originates with the ghost-theory, and the concomitant rise of a belief that some men, having acquired power over ghosts, can obtain their aid. Generally the chief and the medicine-man are separate persons; and there then exists between them some conflict: they have competing authorities. But, where the ruler unites with his power, naturally gained, this ascribed supernatural power, his authority is necessarily much increased. Recalcitrant members of his tribe, who might dare to resist him if bodily prowess alone could decide the struggle, do not dare to do this if they believe he can send one of his posse comitatus of ghosts to torment them. That rulers desire to unite the two characters we have, in one case, distinct proof. Canon Callaway tells us that, among the Amazulu, a chief will endeavor to discover a medicineman's secrets and afterward kill him.
Still there recurs the question. How does permanent political headship arise? Such political headship as results from bodily power, or courage, or sagacity, even when strengthened by supposed supernatural aid, ends with the life of any savage who gains it. The principle of efficiency, physical or mental, while it tends to produce a temporary differentiation into ruler and ruled, does not suffice to produce a permanent differentiation. There has to coöperate another principle, to which we now pass.
Already we have seen that even in the rudest groups age gives some predominance. Among both Fuegians and Australians, not only old men, but old women, exercise authority. And that this respect for age, apart from other distinction, is an important factor in establishing political subordination, is implied by the curious fact that, in sundry advanced societies characterized by extreme governmental coercion, the respect due to age takes precedence of all other respect. Sharpe remarks of ancient Egypt that "here as in Persia and Judea the king's mother often held rank above his wife." In China, notwithstanding the inferior position of women socially and domestically, there exists this supremacy of the female parent, second only to that of the male parent; and the same thing occurs in Japan. As supporting the inference that subjection to parents prepares the way for subjection to rulers, I may add a converse fact. Of the Coroados, whose groups are so incoherent, we read that "the pajé, however, has as little influence over the will of the multitude as any other, for they live without any bond of social union, neither under a republican nor a patriarchal form of government. Even family ties are very loose among them . . . there is no regular precedency between the old and the young, for age appears to enjoy no respect among them." And, as reënforcing this converse fact, I may add that, as I have shown elsewhere, the Mantras, the Caribs, the Mapuchés, the Brazilian Indians, the Gallinomeros, the Shoshones, the Navajos, the Californians, the Comanches, who submit very little or not at all to