From this jumble, as we find it in Crestien and his continuators, we can pick out a story akin to, but not identical with that in the Welsh work. But if Crestien had the hypothetical original of Peredur before him, how comes his own narrative to be so confused and unintelligible; how is it in especial that his continuators go off on half-a-dozen different tracks? Did they know nothing of this original? If not, how comes it that portions of it are to be recovered from them alone, there being nothing in Crestien's portion of the Conte del Graal that could give rise to them? Moreover, the Welsh tale contains incidents (to one of which the only known parallel is in the eighth-ninth century Irish Voyage of Mael Duin) which are absolutely unknown to any existing French romance. Would this be the case if it represented the original of such a famous work as the Conte del Graal? I sincerely welcome Dr. Hagen as a fellow-worker in this obscure field of literary history, but I cannot admit that he has convinced me as yet, and I still hold to my explanation of the Peredur problem, namely, that the Welsh tale is in the main the oldest extant form of the Perceval story, but that, as it has come down to us, it is comparatively late (say 1230- so), and has been influenced by the writings of the leading European poet of the twelfth century.
Now how does Prof Rhys stand with regard to these questions? He analyses the stories of Owain and Peredur minutely (chapters iv, v), and resolves them into variant versions of a nature-myth, an Irish analogue to which he finds in the dealings of Cuchulainn with the Morrigu. But to do this he is obliged to have recourse to considerable modification of the stories in their present form, and he justifies such modification on the ground that the Welsh versions have been influenced by the French ones. In so far he countenances those who contend for the secondary nature of these two Welsh tales. But it will be admitted, I think, that it is perilous in the extreme to postulate modification save when it is vouched for by positive and
