persons who have not won the confidence of the people, nor discovered their real beliefs by long and patient observation. This must be borne in mind when missionaries tell us that savages believe in one supreme deity, in a mediator, and the like. A very pleasing example of inconsistency in reports about the same race may be found in a comparison of the account of the Khonds in the thirteenth volume of the Royal Asiatic Society with the account given by General Campbell in his Personal Narrative. The inquirer in the former case did not know the Khond language, and trusted to interpreters, who were later expelled from the public service. General Campbell, on the other hand, believed himself to possess "the confidence of the priests and chiefs," and his description is quite different. In cases of contradictions like these, the anthropologist will do well to leave the subject alone, unless he has very strong reasons for believing one or other of the contending witnesses.
We have now considered the objections that may be urged against the bias of witnesses.
Mr. Max Müller founds another objection on "the absence of recognised authorities among savages."[1] This absence of authority is not always complete; the Maoris, for example, have traditional hymns of great authority and antiquity. There are often sacred songs and customs (preserved by the Red Indians in chants recorded by picture-writing on birch bark), and there always is some teaching from the mothers to their children. All these, but, above all, the almost immutable sacredness of custom, are sources of evidence. But, of course, the story of one savage informant may differ widely from that of his neighbour. The first may be the black sheep of the tribe, the next may be the saint of the district. "Both would be considered by European travellers as unimpeachable authorities with regard to their religion." This is too strongly stated. Even the inquiring squatter will repose more confidence in the reports about his religion of a black with a
- ↑ Hibbert Lectures, p. 92.
