T. S. Eliot the Poems, Volume 2

by T. S. Eliot

  ts2b (WLFacs 16–21): blue carbon, annotated by TSE. Three leaves, formerly folded in half, with holes from having been twice pinned.

  ts3: first known typescript of Part III, with much of the published text absent. Second typewriter. Assigned to Nov 1921 by Rainey 200–201. Two versions on Verona Linen paper:

  ts3a (WLFacs 22/23, 26/27, 30–35): ribbon copy on five leaves, formerly folded in half. Annotated by Vivien Eliot, Pound and TSE. Given in WLComposite for 229–421. Vivien Eliot’s marginal suggestion “Our” (231) is in the same pencil as her closing comments on ts2a, and the resemblance suggests that she read Parts II and III on the same occasion (these being the only parts that she marked). Her annotations are written lightly in pencil, and often of uncertain intention. TSE’s sharp pencil deletion of “room” at 360 is quite different from Pound’s blunt pencil and bold marks at 245–46, 374–384 which are similar to marks and deletions at 244, 270–73, 282, 284, 326, 335–43. These suggest that several of TSE’s emendations were prompted by Pound’s deletions. When ringing a word or phrase, TSE characteristically began by underlining it before looping around; Pound began at top left and usually formed something more like a box.

  ts3b (WLFacs 38–47): blue carbon on five leaves, formerly folded in half. Annotated by Pound in pencil and ink prior to his annotation of the ribbon copy, ts3a. Grover Smith noted that this order of marking is borne out by Pound’s comment on the fourth page of ts3a, “vide other copy”. At some stage Pound incorporated into ts3b the third page only of ts3a (without removing the cognate page of ts3b), so forming a six-page working draft, which he renumbered 1–6. The cognate third pages, showing 317–56, are both heavily annotated. (WLFacs 30/31, 42/43)

  Although Gallup 1968 described ts3 as “probably a first draft”, TSE had written a previous draft of Part III, in manuscript, at Margate (see Commentary headnote, 1. COMPOSITION).

  [Poem I 53–61, 323–46]

  ts4 (WLFacs 62–69): violet typing of the long first draft of Part IV (475–567), derived from ms5 and annotated by Pound. Assigned to Jan 1922 by Rainey 200–201. Four leaves, formerly folded in quarters, of India paper of the kind Pound used when writing to James Sibley Watson 4 Jan 1923. Both this and ts5 were probably typed on Pound’s typewriter. Although Valerie Eliot states that this was “typed with the violet ribbon used by Pound”, the colour was probably from a carbon sheet not a ribbon, as suggested by a letter and article sent to Watson by Pound on 4 Mar 1922: the letter is typed in black, but he encloses a violet carbon (double-spaced) of his review of The New Therapy by Louis Berman, “knocked off · · · during the morning”, the top copy of which was sent to New Age (Berg). Like Pound’s article, ts4 is double spaced. On the verso of the last leaf is a note in TSE’s hand:

  1. Sunburn salusta lotion

  1. Smallest pot skinfood O.

  -------------------------------------------

  62 Ch.[amps] Elysees

  which confirms a connection with Paris.

  ts5 (WLFacs 82–89): violet typing of Part V, derived from ms6 and annotated by Pound. Typed on the same machine as ts4, likewise with double spacing, but on a different paper: two halves of a cut sheet of double foolscap (WLFacs 82/83 and 88/89) and fols. 1 and 3 of another sheet, folded (WLFacs 84/85 and 86/87). “In violet ink on the double foolscap used by Pound, it may have been typed by him, or by Eliot when he visited his friend in Paris on his way home from Lausanne, where this section was written” (Valerie Eliot). Kenner pointed out that Pound’s “mannerism of hitting the space-bar twice is nowhere visible” (Litz ed. 42). On the other hand, TSE typed page numbers (from “2”) in all the other typescripts. Assigned to Jan 1922 by Rainey 200–201.

  All of these drafts were sent by TSE to Quinn on 21 Sept 1922. The posting label has Quinn’s stamp of receipt dated 13 Jan 1923. Other materials sent to Quinn in the same sheaf as The Waste Land appear in the present edition among the “Uncollected Poems”.

  WLComposite is made up of seven groups of lines:

  1–130 from ts1

  131–228 from ts2a

  229–421 from ts3a

  422–428 from ms1

  429–474 from ms2

  475–567 from ms5

  568–678 from ms6

  2. NON-AUTHORIAL TYPESCRIPTS

  None of the drafts listed so far gives a fair copy of the entire poem. No collation of the earliest extant witnesses to the poem as an organised and edited whole has hitherto been published. Although not by TSE himself, the typescripts made for John Quinn in New York at the end of July 1922 (Q) and for James Sibley Watson in Paris in August (T and W) are copies following two distinct typescripts by TSE, both of which are now lost.

  [Poem I 53–61, 323–46]

  As well as being the first to show the entire poem essentially as it was published, the typescripts made for Quinn and for Watson are the first to show numerous words and phrases in their final form. Q is also the earliest with line numbers. These typescripts also contain the last appearance of other words and phrases, for they were not to occur in any of the subsequent witnesses, as the poem finally reached print. The phrase “Ionian white and gold” ([III] 265), for instance, had read “Corinthian white and gold” in a manuscript (ms1) but not in TSE’s known typescripts. “Corinthian” is also the reading in T and W, but all printed texts have “Ionian”. Both how “Corinthian” came to appear in T and W and how “Ionian” came to be printed are unknown.

  These three typescripts, Q, T and W, differ in small ways not only from the drafts and from any of the published texts but from each other. Even the two copies made for Watson—one at first glance a carbon of the other—turn out to be different. The Quinn copy (Q) is not consistently closer to the published texts than the Watson copies (T/W), but nor are the Watson copies consistently closer than the Quinn, which suggests that just as TSE sometimes worked on more than one copy of a Part of the poem in the draft stage, he continued to do so during the first half of 1922. At [I] 8, the draft (ts1) reads “Summer surprised us, coming over the Königssee”. The first appearance of “Starnbergersee”, the reading in all editions, is in T/W. Yet Q, which was typed by a professional stenographer who made only a very small number of slips, reads “Stainberger”. The typist of T/W, by contrast, does not appear to have been a professional, and a less than perfect grasp of English is suggested by “corps” (for “corpse”, [I] 71, in T); “Onguent” (for “Unguent”, [II] 88); “remenber” ([II] 122 and 124); “voilet” (for “violet”, [III] 220); “madoline” ([III] 261) and “canée” (for “canoe” [III] 295). Yet the unusual word “Starnbergersee” is correct in T/W. Perhaps the typist was more familiar with German, or possibly the copy followed by T/W was typed, whereas in the source for Q it was an unclear manuscript emendation, which would accord with TSE’s letter to Quinn, 19 July 1922: “I had wished to type it out fair, but I did not wish to delay it any longer. This will do for him to get on with, and I shall rush forward the notes to go at the end.”

  Other variants, too, tend to suggest that the source typescript for Q was not the most current when TSE sent it. For instance, the nightingale’s “Jug Jug” ([II] 103) was not enclosed in quotation marks in the draft or in Q, but appeared as “‘Jugjug’” in T/W and as “‘Jug Jug’” in printed editions. Likewise, “I think we met first in rats’ alley” ([II] 115) in ts2 was unchanged in Q, though revised in T/W to the text of printed editions: “I think we are in rats’ alley”. The spelling “fuoco” ([V] 427) seen in ms5 and ts6 had not yet been revised in Q to “foco”, but it was in T/W and in print. In T/W, the Sanskrit “Damyata” appears as “Demyata” in two places, fourteen lines apart ([V] 418, 432), making it likely that this spelling appeared in the typescript being copied, although it does not occur in any surviving draft or printed text.

  Q (Houghton): fair copy, (second?) carbon on 19 leaves (plus title leaf) of Cambrai Bond. Folded in thirds and sent by John Quinn’s office to Jeanne Robert Foster on 31 July 1922 (envelope preserved with “Elliot copy
to JR The Waste Land” vertically on the front). “Made in the office of John Quinn from typescript sent to him by Eliot” (Houghton library card). The Greek in the epigraph on the title-page is in black ink, as is the envelope. Quinn detailed his negotiations with Liveright, receipt of the poem, the stenographic copying of it, disappointment on reading it, and recommendation that as a book it would benefit from some additional poems in a 14-page letter to TSE of 28 July–1 Aug, for which see Egleston ed. He wrote that he had had the poem copied, and was having both TSE’s typing and the copy delivered to Liveright on 1 Aug. On 31 July he sent this copy, Q, to Jeanne Robert Foster who took it to be the original (no accompanying letter has been traced).

  [Poem I 53–61, 323–46]

  After a misunderstanding in March over money (see headnote to The Waste Land, 4. THE DIAL AND THE CRITERION), TSE’s negotiations with Scofield Thayer of the Dial had foundered, yet the journal’s principals remained keen to publish the poem. At the end of July, as impresario to both TSE and the Dial, Pound wrote to TSE on behalf of Thayer (the editor) and James Sibley Watson (the president). TSE replied on 28 July: “I will let you have a copy of The Waste Land for confidential use as soon as I can make one. Of the two available copies, one has gone to Quinn to present to Liveright on completion of the contract, and the other is the only one I possess. I infer from your remarks that Watson is at present in Paris. I have no objection to either his or Thayer’s seeing the manuscript.”

  Here, the term “the manuscript” may have referred to the fresh copy he was promising. On the other hand, a working draft would serve to maintain the interest of the Dial, and two letters from Watson to Thayer (12 and 16 Aug) suggest that by “the manuscript” TSE may instead have meant a copy that Pound still had in hand from his final sieving in January. A fortnight later, TSE did send the promised new typescript to Paris, but such was Watson’s hurry to despatch a copy to Thayer (then in Germany) that T/W may have been made in the interim from the January text. This might help to explain not only why Watson had had time to read the poem three times (as he wrote on 16 Aug) before sending it on, but why the source for T/W read “the walls | Of Magnus Martyr stood and stand and hold”, following the manuscript reading, which was superseded in Q and all printings by “the walls | Of Magnus Martyr hold ([III] 263–64).

  T/W: two versions of a 19-page typing made probably in Aug 1922 for Watson in Paris on a French typewriter (typed grave accent at 130). Both T and W are collated below, with the abbreviation “T/W” being used for variants on pages which are cognate. Faced with these two copies and a secondary carbon of W to correct, Watson thoroughly corrected the ribbon copy, T, to send to Thayer, but in W made only three corrections (on the first three pages), while in the secondary carbon he made none. Most of the leaves of W are carbons from T, but some leaves, and parts of leaves, are not cognate. Perhaps because the carbon paper slipped at the foot of the first leaf, [I] 19–24 were typed twice, with single spacing in T but double spacing in W. The third leaves ([I] 50–76) are not cognate, these lines having been typed twice, with “I had not thought death had undone so many.” ([I] 63) being omitted from T and supplied by hand. On the seventh leaf of each, the final “Good night” of Part II has been typed in error as its own line, but then on W it has been typed also in its correct place at the end of the previous line (without any deletion). The typist also had difficulty with the Rhine-maidens’ song. The two lines first appear at the foot of the eleventh leaf of each, with a first attempt at the line “Weialala leia” ([III] 277) being scratched out on both T and W and separately retyped over the top, the new alignment on W being wrong. At the song’s reappearance, the second line has been wrongly typed (copying the first), but only T has been corrected, by hand.

  [Poem I 53–61, 323–64]

  T (Dial/Thayer papers, Beinecke): ribbon copy, variously on paper watermarked Boar’s Head English Bond, Typewriting Parchment (second leaf) and notepaper of the Hotel Meurice, 228 Rue de Rivoli (seventh leaf). Double-spaced on a distinctive typewriter, with lower parts of characters not printing properly (especially in early leaves). Typed commas often appear as full stops. This typescript omits five lines at the start of Part III (caused by an eye-skip). On the first page, the last six lines ([I] 19–24) are single-spaced. Since the typing on each of these pages ends significantly short of the foot, the typist may have been copying another typescript page-for-page (for which, in Q, see textual note to [V] 345 = 591, 592 on the counting of lines).

  W (Watson papers, Berg): carbon on 19 leaves of lightweight wove typing paper, accompanied by a secondary carbon (formerly pinned to W and similarly folded in three).

  3. RELATION OF Q AND T/W TO THE PRINTED POEM

  Since the source of T/W may have been either earlier or later than Q, it is hard to know whether Q or T/W should come first in a full collation of these typescripts, though it is clear that neither is the direct source of the other. Also unknown is their relationship to the setting copy for each of the four original printings.

  On 25 June, TSE wrote to Quinn that he would send “as quickly as possible a copy of the poem merely for your own interest, and I shall send you later the complete typescript with the notes, in the form to be handed to the publisher”. In the event, on 19 July he sent only a typescript such as would “do for him to get on with” (the source typescript for Q), implying that a fair copy was still to come, and added that he would “rush forward the notes to go at the end”. In his long serial letter of 28 July–1 Aug (NYPL), Quinn wrote “I daresay you’ll send the notes direct to Liveright.” No typescript of the Notes is known, but since TSE must have sent one to New York and there is no further mention of them in the Quinn papers, he almost certainly did send it direct to the publisher in August for setting. Having told Quinn that he wanted to type a fair copy of the poem, and given the importance he attached to “the punctuation and the spacing” in what was then planned as the poem’s American debut, it is likely that he took the opportunity to send a fresh typescript of the entire poem.

  A fair copy sent direct to Liveright would have incorporated revisions which TSE had made since sending the source typescript for Q and would explain some of the textual differences between the Liveright edition and Q, including the absence of the line Vivien Eliot wanted excised, “The ivory men make company between us” ([II] 137a). The addition of the Notes almost certainly required a text with line-numbering, and this too gives an indication that the typescript that was the source for Q was not Liveright’s setting copy. For although Q is the earliest text with line numbers, they do not quite match those printed in Liveright’s edition, which counts

  From doors of mudcracked houses

  If there were water

  [Poem I 53–61, 323–46]

  as the single line [V] 345. TSE could not have written notes to the poem such as “Cf. Part III, l. 204” without giving line numbers, so the line numbering in the edition probably derives from such a typescript sent to Liveright but now lost. Neither the Dial nor the Criterion printing has line numbering, for they have no Notes. When the Hogarth Press edition was set in 1923, the Notes certainly and the text almost certainly derived from the Boni & Liveright edition, yet the poem has no line numbering.

  The relation of T/W to the printed editions is also distant and enigmatic. In these typescripts, the poem is without a title, and is set apart from the printed texts by variants such as “a pack of cards” (for “a wicked pack of cards”, [I] 46), “laquenaria” (for “laquearia”, [II] 92) and “reforms” (for “reforms and bursts”, [V] 372). However, as well as preparing the American book and journal publications of the poem, TSE was arranging its appearance in the first issue of the Criterion in October. During July or August, he therefore sent a typescript to the printer, Richard Cobden-Sanderson, and this evidently overlapped with the source typescript for T/W. Of the four original printings, only the Criterion shares with T/W “in the sunlight” (for “in sunlight”, [I] 10), “The chair” (for “The Chair”, [II] 77
), “Wherefrom” (for “From which”, [II] 80), “she says · · · she says” (for “she said · · · she said”, [II] 158–59) and “Upper Thames Street” (for “Lower Thames Street”, [III] 260). Like T/W but unlike Q and unlike the other original printings, the Criterion text has no epigraph.

  4. PRINTINGS

  The order of publication of the earliest printings of The Waste Land was probably: Criterion (Oct 1922 issue), Dial (Nov 1922 issue), Boni & Liveright (15 Dec 1922), Boni & Liveright 2nd imp. (early 1923), Hogarth Press (12 Sept 1923).

  TSE had the chance to bring the details of the British and American texts into line, since he had proofs from both Boni & Liveright (on 15 Sept he wrote to Pound “Liveright’s proof is excellent”) and from the Criterion (letters of 27 Sept and 3 Oct). His preferences, however, were not completely settled, and many small differences remained. To Gilbert Seldes of the Dial, 12 Nov 1922: “Liveright’s proof was on the whole very good indeed and I have no doubt that the appearance in The Dial will be equally good · · · Nov. no. just received. Poem admirably printed.” As well as confirming that TSE read proofs for Boni (as for Criterion and Hogarth), this indicates that he did not read any for Dial, which was apparently set either from the TSE’s ts or from proofs of Boni. No proofs of Boni or Criterion have been traced. The John Peale Bishop papers at Princeton contain a gathering from the setting of the Dial of Nov 1922 which includes The Waste Land, but it has no variants and is unlikely to be a proof.