n the theory of group marriage their share in the
transaction remains absolutely mysterious.

In connection with these possible explanations of the _pirrauru_ custom,
it is important to observe that there are duties in regard to food owed
by the _pirrauru_ wife to her spouse, when her husband is absent. Now it
is hardly conceivable that in a state of "group marriage" any such
practice should have obtained. A woman would doubtless have collected
food for the man with whom she was actually cohabiting; but in the case
of the _pirrauru_ relation, the absence of the _tippa-malku_ wife of her
_pirrauru_ spouse must coincide with the absence of her own
_tippa-malku_ husband before this position is reached. So long as only
one _tippa-malku_ partner is absent, the _pirrauru_ spouse is under the
obligation of lightening the labours of the woman whose place she
sometimes occupies, and this is very far from what we should expect in
the "group marriage" stage.

On the whole therefore I conclude that the _pirrauru_ relation affords
absolutely no evidence of a prior stage of group marriage. So far from
the quantity of evidence for group marriage having been increased by Dr
Howitt's recent book, it has undergone a diminution. Gason had
stated[180] that tribal brothers had the right of access in the absence
of the husband without first being made _pirrauru_. This, if correct,
would have been much nearer group marriage than the actual facts; the
statement however appears to be incorrect, if we may judge by the fact
that Dr Howitt has silently dropped it.

Of the _piraungaru_ relation but little can be said, mainly for the
reason that our information is so scanty. We do not learn, for example,
if it is temporary or permanent, if the consent of the woman is needed,
if she ever asks her husband for a certain _piraungaru_, or if she
applies rather to her elder brothers. We do not know what becomes of the
_piraungaru_ when the primary spouse dies, whether the brother can claim
a right to his brother's wife as _piraungaru_ on giving presents,
whether married and unmarried alike enter into the relationship, whether
a woman can become _piraungaru_ before she has a special husband,
whether relations of free love are barred between a man and his
prospective wife and permitted with other _nupa_ women, and a host of
other questions. We do not even learn when access is permitted to a
_piraungaru_ spouse. We have, it is clear, far too few data to be