Reviews. 107
of the Australian practices are referable to group-marriage, of which he claims them as a survival. Holding strongly, as I do, that civilization has been on its institutional side a progressive regulation of human affairs, it seems to me that the presump- tion is that he is right. In any case, he has in this volume co-ordinated a mass of evidence, much of it, thanks chiefly to him, previously known. It is a very happy thing that he has been able to collect, put into final shape, and present in the light of his long and valuable experience of the blackfellow, all the interesting and important details to be found in these pages.
Nor is this observation to be confined to the subject of mar- riage. It extends to every part of aboriginal life expounded by the author. For instance, the tables of relationship and of the relationship-terms may perhaps be passed over as dry detail by all but very careful readers. Yet they contain some of the most instructive information in the book. There is, however, one matter here to which I must refer. The table of Dieri marriages and descents, facing p. 159, shews, among the de- scendants of the pair numbered respectively 2 and 6, unexpected changes of totem. The son of the Muluru (caterpillar) woman should, according to the rule of descent prevailing among the Dieri, be a Muluru ; but he is in fact given as a Warogati (emu). The son of the Tidnamara (frog) woman is given as a Kaualka (crow). The explanation seems to be found on p. 161, where we read : " In one or two cases a couple had no ' own ' son or ' own ' daughter, and a 'tribal' son or daughter has been interpolated, there being, from a Dieri point of view, no difference in the relation- ships A stronger case could hardly be found to illustrate the meaning of the relationship-terms. But the explanation is hardly complete. Great as are the pains taken by Dr. Howitt to eluci- date the meaning of the relationship-terms, I cannot find a definition of " tribal son " or " daughter." The table of Dieri relationship-terms gives one {Ngata-mura) which denotes, when a man speaks, child, when a woman speaks, brother's child, and another term {Ngatani) which only a woman uses to denote (her own) child. I suspect it also includes her (own and tribal) sister's child. But if so, what is the meaning of " tribal " in this connec- tion? One would naturally suppose it limited to children not
