University of Virginia Library

The chief contention of this article, that Keats wrote the Induction to The Fall of Hyperion in September and October of 1817, is not new. Amy Lowell argued warmly and extensively for such a view,[1] and Claude Lee Finney continued to support the same position with arguments based solely on internal evidence.[2] But all recent major biographers of Keats—Hewlett, Bate, Ward, Bush, and Gittings[3] —have rejected that dating in favor of either late summer and fall of 1819 or fall and winter of 1819. It is fair to say that the later dating has been repeated so often from such respected authorities that it has hardened into a kind of conviction which is generally accepted as virtual fact. There are reasons to believe that it is wrong, and the case should be reopened.

The Induction to The Fall of Hyperion is distinctly separable from the other version, Hyperion, unlike the remainder of The Fall, which is merely a recasting of passages from Hyperion. The dating of the Induction is vitally important for an understanding of Keats's poetic development because in the last three decades it has come to be regarded as one of Keats's masterpieces, valued comparably with the great odes and Lamia. The difference in the dating of somewhat less than a year which is at issue would probably matter little for poets with lengthy careers like Dryden and Tennyson, but for Keats that time period is almost one-fifth of the scant five years that he had to write, and of those five years the last is by far the most significant. Whether the Induction was the beginning or the culmination of the annus mirabilis is of the first importance.

Since my argument for the earlier dating will require the unravelling of a complicated tangle of confusing evidence, it may be helpful at the outset to sketch the basic pattern as a preliminary guide. The Charlotte Reynolds transcript of The Fall of Hyperion furnishes evidence that Keats wrote the Induction in September and October of 1818. I shall contend that he first wrote the opening twenty-one lines of Hyperion before the Scottish tour in the summer of 1818, turned away from that start to begin afresh The Fall of Hyperion in September 1818, recast the first twenty-one lines of Hyperion as ll. 294-326 of The Fall, laid aside The Fall and returned to writing the rest of Hyperion by April 1819, and in the summer of 1819 returned once again to The Fall to expand it from l. 327 to its close. Two obstacles to this view will be removed: Charles Brown's statements that The Fall was a remodeling of Hyperion and Professor Jack Stillinger's conclusion that Charlotte Reynolds copied Woodhouse's transcript of The Fall. Finally, Keats's own report of the poetry written in the summer of 1819 will show the strong probability that he resumed The Fall with l. 327 at that time.

The fundamental evidence upon which my argument is based is the transcript of poems by Keats in the Reynolds-Hood Commonplace Book in the Bristol Central Library,[4] which has not been taken into account properly by anyone who has studied the relationship between the two Hyperions. It contains fifteen poems by Keats (counting Extracts from an Opera as one), including The Fall of Hyperion, ll. 1-326. Thirteen of them have been dated with certainty, and they were all written in 1817 or 1818.[5] Though "Stanzas: You Say You Love" is not dated on any transcript, no one recently to the best of my knowledge has dated it later than 1818.[6] Aside from The Fall of Hyperion, ll. 1-326, the last written poem among the transcripts is "Sonnet Translated from Ronsard," 22 September 1818.[7] Why wasn't a single poem included from the prolific and golden year of 1819? The easiest, and therefore most probable, inference is that they were copied into the

commonplace book late in 1818. Why wasn't the rest of The Fall included? Again the simplest answer is that it had not been written at the time of the transcribing.

Keats's letter to James Rice of 24 November 1818 corroborates the inference that Reynolds received these manuscripts from Keats and had his sister Charlotte copy them into his commonplace book late in 1818.[8] To illustrate how friends can misunderstand each other, Keats wrote:

Rollins identifies the friend cautiously, "Perhaps Reynolds. But if Keats refers to two friends, the second may have been Woodhouse" (I, 407n). But the caution is unnecessary, since the weight of the evidence indicates two remarks to one friend. The only ambiguous phrase is "to each of my friends' questions," and the context guides one to understand that Keats misplaced the apostrophe and that he meant "to each of the questions of my friend." Ample evidence supports the identification of Reynolds as the friend. He was a friend of Rice's, and Reynolds was very close to Rice. He planned to visit Severn, and Reynolds visited Severn frequently, having written Keats on 14 October 1818, "when shall I go with you to Severn's" (I, 377). Severn did paint Reynolds' portrait. Reynolds had earlier requested that Keats provide him with copies of his poems: on 14 October 1818 he had written, "You will gratify me much by letting me have, whenever you have leisure, copies of what you write" (I, 377). Surely M. B. Forman's conclusion is warranted, "I think this friend—the two slips clearly belong to one occasion—was Reynolds."[10]

There is every reason to believe that Keats performed what he promised by delivering the manuscripts to Reynolds in early December

1818 and that Reynolds had them transcribed, including The Fall of Hyperion, ll. 1-326. The last inference, however, leads to the thorniest problem of all, a problem which has caused me on several earlier occasions to abandon further investigation after being convinced that all the other pieces fit the pattern perfectly. How could ll. 294-326 of The Fall of Hyperion possibly have been written before December 1818? The Induction poses no difficulty, but the last thirty-two lines include a reworking of the first twenty-one lines of Hyperion, and, as Sir Sidney Colvin demonstrated conclusively so long ago as 1887, the passages in The Fall of Hyperion which parallel passages in Hyperion are clearly a revision of Hyperion.[11] Colvin was not aware of the transcription of The Fall in the Reynolds-Hood Commonplace Book, which was not discovered until 1932 and did not become widely known until H. W. Garrod's edition of The Poetical Works of John Keats in 1939, but those 326 lines are reasonably close to the Milnes printed text of 1857, which he knew.[12] Colvin also did not have Keats's autograph of Hyperion, which did not come to light until 1905,[13] but he did have a Woodhouse transcript of Hyperion, which included penciled records of earlier readings from the autograph. His shrewd scrutiny revealed that in every case passages from The Fall were revisions, and he concluded with a thunderclap, "the proof is from all sides absolute; and the 'first version' theory [that The Fall had been written first] must disappear henceforward from editions of and commentaries on our poet" (p. 228). Further evidence unknown to him, however, argues that he was only half right.

Beginning his citations immediately after l. 21 of Hyperion and l. 326 of The Fall, Colvin did not prove that ll. 294-326 of The Fall included revisions of ll. 1-21 of Hyperion because he lacked the full evidence of Keats's autograph of Hyperion (Garrod's A). But careful comparison of the autograph and corresponding sections in ll. 294-326 of The Fall shows indisputably that they too are in fact revisions. The clearest and most emphatic example is the famous adjective "realmless" modifying "eyes" in l. 324 of The Fall. The autograph of Hyperion shows that Keats began with "whitebrowd", shifted to "ancient", and canceled that in favor of "realmless" (Garrod, p. 277n.); The Fall includes only the final, masterful "realmless".

How then could the version of The Fall copied into the Reynolds-Hood Commonplace Book have been written before Hyperion? All Keatsian scholars know that Woodhouse wrote at the end of the transcript which he made of Hyperion, "Copied 20 Apl 1819 from J. K.'s Manuscript written in 1818-19" (Garrod, p. 305n.), and all major recent biographers of Keats have set the time of composition of Hyperion from the fall or winter of 1818 until the spring of 1819. The answer is startling in its simplicity. Everyone for the last twenty-five years since Dorothy Hewlett's biography has been almost right about the time of composition of Hyperion, but not completely so. Keats wrote the first twenty-one lines of Hyperion before the Scottish tour.

George knew before he left England on 24 June 1818 that Keats's next major project was the fall of Hyperion (II, 12). The canceled Advertisement for the 1820 printed version of Hyperion reports that Keats "commenced the Poem just before the Publication of his Endymion; and he abandoned the <idea> intention of proceeding with it, in consequence of the reception that work experienced from some of the reviews.—"[14] Endymion was published sometime from about 27 April until 19 May 1818.[15] Heretofore no one has paid much attention to either the canceled Advertisement or the abbreviated and revised version which was printed because Keats wrote opposite the latter in one copy of the Lamia volume, "This is none of my doing—I was ill at the time. This is a lie" (Bush, p. 189). But the relevant part of the printed version reads, "The poem was intended to have been of equal length with Endymion, but the reception given to that work discouraged the author from proceeding" (Lowell, II, 424); it says nothing of Keats's beginning the poem before the publication of Endymion. Keats gave the lie only to the claim that he broke off the poem because of the adverse reviews. There is no reason to doubt that he "commenced," and only barely so, by writing before 19 May at the latest the first twenty-one line passage, which presents of course a magnificently self-contained scene.

Keats was at first eager to read aloud to his friends the recently completed Isabella (by 27 April); he wrote Bailey that he longed to read it to him in Scotland (I, 294), and he did read it aloud to Brown on the walking tour (Forman, p. 160n.). He had not, however, made enough progress with what his friends understood to be his next major project to justify either reading aloud the opening passage of Hyperion

or mentioning it in his letters, at least not in those which have been preserved. In Keats's day, only Woodhouse would have thought such an apparently small matter important, and it was he who wrote the canceled Advertisement to preserve the truth, unrecognized though it has been for so long.

Two obstacles must be overcome if my view that Keats wrote the Induction to The Fall of Hyperion in September and October 1818 is to be accepted. The first is Charles Brown, who reported that at the time Keats was writing The Cap and Bells late in 1819, "In the evenings, at his own desire, he was alone in a separate sitting-room, deeply engaged in remodelling his poem of 'Hyperion' into a 'Vision.' The change in the conduct of the poem has not, in the opinion of his friends, been regarded as an improvement."[16] Brown also endorsed a wrapper enclosing transcripts of poems made by Woodhouse's clerks in 1833, "Hyperion (remodelled) with minor poems" (Stillinger, p. 29). But Brown was far from infallible. Without going to the extreme of Charles Dilke, who noted that Brown's "Life of Keats" was "a dream on the subject," a recent critic has declined to accept his account of the composition of another poem. Robert Gittings has demonstrated that Brown probably confused Ode to a Nightingale with Ode on Indolence in the famous story about recovering the scraps of paper from behind books.[17] Gittings suggests that R. M. Milnes heard of Dilke's skepticism about Brown's account and therefore modified the story in his edition of Keats's poems in 1876.

In his statements about The Fall of Hyperion, Brown would be only a little more than half wrong if the present argument should be accepted; 45% of the poem would still be a remodelling of Hyperion. When Milnes first published The Fall in 1857, he said that he did not know which of the two versions was earlier. When he republished it in 1867, something had changed his opinion so that he oversimplified by declaring without reservation that The Fall was the first version. It seems likely that he overstated the case in the later edition, because one of Keats's surviving friends—Dilke, Severn, Charlotte Reynolds, Clarke, or Holmes—told him that much of The Fall had been written before Hyperion. In any event, it is clear that Milnes did not believe Brown in 1867, and we too should be wary of him and open to other possibilities.

The second obstacle is much more formidable, Professor Jack Stillinger's conclusion that Charlotte Reynolds copied Woodhouse's

transcript of The Fall (W2). Stillinger's massive study of all Keatsian MSS in The Texts of Keats's Poems is one of the greatest contributions to Keatsian scholarship. His overall success, however, creates a problem: whenever he does make an error, the weight of his deserved authority makes it difficult to see. It is necessary to quote Stillinger at considerable length. After having cited evidence that Woodhouse copied a lost holograph of The Fall of Hyperion made from a prior first draft, and after having proved conclusively that Woodhouse copied from the lost holograph, Stillinger continues:

The essential passages in the cross-reference relating to The Fall of Hyperion would seem to be these:

Before undertaking a detailed analysis of Stillinger's argument, I should note at the outset a basic improbability that lies behind it. It is unlikely that Charlotte would go outside the family to Woodhouse for copies of Keats's poems when it was so much easier to borrow them

from members of her family right in the house in Little Britain. All the borrowing that we have any record of was in the reverse direction: Woodhouse borrowed Keatsian material from the Reynoldses, not the Reynoldses from Woodhouse.

Stillinger's evidence can be divided into three parts: substantives, apparatus of the MSS, and punctuation. The substantive differences are merely neutral; there are some, but they are compatible either with Charlotte's copying W2 or with her copying independently the same Keats holograph which Woodhouse copied. If I understand him, Stillinger does not argue that the substantives prove the case one way or another.

Stillinger's account of the similarity of apparatus between the MSS is misleading because he cites only the similarities and none of the differences. One difference is perhaps not especially important: she signs at the end all the poems but three ("Blue!—'Tis the Life of Heaven," "Hence Burgundy, Claret, and Port," and "Sonnet Translated from Ronsard") "J. Keats" or "J. K." Another difference is possibly significant: Charlotte does not number the lines as Woodhouse does. If she had copied from Woodhouse, it would have been simple and convenient to borrow his numbering of the lines for such a long passage. A final difference is, I think, somewhat significant: Charlotte wrote "Canto 1st" rather than Woodhouse's "Canto 1." Keats preferred the form of "1st, 2nd, 3rd" etc. In the fair copy of Endymion, he wrote "Book 1st," "Book 2nd," "Book 3rd," and "Book 4th." In the holograph Hyperion, he wrote "Book 1st" and "Canto 2nd," though he did shift to "Canto 3—" for the third. In Otho the Great, he used "Act 1st Scene 1st"; "Act 1st. Scene 2nd"; "Act 1st Scene 2nd [the 2 altered to 3]"; "Act 2nd—Scene 2nd"; "Act 2rd Scene 2nd" (obviously a slip in the act for 3); for Act V, Scene 3, mistakenly "Scene 2nd"; and for Act V, Scene 4, mistakenly "Scene 3rd".[18] Woodhouse usually preserved Keats's kind of numbering, sometimes in shortened form: in his copy of Hyperion, "Book 1st", "Book 2d", "and Book 3d." and in W2 of The Fall, "Canto 2d", but he deviated for the first canto of The Fall with "Canto 1". The small probability is that Charlotte copied her "Canto 1st" from the Keats holograph. On balance, though the similarities and differences in apparatus between Charlotte and W2 come close to canceling each other out, I believe that Charlotte's omission of the line-numbering and her writing "Canto 1st" argue mildly for her copying from the Keats holograph rather than from W2.

In his account of the punctuation Stillinger is again misleading with the first example cited: that both "omit the opening quotation mark at I.156." That item is more than negated by Charlotte's including the opening quotation marks at I,148, where Woodhouse omitted them. If, as Stillinger implies, Charlotte had been mindlessly following W2 and omitted the quotation marks at I,156 because she did not find them there, she would also have omitted them at I,148. But she did not, and the probabilities would seem to be that the Keats holograph included them at I,148 and omitted them carelessly at I,156; that Charlotte copied what she saw, including them at I,148 and omitting them at I,156; and that Woodhouse carelessly omitted them at I,148. Crediting Charlotte with a conscious process of deciding that the quotation marks at I,148 were needed and of supplying them when she did not find them in her source is clearly less probable than supposing that Woodhouse carelessly omitted the marks at I,156.

The generalized argument that Charlotte must have copied W2 because the punctuation is the same in many long passages is difficult to assess. The chain of reasoning is that Keats punctuated erratically, that Woodhouse had a "known tendency to add punctuation and otherwise alter accidentals," that Charlotte's punctuation is very close to Woodhouse's, and that therefore the regularized punctuation could not have come to Charlotte from an erratic Keats, but must have come from Woodhouse.

Without any testing of the significance of the evidence, the argument from punctuation might seem so persuasive as to tip the balance. A completely satisfactory testing is not possible, since we have no case where Woodhouse transcribed a holograph fair copy of comparable length without filtering through shorthand. It is important to note that Stillinger decides, and I concur, that the source of W2 was a copy and not the original draft. It is also important to note that Stillinger has proved on the basis of Keats's preparation of some fair copy for Poems of 1817 that Keats "could punctuate and spell correctly when he wanted to" (p. 64). Fortunately an approximate test of the significance of the similar punctuation is possible, since Woodhouse copied the W2 transcript of Hyperion directly from the holograph. This test is only approximate because, unlike the source of the W2 transcript of The Fall, the holograph of Hyperion was the draft and not a fair copy. Though some have doubted that the holograph of Hyperion was the first and only draft of the poem, I agree with Stillinger that we ought to accept Woodhouse's distinct and emphatic assertion that it was. It is not nearly so thoroughly marked up as the original drafts of other

long poems, but it does include many cancellations and insertions above the lines, in the margins, and on the facing pages.[19]

Comparison of the W2 transcript of Hyperion with the holograph reveals that Woodhouse was perfectly capable of reproducing "substantial passages" with the punctuation "exactly the same in every particular." I report only the passages of six lines or more where the punctuation is identical: I, 11-16, 23-29, 31-51, 107-119, 220-225, 299-306; II, 4-9, 51-59, 159-165, 310-315; III, 73-78, a line included in the MSS but omitted from 1820 plus 126-130. I have kept this record as strictly as my scrutiny would allow; in some cases the strictness may not give a full impression of the closeness in long passages. For example, if in I,298 the dot of Woodhouse's semicolon in W2 fell on the crossing of the t as seems very likely from the position, and is therefore indiscernible, then the two MSS would be the same for fourteen lines, I, 293-306. And, except for one faint and almost illegible comma in the holograph at the end of III,131, which Woodhouse may have over-looked, the last twelve lines are the same—III, line deleted from 1820 plus 126-136. But even without the two appended suggestions, there are a twenty-one line passage and a thirteen line passage. The twenty-one line passage is not so long as the thirty-one line passage cited by Stillinger in the MSS of The Fall, and there are considerable differences in the punctuation elsewhere in the MSS of Hyperion, but the differing degrees of similarity must be balanced against the fact that Woodhouse transcribed from a draft and not a copy in Hyperion. The test, I believe, reduces the significance of Stillinger's argument from punctuation to very little. It need not seem especially surprising that Woodhouse and Charlotte could transcribe a Keats holograph independently and preserve the punctuation in unbroken passages for about one-fourth of the document, while they varied somewhat in the rest of it.

Unless we consider the small differences in the apparatus of the MSS sufficiently persuasive to indicate a holograph source for Charlotte, we are virtually back where we started, with both possibilities open: that Charlotte copied from the holograph or that she copied from W2. There is, however, another set of accidentals not considered by Stillinger which, in my opinion, will solve the problem beyond reasonable doubt—the spelling. A few of Charlotte's spelling variants are of doubtful significance, though they incline slightly toward a Keats holograph. "Pendant" for "pendent" (I,198) sounds

a little Keatsian since Keats sometimes misspelled words ending in "dent" like "independent" that way, but so do a great many other careless people. "Ancient" (I,326) agrees with Keats's spelling of the word rather than what looks like Woodhouse's pedantic quaintness in his "antient".

A repeated misspelling is of a different order and is definitely significant: Charlotte included "cans't" for Woodhouse's "canst" twice (I,117,167), though she once spelled it correctly without the offending apostrophe (I,108). Keats had a habit of writing "can'st" or "cans't"; one or the other appears in the fair copy of Endymion, II,125 (from which it even slipped into print), in Hyperion, I,338 (very faint), 339, 343, and III,69. In addition, George Keats, who Stillinger assures us was a careful copyist, almost certainly took from a Keats holograph the three examples in his transcript of Grecian Urn, ll. 3, 15, and 17.[20] For a writer in the early nineteenth century, "canst" was not a contraction and needed no apostrophe. According to the O.E.D., the forms "canest" and "cannest" occurred occasionally in the sixteenth and seventeenth centuries, but the dominant form has always been "canst." It seems highly unlikely that Keats ever saw a "canest" in print; if concordances can be trusted, the word does not appear in the King James Bible, Chaucer, Spenser, Shakespeare, Milton, Words-worth, or Coleridge—the word is always "canst" without any apostrophe because no letter was omitted. Either by false analogy with such contractions as "com'st" from "comest" or by loose association with "can't", Keats developed the habit of writing "can'st" or "cans't".

Woodhouse was sometimes capable of copying Keats's "can'sts" unthinkingly, as he did in the W2 Hyperion, I,338, 339, and 343, but he did not include the superfluous apostrophe for any of the "cansts" in the W2 The Fall of Hyperion. Charlotte could not have seen them there. If she had written "cans't" once, it could possibly be charged to the same process of human carelessness that makes twentieth-century writers who know better and who are contemptuous of the error nevertheless add an apostrophe accidentally to the possessive "its." But hardly twice. The probability is that she was looking at a Keats holograph.

A final misspelling is most convincing of all. In my photocopy of

Charlotte's transcript at I,287, the word which is written clearly "forlorn" in W2 is unclear in the space immediately after the "f". To my request for careful scrutiny of the MS, Mr. Geoffrey Langley, County Reference Librarian at the Bristol Central Library, kindly responded as follows: "I have examined the line in question, and find that the difficulty is caused by an alteration of the spelling from the originally written 'folorn' to 'forlorn.' The over-writing of 'or' is in a darker ink, but as far as can be seen from such a limited sample it is in the same hand."

Keats's distinctive, even weird misspelling "folorn" has long been known; so early as 1939, H. W. Garrod (p. xvii) reported that Keats wrote "not 'forlorn,' but, as his habit was everywhere, 'folorn.'" Keats wrote "folorn" fourteen different times: I Stood Tiptoe, l. 172; Endymion, I,205; II, 859; III, 689; IV,372,373; Lamia, II,49,282; Isabella, l. 497 (the draft); St. Agnes, l. 333; Nightingale, ll. 70,71; Hyperion, II,35; and Otho, I,2.87. Garrod's "everywhere" involves slight rhetorical license for emphasis, since Keats did write "forlorn" twice. The exception in Endymion, III,227 is easily explained: Reynolds, who revised the first draft (sloppily, if he is responsible for missing the other five misspellings of the word), probably corrected the error and Keats copied the correction.[21] The second exception is the Egerton MS (Garrod's E) of Isabella, l. 497, and again it is noteworthy that this is a fair copy—writing more spontaneously in the draft, he had misspelled it "folorn".[22] The exceptions are small and explainable as such; Keats's misspelling of the word is firmly established. Charlotte could have seen the "folorn" which she copied first nowhere else but in the lost Keats holograph; that she (or possibly someone else) came back later to correct the spelling does not alter the fact that she first set down the Keatsian misspelling. I believe that the probability is so strong that it should be generally accepted that Woodhouse and Charlotte copied the same Keats holograph independently.

Before leaving discussion of the MS, I must present one other piece of evidence which supports my central contention that Charlotte copied the first 326 lines of The Fall before Keats had finished writing the rest of the fragment. As Stillinger observes, Charlotte used "for fragmentary pieces . . . dashes and x's at the end to indicate incompleteness" (p. 48). Immediately after I,326, without any break, she

included a line of fourteen x's and two dashes, one beneath the other, under the fifth x. These, I submit, are a clear indication of her meaning: "That is all there is; there is no more to copy." By the time Woodhouse secured the same holograph, probably some time after December 1819, Keats had added the rest of the fragment.

We may turn now to other sources for evidence bearing upon Keats's composition of the Induction to The Fall. After his return from the Scottish tour on 18 August 1818, Keats probably did not write anything for three weeks because of Tom's illness, his own illness, and his resolve to study and reflect before writing further. On 14 September at the party at Hessey's, he disturbed Woodhouse deeply by declaring that he had stopped writing poetry (I,45). But about 21 September he changed his mind, as he reported to Dilke in a letter of that date (I,369), and sought in poetry relief from his feelings about Jane Cox and from the haunting effects of Tom's fatal illness. Instead of resuming with Hyperion, as most have supposed for the last several decades, I maintain that he put it aside and began an entirely new version, The Fall of Hyperion.

Passages in the letters to Dilke and Reynolds have some small significance as a hint pointing toward The Fall rather than Hyperion. He wrote Dilke, "I am obliged to write, and plunge into abstract images" (I,369), and he echoed the word "abstract" with slight variation to Reynolds, "This morning Poetry has conquered—I have relapsed into those abstractions which are my only life. . . . Poor Tom— that woman [Jane Cox]—and Poetry were ringing changes in my senses" (I,370). Ten months later, when he resumed The Fall, he attached the same idea of abstractness to it and associated it with strong feeling for a woman, though a different one. He wrote Fanny Brawne on 25 July 1819, "Forgive me if I wander a little this evening, for I have been all day employ'd in a very abstr[a]ct Poem and I am in deep love with you—two things which must excuse me" (II,132). One must not, however, exaggerate Keats's association of "abstract" with The Fall, since elsewhere he links it generally with beauty and poetry (I,373,403).

In the next three weeks Keats probably wrote all of the entirely new Induction, ll. 1-293, and wove passages from Hyperion, ll. 1-21, into the new version as ll. 294-326 of The Fall. I suspect that he did not find congenial the task of reworking old material. When Reynolds called on 13 October and borrowed a copy of Isabella (I,45), he must have learned also of Keats's work on The Fall, and, I would surmise, he may have received some indication that Keats did not intend to proceed with it, at least for a time. On 14 October Reynolds urged

Keats to publish Isabella because "its completeness will be a full answer to all the ignorant malevolence of cold lying Scotchmen and stupid Englishmen" (I,376; italics supplied). That word "completeness" deserves consideration which it has never received. The implication is that there was another poem available which was not then suitable for publication because it was incomplete, and that poem was, I contend, The Fall, ll. 1-326. My reason for suspecting that Keats had already decided by 13 October to lay it aside is that otherwise Reynolds might not have been inclined even to consider its publication.

Later in October Keats began work again on the Hyperion project, but according to my view instead of continuing with The Fall he returned to the original manuscript of Hyperion and expanded from the first twenty-one lines. On 27 October he wrote Woodhouse of "cogitating on the Characters of saturn and Ops" (I,387). Any mention of sympathetic identification, negative capability, and the egotistical sublime, subjects treated or implied in that letter, is likely to arouse some disagreement because readers have explored these rich and provocative concepts so deeply that different individuals have inevitably assigned differing weights to particular implications, ramifications, and nuances. And yet most will agree, I think, that at least in the obvious surface sense The Fall of Hyperion, which I believe that Keats had just renounced, is the egotistical sublime. An intense effort to achieve the sublime it certainly is, and it is also in a sense egotistical, though the nature and quality of Keats's subjectivity differ of course from Wordsworth's. Just as clearly, the "severe magnificence" of Hyperion is as objective as The Fall was subjective. Keats would have been especially well prepared to develop the critical distinctions in the famous letter if, as I contend, he had just rejected for a time the subjective in favor of the objective.

There is no reason to suppose that his progress on Hyperion differed from the conjectural outlines of it provided by Keats's recent biographers. He completed it by 20 April 1819 when Woodhouse copied it. Shortly before 24 November 1818 he told Reynolds that he would let him "have the Mss next week" (I,407). He could lend Reynolds the manuscript of The Fall, along with the other manuscripts, because he was not then working on it and did not plan to work on it any time soon. Reynolds had all the manuscripts transcribed and presumably returned them to Keats as he had Isabella before (I,376-377).

One final piece of evidence from the holograph of Hyperion should be introduced at this point to support my position that Keats began Hyperion before 19 May 1818, broke it off before the Scottish tour,

and then returned to it once again later. He headed the first section of Hyperion "Book 1st." He shifted from "Book" to "Canto 1st" for the first section of The Fall. When he returned to the Hyperion MS in October, he completed the first section and proceeded to the other two, but, forgetting his initial designation of the first section as "Book 1st" in Hyperion, he continued there with the headings to which he had shifted for The Fall, "Canto 2nd" and "Canto 3". The alternating from one version to the other and then back to the first explains an otherwise inexplicable inconsistency.

A major difficulty for biographers who have maintained that Keats wrote almost all of The Fall of Hyperion in the summer and fall of 1819 has always been insufficient time. Keats's own statistics reveal that he could have written very little of it from 14 June to 14 August 1819. On the latter date, he wrote Bailey, "Within these two Months I have written 1500 Lines . . . . I have written . . . Lamia—(half finished—I [hav]e a[l]so been writing parts of my Hyperion and [c]ompleted 4 Acts of a Tragedy" (II,139). All recent biographers except Ward infer that by "Hyperion" Keats meant The Fall. The first part of Lamia and the first four acts of Otho the Great include a total of 1,860 lines. Although the "1500 Lines" was clearly just a rough estimate, it seems unlikely that he could have written much of the 468 lines of the first canto of The Fall without rounding off the estimate at a higher figure. He told Fanny Brawne on 25 July that he had spent "all day" (II,132) on The Fall, but it seems doubtful that he could achieve more than half of his old quota of fifty lines of Endymion a day in composing the condensed and mellow poetry of The Fall. Grant for the sake of argument that he wrote thirty on that day and another twenty at times not mentioned. That would leave 424 lines yet to be written by 21 September when he quoted Canto II, 1-4,6 in a letter to Woodhouse (II,171). It strains credibility to believe that with all his travelling and his activities in behalf of George he could have written in the thirty-eight days from 14 August to 21 September Act V of Otho (363 lines), Part II of Lamia (311 lines), and "To Autumn" (33 lines) for a total of 707 lines and still have time to write another 424 lines of The Fall.

One senses a strained credibility even in those who have believed that Keats wrote the Induction in the summer and early fall of 1819. W. J. Bate, for example, writes that after subtracting the time for Act V of Otho and other activities, "This leaves little more than three weeks relatively free for other work. During that short time he took off a few days to write the second half of Lamia and tried to go on with the revised Hyperion, possibly writing most of it at this time",

and again, "Keats, left to himself, would be able to find some time for the Fall of Hyperion" (pp. 562, 574). On the other hand, Aileen Ward, who decides that it was impossible for Keats to have written by 21 September 1819 all The Fall through Canto II, ll. 1-4,6, which he quotes in the letter to Woodhouse, strains probability just as badly by contending that Keats composed only those passages quoted and perhaps a few others and afterward from late September through December filled in gaps to complete the remainder of the fragment (p. 434).

It is unnecessary to do so much squeezing and twisting for time in either way. The alternative proposed here is far simpler and more probable: that in the summer of 1819 Keats began again on The Fall with the 326 lines which he had written the preceding autumn and expanded them by 148 lines as he adapted passages from Hyperion for the other version until he reached Canto II, ll. 1-4,6, which he quoted for Woodhouse on 21 September 1819. First he quoted the new passage, and then he added two passages from the earlier Induction, which he had no reason to think Woodhouse had seen. The tense of the verb in the introductory sentence to the last quoted passage is significant, as Amy Lowell (II, 343) saw long ago, "Here is what I had written for a sort of induction—" (II,172; italics supplied). The past perfect tense separates the Induction from his current writing and refers to an interval of almost a year.

Very little remained for Keats to do with the poem. After he had discontinued it on 21 September, Brown reports that he returned in his growing illness in the winter of 1819 to a feeble attempt to adapt additional material from Hyperion,[23] but, as Aileen Ward has remarked (p. 434), he added only five new lines, transferring practically verbatim the remaining fifty lines of the fragment.