more in our own case. If the witness be unlearned, we have, at least, the probability that he is not transplanting to Otaheite or to Queensland ideas and customs which he has read about in Herodotus or Strabo, or theories of Müller or McLennan.[1] Lastly, if all evidence from all quarters and all ages, evidence learned and unlearned, ancient, mediæval, and modern agrees in certain points, and if many of the witnesses express surprise at the occurrence of customs and notions, which our reading shows to be almost universal, then let the undesigned coincidence itself stand for confirmation. To our mind this kind of treatment of evidence is not unscientific. It is permitted to investigators, like Darwin and Romanes. Mr. Max Müller, however, is so far from being satisfied with the method (as we have stated it) that he draws a line between what will content the scholar, and what the ethnologist will put up with. Mr. Müller's criticism deserves quotation in full (Nineteenth, Century, Jan. 1882): "Comparative mythology is chiefly studied by two classes—by scholars and by anthropologists. Now the true scholar who knows the intricacies of a few languages, who is aware of the traps he has to avoid in exploring their history, who in fact has burnt his fingers again and again when dealing with Greek, and Latin, and Sanskrit, shrinks by a kind of instinct from materials which crumble away as soon as critical scholarship attempts to impart to them a certain cohesion and polish. These materials are often supplied by travellers ignorant of the language, by missionaries strongly biassed in one direction or the other, or by natives who hardly under-
- ↑ "Illiterate men, ignorant of the writings of each other, bring the same reports from various quarters of the globe." So the author of the Origin of Rank (Prof. Millar, of Glasgow) wrote in the last century. This argument from undesigned coincidence, or recurrence, must be faced by people who deny the adequateness of anthropological evidence.
