paternal
and the maternal, can be derived from it, whereas it can hardly be
derived from either of them.'

This argument appears to take for granted that the conception of primal
ancestral spirits, perpetually reincarnated, is primitive. But, in
fact, we seem to know it, among Australian tribes, only in these which
have advanced to the possession of eight classes, and have made 'the
great step in progress' (if it is a great step), of descent of the
totem in the paternal line. The Urabunna, with female descent of the