him, and his work deserves an instant's consideration, excellent example as it is of the way in which the Breton lais (themselves, I believe, adaptations of current folk-tales) were turned into long romances. The adventures of Eliduc at the court of the King of Exeter are, to some extent, used twice over: firstly, in the account of how Ille wins the love of his first wife, Galeron, sister to the Duke of Brittany, and then—when Ille, having lost his eye in a tournament, and fearing his wife will love him no more, flees from her—in the account of the help he gives the Emperor of Rome, and of the love he excites in the breast of the Emperor's daughter Ganor. But Galeron, instead of staying quietly at home, as does Guildeluec, seeks her truant husband, and finds him just as he, thinking her to be dead, is about to wed Ganor, out of pity for her great love. Galeron then offers, as does Guildeluec, to retire to a convent, but Ille will have none of the proposal, returns with her to Brittany, and there they live happily for many years. But Galeron, being in sore peril in childbirth, vows herself to the service of God if she wins through. This happens, and Ille, thus released, sets forth in search of Ganor, delivers her from great peril, and finally weds her.
The above brief abstract suffices to show the softening of the original polygamous feature begun in Eliduc fully carried out by Marie's contemporary. In the process, the "villain" has completely disappeared—as was, indeed, to be expected. The fact is, however, instructive to note for any who may hold that the lai of Eliduc is the source of the folk-tales. When we do find a derivative of the Breton lai, the development is in the very opposite direction.
The ersions hitherto cited of the "Husband and Two Wives" can thus throw no light upon the origin of Eliduc; on the contrary, they must be looked upon as mere literary offshoots from the lai stock. There is, however, one version which has never yet been mentioned in this connection to my knowledge, which cannot be directly connected with Eliduc, but which may be, and I believe is, an inde-
