It has been mentioned above that the pirrauru custom, so far from being an extension of the recognised practice of Australian tribes, is in some respects a limitation of it. We may now proceed to illustrate this. Even among the Dieri the tribal festival on the occasion of an inter-tribal marriage is marked by free intercourse between the sexes without regard to existing sexual unions181 (? either tippa-malku or pirrauru). In the same way the Wiimbaio tribal gatherings were accompanied by regulated promiscuity, the class rules being the only limitation. At others wives could be lent or temporarily exchanged by the husbands182. The Geawe-gal held festivals at which wives were lent to young men, subject to class laws183. In other cases the exchange was limited to brothers or men of the same totem184. Among the Kamilaroi a wife was lent to friendly visitors but only with her consent. In all these cases we see a state of things similar to or not unlike the relations of the Dieri pirrauru spouses, and it should be noted that it is at tribal gatherings that the latter can claim to exercise their rights. From this it appears that the Dieri custom amounts to an ear-marking of certain women for the use of certain men, and is consequently a limitation of the common custom; in consideration of the fact that the pirrauru men protect them in the absence of their husbands, they are permitted at the same time to exercise marital rights, provided their own primary spouses are absent.

Among the Wiimbaio, when sickness was believed to be coming down the Murray185, and among the Kurnai, when the Aurora australis was seen186, an exchange of wives was ordered by the old men to avert the threatened evil187. This is explained by Dr Howitt as a reversion to the ancient custom of group marriage. It is however not quite clear on what grounds it is necessary to treat it as a survival at all. If a day of prayer and fasting is ordered in order to avert national calamities, it does not follow that the nation in question was in the habit of perpetual prayer and fasting at some previous stage of its existence. Moreover, if the magical rite was formerly the universal practice we may well ask what induced the tribes which believe in its efficacy to adopt a new form of marriage. Ex hypothesi, it is pleasing to Mungan, or good against disease; knowing this, they have not hesitated to abolish group marriage, but apparently without incurring Mungan's wrath, or bringing any epidemic upon them.

Among the Narrinyeri188, the old men have a right of access to the newly initiated girls, but apparently Dr Howitt does not regard this as a survival. On the other hand the narumbe (initiated youths), who may not at this period take wives, had unrestricted rights over the younger women, those "of his own class and totem not excepted," and this Dr Howitt regards as a survival from the days of the undivided commune, though if it is so it is hard to see why they should have rights only over the younger women. The practice does not appear to differ from the free love found among the Dieri except in the absence of class restrictions and its limitation to the period after initiation which is among many other peoples a period of sexual licence.

Another group of customs, also interpreted by Dr Howitt as a survival of group marriage and an "expiation for individual marriage," calls for some discussion. It is unnecessary to refer here to the explanation of the jus primae noctis suggested by Mr Crawley. It may be that the matter can also to some extent be explained as payment for services, in the same way as the pirrauru relation shows some signs of being a quid pro quo.

In certain tribes access to the bride is permitted to men of the group of the husband. Among the Kuinmurbura they are the men who have aided the husband to carry off the woman189; and the same is the case with the Kurnandaburi and Kamilaroi tribes190. It is very significant that among the Narrinyeri the right of access only accrues in case of elopement and precisely to those men who actually give assistance in the abduction, a fact hard to explain on the theory of expiation191. Among the Mukjarawaint the right seems to belong to those of the same totem, but apparently the young men only192; but here too their position as accessories is quite clear, as indeed it must be in any tribe where the right accrues to men of the same totem. By all the rules of savage justice a punishment may be inflicted in these cases either on the offender himself or on the men of his totem. It is therefore not strange that they require from the abductor some return for the danger to which he exposes them, especially if they actually take part in the abduction. An aberrant form of the custom is found among the Kurnai, among whom the jus primae noctis falls to men initiated at the same jeraeil as the bridegroom.

Among the Kurnandaburi there was a period of unrestricted licence after the exercise of the jus primae noctis, even the father of the bride being allowed access to her. This did not of course violate totem or phratry regulations. Dr Howitt does not comment on the case, but it would have been interesting to hear whether both these customs are to be regarded as survivals and if so what caused the duplication193.

In estimating the value of the custom of jus primae noctis as evidence of a prior state of group marriage, a custom of the Yuin should not be overlooked. If a man elopes with another man's betrothed he is punished by having to fight the girl's father, brothers, and mother's brother; the girl was sometimes punished by being beaten; all the men who pursued her had a right of access provided they were of the right totem and locality. If however the eloping couple were not caught they were not liable to punishment after a child was born. There is no mention of any jus primae noctis where the marriage was the result of betrothal. In this case therefore the right of access is a punishment, so far as the girl is concerned; it is earned by taking part in the pursuit, a fact which confirms the suggested explanation of the right of access at marriage.

It should not be overlooked that this form of punishment is found among some tribes as the penalty for adultery194, when it certainly cannot be interpreted as an expiation for individual marriage. This was the case among the Wotjoballuk, the Kamilaroi, and the Euahlayi.

We may now turn to the customs of the central and northern tribes visited by Messrs Spencer and Gillen. Except in the case of three of the north-eastern tribes the right of access accrues in connection with the ariltha ceremony. It may be said at once that there is among these tribes no trace of access as payment for services; for on the rare occasions when a wife is captured she is allotted to an individual and becomes his property at once, according to a statement in the first work of Spencer and Gillen195. In the same work, it is true, this statement is contradicted by the assertion that on such occasions only the men of the right class are allowed to have access196. But this statement does not seem to be based on any facts within the knowledge of the writers, for they make a definite statement to the contrary with regard to the Arunta customs, and it was with the Arunta that they were specially concerned, and in the later volume no further details are given, as they should have been, if the custom was found among any of the tribes visited on the second expedition.

The association of the right of access with the initiation ceremony is paralleled, as we have already seen, among other tribes. It hardly seems necessary to argue a state of primitive promiscuity from a custom of licence at the period of puberty, which does not in fact differ, except in degree, from the licence normally enjoyed by the unmarried, and is readily explicable on other grounds than those suggested by Spencer and Gillen. If we are not prepared to regard this licence at puberty, which may equally well have subsisted side by side with marriage or group promiscuity, as a mere expression of the newly attained sexual rights, we have as an alternative the magical theory of Mr Crawley. I do not propose to dwell on this but will pass at once to discuss some points which seem to have escaped the notice of Spencer and Gillen when they proposed their hypothesis of promiscuity.

The essential point in connection with these ceremonies is the fact that access is not limited, as in the case of the Dieri, to men who might lawfully marry the woman. The right is restricted to men of six classes out of the eight, including all four of the other moiety and the two of her own half of her own moiety. Now whatever else may be deduced from this, one thing is clear, and that is that the custom in its present form, at any rate, took its rise before the eight classes were introduced but after the four classes were already in existence and a fortiori after the phratries were known. Consequently no argument for promiscuity can be founded on the right of access at initiation. It cannot be a survival from a time when no marriage regulations were known, for the simple reason that the custom itself bears unmistakeable traces of regulations of a comparatively advanced type. It may of course be argued that these limitations are of late origin. How far this is so and why such limitations should have been introduced it is impossible to say; but it is impossible to base an argument for primitive promiscuity on a state of things which is admittedly not primitive unless we have good primâ facie grounds for regarding the custom as a survival. There is nothing in the present case to show that it is not a magical rite.

At other times access is permitted in accordance with class regulations, the husband's consent being necessary, if indeed he does not actually take the preliminary steps himself. We have seen that a similar state of things exists in other tribes. It does not seem necessary to look for the explanation further than the ordinary customs of savage hospitality, the desire to do a favour to men who may be useful. It is difficult to see why Spencer and Gillen regard the fact that women are lent in this way only to their unawa as a proof of the former existence of group marriage. Clearly if intercourse is permitted only between certain persons before marriage and only certain persons are allowed to marry, we can hardly be surprised to find that these latter are restricted in the choice of men to whom they may lend their wives after marriage. The surprising thing would be if it were otherwise.

In addition, as in the tribes we have already considered, irregular access is practised for magical purposes in connection with the performance of ceremonies and the sending out of messengers. It has already been pointed out that we have no grounds for regarding such practices as survivals; for if we put on sackcloth and ashes as a penance for our misdeeds, it does not follow that this was ever the prevailing costume. It is even less possible to interpret the ritual lending of wives to messengers as a survival, for, ex hypothesi, the messengers were not of the group which "group-married," and messengers of any sort point to a stage when inter-tribal relations had made considerable advance and the tribes in question are hardly likely to have been still in the stage of the "undivided commune."

The survey of Australian customs and terms of relationship leads us to the conclusion that the former, so far from proving the present or even former existence of group marriage in that continent, do not even render it probable; on the latter no argument of any sort can be founded which assumes them to refer to consanguinity, kinship or affinity. It is therefore not rash to say that the case for group marriage, so far as Australia is concerned, falls to the ground. Even were it otherwise, even were group marriage proved for Australia or for any other part of the world, we should still be far from having established promiscuity and group marriage as a stage in the general history of mankind. For that at least a scheme of development is needed. Even were the arguments in favour of the group marriage hypothesis much stronger, its supporters might reasonably be asked to give us something more than assertion and reassertion without any attempt to show in detail the process of evolution. To take an example from another sphere, it may safely be said that the general theory of evolution would find few supporters if it were not possible to trace some existing species and genera back to some generalised type in the past. At present the position of a supporter of the theory of primitive promiscuity and group marriage is analogous to that of an evolutionist who can only point to a few more or less useless peculiarities in the anatomy of man without being able to show resemblances between them and the corresponding portions of fossil or actually existing anthropoids. He calls them "vestiges197" and insists that homo is descended from a generalised anthropoid. The mere assertion of the vestigial character of such bones or organs would hardly carry conviction unless they could be shown to exist in some anthropoid in a more fully developed state. Similarly the arguments for promiscuity and group marriage suffer from incurable weakness, and would so suffer, even were the basis far more reliable than I have shown to be the case, unless and until it has been shown by what process and for what reasons man took each upward step. So far only one writer has attempted, and that nearly thirty years ago, to trace the course of human development on the hypothesis of primitive promiscuity, and his scheme is a house of cards.

The student of sociology is at a disadvantage compared with the zoologist in not being able to unearth his fossils for comparison with living forms. He must therefore trace the relationship between living forms, and, in seeking to discover the earlier stages of human progress, rely in part on the sociology of the higher mammals, in part on the possibility of showing a logical scheme of human development. When he examines the living forms he is of course unable to say whether actually existing savage institutions are in the main line of human progress or merely bye-paths embryological or teratological. It may be possible to show that group marriage exists somewhere on the earth at the present time. Even if this is so, the theory of primitive promiscuity and group marriage as stages in the general history of mankind remain mere baseless guesses until we have a systematic account both of the causes which led to the various steps, and of the processes by which the various stages were reached.

APPENDIX.

A certain number of Australian tribes have ceased to adhere strictly to the regulations of their class systems. Thus, in the Kamilaroi tribe a correspondent of Dr Howitt's found intra-class marriage, the totem only being different; in determining the class and totem of the children the ordinary rule held good198. The Wiradjeri on the Lachlan permit Ipai to marry Muri as well as Kumbo, the two classes both belonging to Kupathin; in each case certain totems only, viz. emu, opossum, snake and bandicoot, have the privilege199. The same anomaly is found in the Wonghibon tribe200.

Among the Warramunga and other northern tribes Spencer and Gillen find that the division of the classes, explained in the last chapter, does not prevent marriages from taking place which this division ought to prevent, if the Arunta rule were followed201. A curious feature of these marriages is that the children of the anomalous union pass into the class which would have been theirs if their mother had wedded her normal spouse. It is not easy to say whether this should be regarded as a survival of matrilineal descent; it is, however, clear that only the existence of phratriac names enables us to say definitely that the descent in this tribe is in the male line.

According to the information printed by Mr R. H. Mathews this irregularity is by no means the sum total of anomalies. His information is far from being commonly accepted as accurate; but, as will be shown later, there are correspondences between his statements and those of other observers, which make it probable that his statements have some basis in fact. At any rate they deserve notice, if only that they may be contradicted by competent witnesses, if they are incorrect.

In the Inchalachee tribe, according to Mr Mathews, descent of the classes is reckoned through females. In the place of the arrangement shown in Table Ia, he gives the order 3, 4, 8, 7; 6, 5, 1, 2202. Any man of the first moiety may marry any woman of the second, though certain marriages are normal and one of the remainder more usual than the others. The effect of these rules is to make it possible for a man to marry any woman of his own generation, even if she be of his own class. This is precisely the same as the case reported from the Kamilaroi by Dr Howitt, if we may take it that in the latter case the normal marriages are found side by side with the anomalous ones.

In the Inchalachee marriages the children, as in the Warramunga cases of Spencer and Gillen, take the class which they would have had if the woman had taken her normal spouse. On this Mr Mathews relies for the statement that descent is reckoned in the female line in this tribe. But, as we have seen, such a view is erroneous as regards the Warramunga, among whom anomalous marriages also occur; it is therefore by no means clear that the Inchalachee are matrilineal. We have even more reason to doubt his view as to the Binbinga, for whom we have the evidence of Spencer and Gillen.

Mr Mathews also reports among the Wiradjeri marriages resembling in many respects those mentioned above from the Wailwun tribe203. The table does not seem to be complete; it is therefore useless to enquire on what principle these marriages are arranged. There seems, however, no reason to doubt the substantial accuracy of the information.

More revolutionary is the statement that these cross-class marriages are based on an actual kinship organisation, to which Mr Mathews gives the name of "blood" (Table III, p. 50)204.

Running across the phratries and classes are divisions known as Gwaigullean and Gwaimudthen, Muggulu and Bumbirra, etc., which have the meaning of "sluggish" and "swift" blood respectively. The bloods again are sometimes subdivided. In the Ngeumba tribe Gwaimudthen is divided into nhurai (butt) and wangue (middle), while Gwaigulir is equivalent to winggo (top). These names refer to different portions of the shadow of a tree and refer to the positions taken up in camping by the persons belonging to the different "bloods" and "castes." In this, it may be noted, these organisations follow the parallel of the phratries and classes.

With the correspondences in names shown in Table III. before our eyes, it is difficult to suppose that the statements of Mr Mathews have no basis in fact. In the absence of further information, however, it is clearly impossible to discuss the origin of these divisions. It seems most probable that they are the systematisation of the anomalous marriages already cited. But much more information is needed before anything like certainty can be attained in the matter. Both actual genealogies and tables of terms of relationship must be in our hands before we can come to a decision.

INDEX OF PHRATRY, BLOOD, AND CLASS NAMES.