tradition, and be all the more puzzling for the portion of scientific truth which it really contains. The object of the present chapter is to collect a quantity of evidence, bearing on the problem how to separate Historical Traditions and Myths of Observation from pure Myths, and from one another.
Though it may not be possible to lay down any general canon of criticism by which the historical and mythical elements of tradition may be separated, it is to some extent possible to judge by internal evidence whether or not a particular legend or episode has a claim to be considered as history. It happens sometimes that a legend contains statements which are hardly likely to have come into the minds of the original narrators of the story, except by actual experience. The Chinese legend which tells us the name of the ancient sage who taught his people to make fire by the friction of wood cannot be taken as it stands for real history, seeing that so many nations ascribe this and other arts to mythic heroes, yet it embodies a recollection of a time when this was the ordinary way of producing fire. So, when the same people tell us that they once used knotted cords like the Peruvian quipus, as records of events, and that the art of writing superseded this ruder expedient, we are in no way called upon to receive the names and dates of the inventors to whom they ascribe these arts; but, at the same time, it is hard to imagine what could have put such an idea into their heads, unless there had been a foundation of fact for the story, in the actual use of quipus in the country before writing became general.
In the traditions which the Polynesians have preserved of their migrations in past times, it is likely that some historic truth may be preserved, and with their help, aided by a closer study of the languages and myths of the district, it may be some day possible for ethnologists to sketch out, at least roughly, the history of the race for ages before the European discovery. Much of the historical value of the South Sea traditions is due to their being commonly preserved in verses, kept alive by frequent repetition, and in which even small events are placed on record with an accuracy and permanence that yields only to written history. Thus a question that arose when Ellis was in
