In order to strengthen the case for the exogamous theory it is necessary to show that the same results could not be accomplished, or not so well, by other conditions of equally wide distribution. As a matter of fact, an alternative interpretation has recently been advanced.[44] In the case of the non-exogamous Californian Yahi Dr. Sapir connects the merging of lineal and collateral lines with the marriage regulations obtaining there and suggests that these rules “may no doubt not infrequently be examined as an equally or more plausible determining influence”. The practices referred to comprise the levirate, _i.e._, a man’s marriage with his brother’s widow, and marriage with the deceased wife’s sister. (Why, _deceased_? we may well ask Dr. Sapir, since a man’s preëmptive right to his wife’s younger sisters is a widespread custom in North America.) I do not doubt for a moment that the customs in question have affected kinship nomenclature, but I seriously question whether they constitute an adequate substitute for exogamy as an interpretation of the empirical distribution of the Dakota principle. The levirate, it is true, is an exceedingly widespread institution: Tylor found it among one hundred and twenty out of some three hundred peoples.[45] But the levirate alone will not do since it only explains the extension of the father term to the father’s brother and the correlative extension of the term ‘son’ to the brother’s son (man speaking). It remains to be seen, therefore, to what extent the levirate is united in different regions of the globe with the usage of marrying two or more sisters, which would further explain the classification of mother’s sister with mother and of the sister’s children with the children (woman speaking). So far as I know, the range of the two usages jointly has not been ascertained; pending its determination, the distribution of the Dakota principle is not accounted for, as it approximately is by exogamy. There are certain other objections to the levirate hypothesis. One of them was already urged by Morgan, who examined it under the heading of polygamy and polyandry, which together might obviously lead to the same results as the Yahi usages.[46] These customs do not necessarily take in the entire population. A man may not have a brother to inherit his widow, nor have all women sisters to join or follow them in wedlock. On the other hand, clan or gentile affiliation is an automatic affair not touched by such contingencies.