¹ This classification, undertaken primarily to settle the interpretation of the passages treated in Chapter i (Publ. M. L. A., xxvi, pp. 142 ff.), has been made more extensive in order to serve as the basis for a critical edition of the poem, which the author is now preparing.

¹ As above noted, the material here indicated has been later supplemented by an examination of all these mss., and in this examination the first three branches, the Swan children, the Swan knight and the Godfrey poems have all been included.

¹ Of the Swan-knight branch, there is not included here, nor generally in this investigation, the “Enfances” or swan children. This “Enfances” is not found at all in S, and we have not had copies of it from some of the other mss. The comparison is usually taken up, then, at the point where the Swan-knight appears in the boat drawn by the swan, to become the champion of his future wife.

² Corresponding to the 1909 lines of the Godfrey branch found in S, A has 2026 lines, or 117 more. Most of this difference is doubtless due to the many little additions in A. For examples of such additions, shown by the final classification of the mss. and in some cases by the context also, see ms. 12558 (fo. 47vo, col. 2, vv. 27-30, vv. 34-6, and v. 43; fo. 48ro, col. 1, vv. 2-4 and col. 2, vv. 6-8, v. 22 and v. 28).

³ Examples of original lines omitted by S are to be found in Hippeau, vol. 2: vv. 137-9 (S. fo. 83vo), v. 565 (S. fo. 88ro), vv. 3213-5 (S. fo. 111vo), v. 3370, vv. 3373-6 and v. 3387 (S. fo. 114vo). It should be said that in certain cases, where S does not have passages found in all the others, Krüger has claimed that S represents the original and that these passages have been added in the others. It can, however, be shown from the context alone that S has omitted some of these passages cited by Krüger, which were certainly in the original, and a classification of the mss. makes it clear that other similar passages were omitted by S likewise. It would be out of place to take up here in detail Krüger's notice of this ms.; and this has been done, moreover, in a separate article (see Romania, 1909).

¹ B (fo. 121ro, col. 2, v. 53) in the first of these passages has “Caance,” which is doubtless a mistake for “Maience”; for we find this latter word in B in the second passage. Sp (p. 79, col. 2) has Nimwegen in both of these passages, which would be a possible reading, were not the weight of manuscript authority for Maience.

¹ A (fo. 46ro, col. 2, v. 43); S (fo. 82vo, v. 5); E (fo. 61ro, col. 2. Omits, like S-A); G (fo. 70vo, col. 2. Omits).

² These lines are found in B (fo. 135ro, col. 1, vv. 50-1), C (fo. 64vo, col. 1, vv. 32-3), and the Spanish (p. 89, col. 2). They are:

¹ A (fo. 49ro, col. 1): S (fo. 90vo): G (fo. 73vo, col. 1): E (fo. 64ro, col. 1).

² B (fo. 138ro, col. 1): C (fo. 68vo, col. 1): Sp. (p. 94, col. 2).

¹ A (fo. 48ro, col. 2, v. 11): S (fo. 88ro, v. 18): E (fo. 63ro, col. 1, v. 35): G (fo. 72vo, col. 2, v. 20).

¹ It will be noted that this line was cited in the passage under consideration. It is of importance in deciding which son figured in the nursing scene, and is taken up elsewhere in connection with that question.

¹ It is, of course, more difficult to discover common errors in a translation and another ms. than in two French mss. The evidence, however, seems conclusive. If the ms. that served the Sp translator as a model had contained the mistakes found above, or others, some of them would certainly have been retained; moreover, in cases where they might have been corrected, the corrections often would have still differed from the original, as shown by the ms. agreement.

¹ It should be noted that the expression “D version” is frequently used in this paper in a sense different from “D ms.” By “D version” I frequently mean the first revision, of which the D ms. is the best example, but which really includes other mss. as well.

¹ An interesting proof, if any should be needed, that B has made this substitution and that Godfrey can not be the hero of this episode, is found in B (fo. 139ro, col. 1), corresponding to Hippeau, vol. ii, p. 35, v. 939. Here, in mentioning the sons left at home, the reviser of B forgot to change Godfrey into Eustace, leaving the verse exactly as it is in D; so that we have Godfrey at home and in England at the same time.

¹ The rubric at the beginning of the passage, however, says it tells of the knighting of Godfrey (B fo. 142ro col. 1). In B, Godfrey is knighted by the King of England (fo. 141vo, col. 1, v. 54).

² It is most likely in this way that various laisses mentioned above as belonging to the D revision, but not found in B, were omitted. Because they are comparatively slight additions of the D version, the author of B has failed to note them. He would certainly have found them if he were using a ms. of the D version as the chief basis for his text; and it would be very strange if the B ms., the longest and most burdened of all with additions, should have knowingly omitted anything.

¹ It might be remarked that this would be a natural assumption, even if it were not supported by abundant evidence. It is not likely that the revisers of this period, in basing a version on two mss., often took the trouble to make a line by line comparison in the text common to the two.

¹ I should say here that it is certain that the relations of the various mss. of the Cycle of the Crusade are often different in the different branches, even in those so closely related as the Swan-children, the Swan-knight and the Godfrey branches, of which the last two are usually treated as one poem. Most of the critics have not kept possibilities of this sort sufficiently in mind in treating these mss.

¹ It should not be inferred from the statement above that the importance of E and G in determining the original is by any means as great as the importance of A and S. A and S have not been revised, and stand much nearer to the original both in date and in form. What I mean is that there appears to be an SAEG version, from which the EG and SA versions are descended, and in determining this, EG, although less good than SA in general, is nevertheless not without weight and may give the correct reading in some cases where SA is wrong.

¹ A later examination of mss. Royal 15 E vi and 36615 of the British Museum, and of ms. 320 of Bern, of which the last two were briefly mentioned above as supporting the interpretations proposed, shows at least one more independent derivation from the original in the Godfrey poem. The British Museum mss. are, in this branch, like the unrevised versions Sp, S, and A in not containing any of the features of the later revisions; and the one at Bern has only one such addition, the continuation of the adventures of Cornumarant at the end. In the Swan-children and Swan-knight branches, however, these three mss. have some of the characteristic features of the later versions classified, such as B, E, and G.

In so far as they have been examined, these three mss. do not show any such relation to the S A version as was established for E and G. Their corroboration, then, of the other mss., with which also no close connection has been established in this branch, is of considerable weight.

² For a diagram of this classification, see Chapter i (Publ. M. L. A., xxvii, p. 163).