invasion of Denmark on account of a tribute dating back to the time of Arthur. Gunter is killed, and the queen, who does not appear in the English romance, flees with Havelok. There is an attack by pirates in which the queen is killed. Instead of a king over all England and a usurping earl, as in the English version, there are two kings, one ruling over Lincoln and the other over Norfolk. Havelok is called Coraunt (Cuaran), a name which does not occur in the English romance. Havelok returns to Grimsby by his wife's advice. He does not know who he is, until told by Grim's relative. Edelsi submits after fighting and gives Argentille her heritage. In addition to the omissions noted above, the Interpolation says nothing about Grim's being a fisherman and salt merchant in Grimsby, about Cuaran's being a juggler or fool at the court, about the attack on Havelok and his bride by the six youths, nor about Havelok's fear when led into the hall before Sigar. None of these omissions need cause any surprise. It was almost inevitable that the more extraordinary incidents should be cut down by a matter-of-fact writer, such as this interpolator seems to have been, while the other omissions resulted naturally from the attempt to condense. When these allowances are made, it is evident that the general outline of the story is the same in Gaimar, the Lay and the Interpolation.

A more detailed examination reveals the following points. in which the Interpolation is closer to Gaimar than to the Lay: (1) The invasion of Denmark is for tribute which had been withheld (Lamb. 2-4, Gaimar 410-411).1 In the Lay it is to demand tribute (lines 27-30). (2) Grim in both appears as a mariner, whereas in the Lay he is a baron (Lamb. 14, Gaimar 423, Lay 57). (3) Edulf is defeated in a general battle and not as in the Lay in a single combat (Lamb. 70-72, Gaimar 739-742, Lay 940-970). In all three cases the agreement between Gaimar and the Interpola

1 References to the French versions of Havelok are to the edition of Gaimar in the Rolls Series, London, 1888.

tion seems to point to an earlier form of the story than that contained in the Lay. Grim certainly has no right to be a baron. That this is a modification made by the rather late writer of the Lay is almost self-evident, but is made certain by the slip in line 135 of the Lay where we are told that Grim, when he reached Grimsby, went fishing "as he was accustomed to do." The writer forgot that he had transformed the fisherman or sailor into a baron. The change is due to the fact that the Lay has throughout a more courtly and knightly tone, approaching the form of fiction in vogue during the thirteenth century. The Interpolation, on the other hand, is simpler, and in this respect resembles Gaimar, both these versions preserving what must have been the spirit of the original.

Additional evidence for the close relationship between Gaimar and the Interpolation is furnished by the agreement in geographical details. In both, for instance, Edelsi's kingdom extends from the Humber to Rutland, and Adelbrict's from Colchester to Holland. For the first of these pairs the Lay has Rutland and Stanford, while the second is replaced by "vers les Surois," Surrey being probably intended.*

So far nothing has been presented to disprove Madden's assertion that the Lambeth Interpolation was derived from Gaimar. In fact the evidence has all pointed that way. But there are differences between the two which must not be overlooked or ignored. Most prominent perhaps is the fact that the order in which the events are related is not the same. Gaimar's narrative opens in England. There is no direct relation of the early events in Denmark and at Grimsby, these being recapitulated very briefly by Kelloc and others. The allusions to the early part of the story are so scattered and incoherent that they give the impres

1"Pescher aloit si com il soloit."-Lay, 135. Lamb. 34, Gaimar 75.

Lamb. 30, Gaimar 51. 'Lay 198, 201.

5 Gaimar 359-454, 575–628. Lines 505-528 are related by the author,

but merely as an incidental explanation.

sion that they are echoes of a more complete original which Gaimar modified for the sake of condensation or, perhaps, to secure a sort of epic unity by plunging in medias res. The Interpolation, on the other hand, opens in Denmark and the early parts of the story are related in consecutive order. This order might be made up from the allusions in Gaimar, but that would require more skill and pains than could be expected in a scribe, even though he were clever enough to be an interpolator. The natural thing for a man of his capacity to do is to follow the order of events in his original. This alone would not prove that the Interpolation had a different original from Gaimar, but it raises a question which must be met. The matter is made the more noteworthy from the fact that the order of events is exactly the same in the Interpolation and in the Lay. For this to be accidental is possible but not very likely. Other matters being left out of consideration, it would be reasonable in such a case to suppose some sort of relation between the Interpolation and the Lay independent of Gaimar.

This relationship between the Interpolation and the Lay is made the more evident by certain details which the two have in common, but which are not found in Gaimar, such as the following: (1) Gunter's enemies plan shame for his relatives (Lamb. 9, Lay 79-82). (2) Edelsi, instead of being called merely "Breton," as in Gaimar (line 61), is said to be "of Breton kynde" (Lamb. 31) or "Bret par lignage" (Lay 200). (3) In Gaimar Edelsi forces Cuaran and Argentille to lie together without a formal marriage (lines 167-176), while in the other versions there is a marriage (Lamb. 47, Lay 377-380). (4) The Interpolation says that he brings about the marriage, though many are wroth, which seems to correspond with the account given in the Lay of the anger of the barons at the king's violation of his oath (Lamb. 48, Lay 279-376). (5) In Gaimar there is no description whatever of Havelok's departure from Grimsby for Lincoln and the only allusion to it is the statement of Havelok to Kelloc that

he departed from Grimsby when Grim was dead (line 371). In both the Interpolation and the Lay Grim is alive when Havelok departs, and dead when he returns with his bride (Lamb. 27, 56; Lay 157-192, 565). (6) Gaimar introduces the fight rather abruptly after Havelok's return to England (line 767). The Interpolation and the Lay mention the gathering of a host by Edelsi (Lamb. 75-76, Lay 10071026). It is difficult to imagine that all these resemblances are accidental. The first two and the last might be so, but the others seem to point to details in a source common to both the Interpolation and the Lay. This common source cannot be Gaimar, because in these points Gaimar differs. Moreover, in all three points Gaimar, rather than the other versions, seems to show a change from what must have been the original form of the story. It seems reasonable to suppose that there was a marriage, that Argentille's friends. should become angry at her disgrace and the seizing of the kingdom by a usurper, and that there should be some more definite statement about Havelok's departure from Grimsby. The number of important details common to the Interpolation and the Lay and the exact agreement in the order of the narrative establish a close relationship between the two and a common source independent of Gaimar. It becomes evident, therefore, that the traditional view, hitherto held without question, that the Interpolation is "copied" from Gaimar, must from now on be rejected.

In looking for the source of the Lambeth Interpolation it may be well to set aside at the outset any notion that it may be derived from a combination of two or more versions. Such a combination would of course explain anything except itself. An interpolating scribe, for the sake of inserting into a chronicle an episode of less than a hundred lines, is not likely to take the trouble to compare varying versions of a romance, perhaps in more than one language, and to make out of them a consistent whole. It was hard enough in those days for the most skilful writer of chronicles or romances to

make such a combination without revealing the artifice by a botch or confusion.' The Lambeth Interpolation tells a straightforward, consistent story, and any lack of clearness is due to nothing more than the extreme condensation. There is every reason to believe that it had a single source.

It has already been shown that this source of the Interpolation could not have been either Gaimar or the Lay. The source, however, must have been closely related to both Gaimar and the Lay, and the probabilities are all in favor of its having been in French. There is evidence for this in the fact that Havelok is called "quistron" instead of scullion. Though it is now lost there must have existed at some time a French version of the romance distinct from Gaimar and the Lay. That such a version did exist and was the common source of both Gaimar and the Lay has been effectively proved by Kupferschmidt. As the Lambeth Interpolation

An example of such confusion occurs in the abridgment of Havelok in Thomas Gray's Scala Cronica, the passage being reprinted in Madden, pp. xxxiv-xxxv. Gray failed to recognize that Havelok and Cuaran were the same person.

Kupferschmidt's investigation, already referred to, must be regarded as settling the fact that Gaimar and the Lay had a common source written in French octosyllabic rimed couplets. Ward appears not to have read Kupferschmidt. His attempt to derive the Lay directly from Gaimar cannot be accepted. Every one of his six arguments can be used with equal force in favor of a common source for Gaimar and the Lay. Ward, Catalogue of Romances, Vol. 1, 437-440. With the exception of Madden, who thought Gaimar had merely abridged the Lay, and of Ward, practically every investigator has concluded that the two extant French versions had a common source. The early writers assumed this to have been a "Breton lay;" but the later ones have realized that this source must have been a lost French version.

It would be very hard to defend the possibility that the lost French version was derived from Gaimar, and became in turn the source of both the Lay and the Interpolation. There are too many points in which the Lay and the Interpolation, one or both, point back to a form of the story earlier than Gaimar. Kupferschmidt has mentioned some of these and might have added the narration of the early events in Denmark and Grimsby, the marriage of Cuaran and Argentille, and the opposition thereto, and Havelok's finding Grim and his wife dead when he returns