their roots in the social constitution either of the people who narrate the myth or of those from whom the myth has been derived.
I propose, then, to accept without question the value, as historical evidence, of incidental references to social conditions and of the social setting of mythical narratives. It is not with these aspects of myth that it is the special purpose of this paper to deal, but with that kind of myth the primary motive of which is to account for social phenomena. I may point out at once a feature of such myths which differentiates them from myths having natural phenomena as their subject. When people relate a story which provides an anthropomorphic or theriomorphic interpretation of the moon's periodical changes, or narrate the events which produced the features characteristic of some animal, it is clear that we have to do with myth. When, on the other hand, they give an account of events which led to the appearance of a form of social structure or of the mode of performing a rite, we have no such immediate certainty that we have to do with myth. It is possible that the people may be preserving the memory of an actual occurrence; that we may have to do, not with myth, but with historical tradition. A possibility, then, which must always be kept in mind in dealing with narratives which appear to be social myths is that they may not be myths at all but historical traditions, or combinations of both history and myth.
The first fact which meets us in our inquiry is that narratives of a mythical kind which serve to account for social conditions occur but seldom among the records of savage or barbarous peoples. Accounts dealing with features of the sun, moon, and stars, of land and sea, of winds and weather, of animals and plants, are abundant, but it is only exceptionally that we meet with narratives which serve primarily and especially to account for social conditions. This predominance of natural phenomena as
