social arrangements. It is hard to see how else the dififerences between the institutions of the various tribes are to be accounted for; and the onus of proof lies upon those who challenge the theory.
But it must be admitted that, when we come to apply it, the theory is not without its difficulties. Dr. Howitt suggests (p. 89) that a social change of the kind indicated might be brought about by a dream in which a man of great repute in the tribe — a medicine-man for instance — dreamed he was visited by " some supernatural being, such as Kutchi of the Dieri, Bunjil of the Wurunjerri, or Daramulun of the Coast Murring," from whom he received a command which he would communicate to his fellow-medicine-men ; by these it would be first discussed and, if accepted, afterwards announced to the people. The suggestion may perhaps account for later changes among some of the more advanced tribes of the south-east ; but it is obvious that, put thus, it will not account for a reform among tribes like the Arunta which do not recognize a Bunjil or a Daramulun. In fact, as Mr. Lang has pointed out in Man for January last, the difficulty is to imagine the first step. What was the tribe, or the horde, or whatever we please to call the group, before the first attempt at organization took place.? Of course this is a difficulty that confronts us everywhere as we seek for human beginnings : it is not peculiar to Australia. What is peculiar to Australia (though not without some analogies elsewhere) is the fundamental division of a tribe into two intermarrying moieties. There are few tribes, most if not all of them coast-tribes, in which we do not find at least traces of this dual system. There is, therefore, reason to think that all the tribes were originally so divided. If the theory of a reformatory movement be sound, it must apply to this division. The competing theory, favoured by Mr, Lang is, if I understand rightly, that of a union between two previously exogamous groups, based on mutual rights of connubium. It is plausible ; but it would seem to involve no less difficulty than the other. One cannot help thinking it odd that everywhere the tribe should be divided into two intermarry- ing moieties and no more. If two groups might have united, why not three, why not five, on the footing of mutual conmibium ?