[79] The brother feature appears likewise in Wolfram von Eschenbach, where Parzival's final and hardest struggle is against the unknown brother, as the Great Fool's is against the Gruagach. This may be added to other indications that Wolfram _did_ have some other version before him besides Chrestien's.
[80] I cannot but think that these words have connection with the incident in the English Sir Perceval of the hero's throwing into the flames and thus destroying his witch enemy.
[81] I must refer to my Mabinogion Studies, I. Branwen for a discussion of the relation of this tale with Branwen and with the Teutonic Heldensage.
[82] Another parallel is afforded by the tale of Conall Gulban (Campbell, III., 274). Conall, stretched wounded on the field, sees "when night grew dark a great Turkish carlin, and she had a white glaive of light with which she could see seven miles behind her and seven miles before her; and she had a flask of balsam carrying it." The dead men are brought to life by having three drops of balsam put into their mouths. The hero wins both flask and glaive.
[83] _Cf._ my Branwen for remarks on the mythological aspect of the ballad. It should be noted that most of the ballads traditionally current in the Highlands are of semi-literary origin, _i.e._, would seem to go back to the compositions of mediæval Irish bards, who often sprinkled over the native tradition a profusion of classical and historical names. I do not think the foreign influence went farther than the "names" of some personages, and such as it is is more at work in the ballads than in the tales.
[84] This may seem to conflict with the statement made above (p. 145), that the Mabinogi probably took over the maimed uncle from Chrestien. But there were in all probability several forms of the story; that hinted at in Chrestien and found in Manessier had its probable counterpart in Celtic tradition as well as that found in Gerbert. It is hardly possible to determine what was the form found in the proto-Mabinogi, the possibility of its having been exactly the same as that of Gerbert is in no way affected by the fact that the Mabinogi, as we now have it, has in this respect been influenced by Chrestien. Meanwhile Birch-Hirschfeld's hypothesis that Gerbert's section of the Conte du Graal is an interpolation between Gautier and Manessier is laid open to grave doubt. It is far more likely that Gerbert's work was an independent and original attempt to provide an ending for Chrestien's unfinished poem, and that he had before him a different version of the original from that used by Gautier and Manessier.
