I can command will carefully analyse and compare all known versions of this folk-tale group, and will essay to determine whether the facts compel us to assume radiation from a particular centre within historic times, or allow us to regard the tale as common property of the various Aryan races.
In any case we have here a most beautiful illustration of the theory I have always urged, viz., that the folk-tale as collected in modern Europe is substantially older than the romances which were written down in the Middle Ages; that so far from being abridged and debased derivations from the romances, they are, in the main, derived from the tales upon which those romances were based; that there existed among the various Aryan-speaking races, as far back as we can trace, a stock of mythic narratives which have lived on to the present day.
No one at all familiar, I will not say with the methods of folk-lore research—these methods are and must be those of historical criticism generally—but with the facts disclosed by that research, but readily admits that, whilst we must always take the earliest version as the starting-point of investigation we must steel ourselves against the presumption that this earliest version is necessarily the starting-point of the series of phenomena we are investigating. It may be so, but more frequently it is not. This view is, however, apparently unintelligible to those distinguished students of history or literary history who sometimes do folk-lorists the honour of noticing them, and is held by them to be the result of the uncritical spirit which pervades all folk-lore study. The boot is really on the other leg. It is the non-folk-lorist who is uncritical in applying critical canons, perfectly sound it may be in his own line of study, to another with which he is not familiar, and to which they are not legitimately applicable. It is not often, however, that a principle so important to folk-lore research as that of the capacity of contemporary tradition to preserve facts which are, using the word in a strict sense, pre-
