female kinship prevailed, "the sovereign nominated his successor—naturally choosing a son." Further, it is to be noted that, where, as in these cases, when no nomination has been made, the nobles choose among members of the royal family, and are determined in their choice by eligibility, there may be, and naturally is, a departure from descent in the female line; and this once broken through is likely, for several reasons, to be abolished. We are also introduced to another transitional process. For some of these cases are among the many in which succession to rulership is fixed in respect of the family, but not fixed in respect of the member of the family—a stage implying a partial but incomplete stability of the political headship. Several instances occur in Africa. "The crown of Abyssinia is hereditary in one family, but elective in the person," says Bruce. "Among the Timmanees and Bulloms, the crown remains in the same family, but the chief or headmen of the country, upon whom the election of a king depends, are at liberty to nominate a very distant branch of that family." And a Caffre "law requires the successor to the king should be chosen from among some of the youngest princes." In Java and Samoa, too, while succession to rulership is limited to the family, it is but partially settled with respect to the individual.
That stability of political headship is secured by establishment of descent in the male line is, of course, not alleged. The assertion simply is, that succession after this mode conduces better than any other to its stability. Of probable reasons for this, one is that in the patriarchal group, as developed among those pastoral races from which the leading civilized peoples have descended, the sentiment of subordination to the eldest male, fostered by circumstances in the family and in the gens, becomes instrumental to a wider subordination in the larger groups eventually formed. Another probable reason is, that with descent in the male line there is more frequently a union of efficiency with supremacy. The son of a great warrior, or man otherwise capable as a ruler, is more likely to possess kindred traits than is the son of his sister; and, if so, it will happen that in those earliest stages, when personal superiority is requisite as well as legitimacy of claim, succession in the male line will conduce to maintenance of power by making usurpation more difficult.
There is, however, a more potent influence which aids in giving permanence to political headship, and which operates more in conjunction with descent through males than in conjunction with descent through females—an influence probably of greater importance than any other.
When showing how respect for age generates patriarchal authority where descent through males has arisen, I gave cases which incidentally showed a further result; namely, that the dead patriarch, worshiped by his descendants, becomes a family deity. In sundry chap-
