to its members, I felt it was right, a feeling shared by the Director of the Society and by the Editor of Folk-lore, that these should have the opportunity of seeing what I had to say in my defence.
I may be permitted to add a few general considerations upon the criticism of the Grail romances. No theory concerning the origin and signification of the legend can be acceptable which does not explain the relation to one another of the various romances, which does not account in a fairly intelligible manner for the development of the ideas and incidents contained in them. Nothing is easier than to pick out, as Dr. Gaster has done, this or that feature in this immense body of romance, to adduce parallels to it, and to fancy the problem solved; nothing harder than to fit all the features of all the versions into an orderly scheme of development.
At the same time no theory can, I think, be successful which makes any one existing version the fons et origo of the whole cycle. Even if we had not positive statements, which there is no reason to disbelieve, we should be compelled to assume an earlier written stage than any we now possess. Behind this written stage we discern an oral stage in which the incidents of the legend were singularly vague and formless, but in which they still hung together. I conjectured that they did this because they came to the French wandering minstrels or story tellers, to whom the first spread of the legend in France was due, mainly from one source and connected with one group of personages. The facts that the majority of these personages bear Celtic names, some perfectly recognisable, others greatly disfigured, and that the scene of their exploits is, in the main, lands dwelt in by Celtic-speaking populations, seemed to me to warrant the conclusion that the traditions underlying the romance came to the French from Celts (whether Bretons or Welshmen is indifferent), and were essentially Celtic, i.e., had passed through the mind of Celts (whether
