obviously a means of achieving the quest, has, in the later one, become the end of that quest.
It seems impossible to resist the following conclusions:—The many forms of the incident found in the Grail romances are not variants of one, and that an orderly and logical original; they testify to the fact that in the body of popular tradition which forms the basis of these romances the incident of the visit to a magic castle was a common one, that it entered into the thread of stories, somewhat similar in outline and frequently centered in the same hero, but differing essentially in conception, and that the forms in the romances which are most likely to keep close to the traditional model are those secondary ones with which the innovating spirit, whether due to the genius of the individual artist, or to intruding Christian symbolism, has least concerned itself. There is apparently but one case in the Conte du Graal, that of Perceval's visit to the Castle of Maidens, which has been modified by neither of these influences.
To accept these conclusions is to clear the ground. If we rid our minds of the idea that there is a Grail legend, a definite fixed sequence of incidents, we need not be discouraged if we fail to find a prototype for it in Celtic tradition or elsewhere. We shall be prepared to examine every incident of which the Grail is a feature upon its own merits, and satisfied if we can find analogies to this or that one. And by so doing we are more likely to discover the how and why of the development of the legends as we find them in the romances.
Leaving subsidiary details out of account, we may bring all the instances in which the Grail appears under two formulas: that of the kinsman avenging a blood feud by the means of the three magic talismans, sword and lance and vessel; and that of the visit to the Bespelled Castle, the inmates of which enjoy, thanks to the magic vessel, a supernaturally prolonged life, from which they are released by the hero's question concerning that vessel. The one we may call the feud quest, the other the unspelling quest. The Proto-Mabinogi belonged, as we have already seen (supra, p. 139), to the first class, and accordingly we find that all relating to the question is obviously interpolated from Chrestien. Chrestien's model
