University of Virginia Library

An extant typescript showing Conrad's revisions from April and May 1910 is held at the Philadelphia Free Library. The transmission of text from this revised typescript (TS) to Methuen's first edition has been discussed by David Leon Higdon (Higdon, 1991b). But because Higdon excludes the American texts from his analysis the fails to fully describe the material processes that produced the texts of Under Western Eyes. This essay rectifies this failure by providing a more comprehensive description of the textual transmission.

For more comprehensive descriptions of the extant typescript see Higdon, 1991a; Carabine, 1996; Osborne, 2000; and Osborne, 2002.

The similarities between the characters and events of Under Western Eyes and Conrad's family history have been comprehensively discussed (Carabine, 1996). Conrad's persistent medical problems and ongoing treatment also contributed to the breakdown.

For further discussion on this point see Osborne, 2002.

Robert Garnett was the critic Edward Garnett's brother. Edward Garnett had been Conrad's friend and confidante for many years and may have had some influence on Conrad's decision to shorten the text of Under Western Eyes in April 1910.

Karl and Davies note, "The size of each issue rather than the considerable size of its circulation" (Letters, 4, 372, n., 1).

Both serials began monthly instalments in December 1910 and concluded in October 1911.

Of the approximately 1600 substantive variations from TS found in the published texts, there are 361 agreements between all published texts, indicating the number of changes possibly made by Garnett. A further 28 differences were restored to the TS reading on galley proofs of the English Review. Furthermore, most of the accidentals that are found in the first English edition originate from the typescript "corrected" by Robert Garnett.

As discussed above, Conrad probably saw batches of typescript when they returned with galley proofs of English Review instalments. There is no evidence, however, that Conrad corrected the typescript as a whole.

One substantial variation is the absence of Parts in the American serial. The North American Review chose to signal chapters only, most likely an editorial decision.

Before the final chapter of the novel the largest unique variation between the TS and the American serial is the omission of the sentence "It grew more depressing as one came nearer" (TS 292).

The text of the last page of Under Western Eyes in the North American Review finishes with "THE END" centred on the very last line of page 656.

Examination of the accidentals reveals a story similar to that told by the substantive variants. For example, of the 568 accidentals in the third chapter of Part Three, 18% are unique to the American serial, 32% appear in all published texts and 12% see agreement only between the English texts. Such a pattern appears in most chapters, showing that Robert Garnett (or his typist) was probably responsible for much of the punctuation added to the text when the new typed copy of TS was made.

The most obvious agreement between the TS and the two American publications is found in the last sentence of the novel. The TS and the two American publications all end with "Peter Ivanovitch is a wonderful man", whereas both English publications read "inspired man". Conrad probably changed this on the proofs of the English Review in August or September 1911, but did not transcribe it to the set of proofs sent to Harper.

One of the most obvious examples is the frequent changes like "Yes. Something . . ." to "Yes; something" or "Yes! A . . ." to "Yes, a . . .". These unique North American Review readings are found in almost every chapter, demonstrating that Harper and Brothers drew their text from another source—the galley proofs of the English Review serialisation. The large number of accidentals unique to Harper and Brothers' text indicates active house styling, making it difficult to identify any direct relationship with TS. Nevertheless, regular agreement between the English Review, Harper and Brothers and Methuen suggests a common source: galley proofs of the English Review. The accidentals unique to each publication are probably a result of house styling.

Conrad reiterated this point to Austin Harrison in 1912, writing, "Generally I don't care a bit for serial publication. . . . But in the case of the E. R. my feeling is different, since as in the case of Western Eyes the text of the E. R. would be the final text of the book form" (Letters, 5, 444). As Higdon has shown, and I confirm below, this was not the case.

This letter is undated, but Karl and Davies place it after 29 April 1911 and before 10 May 1911 (Letters, 4, 436).

That is, the North American Review typically agrees with TS, but the texts of the English Review, Methuen and Harper and Brothers contain the same variant. Approximately 39% of substantive variation from TS falls into this pattern. All percentages referred to in this essay were calculated from results of computer collation produced by MacCASE at the Australian Scholarly Editions Centre.

As with the corrected pages for Methuen this batch contained text up to the end of the second chapter of Part Three.

Robert Cunninghame Grahame is the only "friend" who can be situated near Conrad at this time (Knowles, 1990, 81).

On 12 January 1911 Conrad discussed the English Review proofs in a letter to Edward Garnett, saying "Directly I get 3 instalments or so of the novel together I'll send them to you. My copy of proofs is uncorrected as yet" (Letters, 4, 407). There is no evidence to confirm that Garnett received this copy of proofs, nor why he was to receive them. Perhaps Conrad entertained the prospect of using some of the material excised in TS for book publication and had enlisted Garnett to assist. If this was Conrad's plan it did not occur and this set of galley proofs would have been available for use as setting copy in May 1911, requiring only transcription of changes made for the English Review.

Figures relating to the printing of Under Western Eyes have been supplied by the Center for Conrad Studies, Kent State University, from William R. Cagle's forthcoming bibliography of Conrad's works.

Harpers would give no publication date until final copy reached New York (Letters, 4, 467).

Higdon has closely analysed the substantive and accidental variation between TS, the English serial and the Methuen text. He makes three conclusions regarding the correction and revision: "Conrad was a parsimonious writer who much preferred to adapt and readjust a sentence rather than to discard it, or add more than a few words to it. Second, he effected no major structural rearrangements or character redefinitions in this stage but rather concentrated on sharpening existing phrases and clarifying, often highlighting, character relationships. Third, his hand was not always successful in improving his text but sometimes ensnared him in unidiomatic phrasing and flabby sentences" (Higdon, 1991b, 180). Higdon also points out that although the Methuen text "may be the text that Conrad wished finally to be established, the book has moved far away from Conrad's usual practice in accidentals", and concludes that the serial is more authoritative in relation to Conrad's punctuation" (Higdon, 1991b, 181).

Of course, the variation between the two book versions might be considered compositorial error, but the major point, that Conrad inserted the word in both texts, remains.

This agreement accounts for only 2% of the variation from TS. American spellings account for a number of these, leaving only a very small percentage of significant variation.

Examination of the accidentals supports this explanation. Although the two American texts frequently agree, the high percentage of unique readings in both texts makes it difficult to argue that they are directly related. Furthermore, because the pattern of variation that sees the text of the English Review agree with both first editions dominates throughout the novel, galley proofs of the English Review (marked up by Conrad to different degrees) are the most likely setting copy for all three.

Agreement between the two English publications accounts for 13% of the substantive variation from TS.

An example similar to this situation is found in earlier chapters. Conrad originally named his female terrorist Sophia Semenovna (an explicit reference to one of Dostoevsky's characters), but later changed the last name to Antonovna. Because Conrad did not correct this late change on several occasions in TS, Semenovna was probably transmitted to both serials. When English Review galleys were made this error was still incorporated, but missed by Conrad and the English Review editors, because Semenovna appears on six occasions only in the English Review serial. Intervention on galley proofs or later proofs by Conrad or an editor produced the "correct" readings in Harper and Brothers and Methuen.

These variants are concentrated in one gathering of the October instalment. Of the sixteen pages in the gathering marked 2 G, eight pages contain the opening of the third chapter of Part Fourth.

Conrad had another opportunity to revise when Under Western Eyes was being prepared for publication in a series of Collected Editions of the 1920s. However, David Higdon's analysis of the variants in those editions shows that Conrad seemed "to have given no more than a perfunctory nod to the text of Under Western Eyes, trusting in the skills and good will of the firms of F. N. Doubleday and W. W. Heinemann" (Higdon, 1986, 129). Optical collation has confirmed this view.