In connection with this form of marriage there are two points of importance to be noted. The first is that whereas a man may have as many tippa-malku wives as he can get, a woman cannot have more than one tippa-malkuhusband, at any rate not at the same time. After the husband's death she may again enter into the tippa-malku relation. The second point is that the tippa-malku relation must precede the pirrauru relation, of which I shall speak in a moment, and cannot succeed it[157].

There are unfortunately many points in Dr Howitt's narrative which demand elucidation. He says, for example, that noa individuals become "tippa-malku for the time being[158]." This suggests, probably erroneously, that the tippa-malku relation is merely temporary; but I am unable to say whether it in reality means that the tippa-malku relation is terminated by the capture of the woman, or that divorce is practised and may terminate the relationship at the will of the man only or of both parties.

Another point on which we have no information is the position of the unmarried girls and widows. Free love is permitted, the only limitation[159] given by Dr Howitt being that the man (who must of course have passed through the Mindari ceremony) must not be tippa-malku to the girl, but must be noa-mara. It would be interesting to know whether girls in the tippa-malku relation before actual marriage are at liberty to have sexual relations with any men of the right status or only with unmarried men, or whether the privilege is restricted to those who are not yet tippa-malku to any one, and how far the same restriction applies to the men.

Any man who has been duly initiated, whether he is married to a tippa-malku wife or not, and any woman who has a tippa-malku husband[160], can enter or be put into a relation termed pirrauru with one or more persons of the opposite sex. The effect of the ceremony—termed kandri—is to give to the pirrauru spouses the position of subsidiary husbands and wives, whose rights take precedence of the tippa-malku rights at tribal gatherings, but at other times can only be exercised in the absence of the tippa-malku spouse, or, when the male is unmarried, with the permission of the tippa-malku husband of the pirrauru spouse.

The pirrauru relation is, for the woman, a modification of a previously existing tippa-malku marriage; that being so, it cannot be quoted as evidence of a more pristine state of things in which she was by birth the legal and actual spouse of all men of a certain tribal status.

The pirrauru relation falls under two heads of the classification I have given above, according as the man has or has not a tippa-malku wife. In the first case, it is, taken in combination with the tippa-malku marriage, a case of bi-lateral adelphic dissimilar (M. and F.) polygamy. In the latter it is dissimilar adelphic (tribal) polyandry, adelphic being taken here, be it noted, in the sense of tribal, and possibly, but not necessarily, own brother.

Here too our information is unfortunately fragmentary and sometimes contradictory. We learn from Dr Howitt, for example, that a pirrauru is always a brother's wife or a wife's sister (they are usually the same), and the relation arises through the exchange by brothers of their wives[161]. But on the next page we learn that the unmarried (men) can also become pirraurus. It appears further that a woman may ask for a pirrauru, but whether he must be a married man or not is not clear. It is only stated that she has to get her husband to consent to the arrangement. Further we find that important men have many pirrauru wives, but it does not appear how far they reciprocate the attention. Then again we are told that when two new pirrauru pairs are allotted to each other, all the other pairs are re-allotted. Are we to understand from this that the allocation of new pirraurus is a rare event or that the pirrauru relationship is a very temporary affair? Or does re-allotted simply mean that the names are called over? If the latter, the terminology is very unfortunate. Gason's statement is perfectly clear: once a pirrauru, always a pirrauru[162]. Again does it imply that the wishes[163] of the already existing pirraurus are consulted in the matter or not? If, as is stated, there is a good deal of jealousy between pirraurus, especially when one of them (the male) is unmarried, it is difficult to make the two statements fit in with one another. Once more, it is said that a widower takes his brother's wife as his pirrauru, giving presents to his brother. Does this imply that the consent of the husband is not necessary, or that he cannot refuse it, or that it is purchased? Again we read "a man is privileged to obtain a number of wives from his noas in common with the other men of his group, while a woman's wish can only be carried out with the consent of her tippa-malku husband." This latter statement clearly implies that a man can obtain a pirrauru without the consent of the tippa-malku husband, but this contradicts what has already been told us about the exchange by brothers of their wives. Exchange is clearly not the right term to apply; if one or perhaps both have no voice in the matter, it is rather a transfer. These are by no means all the unsettled questions on which light is needed. What, for example, is the position of a pirrauru wife whose tippa-malku husband dies? Does she pass to a new tippa-malku husband? If so, must he be an ex-pirrauru? Does she continue in the pirrauru relation to her former pirraurus, regardless of her new husband's wishes? Can the pirrauru relationship be dissolved at the wish of either or both parties and by what means?

With so many obscurities in the narrative we must esteem ourselves fortunate that we are not left without the information that a special ceremony is necessary to make the pirrauru relation legal; this is performed by the head or heads of the men's totems, and need not be described here.