s,
as we have seen, Dr Howitt declines to accept on the strength of the
philological argument.

We are therefore reduced to the levirate as a proof of the former
existence of group marriage. But there is nothing whatever to show that
it is not a case of inheritance of property. For the Australians, as for
many other savage peoples, the married state is the only thinkable one
for the adult, and that being so it is natural for the widow to remarry.
She has however been purchased by the exchange of a woman in the
relation of sister to the deceased, and if the widow were allowed to
pass to another group, the property thus acquired would be alienated.
Moreover the marriage regulations require the woman to marry only a
tribal brother of the deceased. It is therefore in every way natural for
a brother to succeed to a brother. No arguments for the prior existence
of group marriage can be founded on the levirate, any more than an
argument for primitive communism can be founded on other laws of
inheritance. At most the _maian_ relationship is evidence of adelphic
polygyny[172].

For the Urabunna we depend on the information gained by Spencer and
Gillen on their first expedition. Here the circle from which a man takes
his wife is much more restricted than among the Dieri. Not only is he
bound to choose a woman of the other moiety of the tribe, but he is
restricted to a certain totem[173] in that moiety, and to the daughters
of his mother's elder brothers (tribal) in that totem. Hence although
the _kami_ relationship of the Dieri is unknown among the Urabunna, the
choice among the latter is more limited.

The marriageable group is termed _nupa_ by both men and women; in
addition to the _nupa_ relationship and the unnamed individual marriage,
into which a man enters with one or more of his _nupa_, there is the
_piraungaru_ relationship, corresponding to the _pirrauru_ of the Dieri.
In each case the elder brothers of the woman decide who are to have the
primary and who the secondary right to the female. In the case of the
_piraungaru_ however the matter requires confirmation by the old men of
the tribe. The circumstances under which the _piraungaru_ claims take
the first rank are not stated by Messrs Spencer and Gillen; the
statement that a man lends his _piraungaru_ need not, of course, refer
to times at which he himself cannot claim the right of access[174].

We may now turn to a discussion of the bearing of the facts just cite