Does his theory entirely exclude that of M. Gaston Paris? Does it bear all the conclusions drawn from it, implicitly rather than explicitly, it should be noted? The French verse and prose narratives of the twelfth century may go back exclusively to Breton lais—does this prove that the Arthur saga was originally historic in its essence, and that the later romantic developments are exclusively Breton. In the course of centuries the Breton forms may, indeed must have grown differently from the Welsh ones—does that prove that every specific Breton feature is, if not non-Celtic, at least foreign to the original form of the legend? Thus Prof. Zimmer regards the passing of Arthur to Avalon as specifically Breton, as foreign to the historical spirit of the original Arthur tales. Yet who more than Prof Zimmer in his studies on the Brendan legend has thrown clearer light upon that Celtic presentment of the Otherworld and of the hero's journey thither of which the whole Avalon episode is such an unmistakable variant? Again, Erec and Lancelot are held to be purely Breton. Granted for argument's sake that they do not appear in the Welsh record, does that prove that they cannot be elaborations of old Celtic heroes, that Erec must be derived from the sixth-century Visigoth chief of Aquitaine, Euric, or Lancelot from the ninth-century Carolingian warrior, Lantbert?
I should be sorry indeed if Prof. Zimmer had denied himself these latter hypotheses; in working them out he forces his readers into by-paths of history which the majority would otherwise never tread. But what single shred of positive evidence is brought forward in support of the equation Lancelot=Lantbert? Not one. In what respect does the equation explain the story we find in the twelfth-century French poets? In no single one. Accept every assertion of Prof Zimmer's, and we are as far as ever from realising the nature of the Lancelot episode. For that we must turn to the scholar of whom Prof. Zim-
