the story in 11. 5-28, and ends by saying that the lay has now the name of

The reference to an earlier form of the story is here quite definite, and it also seems to follow that the name was changed,' because the real subject appeared to her the submissive wife rather than the faithless husband. In the outline which Marie then adds there is not the slightest hint of the two episodes of the shipwreck and the resuscitation. It would perhaps help the theory if it could be assumed that this silence is evidence that they represent her additions, but there is not sufficient basis to warrant such an inference.

Gautier refers to his source most directly in I 929-36. In the note to these lines Förster explains the passage as having reference to the vogue of lays in general, but Gaston Paris (Rom., XXI, p. 278) has given another interpretation of it, which seems to me undoubtedly correct. A paraphrase of the whole passage will bring out its meaning. The author comments on his story. Ille has fallen in love with Galeron and she with him. She is of noble station and he a simple knight; how could they ever expect to enjoy each other's love! But they do not think of such difficulties and take pleasure in each other's company. Such is the nature of love. It flatters people to attract them, and later it has no joys to offer. To be sure, they do not think of this and would like the present condition to continue; but if love did not have its sadder side, this lay would not be such a favorite, and knights would not prize it as they do. A fine story is that of Ille and Galeron. It contains no witchcraft nor lengthening, you'll not find anything supernatural in it. There are other lays, which make the one that hears them think that he has slept or dreamed.

2

It is evident from these lines that the lay at present does not have the name which Marie intended to give it.

2 Gautier returns to this same thought, 11. 1532-38. In the first passage it forms a natural introduction to the reference to his source which follows, and this fact makes it impossible to accept Foulet's suggestion (Zs. f. rom. Phil., XXIX, p. 303) that the lines in question represent a later addition, either by Gautier himself or by a jongleur who recited the poem.

Since the poem on Ille et Galeron which we know is no lay, it follows that Gautier here refers to his source and that this bore the same title. The allusion to the absence in it of witchcraft and the supernatural is important, for it contains a distinct reference to the peculiar atmosphere of the lays of Marie de France, the chief exponent and, if Foulet in the article just cited is correct, the inventor of this class of composition.

The further inference that the Eliduc lay itself is meant, and specifically the two episodes not duplicated in I, depends to a certain extent upon the question at issue in this paper. If I represents a reversal of E2 it would probably be exact. Yet, even if the source of I were entirely independent of E, it might still be true; for, granting the earlier date of E2, we should have proof of his acquaintance with this poem in the name Eliduc, which he applies to the father of his hero.

We have thus on either side the author's testimony of the existence of an older form of either poem. The question to solve is whether this earlier form is identical for both E' and I, and whether the points of contact enumerated by Förster contain evidence that I is a reversal of E'. Before going any farther it will be of service to state clearly Förster's position. He maintains that Gautier reworked the Eliduc story, and purposely eliminated all its immoral features. While Eliduc was ready to forget his wife and commit bigamy, Ille remains faithful to Galeron, until of her own will she sets him free. Hence the remorse of Eliduc and Guilliadun could disappear from Ille et Galeron. The fundamental difference in the appearance of the two protagonists-Eliduc as a knight with followers, Ille poor and unknown—is due to Gautier's principle to let his hero create his position through his valor and daring. The motive of the separation of Ille and Galeron, he thinks, is based upon a question debated at the court of Marie de Champagne, and called up presumably by incidents that must have been of frequent occurrence in the tournaments of the time, viz., whether a lady is justified in dismissing her lover when his appearance is changed as the result of injuries received in combat, in this instance the loss of an eye. And he refers to a passage in André le Chapelain's

well-known book' where a decision on this very question in the sense of Gautier's poem is given by the countess Irmengard of Narbonne. To this reconstructed story Gautier then made various additions. He increased the poem to the proper length of a roman d'aventure by adding an account of the youth of his hero and his first marriage. He fills in and lengthens out descriptions of battles, discourses on love in the manner of Chrestien de Troies, and inserts some clever scenes of his own invention, such as the life of Galeron in Rome and her appearance at the door of the church at the moment of Ille's wedding to Ganor, the latter's journey to Bretagne to implore the aid of Ille, and their failure to meet in Vienne, and the like.

Bearing these points in mind, we may now proceed to a critical examination of Förster's argument, which is based partly on alleged similarities between E and I, and partly on contrasts in the main motives of the two stories.

The first points of contact which he notes are contained in the battle scenes. Only one such incident occurs in E2. Here it is related that Eliduc arrives at the court of the king of Exeter with 10 of his own knights (1. 79). He is joined by 14 of the king's men (l. 155), 'and together the 25 (1. 221) set out to give battle to the enemy who is pressing him. The subsequent victory takes place in a destreit, pointed out to Eliduc by one of the king's men upon his question (ll. 166-84); 30 of the enemy are captured (1. 221); and when the king now sees this crowd of 55 knights approaching his castle, his first thought is that the enemy has been victorious (11. 235 ff.). There are five battles described in I, and the similarities are scattered through several of them. In the first of this list Ille has set out to conquer his heritage with 10 knights (1. 319) and two old companions (11. 194 and 329), so that the whole cavalcade numbers 13 men. These expect to be joined by 20 additional knights (1. 340) of Ille's faithful friends, but their plan miscarries. The 20 are attacked by 100 of Hoel's men (11. 400 ff.), and hard pressed, and when Ille and his 12 companions arrive a battle ensues, 13 against 60, and 20 against 40 (11. 494, 495). Of the 20 finally only 13 remain, and of the

1 De Amore, edited by E. Trojel (Havniae, 1892), p. 287.

40 only 16 (11. 512, 513). Ille and his men in the end overcome the other 60 (11. 517-792). He kills 18, and the others flee.

The similarities which Förster sees in these two scenes are rather dim. Eliduc's battle is directed against a rejected suitor for the hand of Guildeluëc, the daughter of the king of Exeter; while Ille fights for his own title, though later, after his marriage to Galeron, he defends her twice-a fact which Förster overlooks -against rejected suitors, first against Rogelion (ll. 954–1176), and later against the counts of Anjou and Poitou and the duke of Normandy (l. 1494–97 and 1538-80). This latter battle takes place in a destreit (1. 1552) as that of Eliduc; yet Förster fails to note the vital difference that in E2 the pass is a part of Eliduc's prearranged plan of attack, while in I it represents the natural road of the enemy in the third of these battles. He maintains further that the 14 knights in E2 (1. 155) play the same rôle as the 20 in I (1. 483) in the first battle; but he overlooks that the 14 aid Eliduc in defeating the enemy, while Ille frees the 20 from the danger in which they are caught. Furthermore, Eliduc's increase comes from the king of Exeter's men, while Ille's 20 knights are friends of his youth, trying to effect a union with him. Finally the one battle takes place for the conquest of Ille's inheritance, while the other is in aid of the father of Eliduc's future wife; and above all, the points compared are divided between two entirely different scenes. It is evident that the significant features of the three battles are quite dissimilar. The mistake of the king in E2, when he sees the larger number of knights returning to his castle, and which is so unique that it should have appealed to Gautier, is entirely lacking. Taking all these variations into account, I think it will be agreed that the few scattered points of contact, meager as they are, must be fortuitous, all the more when it is remembered that they could be duplicated from other poems.

Förster thinks, in the next place, that the circumstances attending the appearance of Eliduc and Ille at the courts of the fathers of their second wives are identical, barring the difference already referred to that the former arrives surrounded by followers, while Ille comes alone in shabby dress and is exposed to ridicule. In both poems the king is described as old and feeble (E2 90, Vielz

huem et anciens esteit= I 2004, Que d'une part l'aqeut viellece), and both have refused a neighbor the hand of their daughters (E' 95-98-I 5400). Here Förster distorts the facts to support his thesis. In E the king is attacked by a rejected suitor, but in I the sole reason for the Greek emperor's aggression is the age (1. 2004) and feebleness (1. 2007) of the emperor of Rome. The emperor of Constantinople is already married to Ganor's cousin, and the question of his suit for the hand of Ganor does not arise until much later, when her marriage to Ille is not thought of. Galeron has reappeared and both have returned to Bretagne. Ganor's father has died (1. 5400) and her cousin has succumbed to the effects of her husband's cruelty. war of the poem, and, like the one preceding it, is in the main a war of conquest in which the idea of a marriage is of secondary consideration.

This is the final

Förster sees further evidences of the indebtedness of Gautier to the Eliduc lay in certain features of Ille's battles against the Greeks after his arrival in Rome, ll. 2201 ff. The comparison is of course again with the single battle in E. The Romans retire to a castle (I 2255), where they are besieged, while Eliduc prepares an ambush (E' 173) for the enemy. Förster notes particularly that in both cases the action is the result of a conseil. He overlooks that in E' the counsel is sought by Eliduc, while in I it is offered with diffidence by Ille to the seneschal (I 2237, and particularly ll. 2274 ff.). The siege which the Romans undergo in this castle Förster compares, if I understand him correctly, with the siege which the king of Exeter suffers at the hands of the rejected suitors, when Eliduc first appears at his court (E2 99)! He then notes that as the result of the victory Eliduc becomes gardein de la tere (E2 270) and Ille senescal (I 2476); but he overlooks that in E2 the appointment is made by the king, while Ille is elected to the position by the knights on the battlefield after the seneschal's death, when they are in need of a new leader (I 2470). The emperor merely confirms the choice (I 3165 and 3237-67).

Förster lays stress upon the fact that in both poems it is the princess who falls in love first with the newly arrived knight, and he points out certain similarities that exist in the description of