the Orinoco, Humboldt says: "It is most considerable among the Caribs, and all the nations that have preserved the custom of carrying off young girls from the neighboring tribes." How, then, can wife-stealing be ascribed to scarcity of women?
A converse incongruity also militates against Mr. McLennan's theory. His position is that female infanticide, "rendering women scarce, led at once to polyandry within the tribe, and the capturing of women from without." But polyandry does not, so far as I see, distinguish wife-stealing tribes. We do not find it among the above-named Tasmanians, Australians, Dakotas, Brazilians; and although it is said to occur among the Fuegians, and characterizes some of the Caribs, it is much less marked than their polygyny. Contrariwise, though it is not a trait of peoples who rob one another of their women it is a trait of certain rude peoples who are habitually peaceful. There is polyandry among the Esquimaux, who do not even know what war is; there is polyandry among the Todas, who in no way aggress upon their neighbors
Other minor difficulties might be dwelt upon. There is the fact that in many cases exogamy and endogamy coexist, as among the Comanches, the New-Zealanders, the Lepchas, the Californians. There is the fact that in sundry cases polygyny and polyandry coexist, as among the Fuegians, the Caribs, the Esquimaux, the Warans, the Hottentots, the ancient Britons. There is the fact that there are some exogamous tribes who have not the form of capture in marriage, as the Iroquois and the Chippewas. But, not dwelling on these, I turn to certain cardinal difficulties, obvious a priori, which appear to me insuperable.
Setting out with primitive homogeneous groups, Mr. McLennan contends that the scarcity of women caused by destruction of female infants compelled wife-stealing; and he thinks that this happened "at a certain stage among every race of mankind" (p. 138). The implication is, therefore, that a number of adjacent tribes, usually belonging to the same variety of man in the same stage of progress, were simultaneously thus led to rob one another. But immediately we think of wife-stealing as a practice not of one tribe only, but of many tribes forming a cluster, there presents itself the question, How was the scarcity of women thus remedied? If each tribe had fewer women than men, how could the tribes get wived by robbing one another? The scarcity remained the same: what one tribe got another lost. Bearing in mind the low fertility and great infant mortality among savages, if there is a chronic deficiency of women and the tribes rob one another equally, the result must be diminished population in all the tribes. If some, robbing others in excess, get enough wives, and leave certain of the rest with very few, these must tend toward extinction. And if the surviving tribes carry on the process, there