studied the subject matter of the romances, not the words in which they have been handed down. Those who seek for philological disquisitions are, therefore, warned that they will find nothing to interest them; and those scholars who are well acquainted with the printed texts, but who are on the search for fresh MS. evidence, must not look here for such. On the other hand, as the printed texts are for the most of such rarity and price as to be practically inaccessible to anyone not within reach of a large library, the writer trusts that his abstract of them will be welcome to many. He has striven to take note of all works of real value bearing upon the subject. He endeavoured, though unsuccessfully, to obtain a copy of M. Gaston Paris' account of the Arthurian romances which, though it has been for some months in print, is not yet published.
The writer has done his best to separate the certain from the conjectural. Like M. Renan, in a similar case, he begs the reader to supply the "perhaps" and the "possibly's" that may sometimes have dropt out. The whole subject is fraught with difficulty, and there are special reasons why all results must for some time to come be looked upon as conjectural. These are glanced at here and there in the course of these studies, but it may be well to put them together in this place. Firstly, whatever opinions be held as to which are the older forms of the legend, it is certain that in no one case do we possess a primary form. All the versions that have come down to us presuppose, even where they do not actually testify to, a model. Two of the forms which there is substantial agreement in reckoning among the oldest, the poems of Chrestien de Troyes and Robert de Borron, were never finished by the authors; sequels exist to both, of a later date and obviously affected by other forms of the legend. A reconstruction of the original story is under these circumstances a task of great uncertainty. So much for the difficulty inherent in the nature of the evidence, a difficulty which it is to be feared will always beset the student of this literature, as no new texts are likely to be found. Secondly, this evidence, such as it is, is not accessible in a form of which the most can be made. The most important member of the group, the Conte du Graal, only exists in one text, and that from a late and
