This article has been published in Notes & Queries, by Oxford University Press. Leonard, S. M. & Marland, R. (2025) Unnoted manuscript leaves of Oscar Wilde’s ‘The Fisherman and His Soul’, Notes & Queries, doi:10.1093/notesj/gjaf078
Illustration of the fisherman and mermaid in ‘The Fisherman and His Soul’ that appears in the first edition of A House of Pomegranates.
The 1895 auction of Oscar Wilde’s possessions resulted in the division and loss of many of his manuscripts. But even after the initial sale, some booksellers and auction houses exacerbated this fragmentation of Wilde’s manuscript corpus by breaking apart collections of letters or whole manuscripts and tucking them into rare editions in order to increase their prices. Martin Birnbaum, Oscar Wilde: Fragments and Memories (New York, 1914), 13.
Such seems to have been the fate of Wilde’s fair copy manuscript of ‘The Fisherman and his Soul’, the third fairytale printed in his collection The House of Pomegranates (1891), a story that is—appropriately enough—about division and reconnection.Oscar Wilde, The House of Pomegranates (London, 1891). There is a patchwork, incomplete draft at the Fondation Martin Bodmer, Geneva that has been placed into a single bound volume along with pages of the printed first edition. A note pencilled on a flyleaf in handwriting that is not Wilde’s explains its contents: ‘Containing the Original Manuscript of “The Fisherman and his Soul”, written on forty-one leaves of foolscap. The M S. is not quite complete’. This note suggests that the bound volume contains a single—though incomplete—draft, but the condition of the pages and numbering system that Wilde used on the individual leaves suggests that it is a composite document, a feature that Ian Small accounts for in the OET edition of The Complete Works of Oscar Wilde, Vol. 8: The Short Fiction (2017) by describing it as consisting of two parts: MBMS and MBMS2.Ian Small (ed.), The Complete Works of Oscar Wilde, Vol. 8: The Short Fiction (Oxford, 2017), xcvi–xcvii. MBMS contains thirty-three leaves which are sequentially numbered in Wilde’s hand as 4–12, 12A, 13–16, 16A, 16B, 17–19, 19A, 20, 23–29, 34B,Ian Small refers to this page as ‘32’ but also acknowledges that, ‘the number is not easily read and has perhaps been corrected to “34 B”’ (note on line 487). Upon close inspection, the handwriting not only better resembles ‘34 B’ but also follows the content of ClarkMS f. 34, as we will discuss below. and 51–54.Wherever possible, all designations of textual witnesses correspond with those in Small (2017). The numbering of these leaves implies that at least twenty-six leaves are missing from this portion of the manuscript (1–3, 21–22, 30–34A, 35–50). MBMS f. 54 ends with text that concludes a major section of the story. In the same volume there are leaves that continue the story with the subsequent section but are not numbered as part of the MBMS sequence. Instead they are numbered in Wilde’s hand as 1–5, 5A, 7, and 10. Small notes that this portion ‘might belong to a different draft of the story than the earlier folios’ and he therefore refers to it as MBMS2.Small (2017), n. 905. MBMS2 f. 10 concludes in the middle of a sentence, 283 words from the end of the tale as published in the first edition. This amount of missing words and the numbering of MBMS2 suggest that at least five pages are missing from this portion: three pages that would fill in gaps in numbering and at least two pages at the end (6, 8–9, 11–12).Since there are only 282 words missing from the end of the first edition and Wilde seems not to have added a good deal of material in the extant proof pages, we propose two concluding leaves (*11, *12) following f. 10. Wilde did not average much over 200 words per page in this draft and the final page would likely have contained his signature, which might take up more room. The numbering and renumbering with the additional use of A and B lettering suggests that MBMS and MBMS2 likely represent multiple copies or drafts that initially had been different drafts and were then combined by Wilde in a stage prior to the galley proofs. A partial set of galley proofs with Wilde’s corrections and annotations exists at the Morgan Library & Museum, New York, (hereafter MP) and corresponds to the first twenty-two pages of the sixty-five-page story in its first edition. These textual witnesses have been collated in the OET edition.Small (2017).
Two additional textual witnesses, unnoted in the OET, likely correspond to leaves missing from MBMS. They may improve our understanding of Wilde’s compositional process in his creation of this story, which is particularly important given that so few of the manuscripts relating to Wilde’s short fiction have survived.
The first additional textual witness is a pair of manuscript leaves sold at Christie’s auction house on 13 February 2025 (Barry Humphries: The Personal Collection; lot 144), referred to here as BHMS.Thank you to Merlin Holland for allowing the publication of unpublished material. Thank you to Christie’s auction house for supplying reference photos for these pages. These leaves, numbered 21 and 22, have been tipped in to a first edition copy of A House of Pomegranates.Barry Humphries: The Personal Collection, Live Auction 22992, 13 February 2025, Christie’s Lot 144. The textual content, paper quality, and numbering suggest that these leaves had originated from the same set of leaves as MBMS prior to their acquisition by the Fondation Martin Bodmer. The final words on MBMS f. 20 conclude a sentence that is at p. 83, line 24 of the first edition (line 349 of the OET edition). BHMS f. 21 begins with the very next sentence on that same line. Additionally, BHMS f. 22 concludes with the words ‘went weeping away over the marshes’, which ends a major section of the story on the last line of p. 85 of the first edition (line 386 of the OET edition). MBMS f. 23 picks up with the word ‘space’ which Wilde surrounds with a hand-drawn box before beginning the next section. This ‘space’ indication in MBMS corresponds to an illustrated spacing device of six pomegranates on p. 86 of the first edition and continues on that page of the edition with the rest of the content of f. 23. This textual continuity as well as the similarity in Wilde’s frequency and manner of corrections in both manuscripts mean that it is likely that these pages were originally part of the same draft.
The second additional textual witness unnoted in the OET edition is a set of four fair copy leaves held by the William Andrews Clark Memorial Library, Los Angeles, CA.Oscar Wilde, The Fisherman and His Soul, Clark Shelfmark, W6721M3 F537 [1891] Bound. The library obtained these pages in November 1950, but an earlier 1923 sale catalogue attests to the fact that these four pages were separated before this date.Valuable Books and Autograph Letters, ‘A Collection of Autograph Manuscripts by Oscar Wilde,’ Sotheby, Wilkinson & Hodge, London, 23–24 April 1923, Lot 537, 63. Thank you to Wolfgang Maier-Sigrist for his assistance in locating this sale catalogue. The Clark library obtained these leaves in November 1950 from British bookseller Winifred Myers, as a part of a collection of eight other manuscripts from the collection of Frederick E. Peters, accession number MS.1950.054. Thank you to Manuscript & Archives Librarian Rebecca Fenning Marschall for this information. These leaves, referred to here as ClarkMS, are numbered 31–34. Small notes in his monograph Oscar Wilde: Revalued (1993) that ‘The Clark Library holds a portion of the original autograph MS of “The Fisherman and His Soul”’ but neglects to refer to ClarkMS in the OET edition.Ian Small, Oscar Wilde: Revalued (Greensboro, NC, 1993), 107. Instead, Small notes that f. 29 of MBMS is ‘not continuous with the next folio in MBMS, which is numbered f. 32 (the number is not easily read and has perhaps been corrected to “34 B”).’Small (2017), n. 487. That the pages of ClarkMS are numbered 31–34 and do not repeat any of the content of MP or any of the other known textual witnesses lends additional weight to the reading of the corrected number of the MBMS folio as 34B. The pages of ClarkMS correspond with pp. 93–6 of the first edition. Like the leaves of BHMS, the text, paper quality, and frequency of annotations in ClarkMS seem to indicate that this set of leaves was also originally part of the same fair copy draft(s) as was MBMS. ClarkMS f. 31 begins with the words ‘his small slanting eyes’, which are at page 93, lines 16–17 of the first edition (lines 504–5 of the OET edition). MBMS does not include an f. 30 and no such leaf has been collated in any scholarly edition. However, there is evidence of its existence in a 1927 sale catalogue alongside lots of other Wilde manuscripts,Autograph ms, p. 30 in First Editions of Modern Authors. Collected by Fred A. Rosenstock, Denver, Colorado, Anderson Galleries, New York, Dec. 12, 1927, lot 698, p. 76. Thank you, again, to Wolfgang Maier-Sigrist for drawing our attention to these sale catalogues. and later in a 1933 sale catalogue this item is described as containing ‘the last 11 lines on page 92 and the first 16 lines on page 93. About 200 words in all.’A Catalogue of Autograph Letters and Manuscripts, James F. Drake, New York, no. 234, 1933, lot 311, p. 56. If this description is accurate, the missing f. 30 would sit between MBMS f. 29 and ClarkMS f. 31. The final page of ClarkMS, f. 34, ends mid-page with text that corresponds with the last line on page 85 of the first edition (line 564 of the OET edition) while MBMS f. 34B corresponds with the first line on page 86 of the first edition (line 565 of the OET edition). Unlike BHMS, none of the pages of MP correspond to the ClarkMS, making ClarkMS the only known textual witness of Wilde’s corrections in this portion of the story.
Because Small did not collate either BHMS or ClarkMS in the OET edition, we would like to offer the following annotations as a complement to the textual apparatus of that edition. All line numbers and lemmata preceding the right bracket refer to the OET edition. These annotations are formatted to align with the OET edition using the same sigla to indicate deletions (< >), insertions (/ \), overwriting ({ }), omitted punctuation (^), and unchanged text preceding punctuation (∼). For clarity, we have indicated both where folios begin and end.
Of greatest note are variants in the ClarkMS, which are not recorded elsewhere. For instance, though the typescript and final draft of the story describe a statue of a god wearing a copper belt, Wilde describes the belt as ‘silver’ in ClarkMS. Additionally, rather than the location being described simply as ‘the temple’, this version describes the location as a ‘rose-red house’ with ‘black marble steps’—a description that is used several lines earlier on line 488 of the OET edition, indicating the manner in which Wilde moved around and repurposed some of his descriptions in the story. Variants in BHMS mostly align with Wilde’s revisions in MP, but Wilde’s revisions in this stage such as his substitution of ‘well-knit’ for ‘beautiful’ and ‘Grecian’ for ‘Greek’ are notable since this stage of his revision process is not reflected in MP.
349 Go wherever] f. 21 of BHMS begins here.
350 thou wilt] <you> /thou\ will{t}
351 piteously,] ∼^
353 ground] ∼,
354 well-knit] <beautiful> /well-knit\
354 Grecian] <Greek> /grecian\
357 shadow, which was] shadow,
359 ‘If indeed thou must] ‘If <you> /thou\ must indeed
359 thee,] <you> /thee,\
360 forth] away
360 a heart] <your> /a\ heart
360 is cruel] is <very> cruel
360 thy heart] <your> /thy\ heart
360 take] bring
362 tossed] <shook> /tossed\
362–3 With what…heart?] <How could I love my love without my heart?> /With what should I love my love if I gave thee my heart?\
364 thy] <your> /thy\
364 is very] is /very\
366–8 ‘My heart…his Soul.] These words do not exist in f. 21 of BHMS.
370 the little] f. 21 of BHMS ends here. The next folio is continuous and is numbered f. 22.
373 crept] started
373 thrust the knife into his belt] the knife fell from his hand
375 Soul] s{S}oul
376 spake] spoke
377 wilt] <canst> /wilt\
379 will] shall
379 this place] thee{is} place
379 Soul] s{S}oul
381 young Fisherman,] /young\ fisherman:
382 and the] and the <little mermaid came to meet him>
384 around] round
384 kissed him on the mouth] kissed <them> him
385 them. And] them, and
386 over the marshes.] f. 22 ends here, and this ends BHMS. The next line occurs in f. 23 of MBMS after the indication of ‘Space.’
***
504 his small slanting eyes.] f. 31 of ClarkMS begins here.
506 forest,] ∼^
509 couch,] ∼^
509 by him] watch <with> /by\ him
510 feast,] ∼^
510 cried.] cried, ‘and none may see him.”
512 was my answer] I answered
513 and,] ∼^
513 hand,] ∼^
514 up, and led me into the temple] <up>, and led me <into the> up the steps of black marble into the rose-red house
516 ebony,] ∼^
519 copper] silver
521 priest, “Is] f. 31 of ClarkMS ends here. The next folio is continuous and is numbered f. 32.
523 or I will surely] <and> or I will
524 hand,] ∼^
525 besought me,] besought /me\^
526 will show] will surely show
527 with my breath] /with my breath\
527 hand,] ∼^
528 into the] into <a> /the\
530 was twice] was <of> /twice\
531 chrysolite,] ∼^
533 ware] had
533 its thick] its
534 selenites] <opals> selenites
535 god?”] God?”.
535 me,] ∼^
537 surely] <surely>
538 eyes,] ∼^
538 and they] and he
539 me,] ∼^
539 his servant,] f. 32 of ClarkMS ends here. The next folio is continuous and is numbered f. 33.
540 will] will surely
542 them] /t\hi{e}m Wilde also scores through the superscript dot of the ‘i’.
542 again,] ∼^
544 metal] ∼/,\ The comma is darker than the surrounding words and of a different shape than other commas, suggesting that it was inserted on a later pass.
544 altar of stone.] altar
545 said to the priest] said
546 me:] ∼,
547 Mirror] mirror
550 there,] ∼^
550 Mirror] mirror
552 them. And] Them, and
553 And I looked] Wilde begins a new paragraph before these words.
554 he had said to] he <told> /had said to\ f. 33 of ClarkMS ends here. The story continues on the next folio, which is numbered f. 34.
556 have I] I have
556 Mirror] mirror
557 thee again] thee
557 servant,] ∼^
559 none will be as wise as thou] no man s{w}ill be as wise as thou shalt be The overwritten ‘s’ was presumably intended to begin ‘shall’: Wilde repeatedly admitted to muddling his ‘shalls’ and ‘wills’.Merlin Holland and Rupert Hart-Davis (eds), The Complete Letters of Oscar Wilde (London, 2000), 473, 604, 1105, 1135.
564 the marshes.] f. 34 of ClarkMS ends here, slightly below the middle of the page. Small notes that MBMS picks up on line 565, which makes for a seamless transition between these two documents.