In spite of all these proofs, exogamy and the capture of women from strange tribes differ as regards _one_ feature of paramount importance. In both legend and history the captured woman is universally of a _strange_ tribe, whereas totemic exogamy never occurs except between clans of the same tribe. Added to this is a further consideration. The above-mentioned passage from Deuteronomy certainly presupposes that the Israelite who captures a wife in warfare with a strange tribe already possesses a wife from among his own tribe. This is his chief wife, in addition to whom he may take the strange woman as a secondary wife. We may refer to Hagar, the slave, and to Sarah, Abraham's rightful wife, who belonged to his own tribe. The resemblance between exogamy and the capture of women in warfare is so far from being conclusive that exogamy is permitted only between clans of the same tribal group; hence, in cases where there are four or eight subgroups, it is not even allowed between members of the two tribal halves. Indeed, the essential characteristic of exogamous tribal organization, marriage between _specific_ social groups, is entirely lacking in the marriage by capture that results from war. Moreover, the woman married under exogamous conditions is either the only wife or, if she is the first, she is the chief wife; in the case of marriage by capture in war, the captured woman is the secondary wife. 5. MODES OF CONTRACTING MARRIAGE. Though the theory that exogamy originated in the capture of women in warfare is clearly untenable, it has without doubt seized upon _one_ element of truth. Marriage by capture may also occur within one and the same tribe, and under relatively savage conditions this happens very frequently. Indeed, it is precisely in the case of the Australians, to judge from reports, that such marriage is probably as old as the institution of exogamy itself, if not older. Early accounts, in particular, give abundant testimony to this effect. That later writings give less prominence to the phenomenon does not imply its disappearance. The decreased emphasis is due rather to the fact that in more recent years the attention of investigators has been directed almost exclusively to the newly discovered conditions of tribal organization. Even on a more advanced and semi-cultural stage we find struggles for the possession of a wife. The struggle, however, is regularly carried on, not between members of different groups, much less between entirely strange peoples of widely differing language and culture, but between members of one and the same tribe. Two or more members of a tribe fall into a quarrel for the possession of a woman who, though not belonging to their own clan, is nevertheless a member of a neighbouring clan of the same tribe. Such conditions are doubtless to be traced back to earliest times. The victor wins the woman for himself. The custom of marriage by capture has left its traces even