returns to his own dominions, after surrendering Guinevere to her husband, so the main framework of his subsequent narrative was supplied the author by Wace not the Wace of our manuscripts, but the expanded version of that poet's Brut, which, though lost in its original French form, has been preserved in Layamon's English paraphrase.208 With even greater boldness than in the first part of his work, however, the writer has here reshaped the materials of his source, so as to fit them to his new design of connecting the adulterous relations of Lancelot and the queen with the destruction of Arthur and the Round Table. Thus, in his romance Lancelot usurps the rôle of Mordred as Guinevere's lover.209 Furthermore, he fills the place of Frollo,210 Arthur's enemy, before his death, and of Constantine, the same monarch's avenger, after his death.211 This double rôle suits the conception of Lancelot which runs all through the Mort Artu. Having wronged Arthur through his guilty passion for Guinevere, he becomes, against his will, the king's enemy. On the other hand, he has never faltered in his personal loyalty, and so, being the greatest of Arthur's knights, he is the proper avenger of his sovereign.

206 On this subject cp. pp. 30 ff., above.

109 This and the following four sentences are taken, with some slight changes of wording, from pp. 452f. of the present writer's article, "The Development of the Mort Arthur Theme in Mediaeval Romance", RR, IV, (1913). In the section of that article, pp. 451ff., entitled "The Verse-Chronicle Source of the Vulgate Mort Artu", there is a full discussion of the relation of our romance to this lost source. 210 Cp. Geoffrey's Historia IX, 11, Wace, II. 10158 ff., Layamon, 11. 23393 ff. All three represent Frollo (Frolle, etc.) as a Roman, by birth, who ruled France as a vassal of the Roman emperor and was slain by Arthur in the latter's invasion of France. In the Lancelot, V, 374, and Mort Artu, VI, 346, however, he is a German prince whom Arthur kills in France. He is an ally, moreover, of the Romans

hence, according to the Mort Artu, the Roman emperor began his war against Arthur. Otherwise, Lancelot, in the present romance, has annexed the part which Frollo plays in the Geoffreyan tradition.

211 Constantine was the son of Cador, Duke (or Earl) of Cornwall. On embarking for Avalon, Arthur named him as his successor and he defeated and slew Mordred's sons. Cp. Geoffrey, XI, 2-4, Wace, 11. 13601 ff., Layamon, 11. 28590 ff., 28652 ff.

With this substitution of Lancelot for the characters just named, the old narrative of Geoffrey and the metrical chronicles gains immeasurably in dramatic force, and every incident in it is given new life. Arthur's continental war 212 is no longer inspired by a mere idle lust of conquest. Though ready to condone the wrongs which he himself has suffered, he is driven, contrary to his better judgment, into this disastrous conflict with the best beloved of his followers by Gawain, who, in his eagerness to avenge his own brothers' death, pursues unrelentingly his lifelong friend and companion-in-arms. Again, under the new circumstances, Gawain's death is no longer simply one of the ordinary chances of war, but the penalty which he pays for his inexorable spirit of vengeance.213 Similarly, in the new form of the story, Arthur's avenger is not merely a relative (Constantine), whose chief aim is to secure his own succession to the throne, but the friend whose affection for his dead lord remains undiminished, in spite of the hostilities with which the latter had, in his last days, continued to visit him. Even Guinevere profits by these changes; for the object of the passion which has caused her to violate her marriage vows is not the false and treacherous Mordred, as in the Geoffreyan tradition, but a man who was the lodestar of chivalry in his age. Thus, in the climax of the action, she appears as a fugitive from the traitor's embraces, not as his partner in crime.

It should be observed, moreover, that this climax, as, indeed, the whole story of Arthur's ruin, is invested with a greater tragic intensity through the author's original conception (not inherited

212

Geoffrey, IX, 11, X, 2, Wace, 11. 10146 ff., 11452ff., Layamon, 11. 23397 ff., 25 465 ff., distinguish two expeditions of Arthur to the continent the first against Frollo, the ally of Rome, the second against the Romans, who are led by their emperor. In the Lancelot, V, 336, 370, the order of the expeditions is reversed and Arthur does not go with the first. In the Mort Artu we have the war with the Romans repeated, but Frollo is mentioned only once, VI, 346, and then as dead.

913

The wound which he received from Lancelot was ultimately the cause of his death. This was the invention of the author of the Mort Artu, like the whole story of the conflict between the two characters.

from his predecessors) of the traitor who was the cause of that ruin as being Arthur's own son the child of an unwitting act

of incest which he had committed with his sister. Thus, as in the legend of Oedipus, the sin of the father, though unconsciously perpetrated, brings with it the blind and terrible retribution of the Fates.214

We have already discussed above 215 our author's sources for the concluding episodes of Arthur's career and his treatment of the same also, his awkward attempt at combining the ancient tradition as to the fate of the king, after he received his fatal wound, with the late fabrication of the monks of Glastonbury to

214

G. Paris, Huth-Merlin, I, p. XLI (including note 3), thougth that the idea of Mordred's incestuous birth was imitated from the story (Plutarch's Brutus, ch. 5) that Caesar was Brutus's father. He suggests, also, the possible influence of the mediaeval legend of Pope Gregory. On the other hand, Lot, Lancelot, p. 444, believes that the story of Roland's being the child of Charlemagne and his sister (Cp., for example, G. Paris's edition of the Vie de Saint Gilles, pp. LXIV ff., LXXVf. in the publications of the Société des Anciens Textes Français) is the source. The conception is, however, probably derived from the legend of Pope Gregory, which was much the most widely diffused of all the numerous tales of incest in the Middle Ages. Moreover, other Arthurian romances show contamination with this legend, which is not the case with the Brutus or Roland stories. Such romances are the Latin romances, De Ortu Waluuanii (thirteenth century) and its cognates. Cp. my edition of this and the Meriadoc romance, pp. XXXV ff. Still another is the Chevalier à la Manche. The Middle English romance, The Awntyrs of Arthur at the Terne Wathelyne, also, uses a legend concerning Gregory, but not the one which told of his incestuous birth.

G. Paris, loc. cit. states that this conception, as applied to Mordred, first appeared in the Lancelot, V, 284f. That passage, however, was certainly written with reference to the Mort Artu, (VI, 325, 349, 377) and I believe was interpolated by the author of the latter. Cp. the present writer, RR, IX, 382 ff., on this question. The detailed account of Mordred's conception in the Vulgate Merlin II, 128 ff., was, of course, inspired by the pass. es just cited from the Lancelot and the Mort Artu.

L

J. Rhys's contention, Arthurian Legend, pp. 21 ff. that the story is of Celtic origin is unsupported by any evidence.

215

Cp. pp. 439f., above.

the effect that the monarch was entombed in their abbey.216 There was an epic grandeur about the great tradition which the writer was here handling, but the tradition was not Christian in origin. On the other hand, to a mediaeval author, however secularly inclined, it would have seemed unfitting to end the story of such high and serious matters on a non-religious note, and so, apart from the touches of Christianization which he has imparted to the last scenes of Arthur's life, he makes the passion-tost Lancelot and his surviving kinsmen, like so many other knights of the Middle Ages, both real and fictitious, seek a final calm in religious devotion and a hermitage.217

216

217

Cp. p. 431, note 179 above.

As I have already observed, the author was here influenced especially by the Queste, VI, 198, where Perceval, after the conclusion of the Grail quest, retired to a hermitage, accompanied by Bohort. Among the heroes of mediaeval romance who have ended their lives in a similar manner, Guillaume d'Orange, Robert le Diable and Guy of Warwick are conspicuous. On the whole subject cp. T. Walker, Die altfranzösischen Dichtungen vom Helden im Kloster. Tübingen, 1910.

The main sources of the Mort Artu, as we have seen, are the Tristan poems, the Lancelot, and Layamon's lost French original. I have, also, noted Robert's Merlin, the Queste and the two chansons de geste (one or the other, or possibly both), Gaydon and Parise la Duchesse, as minor sources of the branch. I do not think that it is necessary to assume with Lot, Lancelot, pp. 194, 201, note 2, that our author drew directly from Geoffrey's Historia or Vita Merlini. The features of these works in question, I doubt not, reached our author through his lost verse-chronicle source.

As far as Chrétien (whose works must have been known to him) is concerned, I agree with Lot, p. 199, that the incelot 3d Perceval of that poet have had no perceptible influence on the Mort Art His Erec, 11. 6865 ff., however, suggested, I believe, the crowning of Bohort and Lionel, VI, 377. Possibly, as Lot suggests, p. 201, note 2, it may have influenced, too, his conception of Avalon, although, personally, I question this. The same scholar has also suggested, pp. 199 ff., the influence of Pseudo-Wauchier and Wauchier in certain matters correctly, I believe, save in respect to the episode of the Maid of Ascalot's dead body and the boat, which, according to Lot, he derived from Pseudo-Wauchier, 11. 20857 ff. The immediate original of the Mort Artu, here, however, seems plainly the Queste, as I have observed above.

Chapter IV.

Variant Versions of Parts of the Vulgate Cycle.'

We have so far discussed the successive members of the Vulgate cycle, as they have descended to us in the numerous MSS. of what we may call, roughly speaking, the textus receptus of the five branches. Before we leave the subject, however, it should be stated that, after the completion of the Vulgate cycle, various attempts were made to substitute new narratives for different parts of the cycle. Thus, leaving aside variants of more limited extent, we have:

1. The so-called Livre d'Artus of the unique thirteenth century MS. 337 (fonds français, Bibliothèque Nationale)3 a long Merlin-continuation which was intended to supplant the usual Vulgate version. The marked variation from the Vulgate commences just after the episode in which Guinevere excites Morgan's enmity

1

As will be seen later on, in the opinion of the present writer, the whole pseudo-Robert de Boron cycle is derived directly from the Vulgate, but, inasmuch as that is a separate cycle, which will be discussed below, I do not include its various branches in the present list of variants.

2

Paris, TR, II, 392ff., was the first to use the term, vre d'Artus, as an equivalent of "Merlin-continuation." Latterly, however, there has been a tendency to apply it only to the particular Merlin-continuation which is preserved in MS. 337. When Sommer in his edition of Malory III, 176, says that the last three branches of the Vulgate cycle combined are sometimes called the "Livre d'Artus," He has misled W. H. Schofield, English Literature from the Norman Conquest to Chaucer, p. 236. (London and New York, 1906).

this is an error.

3 The MS. is defective at the beginning and at the end. It belongs to the last quarter of the thirteenth century.