page 1079 note 1 Janet Spent, Spenser's Faerie Queene (London, 1934), pp. 13–37; J. H. Walter,“ ‘The Faerie Queene’: Alterations and Structure,” MLR, xxxvi (1941), 37–58, and “Farther Notes on the Alterations to the ‘Faerie Queene’,” MLR, xxxviii (1943), 1–10; Josephine Waters Bennett, The Evolution of “The Faerie Queene” (Chicago, 1942), passim.

page 1079 note 2 It should be dear that, In order to give substance to the view that an earner version is detectable with the aid of this evidence, the earlier version must be defined with more or lest precision: the statement that the poem, or this or that passage in the poem, has been revised, without the support of the further statement that the present poem or passage developed from such and such a version of poem or passage, has no meaning that I can discover. Yet such unsupported statements have been made: see, e.g., Mrs. Bennett's discussion of difficulties connected with the portrayal of Ate, op. cit., p. 165.

page 1079 note 3 According to Mrs. Bennett, the poem contains “a history of struggle, of shifting aims and plans”; it is to be regarded as “a growing and developing conformation of ideas”; and her book is “an attempt to discover how [the poem] might have grown ... to Its present state” (pp. 1, 3). The titles of Mrs. Bennett's book and Mr. Walter's articles imply such a view.

page 1080 note 4 “A Spenser Note,” MLR, xliii (1948), 239–241; “Spenser's Letter to Ralegh,” MLR, xlv (1950), 511–512; “‘In These xii Books Severally Handled and Discoursed’,” ELH, xix (1952), 165–172; cf. also Janet Spens, “‘The Faerie Queene’: A Reply,” MLR, xliv (1949), 87–88.

page 1080 note 5 I do not discuss the October Eclogue, since its significance is not in doubt.

page 1080 note 6 Smith's edition (Oxford, 1909), i, xi–xii (the quotation “snows that by 1588 the F.Q. had not only bam composed, but disposed into its present arrangement of books and cantos so far at least as ii.iv”); Spens, p. 37 (“the first three books were nearly in their present shape by 1588”); Bennett, p. 239 (the “citation of book, canto, and stanza shows that [Fraunce] had seen a manuscript of the poem, complete through at least ii, iv”).

Probably by the summer of 1587, as Mrs. Bennett shows, op. cit., pp. 239–241.

page 1080 note 8 That is, he already had in mind, or had adopted, the repetitive structure which is de scribed in the letter and to which the poem partly conforms; see below, Sections iii and iv.

page 1081 note 9 A Discourse of Ciuill Life (London, 1606), pp. 26–27.

page 1081 note 10 It is claimed by John Erskine, “The Virtue of Friendship in the Faerie Queene,” PMLA, xxx (1915), 831 ff., that the meeting described never took place. Plomer and Cross, Life and Correspondence of Lodowick Bryskett (Chicago, 1927), pp. 77 ff., undertake to show that it could have taken place, while admitting that Bryskett is guilty of at least one major historical inaccuracy. This, I believe, is as far as discussion can proceed: it is not possible to show that, if the meeting took place, Spenser necessarily said what he is reported to have said, and that Bryskett did not derive his account from the letter or tie poem or both; nor, conversely, that, if the meeting did not take place, Bryakett had not heard from Spenser, in other circumstances, some such account as he gives here.

page 1081 note 11 Mrs. Bennett's suggestion that Spenser's starting-point was some (Ariostan) hero's quest for the Fairy Queen, with, seemingly, the implication that the Fairy Queen once had a more important role than she has in the present poem, is of this nature (op. cit., pp. 10–11, 15 ff., et passim).

page 1082 note 12 Ibid., p. 18.

page 1082 note 13 F.Q. x.ix. 12–15. The view that Sir Thepas is Spenser's main source to me inevitable (Bennett, pp. 11–15).

page 1082 note 14 For instance, in the case of Book iii, the “original,” rational version it for Mr. Walter a book resembling Book i in structure (MLR, xxxvi, 48–49); for Mrs, Bennett, it is two Ariostan, stories which Spenser had already written when ht began to assemble the book (op, cit., l43).

page 1082 note 15 See, e.g., Bennett, pp. 165–166, where narrative difficulties in a particular passage are attributed merely to “replanning and imperfect combination of materials,” which are not further defined, Cf. above, n. 2.

page 1083 note 16 I do not know that any of Miss Spens's reviewers pointed out that her “one piece of direct evidence in favour of an original eight books plan” (p. 27), the Red Croat Knight's seven years' service of the Fairy Queen, it a romance commonplace which does not require any particular explanation: see Hall's edition of King Horn (Oxford, 1901), p. 139, note to line 732.

page 1083 note 17 Op. cit., pp. 103–104. Mrs. Bennett's examples are baaed mainly on the valuable, but admittedly conjectural, analyses of Professor Renwick. The only certain example (apart from the mere insertion of topical passages) it the revision of Spenser's contributions to van der Noodt's Taheatre, and this is a matter of revising words and idioms, which is not the sort of revision postulated for The Faerie, Queene by the critic Nor is the assembly of isolated fragments which is thought to be involved in some of the minor poems really analogous to this sort of revision; though it is, in fact, analogous to the method by which I conceive The Faerie Queene to have been put together. As far as I know, the only work in which Spenser can be clearly seen revising and rearranging on a large scale is the View. Here the errors of logic and fact which occur in the P.R.O. MS. are not unlike the narrative discrepancies of The Faerie Queene; the methods which have been applied to the poem ought to be capable of showing that this MS, represents the latest version, arrived at by tampering with the more logical version found is other MSS.; yet no one, I suppose, will quarrel with the conclusion of Professors Renwick and Gottfried that the offending passages represent an early version. See Variorum Prose Work, (Baltimore, 1949), pp. 509–510.

page 1083 note 18 MLR, xxxvi, 37. I am indebted to Professor C. J. Sisson for permission to quote from MLR>

page 1084 note 19 Ibid., p. 47; cf. esp. p. 39: “It would seem that after would be more effective in its present position”; and p. 44, the discussion of the rape of Amoret.

page 1083 note 20 Ibid., p. 37.

page 1084 note 21 Mrs. Bennett in fact does so, by interpreting the 1590 version as a “fragment of a continuation of Sir Thopas” (op. cit., pp. 18 ff.). But in such a continuation such an episode would surely be discussed at length; a seven-word summary can hardly be a fragment of the story which it summarises.

page 1084 note 22 iii.iii.53, an unsuccessful attempt to tighten the grammatical structure.

page 1084 note 23 See Smith's edition (Oxford, 1909), i, xvi; Variorum i, 516; Bennett, p. 161, for a consensus, of opinion to this effect.

page 1084 note 24 Spens, p. 16; Walter, MLR, xxxvi, 52; Bennett, Ch. xii.

page 1086 note 25 Walter, MLR, xxxvi, 44; Bennett, p. 159.

page 1087 note 26 ELS, xix, 166–167: “... twelve book-units each of which, after the first, repeats in essentials the narrative structure established by the fait, and makes no major contribution towards developing the matter of its predecessor.”

page 1087 note 27 In an epic structure there would have been the excellent reason of authority; but the repetitive structure is not epic.

page 1088 note 28 No doubt for the sufficient reason that the Arthur of Geoffrey and his followers became king at fifteen. “Historical” logic applied to The Faerie Queene would thus make Spenser's Arthur at most a precocious adolescent, if not a forward infant not yet in his teens.

page 1088 note 29 It is thus not so dear as has been urged (e.g., by J. W. Draper, PMLA, xxxix, 319–320) that Spenser was much concerned with an historical basis for his poem. Mrs. Bennett accounts for the scantiness of Arthur's story bv suggesting that he is a late addition to the poem (op. cit., Chs. v and vi). “Lack of connection between Arthur's different appearances” (p. 58) proves nothing except Spenser's failure to provide the connection, and has no particular relevance to date; the appearances could have been logically linked whether Arthur wan a “late” or “early” idea in Spenser's mind. Dating does not affect the argument above, that Spenser was inevitably short of material recognisably Arthurian. Mrs. Bennett recognises this herself (p. 10), yet devotes a chapter to showing that Spenser was likely to avoid the tradition material because its historicity was in doubt, rather than for literary reasons. The point of this demonstration would be clearer if it could be shown that Spenser chose to writ* on Arthur “before he was king” because he wished to avoid the traditional material, rather than that be avoided the traditional material because he wished to write en Arthur “before he was king”; or if it were obvious from the letter to Ralegh that Spenser regretted having undertaken The Faerie Queene rather than the projected epic on Arthur “after that hee came to be king.” It is not obvious; and, in spite of remarks about “historicall fiction” and “being made famous by many mens former workes,” the letter implies an attempt to use a hero who resembles the traditional Arthur in little more than name, a hero without a story; and this is borne out by the poem. The utmost use Spenser can make of the traditional material is to snatch from it a few hints meant to assure the reader that this is indeed Arthur of Britain (e.g., the dragon-helmet, i.vii.31); for the rest, the traditional material must be carefully avoided. The matter from the chronicles in ii.x and iii.iii thus deals with most things in Geoffrey except Arthur. The time of the action is meant to be Uther's reign (iii.iii), but the scene is in Fairyland, not in Uther's Britain. The account of Spenser in H. Maynadier, The Arthur of the English Poets (Boston, 1907), Is punctuated with comments on the failure to use traditional material (pp. 258, 253, 266, 270–271, 274); the obvious explanation of the failure, that for Spenser's Arthur no traditional material was available, does not seem to have occurred to the author.

page 1089 note 30 Edmund Spenser (London, 1925), pp. 51–55.

page 1090 note 31 “The third book shows many signs of confusion. The fourth book is chaos” (H. E. Cory, in Variorum I, 344); Books i and ii “have a form and unity . . . which the others, save the Fifth, arc almost without”; “the weakness of [Book iv] may be chiefly traced to its utter want of form . . . In the third book. . . there is little unity of form . . . the fourth book is a riot of formlessness. . . this book it a piece of patchwork” (K. M. Warren, in Variorum iii, 312; iv, 283); “the fourth [book]... it... the least satisfactory at a story” (John Erskine, in Variorum iv, 289); “To many critics the fourth book ... hat appeared to be the moat loosely organized of the entire poem” (H. S. V. Jones, ibid., p. 301); Book iii was “hastily concocted”; It was “arranged rather than written at a book of Britomart. . . with something like a scissors-and-paste technique,” and “shows every evidence of hasty workmanship”; Book iv “was made out of leftovers from the material out of which Book iii was hastily constructed” (Bennett, pp. 144, 152, 153, 155, 163).

page 1090 note 32 Neil Dodge had sufficient material to demonstrate this thesis in his classic paper of 1897. He was more concerned to show and describe Spenser's borrowing of episodes than to discuss large-scale poetic form, yet he often came near to eatablishing the specific point at issue: ate especially his discussion of Britomart's story, PMLA, xii, 175–178, and his deseription of Book iii, ibid., pp. 190–195. A. H. Gilbert, “Spenser's Imitations from Ariosto: Supplementary,” PMLA, xxxiv (1919), 230, arguing mainly from Spenser's use of Ariostan devices of rhetoric, states the view I wish to street: “The narrative of Books Three, Four, and Five reminds one, in its construction, of fifteen cantos or so of the Orlando.” There are hints of a similar view in Professor Gilbert's On the Composition of “Paradise Last” (Chapel Hill, 1947), pp. 165–167.

page 1091 note 33 Here too may belong the episode of Braggadochio and Belphoebe in ii.iii, which in it-self is pointless as far as Guyon's story it concerned, though it is well cued into Book ii. See Bennett, pp. 49–50. Except for matter rising out of the linking of Books ii and iii, and Arthur's story, Braggadochio and Belphoebe are the only major characters of iii and iv whose stories are back-referenced to ii (the Timias of iii and iv is hardly to be recognised in the squire of i end ii).

page 1091 note 34 Discussion of the making of Boob iii and iv, regarded at more or less self-sufficient unite of the poem, at in Mrs. Bennett's Chs. xi and xiii, is thus not to the point, even though Mrs. Bennett recognises the Ariostan basis of the hooks.

page 1092 note 35 ELH, xix, 166.

page 1092 note 36 I group the books of 1596 with those of 1590 (even though they were published after the letter) since much of them may have been written by 1590, but more especially since they show no more advance beyond the second stage of this series than do Books i–iii. I am of course aware that parts of iv (probably) and of v (certainly) must have been written or adapted to suit historical events which occurred after 1590, and that vi may have been written entirely after 1590. This does not disturb the drift of the argument, since only limited areas of the poem are concerned.

page 1094 note 37 See especially Bennett, p. 87. “The earliest probable date for the beginning of the story of Britomart” is thought to be “late 1585 or 1586,” mainly because iii.iii.35 is said to contain matter borrowed from Powel's History of Cambria (1584). (See C. A. Harper, The Sources of the British Chronicle History in Spenser's Faerie Queene [Bryn Mawr, 1910], pp. 153 ff. The use of Camden [in sts. 7–8?] seems less certain, and I do not understand Mrs. Bennett's reference to Warner.) Two pages later, however, Mrs. Bennett admits that “the substance of the chronicle matter . . . could have been written by 1580,” and it if obvious that the matter from Powel could easily have been inserted at any time between 1584 and 1590. But Spenser could have used a MS. source much earlier. Humphrey Llwyd's MS., the basis of Powel, fat identical with Powel's text in all relevant essentials (see Cotton MS. Caligula A.vi [1559], fob. 19^v–20^r; Powel's edition, pp. 15–16); Powel (sig. V) says that Sir Henry Sidney had a copy “lieng by him a great while”; and other copies were available, since Powel had one presented to him, too late for him to make use of it (sig. [vi]). Miss Harper (pp. 24 ff.) declined to consider the possibility that Spenser used MSS.; arbitrarily, as it seems to me, and especially so in this case, when a contemporary MS. was in the hands of a family with whom Spenser was acquainted —Mrs. Bennett seems sufficiently convinced that the matter from Powel Indicates wholesale revision of the chronicle matter, not only in iii.iii but also in ii.x, to date the addition of ii.x to Book n by reference to the publication of Powel's book (p. 136).

page 1095 note 38 Op. cit., pp. 81 ff. It does not seem to follow that, as Mrs. Bennett urges (p. 92), the appearance of Elisabeth in iii.iii Is necessarily an afterthought: the chronicle matter can be Interpreted as a double-barrelled compliment deriving the Queen from Geoffrey's Arthgall of Warwick via the British kings.

page 1095 note 39 Bennett, p. 92. See previous note.

page 1096 note 40 I hope to treat this aspect o(the subject more fully in a later paper.

page 1096 note 41 See above, p, 1080. Since it is hard to believe that Spenser could have had the plan of the letter In mind for two or three years without either seeing that it was ultimately unworkable or coming nearer to achieving it in Books i-iii by 1590, I should suppose that in 1587 he bad reached only the notion of repetitive structure, and that the super-epic structure described in the letter was the ill-considered invention of late 1589.

page 1096 note 42 By bringing to a close the unexciting “quest” of Scudamour.

page 1096 note 43 Whence, presumably, arise difficulties at the beginning of Book iii: the violently achieved conjunction of Arthur and Guyon—though it is worth observing that Guyon in ii.xii.84, 87 is obviously going back on his tracks (to the house of Alma?); Guyon's possession of a horse, and the attribution of anger to Guyon. For it is clear that, if the episode at the beginning of iii.i belongs to a stratum of the poem different from that of Book ii, the Guyon of Book ii has in iii.i been inserted into a context to which be does not belong. Assumptions of large-scale rearrangement at the end of Book ii (Spens, p. 18; Walter, MLR, xxxvi, 39; Bennett, pp. 130, 145) are unnecessary. The two adventures are distinguished in: “long wayes” refers to Guyon's expedition, “sory wounds” to Arthur's acquired in hence it does not appear likely that iii.i.l is older than ii.xii, and the reference to Acrasia in iii.i.2 surely cannot be (Bennett, p. 145). iiii.l is therefore patching at the junction of the two strata of the poem, and so is iii.i.2, as Miss Spent says, but not for (be reasons she suggests. The change in grammatical number which disturbs her can be paralleled in i.iii.23. The story of “Duessaea traines” (iii.i.Arg.; cf. Walter, MLR, xxxvi, 40, and Bennett, pp. 146–147) may be taken as one which Spenser proposed to insert in proof, as an additional link between Books i-ii and iii; in fact be went no further than correcting a misprint in the versicle in Faults Escaped. This explanation is guesswork, as is any other; but there is certainly no evidence that Duessa ever appeared in the canto.

page 1097 note 44 Amoretti lxxx.

page 1097 note 45 The fact that Britomart is first introduced into the poem in Book iii (in the first place, as Bradamante it introduced in O.F. i, but in the second place at Red Cross is introduced in F .Q. i and Guyon in F. Q. ii) makes it easier for the reader to see Book iii as bearing tome excited by the fairly similar openings of i, ii, and iii is, of course, disappointed when he finds that iv merely continues the narratives of iii without introducing a new major character. In this sense the Ariostan matter of iii is more easily adaptable to the repetitive struc-ture than that of iv. Possible reworkings in iv are : the appearance of Arthur in canto viii, an attempt to match the structure of i and ii in this respect (Bennett, pp. 164–165); and the story of Timias, written or adapted to suit the current affairs of Ralegh (Bennett, pp. 168–173). But there is Ariostan matter in this episode which suggests that it is at most a rewriting: see W. J. B. Owen, “Spenser and Ariosto,” N & Q, cxciv (1949), 317. Elsewhere Mrs. Bennett seems to envisage the possibility that “the poet bad on hand a story which he found it possible to fit to the allegory of Ralegh's disgrace with almost no revision” (p. 279). The identification of Timias with Ralegh has, of course, been doubted.

page 1098 note 46 On the structure of Books V and VI see Bennett, Chs. xiv, xvi. This paper is not concerned with their structure or evolution except in so far as v contains some of the Ariostan stratum and both are, as wholes, intended as units of the repetitive structura. I should suppose that they were modelled, more or less, on i and ii, but on i and ii as stories of knightly quests, not as miniature epic structures.

page 1098 note 47 It may be useful to indicate briefly the differences between this scheme and that proposed by Mrs. Bennett. I give references to her summarising Ch. xviii, where the order of pages conveniently parallels the order in which, according to Mrs. Bennett, the poem was composed: (I) I find no satisfactory evidence for an Ariostan beginning based on Sir Thopas or the order of Maidenhead (Bennett, p. 231; above, pp. 1081–82, 1085 n. 21, 1095); further, neither of these plot-dements can be satisfactorily connected with the fragment of a poem called The Faerie Queene which Harvey saw is 1580 (Bennett, p. 232; above, pp. 1081–42, 1094–95). (2) I agree, in general terms, that at some stage Spenser shifted his attention from an Ariostan imitation to the illustration of virtues (Bennett, pp. 235–234); but I do not accept Mrs. Bennett's view of the nature of that Ariostan imitation (above, pp. 1095–96), nor Bryskett's Discourse as certain evidence for date (Bennett, pp. 235–236; above, pp. 1081, 1096). (3) I do not accept Mrs. Bennett's notion of a second period of Arioetan imitation concerned with Britomart (pp. 237-238); on the contrary, I see this as Spenser's beginning (above, pp. 1092, 1094); nor do I accept her method of arriving at this conclusion (above, n. 37). I agree that Britomart's story is probably designed to compliment the Dudley family (above, p. 1095). (4) I find no evidence tending to show when Arthur was introduced into the poem, but, having regard to the Arthur Spenser seems to have wished to present, I do not think that Mrs. Bennett's argument for a late date is necessarily valid (Bennett, p. 238; above, n. 29). (5) I do not think that large-scale revision of Boats i and ii (or indeed, of any part of the poem) can usually be detected (Bennett, pp. 239–240; above, Section ii of this paper). I agree that minor revisions (of a stanza or two) may have been made in Books i and ii (or elsewhere) at a comparatively late date (Bennett, p. 242). I see no evidence of major revision in Book iii (Bennett, pp. 239, 241): the only points of importance are the cuing-in to the end of ii at the beginning of iii, and the completion of the Scudamour-Amoret-Britomart episode at the end of iii in the 1590 text (above, p. 1096). (6) This paper is not concerned with the evolution of Books v and vi (Bennett, pp. 243–244) except as is indicated in n. 46, above. (7) I agree that roost of Book iv was in existence by 1590 (Bennett, p. 243); but I date It earlier than does Mrs. Bennett, and I find no certain evidence of major revision in it (above, n. 45).—There emerges a wide difference of opinion on the nature and date of the Ariostan part of the poem. This is seen by Mrs. Bennett at two strata of writing: one minor and early, the other major and fairly late; the whole assembled from narrative elements, Book iii in 1585–1586 and later, Book iv in 1592, with consequent and often unsatisfactory revision. The result is seen, on the whole, as a poorly integrated composite of Ariostan episodes; signs of Ariostan “organization” in iv are attrihatad to revision (pp. 175–176). The scheme Jescribed above sees in the Ariostan cantos a single stratum of writing, well enough integrated to be recognisable as an imitation of Ariosto's characteristic structure, written early, and showing signs, not of major revision, but rather of the need for it.

page 1099 note 18 For instance, the insertion or substitution of material from Powel's Historie in ii.iii.35, if that matter is really borrowed from Powel. See above, n. 37.

page 1099 note 49 Say from 1584 or 1585 onwards.

page 1099 note 50 For instance, the story of Burbon in v.xi.

page 1100 note 51 There are norefernces to Ariosto in Miss Spens's chapter on the structure of the poem, and no significant ones in Mr. Walter's articles except an attempt to minimize the influence of Orlando Furioso (MLR, xxxvi, 56). Mrs. Bennett is well aware of the importance of Ariosto to Spenser, but, because of her concern with the notion of “development,” her treatment of the Ariostan books suffers from the ambiguites pointed out in Section ii above.