Genetic Joyce Studies - Issue 11 (Spring 2011)
A First Foray into the National Library of Ireland's Joyce Manuscripts: Bloomsday 2011
Luca Crispi


This essay and guide is the first part of several longer and more complete studies that I currently have underway.[1] It opens with a General Survey of Joyce's Ulysses Manuscripts, which is a non-technical, summary overview of the various ways in which Joyce used different kinds of manuscripts to write Ulysses.[2] The general survey is a preamble to the eleven sections that follow; these are more specialised, partial introductions to about half of the NLI's 'Joyce Papers 2002'.[3] Finally, I have compiled a Census of the Extant Ulysses Holograph Manuscripts that appears here as an appendix.

This opening foray into the NLI's collection of Joyce manuscripts includes his juvenilia, but the principal focus here is on the Ulysses draft manuscripts that Joyce wrote from 1917 to 1919, as well as on all of the newly discovered Ulysses notebooks that he compiled in 1917 and then in 1921. This initial critical examination of the documents is part of my continuing effort to catalogue as well as establish digital and print editions in accordance with copyright. This kind of analysis is fundamental to subsequent interpretive work on the genesis of Ulysses. The following manuscripts are dealt with here in varying degrees of comprehensiveness and detail:


Manuscript Name:

NLI MS Number:

Notes on Dante and Italian Vocabulary: 1897–8

MS 36,639/01

Early Commonplace Book: 1903–12

MS 36,639/02/A

Early Ulysses 'Subject Notebook for Drafts: 1917

MS 36,639/03

I. Later Ulysses Notebook: February–May 1921

MS 36,639/04

II. Later Ulysses Notebook: February–May 1921

MS 36,639/05/A

III. Later Ulysses Notebook: January–February 1921

MS 36,639/05/B

Earlier Typescript Schema for Ulysses: November 1921

MS 36,639/06

Earlier Partial Drafts of 'Proteus' and 'Sirens' and Notes: 1917

MSS 36,639/07/A–B

Complete Earlier Draft of 'Scylla and Charybdis': 1918

MSS 36,639/08/A–C

Partial Later Draft of 'Sirens': 1919

MS 36,639/09

Scenes and Fragmentary Tests for 'Cyclops': June 1919

MS 36,639/10


This first piece ends with Joyce's work on the NLI's proto-draft of 'Cyclops' in 1919 because it is one of several pivotal junctures in the development of Ulysses. This transitional period remains a significant turning point in the transformation of the novel, even though we now know that what Michael Groden designated as the 'middle stage' is actually composed of a series of much more complex and nuanced incremental phases.[4] Similarly, the later phases of Ulysses in process are also better understood as a series of gradated innovations rather than distinct breaks with what Joyce had already accomplished.

Therefore, other studies will take up what follows: the two draft levels of the 'Oxen of the Sun' manuscripts (NLI MSS 36,639/11/A–F); the two new NLI 'Circe' manuscripts (MS 36,639/12 and the so-called 'Quinn draft' of 'Circe' [MS 35,958]); the earlier proto-draft of 'Ithaca' (MS 36,639/13); the complete, earlier draft of 'Penelope' (MS 36,639/14); as well as the NLI's relatively disparate Finnegans Wake manuscripts (MSS 36,639/15–19 and the 'Joyce Papers 2006' [MS 41,818], among others). Each of these groupings of manuscripts require an individual in-depth analysis, in part due to their complex inter-relationships with other manuscripts in the Poetry Collection, University at Buffalo, the Rosenbach Museum and Library, the British Library, and elsewhere. I have already catalogued all the related Buffalo manuscripts, and I will treat these other, later NLI Joyce manuscripts in forthcoming publications.[5]

The focus on each of the manuscripts I discuss below is purposefully diverse and eclectic; there are many other topics that need to be analysed and so this first foray is meant to encourage further study and discussion. These particular analyses are concerned with a range of specific (historical-material-textual) genetic issues that have a bearing on our current understanding of Joyce's compositional practices, specifically regarding his work on Ulysses from 1917 to mid 1921–about which we know so much more from these new manuscripts. For now the goal is to determine what this kind of information can tell us about the state of Ulysses in 1917–19. The scholarly perspective here is necessarily more bibliographical than my current (critical-interpretive) genetic work: Becoming the Blooms: Joyce's Art of Storytelling in 'Ulysses'.

The various distinct instantiations of Ulysses are indeed fixed and at least temporarily 'finalised' at each particular juncture in manuscript and in print. This is most obviously the case when Ulysses appeared as a published work from March 1918 first in parts and then as editions as well as further printings. Nonetheless, for at least some of Joyce's readers, the book as a product is intrinsically embedded in the creative process; in fact, publication is only a momentarily distinguishable event in the process of genetic readings.[6] Therefore, to better understand what the text has come to mean as published, the genetic critical endeavour involves disentangling the text's many distinct versions along the way.

The descriptions and analyses below try to isolate the text of Ulysses at particular points in time during the work's seven-year evolution; this type of study permits readers to discriminate between those features of the work that Joyce had already set in place and those that were yet to come, without privileging the earlier ideas and texts over later ones. In general, by focusing on the creative process of a text's genesis, readers and critics can avoid essentialising characterisations of what Ulysses is and thereby avoid succumbing to the fallacy that the work we have in its published forms was something both timeless and necessary.

Although the approach is variously defined and practiced,[7] I maintain that a methodical understanding of the historical, material, and textual aspects of Joyce's manuscripts is a necessary foundation for genetic criticism to be an effective tool in the critical interpretation of his works. Genetic criticism is indeed founded on an assemblage of the verifiable information that happens to be documented in the necessarily partial material traces of a work's evolution. Some literary critics still persist in marginalising the 'scientific' aspect of genetic criticism at a time when historical and material textual approaches have assumed once again a central position in Joyce studies and more generally in modernist studies. Nonetheless, it is precisely these grounded approaches that are the basis of the interpretive insights and hypotheses that a genetic-historicist critic is able to put forward. Furthermore, while generally rigorous in terms of the evidence it marshals, genetic criticism encourages rather than hinders a multiplicity of perspectives and a broad range of readings and interpretations. In fact, one of its strengths as a form of critical interpretation lies in its ability to destabilise seemingly categorical statements based solely on the evidence of a singular, unitary, published text and the monolithic critical conclusions it encourages.

An example here is the best way to illustrate these more abstract points. It was as early as July 1918 that the first readers of Ulysses in The Little Review discovered some details about Leopold Bloom's father, Rudolph Virag, who changed his surname in Ireland. The information is given in an oblique way in 'Lotus Eaters'. Shortly after collecting Martha Clifford's letter at the Westland Row post office and a meandering conversation with C.P. McCoy, Bloom sees an advert for Leah with Mrs Bandmann Palmer and thinks:


Leah tonight: Mrs. Bandmann Palmer. Like to see her in that again. Poor papa! How he used to talk about Kate Bateman in that! Outside the Adelphi in London waited all the afternoon to get in. Year before I was born that was: sixtyfive. And Ristori in Vienna. What is this the right name is? By Mosenthal it is. Rachel, is it? No. The scene he was always talking about where the old blind Abraham recognises the voice and puts his fingers on his face.

Nathan's voice! His son's voice! I hear the voice of Nathan who left his father to die of grief and misery in my arms, who left the house of his father and left the God of his father.

Every word is so deep, Leopold.

Poor papa! Poor man! I'm glad I didn't go into the room to look at his face. That day! O dear! O dear! Ffoo! Well, perhaps it was the best for him.


This is how the scene read in The Little Review (see U 5.194–209). As far as the fictional biography of Leopold Bloom is concerned, the most important fact readers will glean from this scene is the year of Bloom's birth: 1866. As so often happens in Ulysses, Bloom tends to date the events in his own life by correlating them with other personal and historical events, though in this case the associations are partially erroneous. On the one hand, the popular American actress Millicent Bandmann-Palmer did in fact play in Leah the Forsaken at the Gaiety in Dublin the week of 16 June 1904. On the other hand, Bloom recalls his father waiting to see Kate Bateman, another famous actress, in Leah in the Adelphi theatre in London in 1865, the year before Bloom was born. Actually, Bateman appeared in Leah at the Adelphi in October 1863, a fact that Joyce was most likely aware of given all the other accurate historical details that structure this passage and Ulysses more generally. This is just one instance when Joyce was willing to alter historical details for the sake of the fictional histories of his characters.

Furthermore, once readers are able to piece together the story of Bloom's father, the play, which is based on Salomon Hermann Mosenthal's Deborah, has obvious resonances for the story of Rudolph Virag and Ulysses. Simply put: the play is about Leah, the leader of a band of wandering Jews, who have fled from religious persecution in Hungary. Along the way, they stop in an Austrian town where she falls in love with a Catholic boy named Rudolf, but Rudolf's father, Nathan, a converted Jew, eventually breaks off the romance between the two lovers. It seems that the important facts readers are to gather from this passage in 'Lotus Eaters' are: 1) this is the first time we are told the year of Leopold Bloom's birth; 2) the fact that Rudolph Virag was not yet in Dublin in 1865, the year before Bloom was born; and 3) the thematic link between Rudolph Virag Bloom's life and the play, Leah.

But, as usual, when Joyce revised and amplified this passage, he did so more than once. The first time was about three years later (in early 1921) when he added the following bit of historical detail on the typescript for the publication of Ulysses:


Leah tonight. Mrs Bandmann Palmer. Like to see her again in that. ^Hamlet she played last night. Male impersonator.^ Poor papa! [...] (see U 5.194–7)


As well as having the leading role in Leah, Bandmann-Palmer did in fact also play Hamlet at the Gaiety during the week of 16 June 1904, and so she was indeed a male impersonator. It seems that here, at least initially, Joyce was primarily concerned with adding further realistic historical details. But, then at the end of June 1921 (on the second setting of this text in proofs), he made yet another addition to this scene. Among other things, this subsequent addition further strengthens the work's structural parallels to Shakespeare's works, but it also more overtly introduces the sombre issue of 'death by misadventure' as it is called in the next episode (U 6.634):


Leah tonight. Mrs Bandmann Palmer. Like to see her again in that. Hamlet she played last night. Male impersonator. ^Perhaps he was a woman. Why Ophelia committed suicide.^ Poor papa! [...] (U 5.194–7)


Following only the logic of the associations in the published text, readers of Ulysses would be correct to presume that it was Bloom's thoughts about Ophelia's suicide that prompted him to think of his own father's death, but reading the work genetically (that is, as it evolved in a piecemeal manner over an extended period of time on several different manuscripts), we see that Joyce actually worked the other way round: it was the more basic description of Bloom's thoughts about Leah and Virag's suicide that prompted Joyce three years later to reinforce an already complex inter-textual parallel to Shakespeare's plays.

The overlaid network of Shakespearian allusions in the published text seems to encourage the view that they prompted Bloom's sombre thoughts about his father's suicide, a view that has been held by generations of readers and critics; however, by disentangling the creative prompts that generated these additional texts, this singular critical approach to the scene is destabilised and multiple other avenues of interpretation are thereby opened up. It is worth noting that all of the information about the Shakespearean additions to this scene was fully documented in Gabler's Ulysses: A Critical and Synoptic Edition since 1984 and was therefore available to any reader who wanted to examine the later evolution of the work.[8] This is a relatively late example of Joyce's more general tendency to continue to move away from 'character psychology' towards various forms of inter-textual subjectivity; which is one of the issues I will be investigating in Becoming the Blooms.

We can all agree that, if we let them, the various published versions of Joyce's works are more than enough to keep us busy for the rest of our lives. But, as the works make abundantly clear, Joyce incorporated issues about textual production and reception into the very fabric of his writing. Therefore, by concentrating the critical reading of the text on how it is documented in the material traces of its different manuscripts and printed states, we see what Ulysses actually was in process at various stages. This form of genetic criticism also explores what the work could have been and therefore historically re-contextualises the published work we all read. By following this evolution we simply have more versions of Ulysses to study and read; we inherit a wider, even more complex textual canvas for our critical interpretations as well as for our enjoyment as readers.


A General Survey of Joyce's Ulysses Manuscripts[9]


The text of the first edition of Ulysses in 1922 was notoriously flawed, and from then on there have been various efforts by Joyce, his printers, publishers, editors, collaborators, friends, and others to 'fix' or otherwise 'correct' the text. More recently, earlier manuscript evidence has come to play an increasingly decisive role in the textual debates surrounding Ulysses. Until 2000, only twenty-two of Joyce's Ulysses early (pre-faircopy) holograph manuscripts were known to survive, all but two of which were at the University at Buffalo.[10] That year the NLI acquired a new draft of 'Circe'.[11] Less than a year later another manuscript, this one for 'Eumaeus', came to light and was acquired by an anonymous private collector.[12] In 2002, the NLI acquired fifteen further new manuscripts for eight of the eighteen episodes of Ulysses, along with other manuscripts; these are the 'Joyce Papers 2002'. Assimilating all of this new evidence is an ongoing collaborative endeavour.

?? Joyce was not the kind of writer who paid much attention to the quality of his tools; his notebooks, paper, pencils, pens, and crayons were almost always of the most ordinary and inexpensive types. Like his notebooks, most of the copybooks in which he wrote the various drafts of Ulysses were also simple jotters and children's exercise books that he could easily acquire at any local stationer's shop in Trieste, Zurich, or Paris. He also used loose sheets of paper. For Ulysses, there are five broad categories of manuscript kinds: Notes, Drafts, Faircopies, Typescripts, and Proofs.

Roughly speaking, today there are over a hundred pages of notes for Ulysses (in the BL, the NLI, and in Buffalo); there are about thirty-nine holograph drafts (twenty in Buffalo, two at Cornell University, sixteen in the NLI, and at least one in a private collection); there are over eight hundred pages of the Rosenbach 'faircopy' manuscript in Philadelphia; and there are also over one thousand four hundred pages of typescript (almost all of the surviving typescript pages are at Buffalo), of which over one thousand pages are for the 'Circe' and 'Ithaca' episodes alone; as well as over five thousand pages of galley and page proofs for the first edition of Ulysses that are housed in the Houghton Library (Harvard University), Buffalo, the Firestone Library (Princeton University, New Jersey), and the Harry Ransom Humanities Research Center (University of Texas, Austin). There are further Ulysses manuscripts elsewhere.

Although the physical documents are spread out in Europe and across the United States, the Rosenbach manuscript was reproduced in colour facsimile in 1975 and subsequently all of the other then known Ulysses manuscripts were reproduced in black and white photo-facsimile in the JJA in 1977–8. This publishing and scholarly effort culminated in the production of the three-volume Ulysses: A Critical and Synoptic Edition of Ulysses, prepared by Hans Walter Gabler, Wolfhard Steppe and Claus Melchior in 1984. This monumental achievement has made the study of the evolution of Ulysses from the Rosenbach manuscript to Ulysses accessible to scholars who do not have access to the actual manuscripts or the published facsimiles.




Joyce built all of his works from words, phrases, and fragments that he culled from a myriad of printed sources and then slowly and carefully made his own.[13] As far we know, it seems that he first compiled his notes on slips of paper or simple pocket notebooks, usually in pencil; throughout his writing career Joyce rarely recorded his sources. Although there must have been many such documents, only one such notebook survives. Joyce compiled this Ulysses notebook in Zurich in early 1918. His English friend there at the time, the painter Frank Budgen, vividly captured the writer's methods:


In one of the richest pages of Ulysses Stephen, on the sea shore [in 'Proteus'], communing with himself and tentatively building with words, calls for his tablets. These should have been library slips, acquired by the impecunious and ingenious poet from the library counter [at the NLI]. On that occasion he had forgotten to provide himself with this convenient writing material, and was forced to use the fag-end of Mr. Deasy's letter. As far as concerns the need for tablets, the self-portrait was still like, only in Z?rich Joyce was never without and they were not library slips, but little writing blocks specially made for the waistcoat pocket. At intervals, alone or in conversation, seated or walking, one of these tablets was produced, and a word or two scribbled on it at lightning speed as ear or memory served his turn.[14]


The surviving notebooks document the artisan-like way in which Joyce gathered material and then assembled Ulysses. Joyce wrote these notes hastily and only for himself, which accounts for the appearance of his handwriting in the notebooks (as opposed to the meticulous way in which he wrote his manuscripts for his typists and other readers). Then he habitually crossed through the words he had incorporated into his writings with variously coloured crayons. He did not use one particular crayon colour for different episodes as some scholars had thought previously. Instead, whether working on Ulysses or Finnegans Wake, Joyce systematically would use one colour crayon to cross through the notes he used on a particular draft level (that is, during the same period of composition or revision of the text). Budgen's account of Joyce's note-taking practices continues:


No one knew how all this material was given place in the completed pattern of his work, but from time to time in Joyce's flat one caught glimpses of a few of those big orange-coloured envelopes that are one of the glories of Switzerland, and these I always took to be store-houses of building material. The method of making a multitude of criss-cross notes in pencil was a strange one for a man whose sight was never good. A necessary adjunct to the method was a huge oblong magnifying glass.[15]


As Joyce's work on Ulysses progressed, he copied in ink his earlier notes from various notebooks into new ones or on to large sheets of paper; these are what I call 'second order' note-repositories. He generally organized his notes under headings; usually these were the names of the episodes, but in at least one early instance he used character names or thematic subject headings. This 'subject' notebook is the earliest surviving compilation notebook. Some of the headings are obviously relevant, while others play more subtle roles in Ulysses. On the last page of the notebook, Joyce simply resorted to a catchall category he would use again in his notebooks: 'Words'. He incorporated these words in every episode of Ulysses in different drafts, typescript, and proofs from 1918 through 1921. He used at least a dozen of these notes word-for-word in 'Cyclops' and at least three each ins 'Telemachus' and 'Scylla and Charybdis'.

As far as we know based on the later note-repositories that survive, Joyce seems to have found that sorting his notes by episode headings was a more effective procedure for writing and revising Ulysses. In general, he was quite methodical in the way he organized these later 'episode' notebooks and notesheets. Joyce would write all of the headings first, underline them in crayon, and then fill the pages as he came across words and phrases elsewhere that he thought were appropriate to one or another episode, although it was not uncommon for him to ultimately use a note in a different episode. Joyce sorted his older notes in the main body of the page and left himself an ever-expanding left-hand margin because (just as with his drafts) he knew he would use that space to add yet more words and phrases. Joyce filled the margins of these note-repositories these words and phrases in any open space, and often in several different directions. There are many examples of incredibly cluttered and wonderfully colourful pages in Joyce's Ulysses notebooks and notesheets.




Joyce did not write Ulysses from its first to its last word consecutively, rather at least initially he wrote the episodes in a non-sequential order as the contours of the work evolved and expanded over many years.[16] Then, with the prospect of having his work published, he wrote and then rewrote the individual episodes of Ulysses successively from late 1917 through mid 1921, elaborating the text at each stage. It is unlikely that any of the Ulysses manuscripts that are known to survive are complete first drafts, although there are quite a few first draft sections in many of these manuscripts. Predictably enough, the earlier versions of an episode's manuscripts are the most chaotic because Joyce would often fill whatever open space he found on the page with more and more text. These earlier drafts are usually the messiest and most difficult to decipher, and seem also to have been so even for Joyce. He wanted to make sure he had incorporated all of the material he had written, including the marginal and left-hand page additions, so he would cross through methodically the text in coloured crayons as he re-wrote it in later versions. In fact, this practice is just an extension of Joyce's method when he used word and phrases from his notebooks and, unsurprisingly, it is a common procedure with other writers as well. In general, Joyce's cluttered earlier drafts were stages toward more legible drafts.

?? Hampered by poor eyesight, but prompted by his creative impulses, every stage of re-writing was an occasion to develop the text. As far as we know Joyce recopied all of the Ulysses manuscripts by hand more than once before they were given to a typist. From his earliest works onwards, from Stephen Hero to Finnegans Wake, Joyce developed certain writing habits that he used throughout his writing career. An example of this is that he usually filled the right-hand page of a copybook first, leaving himself a wide and expanding left margin. He then filled the margins as well as the left-hand page with further revisions and additions, returning to the same manuscript again and again to add a word or two in pencil, as well as many phrases and sentences and even paragraphs in ink, between the lines, in the margins, and then in any available space.

?? His first compulsion was to transfer all of the text he had already written on an earlier draft to the next version, but as the author he was not constrained to act simply as a scribe when he copied out the older versions of the text. In fact, Joyce regularly revised the text as part of the process of rewriting the earlier version; this makes attempts to 'fix' the text particularly problematic for editors. These later drafts are more uniform: the margins are more fixed; and the additions are generally less numerous, but even these 'later' drafts are just the next stage in the creative process. At every occasion, Joyce added further words, sentences, paragraphs, whole sections, and even episodes to Ulysses. For example, Joyce only conceived the pivotal episode, 'Wandering Rocks' (at least as we know it) at the start of 1919. Quite ill at the time but with pressing publication deadlines, he dictated (probably from previously written but as yet unused fragments) the first complete draft of the episode to Frank Budgen. Unlike other writers (Beckett, for example), Joyce rarely deleted anything that he had already written.




From September 1917 to March 1921, Joyce quite systematically re-wrote his drafts yet again, sequentially episode by episode, but now he was doing so in his most legible hand. The majority of these 'faircopy' (Rosenbach) manuscripts were intended to provide a more readable copy of the text for the typists, although Joyce still made further alterations to these 'clean' copies. It is important to note that (due to various exigencies) the extant faircopy manuscript of Ulysses is a mixed document. For some of the episodes, the faircopies were certainly used to produce the extant typescripts,[17] and these in turn were used to set up The Little Review instalments of Ulysses.[18] On the other hand, for other episodes the Rosenbach manuscript versions were clearly not used to produce the typescript.[19] Finally, only some individual pages (or sections) of the Rosenbach manuscript were used to produce the typescripts of 'Nausicaa' and 'Oxen of the Sun': Joyce presumably recopied the other pages because they had become too messy with additions and changes to be suitable for sale. Therefore, the Rosenbach manuscript has been central to the debates concerning the 'critically edited' texts of Ulysses since 1984.

??????????? Joyce's 'faircopy manuscript' of Ulysses comprises over eight hundred leaves. Almost seven hundred are loose sheets, the rest are in two notebooks (for 'Ithaca' and 'Penelope') that are similar to Joyce's draft manuscripts. The two new NLI manuscripts for those episodes reveal that the 'faircopy' versions of these final episodes are much more complicated and problematic than was previously thought, although the Rosenbach manuscript versions of these episodes were also used to produce the typescripts.

?? Another reason Joyce re-copied his manuscript in a clear and legible hand, in this case on relatively expensive paper, was because he was selling it piecemeal to John Quinn. Modern manuscript collectors had always prized these relatively uniform and more traditional 'holograph manuscripts' (that is, manuscripts that are 'hand-written' by the author). Quinn, the New York, Irish-American lawyer who had unsuccessfully defended the editors of The Little Review in court, was also a well known patron and collector of the arts, specifically of modernist literature and painting. He had also bought Joyce's holograph manuscript of Exiles in March 1917, and then he wrote a laudatory review of A Portrait for Vanity Fair. In 1923–4 Quinn put his vast collection of rare books, manuscripts, and art for sale at auction. A.S.W. Rosenbach, one of the most influential manuscript and book dealers of his era, acquired Quinn's Ulysses manuscript, and it is now at the Rosenbach Museum & Library in Philadelphia, Pennsylvania.




At the end of 1917, the prospect that Ulysses would begin to appear serially in The Little Review in the US and The Egoist in the UK prompted Joyce to have his manuscripts typed. He gave the typists precise instructions to leave a wide left-hand (and almost no right-hand) margin on the page because he knew he would use that space for corrections and further additions, just as he usually did when he wrote by hand. Most of the typescripts of the first fourteen episodes that appeared serially in magazines were typed in at least three copies. Joyce similarly corrected and revised at least two of these copies at the same time, but he only cursorily revised a third, and we think this was usually the copy he retained. He would then send at least one copy to Ezra Pound, who passed it on to The Little Review for serial publication. Joyce expected to be able to further revise the published The Little Review text for the printing of Ulysses, but this sensible plan became untenable for various reasons.

Only five instalments of Ulysses appeared in The Egoist from January-February through December 1919: 'Nestor', 'Proteus', 'Hades', and a portion of 'Wandering Rocks'. On the other hand, Ulysses appeared in twenty-three issues of The Little Review from March 1918 to September-December 1920, but its editorial policy of 'Making No Compromise with Public Taste' caused the magazine and Ulysses to be censored and then banned in the United States. The Little Review did manage to publish 'Telemachus' to the first section of 'Oxen of the Sun', although four issues were seized by the US authorities: the January 1919 issue, with 'Lestrygonians'; the April-May issue, with the second instalment of 'Scylla & Charybdis'; the January 1920 issue, with the middle portion of 'Cyclops'; and, finally, the July-August issue, with the concluding portion of 'Nausicaa'. In December 1920, The Little Review was suspended. For some reason only a few pages of the typescripts for the first three episodes are known to have survived, even though they too were used to set up the proofs for Ulysses, just like the typescripts for all the later episodes.

Joyce wrote all of his works by hand. His friend, Frank Budgen, recounts the following scene:


In leaving the caf? I asked Joyce how long he had been working on Ulysses.

'About five years,' he said. 'But in a sense all my life.'

'Some of your contemporaries,' I said, 'think two books a year an average output.'

'Yes,' said Joyce. 'But how do they do it? They talk them into a typewriter. I feel quite capable of doing that if I wanted to do it. But what's the use? It isn't worth doing.'[20]


Although later in life Joyce (reluctantly and unsuccessfully) tried to learn to use a typewriter, this technique of writing did not suit his creative methods. Therefore, he relied on an odd assortment of typists, often friends, as well as friends of friends, only some of whom even owned their own typewriters. Beginning in late 1917, the typescripts for the earlier episodes were prepared by one of Joyce's friends, the 'English Players' actor, Claud Sykes. Since Joyce was staying in Locarno, as Sykes prepared the typescript in Zurich, Joyce's obsessive inclination to revise and alter the text prompted him to write several postcards to Sykes (see, for example, LI 108–9) with instructions for changes to the text as it was being typed.

Many factors contributed to the problem of getting Ulysses into print accurately as Joyce wrote it (or wanted it written). Not only was his handwriting often difficult to read even on the faircopy (especially for amateur typists), but Joyce also continued to make numerous and substantial changes to the text on the typescripts as they were returned to him. He kept at least one copy of the early typescripts, and he further revised these same 1917–19 typescript copies in 1921 in final preparation for the book's publication.

It was only during the final stages of preparing Ulysses to appear in book form that Sylvia Beach and Joyce began using secretaries or other professional typists. For different reasons, the process of preparing each of the typescripts for the last episodes of Ulysses was unique. Getting 'Circe' typed presented even more problems than Joyce had faced while writing the episode. Four typists refused to undertake the work, some because they were unwilling to grapple with the task and others because they objected to the episode's content; in fact, in one instance, a typist's husband read Joyce's manuscript and threw a portion of it in the fire. At this point, a troupe of typists was finally recruited to get this typescript ready.

?? The typescript of the 'Eumaeus' episode was odd because, while the professional typist did a relatively accurate job of reading Joyce's manuscript, he or she tried to 'clarify' the episode's convoluted grammar by adding around six hundred commas, which Joyce had to methodically remove. This typist was also prudish to the point of leaving blank spaces where s/he disapproved of certain words. Not only did Joyce make his usual rounds of corrections and additions to this typescript, but he was also compelled to fill in such words as 'shite' and 'bloody' on the typescript to prepare it for the prospective printer of Ulysses.

Another of Joyce's friends, the American author and publisher, Robert McAlmon, typed 'Penelope' in Paris in mid August 1921. According to an anecdote Richard Ellmann recounts in his biography, the 'manuscript was so complicated and Joyce's insertions so numerous that occasionally McAlmon got some of Molly's thoughts out of place; he told himself it didn't much matter in what order her unsystematic mind took them up' (JJII 514). Actually, the few changes McAlmon made were minor and the care with which Joyce constructed the episode is well documented in the manuscripts.

On 30 October 1921, Joyce announced that he had finished writing 'Ithaca' and so the composition of Ulysses was complete (see LIII 51). Now, with everyone anxious to get the book published on 2 February 1922, Joyce's fortieth birthday, Beach and Joyce used two different typists, each with their own typewriters, to prepare the episode for the printer. Not only were parts 'Ithaca' re-typed four separate times, but Joyce also revised each version, adding sixteen percent more text from notes to the typescript, before he sent it to the printer a full month later. With the later 'Ithaca' typescripts, Joyce would sometimes use the back of a preceding page for yet more additions to a particular typescript page. He then tagged the additions and drew lines to indicate where they belonged, just as he would with his own hand-written manuscripts. Joyce then had some of the earlier, heavily revised typescripts re-typed, often several times.




A publisher usually supplies the printer with a complete, final working document (either an author's handwritten manuscript or preferably a clean, professionally prepared typescript) from which the process of setting the work in print begins.[21] Joyce's manual artistry was matched by his printers' artisanry preparing the book: Ulysses was all set, gathered, and bound by hand. The process of getting a book printed is usually comprised of two basic stages, the first of which is setting the text in 'galleys' (these are long metal trays of typeset text from which proof sheets are pulled). With Ulysses this initial phase of proofs is more accurately understood as a setting of text as 'galleys in page' because, although the text is set continuously down the sheet, it is already separated in page-length blocks. The proof sheets pulled from this setting are known in French as 'placards' and for Ulysses they usually comprise eight pages of text, set in four vertical columns of two pages each, that are only printed on one side of the sheet, on inexpensive, often pulp paper. These galley proofs are returned to the author, who is supposed to correct any typesetting errors and make what relatively few changes are considered necessary. On the other hand, Joyce made many significant alterations to the text on the placards of Ulysses from how it had appeared in print in the Little Review and on the typescripts Joyce submitted to the printer; the most obvious examples are the 'crossheads' to 'Aeolus' that Joyce only added as he revised the first setting of proofs.

Usually the author returns several sets of corrected galleys to the printer, who then sets up the 'gatherings' of page proofs for the final printing of the book. For Ulysses, these gatherings were printed as sixteen non-consecutive pages; that is, these sheets were printed with eight pages on each side that could be folded in such a way that the pages become sequential in the published book. Then, the author checks these proof sheets again. Finally, the author indicates what corrections still need to be made, or else the author and the publisher sign the page proofs as ready to be printed as the published book. Based on the contract between the publisher, the printer, and sometimes the author, a certain number of further proofs are provided until the proofs are all 'signed off'. Then, in general, the printed gatherings are assembled, the covers are attached, and so the book is ready.

Very little about the production of Ulysses was straightforward. No established English-language publisher was willing to take the risk of publishing Ulysses in book form after the problems The Egoist had faced finding a printer willing to set the work in the UK and after the editors of The Little Review had been fined for publishing obscenity in the US. With little prospect of seeing Ulysses appear, Joyce arrived in Paris from Trieste on 8 July 1920. Soon thereafter, Joyce met Sylvia Beach in her lending library and bookshop, Shakespeare and Company. Beach's own account of the events is quite colourful:


All hope of publication in the English-speaking countries, at least for a long time to come, was gone. And here in my little bookshop sat James Joyce, sighing deeply.

It occurred to me that something might be done, and I asked: 'Would you let Shakespeare and Company have the honor of bringing out your Ulysses?'

He accepted my offer immediately and joyfully. I thought it rash of him to entrust his great Ulysses to such a funny little publisher. But he seemed delighted, and so was I.[22]


In mid April 1921, Beach and Maurice Darantiere, Master Printer, Dijon, signed a contract to print and publish Ulysses, but the book was far from done. Joyce still had to finish writing the last two episodes and these would then have to be typed as well. Furthermore, Joyce would continue to make corrections, revisions, and additions to the earlier episodes on the typescripts that had already been made starting in 1917.

Joyce was supposed to make all of his changes on the typescripts before the printer began the process of typesetting. Nonetheless, the printer began setting the galley proofs of the first five episodes of Ulysses, 'Telemachus' through 'Lotus Eaters', between 11 and 17 June 1921. Darantiere assumed that the second set of corrected and revised galley proofs Joyce had returned would suffice, so the printers moved on to the next stage of setting the text in page proofs as gatherings, which they sent to Joyce and Beach for their approval. When Darantiere realised the amount and kinds of changes Joyce had requested again on these proofs, the printers took the unexpected and costly step of reverting back to galley proofs in late August 1921.

Although the contract stipulated that the printer would prepare up to five sets of proofs for the author to correct any typesetting errors, Joyce needed from five to eleven sets of proofs to accommodate his changes and additions. No one, including Joyce, anticipated the amount of revisions as well as substantial additions he would make on the proofs. In all, Ulysses grew approximately one-third longer on the proofs and, shockingly for Beach,[23] all of these changes on proofs accounted for almost a quarter of the entire printing costs of the first edition. Obviously, without the active cooperation of Joyce's publisher and printer, we would not have Ulysses.


[back to top]


Notes on Dante and Italian Vocabulary: 1897–8

(MS 36,639/01)


Joyce compiled these notes in his last year at Belvedere College. They are transcriptions from and related commentary and annotations on 'The Inferno' of Dante's Divina Commedia, as well as seemingly unrelated notes on Italian vocabulary in Italian and English. These pages are the earliest extant record of Joyce's student reading notes and, as far as we know, he did not use them directly in his writings. His earliest literary endeavour is 'Trust Not Appearances' (probably written in 1896; Cornell MS 1).[24]

Unlike his later practice, here Joyce also noted the author's name, 'Dante' on p. [1r], possibly after he had already started taking the notes on some of the later pages. In 2004, Dirk Van Hulle was the first scholar to ascertain that the text and notes are from Eugenio Camerini's La Divina Commedia di Dante Alighieri: con note tratte dai migliori commenti (Milano: Edoardo Sonzogno, 1884).[25]


[back to top]


Early Commonplace Book: 1903–12

(MS 36,639/02/A)


I prepared a more comprehensive commentary on this manuscript that was published in GJS Issue 9 (Spring 2009).[26] Then, at the start of 2010, Frank Callanan discovered the precise sources of the two lists of books Joyce transcribed on pp. [16], [17], and then p. [30] here; see his 'James Joyce and the United Irishman: Paris 1902–3' in the Dublin James Joyce Journal 3 (2010), edited by Luca Crispi and Anne Fogarty (UCD James Joyce Research Centre in association with the National Library of Ireland), pp. 51–103.


[back to top]



Early Ulysses 'Subject' Notebook for Drafts: 1917

(MS 36,639/03)


This manuscript is the earliest extant notebook Joyce prepared specifically to write Ulysses. The printed label on the front cover is virtually identical to the label on Buffalo MS V.A.3 (in which Joyce wrote the later 'Proteus' draft),[27] but notably the label on the Buffalo draft lists the stationer that produced both these copybooks: 'Eredi fu D. Pellanda – Locarno'. Joyce was in Locarno, Switzerland, from 12 October 1917 to January 1918. This external evidence and the source material for these notes that Wim Van Mierlo has identified make it clear that Joyce began compiling this notebook no earlier than mid October 1917.[28] But, as was his usual practice with this kind of notebook, Joyce almost certainly had already gathered some or most of the individual notes here beforehand.

It seems that throughout his career Joyce most often began making use of the notes just after he had started compiling the notebook and this is clearly the case here as well. He started to use some of these notes to write early drafts of Ulysses in mid October 1917. On the other hand, he also regularly returned to a notebook (even many years later), sometimes for further entries for drafts but also to transfer them to other note-repositories; therefore, many of the notes here definitely entered the text of Ulysses via other note-repositories (possibly various kinds of notebooks and notesheets), only some of which are known to be extant.

Although there are texts and notes that Joyce used to write Ulysses that predate this notebook, this manuscript is the earliest extant document solely devoted to his work on Ulysses. For example, the so-called 'Alphabetical Notebook' (Cornell MS 25) is an even earlier notebook, which Joyce compiled in 1910 in Trieste to write A Portrait, but he then used other notes from it to write several early drafts of Ulysses as well.[29] There is also another comparatively early notebook that Joyce used to write Ulysses, Buffalo MS V.A.2.a,[30] which he compiled in 1918 in Zurich.[31] This notebook is unlike all the other extant Ulysses notebooks in that it is a 'first-order' notebook; that is, Joyce compiled it directly from the various sources he was reading.[32] As such, this early notebook is different in kind from all the other extant Ulysses notebooks that are compilations of diverse notes, which Joyce re-sorted in new constellations in these extant notebooks.[33] Joyce then began compiling the first part of a further grouping of notes, the so-called BL Ulysses Notesheets (ADD MSS 49975, fs. 6–29). He began compiling some of these sheets in June 1919, but continued to compile further BL notesheets into 1921.[34] Furthermore, there are three later Ulysses notebooks at the NLI (MSS 36,639/04, 5A, and 5B, see below) that Joyce compiled from January to May 1921. Finally, there is one other later Ulysses notebook at Buffalo (MS V.A.2.b [V.A.2]); it is possibly the last extant Ulysses notebook and it too was compiled in 1921.[35] Joyce used virtually all of these later note-repositories in the final stages of writing and revising Ulysses in manuscript, typescript, and proofs throughout 1921.

Joyce had written drafts of some episodes of Ulysses (certainly parts of 'Telemachus', 'Nestor', 'Hades', 'Scylla and Charybdis', and probably some parts of other episodes as well) prior to 1917, but no early manuscripts for any of those episodes are known to be extant.[36] At that stage in 1917, these drafts may simply have been various arrangements of fragmentary texts that may have been quite different from the versions of the episodes as they evolved in the extant drafts.[37]

Here Joyce compiled the notes under nineteen subject headings that range from the names of some of the principal characters (such as 'Simon', 'Stephen', and 'Leopold') and themes of the work ('Blind', 'Art', 'Jesus', and 'Homer') to more abstract headings ('? ? ?', 'Choses vues', and 'Names and Places'). The notes were compiled from a variety of printed sources, some of which have been analyzed in detail by Van Mierlo, though others still need to be determined. The physical appearance of the notes here resemble what we find in most of the later Ulysses notebooks: the handwriting is relatively small, the notes are in series, usually separated by punctuation, and virtually all are in black ink. Along with their material disposition, the wide range of sources for these notes indicates that Joyce compiled the individual notes elsewhere (this is what is known as a 'second-order' notebook), presumably in 'first-order' notebooks and/or loose sheets, before he organized them under the headings here.

It is the only known example of Joyce using what I have called 'subject' headings for Ulysses (as opposed to the Homeric episode titles that head the BL Notesheets and almost all of the pages of the later notebooks, all of which are tellingly second-order note-repositories). It is almost certain that there were also other contemporaneous Ulysses notebooks that Joyce relied on to write and revise drafts during this period. Since they have not survived, we have no information about whether they were organized under headings at all and if so what kind they may have been. This notebook is also unusual because it contains relatively few notes, which I would argue indicates that Joyce did not find this method of organizing notes under this kind of headings particularly useful.

Joyce compiled this 'Subject' notebook at a crucial juncture in the development of Ulysses, when Leopold Bloom's character and the plot of Ulysses were still at a relatively early stage of development;[38] this may explain the expedient of using topical headings rather than episode names, some of which Joyce had not yet even conceived. He probably wrote out most of the subject headings before he began organizing the notes. Not only the kind but also the arrangement of the headings in this notebook differs from those in the later Ulysses notebooks. Since the headings appear on both the recto pages (on pp. [2]–[4], [6], and [10]–[13]) as well as on some of the versos, this indicates that Joyce was not following a regular method of organising this notebook, which is unusual among the extant notebooks.[39]

The most notable instance where the headings appear on both the recto and the verso are on pp. [8v]–[9r]; they are headed 'Irish' and 'Jews'. This particular juxtaposition of notes is not coincidental since the headings occupy the central pages of the notebook and the topics are obviously thematically linked in Ulysses. Joyce also wrote distinct headings on both the recto and verso pages on pp. [4v]–[5r] and [13v]–[15r].[40] Another oddity with these subject headings is that although both pp. [14v] and [15r] are blank, the final page (a verso) is headed 'Words'. This seeming exception is easily explained by the fact that (like 'Eventuali') it is a typical catchall heading Joyce regularly used; therefore, he simply put this grouping in a very convenient place.[41]

As he used entries from a notebook, Joyce almost always crossed them through in a coloured crayon. Only some of the crossed-through notes that he used directly from this notebook have been located in the extant contemporaneous drafts as far as we know; these are the first notes Joyce used from this notebook. On the other hand, many other entries have been located in the BL Notesheets and the later Ulysses notebooks (under a variety of Homeric episode headings, which is not unusual), though some notes may have been transferred to those BL Notesheets from yet another intermediary note-repository (or repositories) that is now missing.

It seems likely that Joyce compiled the entries in the notebook over a relatively short period of time and then began using some of them to write and revise drafts almost immediately. Furthermore, over several years Joyce used entries from this notebook to write or revise every episode of Ulysses. The first confirmable Joyce's use of the notebook was to write some of the fragments of the 'Proteus' proto-draft (NLI MS 36,639/07/A, see below). Joyce wrote that manuscript from early to mid 1917 and it is the earliest surviving Ulysses draft, although it is likely that actually he used the notes to write those fragmentary texts on a preceding, now missing document (or documents). The notes can be found on pp. [1r], [3r], [5r] (in this order in the 'Proteus' manuscript) from pp. [15r] ('Weininger'), [8v] ('Irish'), as well as the final page of the notebook, [15v] ('Words'), which indicates that the entire notebook had been compiled before Joyce wrote that draft.

Several notes from this notebook then appear in the next surviving draft of 'Proteus' (Buffalo MS V.A.3) that Joyce completed in the fall of 1917. The earliest surviving 'Sirens' draft (NLI MS 36,639/07/B; notably, this draft is also part of the same copybook as the earliest 'Proteus' manuscript, see below) is the next extant manuscript in which notes from this notebook have been located. These notes can be found on pp. [7r], [10r] and [13r], [6r] and [5r], as well as [7v] and [7r] of that manuscript and are from pp. [1r] ('Simon'), [2r] ('Leopold'), [4r] ('Recipes') and [15v] ('Words') in this notebook, although again Joyce may have used the notes from this notebook in a document that preceded that 'Sirens' manuscript.

Entries from this notebook are also found in all the episodes for which the Rosenbach faircopy manuscripts are the earliest surviving drafts (that is, 'Telemachus', 'Nestor', and 'Calypso' through 'Lestrygonians') as well as the earliest surviving draft of 'Scylla and Charybdis' (NLI MS 36,639/08/A, see below), all of which Joyce had written by mid 1918. Entries from this notebook are also found on the Rosenbach manuscript of 'Wandering Rocks', the earlier 'Cyclops' manuscript (Buffalo MS V.A.8), and the later 'Sirens' manuscript (NLI MS 36,639/09, see below); all of which were written by August 1919. It is possible that Joyce used these notes directly from this notebook as their source to write these drafts as well as earlier versions of the extant drafts.

It seems that thereafter Joyce returned to the notebook primarily to disperse some of its entries into other note-repositories; specifically, they can be found on BL Notesheets 'Cyclops' 8 and 10, particularly the latter. Interestingly, Joyce also transferred notes from Buffalo MS V.A.2.a (as well as from the early Ulysses notebook that only survives as transcribed by an amanuensis in Buffalo Finnegans Wake MS V.C.16) to many of the BL 'Cyclops' Notesheets as well. Since those notesheets were all used to write and revise both levels of the extant 'Cyclops' drafts (see below), the transfer of notes in these cases must have occurred before June 1919 while Joyce was in Zurich. Further notes, though only possibly taken directly from this notebook, were also transferred from here to BL 'Oxen' Notesheet 6 by the start of 1920 in Trieste. Later, further notes were transferred to BL 'Ithaca' 12 and 'Circe' 3 Notesheets, as well as NLI MS 36,639/05/B, p. [6r], but presumably these notes passed through one or more intermediary note-repositories before ending up in that notebook.

Finally, in a further unusual turn in the afterlife of this early Ulysses notebook, in early 1935 while he was writing 'Work in Progress'/Finnegans Wake, Joyce passed it on (along with NLI MSS 36,639/04, 5A, and 5B) to his amanuensis, Mme France Raphael, who transcribed it (as Buffalo MS VI.C.7.255–69).[42]


[back to top]



I. Later Ulysses Notebook: February–May 1921

For Drafts, Typescripts, and Proofs (MS 36,639/04)[43]


Joyce compiled the notes in this manuscript at the start of 1921 (sometime between February and May, probably earlier rather than later) and used it continuously in tandem with his other, then current note-repositories until January 1922, just before Ulysses was finally published. He first used this notebook to continue writing the last episodes of Ulysses, then to revise the typescripts of the earlier episodes for the printer, and subsequently to revise the various settings of proofs of virtually all the episodes. It is one of the later 'second-order' Ulysses notebooks (that is, Joyce compiled and sorted it from previously gathered notes), along with the extant BL Notesheets (for 'Cyclops' through 'Penelope' only), the NLI MSS 36,639/05/A and 5/B, and the Buffalo MS V.A.2.b.

For now this is still a preliminary assessment of Joyce's uses of the notebook; a more comprehensive analysis will be part of proper digital editions of all the NLI Ulysses notebooks that will set out the sources of the notes, the complete draft usage, as well as the notebooks' relationships to one another and to the other Ulysses note-repositories and manuscripts. So far I have begun to source the notes; like most of Joyce's notes, they are based on a variety of printed sources. Given the impetus that research tools like Google books and online database have given to notebook source-studies, this work will be much less difficult than it was in the past.[44] Therefore, I have concentrated for now on determining the earliest draft usages of the individual notebook pages in order to establish the terminus a quo of the notebook's compilation. I have ascertained most of the draft usage on the extant manuscripts, typescripts, and proofs (the remaining draft usage can be ascertained by collating the extant manuscripts), and I am investigating the inter-textual connections between the notebooks and manuscripts as Joyce continued to write and revise Ulysses.[45]

Aside from the extant manuscripts themselves, we have only minimal further information about Joyce's notes from his correspondence. For example, three days after first arriving in Paris, Joyce wrote to Harriet Shaw Weaver on 12 July 1920:


My intention is to remain here three months in order to write the last adventure Circe in peace (?) and also the first episode of the close. For this purpose I brought with me a recast of my notes and MS and also an extract of insertions for the first half of the book in case it be set up during my stay here. The book contains (unfortunately) one episode more than you suppose in your last letter. I am very tired of it and so is everyone else. (SL 265–6)[46]


Ulysses actually appeared a full year and a half later and Joyce in fact stayed in Paris for about twenty more years. The BL Notesheets for 'Cyclops' to 'Oxen' (and probably some of the other sheets as well) are presumably remnants of the 'recast of my notes'. But Joyce probably did not revise the earlier episodes from the 'extract of insertions for the first half of the book' at this time; he was simply too busy with 'Circe' and then 'Eumaeus' to do any other work. Four months later, Joyce referred to his notes again in a letter to John Quinn:


I began Ulysses in 1914 and shall finish it, I suppose, in 1921. [...] The complete notes fill a small valise, but in the course of continual changings very often it was not possible to sort them for the final time before the publication of certain instalments. The insertions (chiefly verbal or phrases, rarely passages) must be put in for the book publication. Before leaving Trieste I did this sorting for all episodes up to and including Circe. The episodes which have the heaviest burden of addenda are Lotus-eaters, Lestrygonians, Nausikaa and Cyclops. (24 November 1920; LIII 30–1)


The extant Ulysses notebooks do not constitute what Joyce referred to here as the 'complete notes' because they were all compiled in Paris in 1921 (some of the later BL Notesheets were also compiled in Paris in 1920–1), but what survives in these notebooks for the earlier episodes were almost certainly based on that bulky assortment. Besides the NLI and Buffalo notebooks, no other notes for the 'Telemachus' through 'Sirens' episodes survive. A collation of the additions and revisions to the typescripts for those episodes (and, when they are missing, to the first settings of proofs) to determine how many of them cannot be traced to the extant notesheets and notebooks gives a clear indication of the relatively large quantity of note-repositories that are not known to have survived.

On 5 January 1921, after having finished 'Circe' (or so he had hoped), Joyce wrote to his Triestine friend Ettore Schmitz (Italo Svevo):


The Eumeus episode, which is almost finished, will also be ready around the end of the month. [...] Now for the important matter: I cannot leave here (as I had hoped to) before May. As a matter of fact, for months I have not gone to bed before 2 or 3 in the morning, working without respite. I shall soon have used up the notes I brought with me here so as to write these two episodes ['Circe' and 'Eumaeus']. [...] Having urgent need of these notes [further notes that Joyce had left in Trieste] [...], I address this petition to you, most honourable colleague, begging you to let me know if any member of your family intends to come to Paris in the near future, in which case I should be most grateful if the above-mentioned person would have the kindness to bring me the briefcase ['mappa'] specified on the back of this sheet.[47]


There is a great deal of play and exaggeration in Joyce's letter to his friend Schmitz, presumably this is also the case with the description of his 'briefcase' of notes, but we have no other, more precise information about what Joyce refers to as his 'mappa' of Triestine notes. Presumably this bulky and mysterious consignment was delivered to Paris (when or how is also not known), but I suspect that the extant notebooks (along with others) are a distillation of precisely that earlier hoard of notes.[48] Generally, there are relatively few notes for most of the early episodes, especially for the first three (in any of the notebooks that survive). Since even those episodes were more than just lightly revised in typescript and proofs, there obviously must also have been other note-repositories that are presently not known to be extant.

As I have already discussed above, throughout his career Joyce typically used his notebooks shortly after he had finished compiling them. If this pattern holds here as well, then I suggest that this notebook was compiled from February to May 1921, since its first draft usage was writing and revising the earliest extant 'Penelope' manuscript (NLI MS 36,639/14), which so far I have not been able to date any more precisely than early summer 1921. Nonetheless, Joyce had certainly completely finished compiling this notebook by early June 1921, since he relied on it to revise the first placards of the initial five episodes ('Telemachus' to 'Lotus Eaters'), which Darantiere set from 11–17 June. Those revisions were incorporated in the second setting of the placards by the end of the month. Joyce continuously returned to this notebook as he further revised the early episodes and began revising the later ones. Again based on the immediacy of draft usage, I do not believe this notebook is the earliest or likely the last of the extant Ulysses note-repositories Joyce compiled in 1921: presumably Joyce compiled both MSS 36,639/05/A and 5/B before this notebook, and Buffalo MS V.A.2.b is either contemporaneous with it or an even later compilation.

The arrangement of the episode headings in each of the extant notebooks is different. Here they proceed from recto to verso to recto sequentially (as a codex-copybook is usually used) covering the entire work from the first to last episodes (with the notable exception of 'Sirens' which is not represented in this notebook), and the usual 'Eventuali' heading at the end of the run of episodes. Joyce continued filling five further pages with 'Penelope' notes, often directly in mid phrase; writing this episode was clearly the major focus of Joyce's attention as he compiled this notebook.[49] Joyce did not bother to head these last 'Penelope' pages as he proceeded (again unusually) still from recto to verso on subsequent pages. At a later stage, Joyce used some of the open spaces on the versos towards the front of the copybook primarily for further 'Circe' notes (including the last page, [12v]) and once for further 'Cyclops' notes (p. [6v]).[50]


[back to top]


II. Later Ulysses Notebook: February–May 1921

For Drafts, Typescripts, and Proofs (MS 36,639/05/A)


Joyce also compiled this notebook sometime between February and May 1921. In the same letter to John Quinn quoted above, Joyce continued:


Therefore, I must stipulate to have three sendings of proofs (preferably a widemargined one must be pulled), namely:

(1) A galley-page proof of all the book up to and including Circe.

(2) A similar proof of the three chapters of the Nostos.

(3) A complete proof of the book in page form. (24 November 1920; LIII 30–1)


Joyce's first use of this notebook was to revise a printers' typescript of 'Hades' in late July 1921. Even though other extant notebooks were also used for that level of revision, this provides further evidence that this notebook was not the earliest notebook Joyce compiled or used in 1921. Its compilation falls in the middle of this extremely busy period in the composition and transformation of Ulysses.

Joyce's usual procedure in the later Ulysses notebooks that survive was to write the headings at the top of just the recto page and underline them first all the way through the notebook. Then, when he had filled that page, he would use the facing verso page for other notes for that episode (usually not bothering to repeat the heading).[51]


[back to top]


III. Later Ulysses Notebook: January–February 1921

For Drafts, Typescripts, and Proofs (MS 36,639/05/B)


This is most likely the earliest of the extant later 'second-order' Ulysses notebooks.[52] As I have already argued, throughout his career Joyce typically used his notebooks shortly after he had finished compiling them. If this pattern holds here as well, then he compiled the notebook just before mid February 1921, presumably in January or early February (but certainly no later), since its first draft usage was the earliest extant 'Eumaeus' manuscripts, which Joyce was certainly also writing in February. He next used this notebook to write and revise the 'Ithaca' and 'Penelope' episodes.[53]

The 'Eumaeus' notes here are of particular interest because they help to understand and date that episode's various drafts (and thereby some of the relevant notesheets and notebook pages as well). The episode's earliest extant draft is the so-called 'Eumeo' manuscript, which is now in private hands.[54] Based on a preliminary analysis of the available reproductions of the 'Eumeo' manuscript, it appears to be the direct antecedent of the only other pre-faircopy manuscript, the partial Buffalo MS V.A.21.[55] Similarly, the Buffalo manuscript–plus a now missing complementary manuscript (or manuscripts)–was the immediate antecedent of the Rosenbach (faircopy) manuscript of the episode. Again based on a preliminary analysis, it is clear that Joyce used some of the BL 'Eumaeus' Notesheets to both write and revise the 'Eumeo' manuscript, presumably in Trieste in 1920.

Furthermore, it is now also clear that Joyce used two of the NLI later Ulysses notebooks (MSS 36,639/05/A and 5/B) to further revise the 'Eumeo' manuscript in Paris in February 1921 as well. A more comprehensive analysis of Buffalo MS V.A.21 confirms this scenario: BL 'Eumaeus' Notesheet as well as NLI notebook entries appear throughout the draft as part of the main text and as additions and revisions.[56] Subsequently, Joyce used the notesheets and notebooks to continue to write and revise every level of the episode's text.

Another interesting feature here is the transfer of notes (at times sequentially and hence directly) from the BL 'Eumaeus' Notesheets to this notebook. The transferred notes are sporadic on p. [6v] but quite consistent on p. [9r] and this page comprises the most concentrated transfer of notes from the BL Notesheets to any of the extant notebooks. If that page can be taken as paradigmatic of Joyce's methods of compiling subsequent note-repositories (in this case it is an at least 'third order' notebook page), then a clear (and not unexpected) pattern emerges. Joyce must have relied on several different prior note-repositories to compile this page, only some of which are known to be extant. Although there is a relatively long series of notes that come from a (broken) sequence of notes on BL 'Eumaeus' 1.87–137, they are preceded by other notes from BL 'Eumaeus' 5, and immediately followed by notes from BL 'Eumaeus' 1.10, 11, 17 and 111 (in that order), and then several further notes from BL 'Eumaeus' 4. None of these transferred entries are completely sequential, there are always intervening notes that have not been located on any BL 'Eumaeus' notesheet, which supports the contention that Joyce used further note-repositories to compile this page. Presumably, this is the case with all of the later Ulysses notebooks.

The notes look as though they were taken hurriedly and the headings are often abbreviated (for example, 'Cycl' (three times), 'Naus', 'Lotus', 'Scy & Caryb' and 'Eol'). No clear pattern of organizing the headings is apparent throughout the notebook, although some pages are related to one another.[57]


[back to top]


Earlier Typescript Schema for Ulysses: November 1921

The Harriet Shaw Weaver-Paul L?on Typescript Copy

(MS 36,639/06)


I have prepared a complete analysis of the various earlier and later 1921 Ulysses typescript schemata as part of my catalogue description of the 'Beach Schema' (Buffalo MS V.A.1.b.i [V.A.1.b]); see


[back to top]


Earlier Partial Drafts of 'Proteus' and 'Sirens' and Notes: 1917

(MS 36,639/07/A–B)


This copybook contains partial drafts and further texts for two distinct episodes: 'Proteus' (36,639/07/A) and 'Sirens' (7/B/1) as well as a loose sheet with fragments that are part of the 'Sirens' texts at this earlier draft level (7/B/2). For a later, complete draft of 'Proteus' see Buffalo MS V.A.3;[58] for a later draft of 'Sirens' see MS 36,639/09 below and its continuation in Buffalo MS V.A.5.[59]

This document is unique among the extant Ulysses manuscripts for several reasons. Firstly, it is the only known instance in which Joyce wrote texts, fragments, and drafts for two different episodes in the same copybook. Furthermore, the 'Proteus' fragments (pp. [1r]–[5r]) are the earliest extant texts Joyce wrote specifically for any episode of Ulysses. Also, the 'Sirens' portion of the copybook is composed of two distinct parts: 1) a continuous early draft (pp. [5v]–[10r]) of most of the first half of the episode (see U 11.98–540) at this stage of its draft development; and 2) a disparate collection of fragments (pp. [10r]–[14r] and the loose sheet) that Joyce variously incorporated into the next extant draft of 'Sirens' and elsewhere.

This 'Sirens' manuscript is the earliest extant draft to feature Leopold Bloom and include references to Molly Bloom (but only in the second part of the draft).[60] Presumably, before beginning the 'Sirens' portion of the manuscript, Joyce also compiled a list of notes based on an as yet undetermined Homeric source on its final page (p. [14v]).


An Earlier Proto-Draft of 'Proteus': (MS 36,639/07/A)


The manuscript is composed of seventeen discrete textual fragments (on nine pages) that are almost all separated from one another by horizontal lines of Xs or asterisks.[61] Joyce probably compiled these fragments (or composed this draft) after mid October 1917 in Locarno, Switzerland, and subsequently wrote the next surviving draft of the episode (Buffalo MS V.A.3) from late October to December 1917. The only known note-source for this text is the Early Ulysses 'Subject' Notebook (NLI MS 36,639/03; see above), but there were almost certainly also other note-sources that Joyce relied on to write and revise these texts. As the entries from the 'Subject' notebook are part of the main text of the draft throughout (rather than just marginal or interlineal additions), it is likey that Joyce actually used these notes on some missing preceding versions of these discrete fragmentary texts, which he must have written (or more likely revised) just before writing this version.

Furthermore, although the texts are revised currente calamo, interlineally, as well as in the margins, the appearance of the basic text-fragments in this manuscript suggests that Joyce copied each fragment as such from another written source (or more likely several other sources), rather than writing them directly from word- or phrase-based notes for the first time.

Joyce probably had some other assortment of texts along with this manuscript that together comprised the episode as it then was in early October 1917, but whether these texts resembled this kind of manuscript, or whether there was another, more cohesive narrative draft into which these fragments were to be incorporated is not known, though I believe that the later possibility is less likely. Also, another draft (or drafts) probably intervened between this manuscript and the next surviving draft of this episode, although no other early manuscript is known to be extant.

Based on the various ways these fragments were incorporated in the subsequent extant draft, it can be argued that Joyce generally used these fragments in three kinds of ways: 1) as integral, epiphany-like set-pieces that remained quite similar on the subsequent extant manuscript;[62] 2) as isolated fragments that were substantially altered and expanded from this manuscript to Buffalo MS V.A.3;[63] and 3) as individual fragments that were combined to form a longer narrative scene on the later manuscript.[64]

In general, it seems almost certain that Joyce must have written all these integrated texts out together somewhere else before they appeared on the next extant draft because (although most of the corresponding texts are basically similar between the two manuscripts) there are several more texts that Joyce probably would not have written in such a relatively unrevised manner on MS V.A.3 for the first time. These are significant blocks of text that first appear on the later manuscript, including the opening (U 3.1–29) and a long early section (U 3.120–208), which among many other scenes includes Stephen's description of his epiphanies, both of which have the distinct appearance in MS V.A.3 of having been written before (possibly more than once) and copied there.

This manuscript has been controversial since it was first discovered in 2002 and Daniel Ferrer's pioneering essay in the JJQhas been central to the debate.[65] Ferrer rightly claims that it 'is quite unlike anything in the published text or elsewhere in the archive' (p. 54). Although my understanding of the status of the 'Proteus' section in this manuscript differs from Ferrer's, there is not enough (internal or external) evidence to resolve the debate at present. Ferrer presents the crux of the debate about the nature of this manuscript as follows:


The fragmentary nature of these short segments is surprising and raises an important question. Is it simply the accidental appearance of a composition in process, a succession of passages that Joyce wrote in this copybook? Or is it a stylistic device reflecting a deliberate aesthetic choice, a mode of presentation that was later discarded in favor of a different option? (p. 55)


Ferrer suggests that it is the latter, while I believe it is the former. I would argue that this manuscript is a transitory collection of blocks of text, based on previously written material (possibly in more than one manuscript or on disparate sheets of paper) that Joyce merely consolidated in this document. I maintain that this manuscript served as a temporary repository for these discrete textual fragments on their way towards a more fully elaborated draft (presumably some form of narrative draft that preceded MS V.A.3), along with other texts in some form.[66] These other 'Proteus' texts could either have been another collection of textual fragments like these (on loose sheets or in a copybook) or possibly a (partial though more fully-elaborated narrative) draft into which these texts were meant to fit.

Ferrer argues that, '[a]esthetically, the episode makes perfect sense in this fragmented form. The poetic prose of Stephen's musings is, in a way, even better set off than in the final version. The underlying narrative is at least as easy to follow, starting as it does with Stephen's decision not to go back to the tower and alternating observations of what he sees around him on the strand, reminiscences, fantasies, and speculations.' (pp. 55–6). Such an 'aesthetic' argument is subjective and certainly founded more on Ferrer's thorough and nuanced understanding of Joyce's later stylistic innovations in Ulysses (as well as more obviously in Finnegans Wake),[67] than on the historical, material, and textual evidence provided by this manuscript. Then, Ferrer continues, 'The episode does seem truncated: it is certainly odd that it would end with the waking of Paris' (p. 56).

It seems unwarranted to look for a narrative in these discrete textual fragments since either by the time Joyce had assembled these fragments here or quite shortly thereafter in the next extant manuscript, he had determined the bare narrative of 'Proteus'.[68] As we know, it is simple enough: Stephen walks along Sandymount Strand, writes a short poem, considers stopping by his Uncle's home in Strasburg Terrace but decides not to, and everything else happens in Stephen's imagination. The associative elements and transitions that bind the episode's narrative in the next draft version (and then in the published texts) are completely absent here, and it is primarily through our understanding of their relationship in the subsequent versions that some cohesive sense can be ascribed to the fragments. Rather, it seems more likely (although, admittedly, the internal evidence does not support this claim any more than it does Ferrer's) that in this manuscript Joyce was simply gathering textual elements that he knew would be components of a draft, with the typical self-assurance that he could arrange and fill in the transitional material as needed when he set about composing the narrative form of the draft.

Famously, Joyce would later claim that he worked both as a 'scissors and paste' man and 'engineer' and here we have an early example of the artist and technician's tools. This manuscript may simply seem unique because it is the earliest and most extreme example of this kind of Ulysses manuscript that has come to light. Besides the early 'Sirens', 'Cyclops', and 'Ithaca' manuscripts, the more complete textual history of the earliest drafts of chapters of Finnegans Wake suggest that it was a common practice for Joyce to gather fragments and then assemble (or re-assemble) them when the time came to produce a narrative. I would argue that this manuscript is a natural component of the continuum of Joyce's writing methods from notes to fragments and then on towards a narrative draft. They function like Joyce's 'epiphanies' that he mined for virtually all of his works.[69] Therefore, these 'Proteus' fragments, though much further developed, written in a clearer hand, and more textually stable, are also generically similar to the 'Sirens' fragments at the end of this manuscript and its loose sheet (see below).

A further interesting issue is the relationship of this manuscript to the next extant draft. It seems unlikely that this NLI 'Proteus' manuscript was the direct precedent of the Buffalo manuscript, even in conjunction with other manuscripts of this kind or else a more comprehensive draft of the episode of which this manuscript was a supplement. The text in this manuscript begins with Stephen's vision of the scene in the Martello Tower in Sandycove after he had left and his final decision not to sleep there that night. On the later manuscript, this fragment is rearranged, expanded, and embedded in a longer and more developed narrative that precedes it, suggesting that it was based on a draft version Joyce wrote in between this NLI manuscript and MS V.A.3. Interestingly, on the later manuscript, Joyce reverted to earlier readings that he had revised here. On the later manuscript, this scene is followed by a transitional intermediary paragraph (U 3.282–5), which is a mixture of first and third person narration that is completely absent on this earlier manuscript. Then, after some revision on the later manuscript, the text is virtually identical to its appearance in the published versions in The Little Review and Ulysses.

The section below this in the manuscript follows next in MS V.A.3 as well (and this may not be coincidental). The scene describes the appearance of the dog's carcass by the edge of the sea and Stephen's musings on language. Although some of the passage is already present in this manuscript, it is both changed and further developed to a point that suggest that another draft may have intervened between the two extant versions. Another fairly substantial section is missing from this draft but is present in large measure on MS V.A.3 (U 3.291–303). What follows marks the first instance where a fragment from a later page is integrated into the text of MS V.A.3 (or whatever may have intervened between the two extant drafts). The next section (U 3.303–6) comes from p. [4r] in this manuscript. Although it is in a fairly rudimentary state here, the way in which Joyce elaborated it on the later manuscript is remarkably similar to its published form. On the other hand, U 3.48–52 is significantly altered and embedded in the text in MS V.A.3.

Finally, if (as is most likely) Joyce had written other texts destined for 'Proteus' by the time he compiled the texts in this manuscript, it is not clear why he did not fill the remaining pages of this manuscript with those texts.


?An Early Partial Draft of 'Sirens'


The first part of MS 36,639/07/B/1 is a partial early draft of the 'Sirens' episode (pp. [5v]–[10r]). It is followed directly by discrete fragments of text (not all of which were used for this episode; on pp. [10r]–[14r]), and there is also a related loose leaf of paper with further fragments that is part of this draft stage (NLI MS 36,639/07/B/2). Finally, the last page of the copybook (p. [14v]) contains notes headed 'Lacedemon' that may or may not be specifically related to either draft in the copybook. As we saw with the notebooks, Joyce may have simply used the last page of this copybook as a convenient place to record general notes, some of which he crossed through in blue crayon, though it is difficult to ascertain how he used them in Ulysses.

Like the preceding 'Proteus' draft in this copybook, this portion of the manuscript is astounding in its own fashion. Quite surprisingly, this draft confirms that 'Sirens' was one of the earliest episodes of Ulysses Joyce wrote.[70] This portion of the copybook contains the earliest extant draft of most of the first half of the published episode (U 11.98–540), though the text is in a considerably different state than the next extant version of this part of episode (and therefore is also quite different from how it appears in Ulysses). Joyce wrote this draft in a clear and continuous manner and in a particularly neat hand that is well laid out on the page; the text is not crowded and there is ample space between the lines of text. This suggests that Joyce was copying from a fairly well developed, earlier draft. Besides the arrangement of the text on the pages, the fact that he was using versos and rectos sequentially also suggests that this section was not a first draft of this material.[71]

The narrative of the first part of the episode is a fairly straightforward description of the scene at the Ormond Bar, with dialogue, and some musical elements, but little of the psychological depth readers expect from 'Sirens'. The 'fugal' (or varied contrapuntal) narrative style of the episode as it is known from later versions of the episode–where events occurring in different places and the various thoughts of the characters are rendered simultaneously–is not evident at this draft stage. On the other hand, the latter portion of the manuscript, which is comprised of a variety of fragments (of descriptive scenes, dialogue, and interior monologue) marks the multi-faceted beginnings of the episode's stylistic breakthrough.

This draft exhibits Joyce's usual, subsequent revisions and marginal additions. He substantially and significantly revised almost all of the text in this portion of the manuscript both in ink and in pencil, which indicates several rounds of revision and expansion. Joyce's revisions of the opening of 'Sirens' are particularly illuminating and are exemplary of the stylistic transformation of the episode in general. He wrote at least several sentences, some dialogue (and probably more of this page and possibly still more), but then revised it, only slowly establishing the opening as it appears in Ulysses. Joyce described both barmaids' heads as 'bronze' at first, but then changed Miss Douse's (she is not yet 'Miss Douce' as we know her in Ulysses) hair colour to 'gold'. He then rewrote the entire opening to start with the well-known 'Bronze by gold', switched the order of the barmaids' names to coincide with the newer description, and specifically reinforced the aural sense over the visual (presumably for stylistic reasons).

The story at the Ormond Hotel here begins with the banter between the barmaids as they watch the viceregal cavalcade go by the blinds of the barroom down the quays. The young women's overt sexuality seems to be the most basic point of the scene so far. Then, quite early on in this version of the events, Simon Dedalus makes his first appearance on the stage of Ulysses as he walks into the Ormond bar preening his 'rocky thumbnails'. Lenehan enters next and asks whether Boylan has been in looking for him; this is the first mention of Blazes Boylan in any surviving document for Ulysses, but obviously Joyce had already established his role as Molly's suitor, presumably several years before.

Blazes enters the Ormond and all the attention turns to him as he tells them of Bantam Lyon's tip for the Ascot Gold Cup race. He is overly concerned about the time, not just because of the race results, but because he has an 'appointment he can't miss'. Lenehan presses Miss Douce to snap her garter, which she does reluctantly but with the purpose of attracting Boylan's attention. Boylan throws back his drink and sets off with Lenehan trailing close behind. So far this follows the action of the published versions of 'Sirens', though much is also missing.

Ben Dollard and Bob Cowley then enter the Ormond discussing the rent Cowley owes his landlord and then Simon convinces Dollard to sing. The song 'Love and War' brings up memories of the past and specifically a particular night when the song was played. At first Joyce wrote that Father Cowley was the pianist but almost immediately he changed this to 'poor old Goodwin' and Simon comments 'a nice hash he made of it'. As far as we know, Joyce had not yet written the other references to Goodwin's concert in the earlier episodes so this may be a new idea for Ulysses.

Next, Father Cowley recalls the night as well and asks about a certain 'Marie Fallon'. Unknown to readers of Ulysses, Joyce quickly changed her name to 'Marie Powell'. Now the name Powell immediately connects this Dublin prima donna to the stories of Ulysses because 'Major' Malachy Powell was one of the sources for Major Brian Tweedy, Molly's father. So now with just a change of name a very significant storyline begins to emerge.

At this point Joyce made several pivotal changes: he added a reference to the song lyric 'My Irish Molly O' that he connected with another bit of information about 'a soldier's daughter ^from rock of Gibraltar^'. Only now does Molly, who is also described as 'a buxom piece', become the main subject of the men's conversation. It is interesting to note how much of Molly's background was already fixed in place at this relatively early stage. Of course, some of it has to do with the real life prototypes of Molly Tweedy, but what is notable is the way in which the various threads are grouped together and are only slowly disentangled as Joyce kept writing 'Sirens'.

It was not the case that Joyce was simply conflating Josie Powell and Molly because Joyce had fixed on Molly (or 'Mollie') as a main character in Ulysses by at least 1917 when the notebook that served as source for this draft was compiled. Molly is distinctly mentioned in another addition on this page, when Bloom notes that she 'has a devil of a quick eye to notice if anyone is looking at her'. Then Cowley asks: 'what became of her Simon? I never see her name is she still alive'. Simon replies that she is indeed 'alive and kicking' but that she has married.

What is most interesting here is that these additions mark the end of the continuous draft of 'Sirens'. Joyce simply ran out of background information just as Dollard is about to start his song. The continuous draft state of this manuscript just stops in the middle of p. [10r]. It is likely that this rupture coincides with the extent of the earlier draft. Later, Joyce cancelled the phrase 'a moment', and replaced it with 'Half time' (obviously intended as a musical term, but it could also have served as a tag pointing to the second half of the episode as it was then structured, though he did not follow through with that plan). He drew a line separating everything that came before, and only then introduced Bloom, already seated alone at the Ormond Hotel.

As Daniel Ferrer was the first to note, the most striking difference between this draft and all other versions of 'Sirens' is that Bloom is completely absent from the first half of the narrative scene. It might be that at this juncture Joyce planned that 'Sirens' would function with two perspectives–much like 'Nausicaa' in fact does, though, as far as we know, Joyce only wrote that episode about a year later–with a Dubliners-like bar scene with dialogue first, counter-balanced by Bloom's interior monologue in the second half of the episode. It is even possible that Joyce could have intended that the second half of the episode would take place at exactly the same time as the first half, more like some of the interpolations in 'Wandering Rocks' rather than the more traditional sequential timeframe of the episode in 'Nausicaa'.

The disjunction between the parties at the Ormond is particularly conspicuous in this version of the scene: there is a convivial crowd at the bar on one side and then there is Bloom, the outsider, who observes the others 'in silence'. The manuscript suggests that the transition to Bloom may have been initiated by Joyce's addition, 'My Irish Molly O'. This is the earliest version of Bloom's interior monologue we have in any manuscript. Then, piece-by-piece, phrase-by-phrase, Joyce began to develop the point of view and thought patterns that are so characteristic of what is commonly (and reductively) described as Bloom's perspective: as the open-minded, intellectually curious, 'modern hero'. Joyce continued Bloom's sexualised reverie about the barmaids at the beerpull here for a few lines and then stopped once again (this scene is all on pp. [10r]–[10v]). The rest of the manuscript is composed of further, non-consecutive, and even less cohesive fragments that in general are certainly more similar to the 'Proteus' material than the preceding narrative-portion of this 'Sirens' draft.

The physical disposition of the texts in the 'Sirens' sections of the manuscript allows us to refine the date of its composition. Firstly, since both parts follow the 'Proteus' fragments in the copybook, they must obviously have been written after Joyce had written them, so after mid October 1917. As with the 'Proteus' portion of the manuscript, notebook usage in these drafts corroborates the end of 1917 as the earliest date for the composition of both parts of the 'Sirens' portion of the copybook. Joyce used several clusters of notes from the Early Ulysses Subject Notebook (NLI MS 36,639/03) to write the first part of the draft. This indicates that whatever form the draft may have had prior to this manuscript, Joyce wrote or revised it with entries from that notebook; therefore, we have further proof that he did so after mid October 1917 when that notebook was compiled. He also returned to it for further entries for the second part of the 'Sirens' portion of the copybook.

Joyce was busy writing the Rosenbach faircopy drafts of 'Telemachus' and 'Nestor' from October through November 1917 and he had certainly finished writing the next surviving 'Proteus' draft (Buffalo MS V.A.3) by mid December that year. Given all of his work on the 'Telemachiad' (although we do not have sufficient evidence one way or the other), it is unlikely that Joyce was also able to write a completely non-Stephen related episode at this time as well. It is not known how much more time elapsed between the writing of the 'Proteus' fragments on the previous page ([5r]) and Joyce's work on 'Sirens', but the evidence suggests Joyce began writing this earliest version of 'Sirens' at the start of 1918 at the earliest.

In his James Joyce and the Beginnings of 'Ulysses', Rodney Wilson Owen presciently wrote 'that the first third of 'Sirens' was earlier and more complete than the rest' (JJBU 67). Furthermore, Owen states that although he is


not trying to argue that the first section of 'Sirens' was actually drafted before Joyce left Trieste in 1915; one would suppose, though, that the notebook [Buffalo MS V.A.2.a] correspondence and relative lack of later additions to the first section indicate that Simon's flirtation with the barmaids, his piano playing, and the presence of both Bloom and Blazes Boylan, were part of the original nucleus of the episode, and as such might have existed in Trieste in a short sketch, an outline, or a compilation of notes. (JJBU 67)


This manuscript provides clear evidence that Owen's intuition was astoundingly accurate: Joyce must have written a very early version of most of the first half of the episode between 1915 and 1917 and it (and possibly other drafts) preceded this manuscript.[72]

Joyce used the Early Ulysses Subject Notebook (NLI MS 36,639/03) to develop key descriptions of Simon Dedalus, Lydia Douse, Mina Kennedy, and Leopold Bloom. Joyce first integrated an entry from the 'Subject' notebook into a relatively well-elaborated exchange about Miss Douse's sunburn and Miss Kennedy's remedy (on p. [6r]; tellingly, it appears under the heading 'Recipes' on p. [4r] of the notebook). The text also clearly appears to have been drafted previously elsewhere. Joyce then used entries from the very start of the 'Subject' notebook (taken from notes under the heading 'Simon' on p. [1r]) to describe the elder Dedalus' telltale gesture as he enters the Ormond Bar (see U 11.192–3). This scene occurs on p. [7r] of the manuscript. Joyce then used another entry to describe the manner with which Miss Douse serves Simon Dedalus: 'With the greatest alacrity' (U 11.213). The entire phrase as such is on the last page of the 'Subject' notebook (under the heading 'Words' on p. [15v]).[73] Another entry from the 'Subject' notebook describes Miss Kennedy's concentrated efforts to continue reading and so ignore Lenehan. It too was under the heading 'Words' and Joyce even tagged it with Lenehan's name. Again, the fact that so little of the text can be traced to extant notebooks suggests that there were other contemporaneous notebooks that Joyce relied on to write this (and earlier) drafts.


?Fragmentary Texts for 'Sirens'


Unlike the punctilious way Joyce segregated the previous 'Proteus' texts from one another with asterisks, the 'Sirens' fragments that comprise the final part of this copybook here are only sometimes separated from one another, and then only with short, hastily-drawn horizontal lines. This suggests that these later fragments were much less developed; in fact, their physical appearance indicates that Joyce was writing some of these texts here for the first time. Later, Joyce (as usual) crossed out with Xs in different coloured crayons the text-blocks he was incorporating in the next draft of 'Sirens' or was transferring elsewhere. Interestingly, although Joyce used some of the blue-crossed-through texts in 'Sirens', others appear in 'Lestrygonians', in 'Cyclops', and even later in 'Circe' (the latter two uses at least came about via intermediary texts). It is likely that after Joyce had written the next draft of 'Sirens' (NLI MS 36,639/09 and Buffalo MS V.A.5 together), he returned to this manuscript and transferred the unused fragments to other repositories and from there they ended up in those other episodes of Ulysses.

For example, the first full fragment on p. [10v] is just one of several snatches of malicious (specifically male Dublin pub) gossip about Bloom and Molly that appears in this manuscript, but Joyce actually ended up using it as a scene in 'Cyclops' (see U 12.1566–9). Based on the state of the 'Cyclops' drafts that survive and other interpretive considerations, I would suggest that Joyce had not yet envisioned what became episode twelve at this stage (at least as we know it from Ulysses) and that it both materially and conceptually grew out of Joyce's continuing elaboration of 'Sirens'. It could be that Joyce came to feel that any sort of altercation between Bloom and the other Dubliners at the Ormond Bar might be psychologically too traumatic an event for Bloom at this crucial and difficult part of his day. Joyce may also have come to realise that it was structurally too complex and demanding from a narrative perspective to have Bloom confront his repressed feelings as well as confront other men (who somehow seem to know about Blazes's impending rendezvous with Molly) all in just one episode. So, as he elaborated his ever-expanding novel, he shunted off the more aggressive and confrontational aspects of 'Sirens' (focusing instead on its musical and stylistic aspects). Although they were previously intended as part of just one barroom scene, at some later stage Joyce decided that the various external disputes would happen later in the day in another episode and elsewhere.[74]

The next fragment (also on p. [10v]) describes the coincidence of Simon Dedalus singing an aria from Martha as Bloom is about to write to his own 'Martha Clifford'. Joyce incorporated the textual fragments on p. [11r] both to fill in what he had already written as well as continue the narrative in the next version. The first fragment describes the Blooms in Holles Street, when he had lost his job at Hely's and 'Mrs Marion Bloom' sold used clothes. It ends with Simon's witty comment that Molly has 'left off clothes of all descriptions'. Another fragment describes Bloom's sympathetic, though ironic, attention to Richie Goulding's poor health and finances. The first mentions of the Blooms coming to Ben Dollard's rescue by getting him a dress suit for his concert occur as marginal additions.

The next short fragment Joyce squeezed in the middle of p. [11v] is also exemplary of the stylistic innovations the episode underwent in this draft on its way to Ulysses. Joyce had tagged this fragment with the overt stage direction '(He thought)', but by simply though determinedly crossing it out, Joyce's text plunges the reader into Bloom's thought processes as he ponders his life. He wonders: 'If she then I would be different then. But she has? Or has she?' That is, Bloom asks himself how would their lives be different if Molly did not keep her rendezvous with Boylan? Tellingly, afterwards, Joyce dispersed and dissimulated this crucial scene in various other episodes of Ulysses, presumably because it could not be part of 'Sirens' for the same sort of psychological and structural reasons I alluded to above. I would argue that given the importance of this question for Bloom that day, such a direct, unqualified statement of his fears as a husband, father, and as a man, would simply have been out of place in 'Sirens' once Joyce had written more of the Bloom-oriented episodes in the first half of Ulysses. On the other hand, such a complex and conflicted scenario does make sense from the perspective of 'Circe', but Joyce only started elaborating the later episode over a year later in mid 1920 (as far as we know).

Several of the fragments here are explicitly from a narrator's, third-person perspective on Bloom in the Ormond, describing Bloom's observations and thoughts. But, in the additional text in the margins, Joyce slowly but surely developed and refined Bloom's interior monologue; he began to explore the workings of Bloom's mind and to present them directly to the reader, thereby giving 'Sirens' the immediacy that is a hallmark of Ulysses.

More fundamentally, this part of the 'Sirens' manuscript demonstrates that Joyce initiated some of the other experimental aspects of Ulysses much earlier than scholars of the genesis of Ulysses had presumed; that is, in 1917 rather than in 1919. For example, the central fragment on p. [12r] presents Bloom's disjointed, half-articulated, but coherent thoughts as he recognizes a pun on chamber music, tries to formulate his own 'science' of acoustics, thinks of Listz's Hungarian Rhapsodies, and then composes a musical arrangement for urine tinkling into a chamber pot: 'Diddleiddle addleaddle ooddleooddle'.

Joyce wrote most of the fragments for the ending of 'Sirens' on p. [12v], but not in a sequential order. They alternate between scenes after Bloom has left the Ormond and Ben Dollard has sung The Croppy Boy. At the bottom of the page Joyce explored Bloom's thoughts about the man who first discovered drums. This prompts him to think about the asses' 'point of view': they are beasts of burden in life and then their skins are used for the drums. Bloom then tries to recall the Arabic word 'Kismet' ('Fate') but he confuses it with 'backsheesh' ('gratuity'), rather than the word 'yashmak' (the veil worn by Muslim women in public) that readers know from Ulysses.

Joyce wrote the first known version of the finale of 'Sirens' on the top half of p. [13r], even though two more pages of fragments follow. The first fragment begins with the 'frowsy whore' Bloom wants to avoid (which Joyce also took from entries in the 'Subject' notebook, under the heading 'Leopold'), then Joyce wove together the closing words of Robert Emmet's speech and, in the margins, Bloom's analysis of the causes of his gas: 'must be the cider or the Burgan.' and, finally, 'Let my epitaph be written. I have. Bfffffff Done.'

The various fragmentary texts that began on the lower half of p. [10r] and follow in this part of the manuscript are difficult to separate from one another at this draft stage and even harder to place as such in the much more developed next draft stage and so in Ulysses. Joyce himself did not always separate the texts with horizontal lines, and even those texts that he wrote as units here he broke up and dispersed at the next extant draft level. It is likely that in this part of the manuscript Joyce was simply collecting some previously written, isolated texts, as well as writing new texts, all without a clear sense of where they would fit into the episode. Obviously, Joyce's sense of the episode as a whole changed radically from this stage to the next.

The transition from this draft stage to the next (see below) is a crucial moment in the evolution of Ulysses. As this episode developed, Joyce presumably realized two major points that altered the course of Ulysses. On the one hand (at a later stage after writing the first part of this draft), Joyce realised that the one-sided approach to 'Sirens' was too simplistic for such a pivotal episode in the work, possibly because it isolated Bloom too obviously at this delicate juncture of the day. Also, Joyce must have realized that it was precisely the musical style of the episode that would allow Bloom to be both on the margins (both structurally and socially) of what was happening in the Ormond and yet not completely cut-off from human camaraderie.


[back to top]


Complete Earlier Draft of 'Scylla and Charybdis': 1918

(MSS 36,639/08/A–C)


This draft consists of three separate copybooks; they contain a continuous text, which together comprise the earliest extant complete draft of the episode. Joyce wrote this draft in early to mid 1918 and had finished the next extant version of the episode (the Rosenbach manuscript) on 'New Year's Eve | 1918'.[75]

Along with 'Proteus' (NLI MS 36,639/07/A and then Buffalo MS V.A.3) and 'Sirens' (MS 36,639/07/B), 'Scylla and Charybdis' was one of the earliest episodes for Ulysses Joyce conceived and presumably wrote.[76] He had gathered some of the elements he would subsequently use for the start of this episode as early as 1903–4 in the Early Commonplace Book: 1903–12 (NLI MS 36,639/02/A), although it is unlikely that he needed to consult that manuscript for those precise phrases when he actually wrote this episode.[77]

It is probable that Joyce considered including a scene that takes place at the National Library in A Portrait of the Artist as a Young Man, but there are no relevant documents that would confirm this hypothesis.[78] On 9 April 1917, Joyce wrote to Ezra Pound, '[a]s regards excerpts from Ulysses, the only thing I could send is the Hamlet chapter, or part of it––which, however, would suffer by excision' (LI 101). What form the episode may have had at this stage is not known, but it was almost certainly an earlier draft or proto-draft of the episode that is now lost (see below).

As is often the case, Joyce's notebook usage helps date this draft. The earliest notes Joyce compiled and then used in this draft are from the so-called 'Alphabetical Notebook' (Cornell MS 25).[79] It dates from 1910 and Joyce first used it to write A Portrait, but then he used other notes from it to write several early drafts of Ulysses, including 'Scylla and Charybdis'. Entries from that notebook–from under the heading 'Stephen', obviously enough–appear both as part of the main text and as additions on this manuscript.[80] Joyce also used entries from the 'Subject' notebook (NLI MS 36,639/03)–from both the headings 'Theosophy' and 'Words'–to write the main text and make additions on various pages of this draft.[81] Joyce also compiled several sets of notes for the episode at the end of this draft (see MS 36,639/08/C, p. [12r]) and used them here, but notably only for additions to the text, which suggests that he compiled the list after he had finished at least a first round of writing this draft.[82]

What we presume was an earlier manuscript of this episode was part of the La Hune 'James Joyce' Exhibition and auction in Paris in 1948. It was part of the collection acquired by the University at Buffalo (as it is now called), but it did not arrive in Buffalo with the other La Hune Joyce material in 1950.[83] What became of that manuscript is not known and it has never reappeared. The La Hune catalogue describes it as:


CHARYBDE ET SCYLLA. / (Neuvi?me episode) 10 grandes feuilles de papier blanc uni, manuscrit ? l'encre uniquement recto. Fragments de conversations qui r?apparaissent, sous une forme tr?s diff?rente, dans la sc?ne de la Biblioth?que. Nombreuses marques au crayon rouge.[84]


[Scylla et Charybdis. / (Ninth episode) 10 large white unlined sheets of paper, holograph in ink only on the rectos. Fragments of conversations that reappear, in a very different form, in the Library episode. Numerous red crayon markings.]


Interestingly, this description suggests that it was similar to the fragmentary texts in the newly discovered 'Proteus' and 'Sirens' manuscript (NLI MS 36,639/07/A–B), which is further evidence that some of the early drafts were composed from disconnected, fragmentary textual blocks. Like other writers, Joyce regularly crossed out (usually in coloured crayons) the text of a draft when he had copied it into another, later document. This manuscript is unusual because none of the main text on any page is crossed out in this way. On the other hand, most but not all of the additional texts in the left margins of the rectos as well as the versos are crossed out in red and sometimes in blue crayon.[85] A possible explanation for this is that this manuscript did not represent the most advanced textual state of the episode when Joyce was recopying it. If that was the case, then here he was primarily concerned with incorporating the additions from this manuscript on some (now missing) subsequent or collateral draft. In general, this suggests that Joyce worked with several drafts of this episode at different stages of its development simultaneously (and this may also account for the different ways in which Joyce wrote out some of the collateral Rosenbach manuscripts, including the one for this episode).

The next extant version of the episode is the Rosenbach manuscript. It is a demonstrably more advanced (and therefore later) draft stage than this manuscript; that is, there are many emendations and additions that appear as parts of the main text of the Rosenbach manuscript that are not present in any recognisable form on this draft. Most notably, there are several lines as well as blocks of text that are completely absent from the text of this manuscript but which appear as part of the Rosenbach version of the episode.[86] Therefore, we can postulate at least one (now missing) draft stage between this manuscript and the Rosenbach version of the episode.

A preliminary collation of the texts on this manuscript and the Rosenbach version demonstrates (when it is possible to adjudicate between variant readings) that this manuscript is closer to the text on the Rosenbach manuscript than to the so-called '(Lost) Final Working Draft' of the episode that produced the extant typescript, particularly for the first third of the episode.[87] The most obvious example is: 'Do you know what you are talking about? Love, yes. ^Word known to all men.^ Amor vero aliquid alicui bonum vult undea et ea quae concupiscumus' (U 9.429–31). It appears in the main text (revised as indicated) on MS 36,639/08/B, p. [2r] and so is identical to the reading on the Rosenbach manuscript rather than the typescript (and so in most versions of Ulysses).

The precise relationship of this manuscript to the Rosenbach manuscript version of 'Scylla and Charybdis' is complicated by the fact that the extant typescript, which in turn produced both The Little Review text (published in April and May 1919) and the episode as published in Ulysses, was definitely not produced directly from the Rosenbach manuscript. The Rosenbach manuscript of this episode is therefore a mixed document, which is relatively unusual amongst Joyce's extant manuscripts.[88] Furthermore, the textual evidence suggests that Joyce produced another now missing draft of the episode, after the Rosenbach manuscript,[89] which was then used as the copy text for the extant typescript. This other (missing) document contained both additional sentences and blocks of text as well as variant readings that are not present on this manuscript or on the Rosenbach.


[back to top]


Partial Later Draft of 'Sirens': 1919

(MS 36,639/09)


This manuscript is a copybook containing the text of about the first sixty percent of the 'Sirens' episode as published in Ulysses (11.01–785). This draft stage is continued directly on Buffalo MS V.A.5 and concluded there.[90] As Michael Groden and Daniel Ferrer have already noted, the discovery of this portion of what we now know is a later 'Sirens' draft did not come as much of a surprise to scholars of the episode for the simple reason that p. [1r] of the continuation of this draft stage tellingly starts in mid sentence on p. '21'. So it was logical to presume that Joyce had already written twenty pages before starting the Buffalo manuscript. Although Joyce continued to refine the text at every subsequent juncture, by this stage the episode had already achieved the formal structure and thematic characteristics that closely resemble its published instantiations, except for the extant version of overture at the start, which Joyce probably only wrote after finishing the rest of the Rosenbach faircopy manuscript.

On the other hand, what is surprising is the radical conceptual and stylistic overhaul the episode underwent between the newly discovered earlier extant 'Sirens' manuscript (NLI MS 36,639/07/B; see above) and this draft stage. The internal textual evidence is inconclusive about when and how Joyce refashioned the episode's (by now) characteristic style: 1) Joyce could have written this draft directly from the earlier version (along with further fragments like the ones at the end of NLI MS 36,639/07/B) and reworked and elaborated them directly here as he wrote this draft; or else 2) Joyce consolidated his earlier work and refashioned the episode's style in a piecemeal manner in one or more intermediary drafts (now missing), and then simply continued that process here. The latter possibility is much more likely but not certain.

Here the fixed, continuous narrative from the earlier manuscript version ends on p. [7v], but is significantly expanded and revised. It is difficult to suppose that Joyce was able to fluidly recast the entire episode, both in terms of its style and content–from its quite different and more primitive state in the earlier draft to this later and well-established version–without positing at least one (but possibly more) intermediate draft stage(s). While Joyce's experience of writing the first ten episodes at least to their 'faircopy' states may account for the first possibility, I consider that scenario quite unlikely.

On the one hand, the fluid progress of the narrative in the last five pages of the copybook clearly suggests that some missing intermediary draft of the whole episode extended beyond what Joyce had written in 1918, especially since so little of the text can be traced to the earlier version. On the other hand, the various ways in which Joyce incorporated the extant fragments from the latter half of the earlier version are ambiguous. Most of the earlier, distinct fragments are consolidated on p. [8v] in this draft. Joyce moulded them together mostly as part of the main text from MS 36,639/07/B, pp. [11v]–[11r]–[11v] (in that order) but, significantly, he also added yet another fragment from p. [11r] in the left margin of p. [8v]. All of these various fragments were crossed through in red crayon on the earlier draft and so one would expect that if Joyce had transferred them all to an intermediary draft, they would all be part of the main text in this later version.[91]

Based on the manuscripts that survive, it appears to have been a regular practice for Joyce to fill the pages of a copybook from recto to verso on later, more stable drafts, but only on rectos on earlier, more fluid drafts (so that he would have ample room to add further text on the facing versos). Though there may be other explanations for his choice here, a notable difference between the ways Joyce used the two copybooks that constitute this same draft stage is that he wrote on the rectos and versos consecutively on the NLI manuscript (the start of the episode), but wrote directly from recto to recto (leaving himself plenty of blank space) on the Buffalo manuscript, reserving the versos for further additions and revisions. Joyce had in fact written the first third of the episode at least twice before and presumably he believed it had reached a fixed (though certainly not final) state. On the other hand, all that survives for the latter portion of the episode before this draft stage are unconnected fragmentary texts. Still, as we have seen in the earlier version of 'Sirens', Joyce's confidence in the constancy of his work was sometimes misplaced.[92]

Although most of the narrative and language of the earlier draft remained, as usual Joyce expanded and reworked much of it. The radical alteration of the conceptual framework of the episode by this draft stage is immediately evident because Bloom's meandering journey along the southside bank of the Liffey to the Ormond has already been woven into the opening dialogue between the two barmaids. Joyce must have decided to include the account of Bloom's arrival at the bar as part of the opening of the episode before he started writing this draft, though it is interesting to see how rudimentary Bloom's itinerary is at this stage (see p. [1r]).[93] For example, at first Joyce simply wrote: 'Mr Bloom went by –– bearing in his breast pocket the sweets of sin', but it was only when he added 'by Moulang's pipes' beside it in the left margin that the blank become a real Dublin locale.[94]

Joyce's other preoccupation here (as it was on the earlier draft and it continued at every subsequent stage) was the ordering of the phrases he had already decided upon.[95] Again from the same paragraph that introduces Bloom in 'Sirens' on p. [1r], Joyce first wrote 'bearing in his breast pocket the sweets of sin, bearing in his memory sweet sinful words. For Raoul.' Unusually, Joyce simply circled the word 'pocket' and it does not appear again; perhaps this is because he began shifting the emphasis here from the physical to the psychological dimension of the scene; this shift in emphasis is also evident in the ways Joyce rearranged the order of the phrases. At a later stage, Joyce also added in the left margin the names of the shop fronts that interrupt Bloom's sensual reverie.[96]

Furthermore, the aquatic ambience of the Ormond is established, for example, with descriptions that liken the counter of the bar to a 'reef' (p. [1v]) and the 'eau de Nil' motif Joyce added on p. [1r]. There are also many musical allusions that appear here already in the text that were not present in the earlier draft and Joyce continued to add even more. He also reassigned some of the dialogue between the barmaids and Dedalus, Lenehan and Boylan, but the same kind of character instability was already evident on the earlier draft; this issue requires further examination in 'Sirens' and in Ulysses more generally.

This manuscript also documents another transformation of the episode that is as momentous as it is problematic: the imposition of the so-called 'fugal' structure on the episode and the development of the episode's overture.[97] Although the conceptual note on p. [1r]: 'Repeat | episodes | phrases' signals the origin of the overture, its first extant version is as part of the Rosenbach manuscript.


[back to top]


Scenes and Fragmentary Texts for 'Cyclops': June 1919

(MS 36,639/10)


This manuscript contains at least four relatively lengthy, isolated scenes (mostly parodies), the episode's new opening, several unconnected fragmentary texts, as well as, unusually, an exchange of comic notes between James and Nora Joyce. Michael Groden postulated the existence of this copybook in his pivotal 1977 study of the evolution of the episode in 'Ulysses' in Progress.[98] Its discovery confirms many of Groden's insights and conclusions.[99] Naturally, it will also prompt further research and discussion of Joyce's creative methods and the several transformations Ulysses underwent in what Groden has designated as the 'middle stage' of the genesis of the work; this brief analysis is just one step in that direction.

As in the earlier 'Sirens' manuscript (NLI 36,639/07/B), presumably here too Joyce was compiling previously written distinct and fragmentary scenes and writing new fragmentary texts without a clear sense of how he would integrate this material into a continuous narrative of the episode.[100] In his discussion of Buffalo MS V.A.8–the sibling copybook of this manuscript stage of 'Cyclops'–Groden describes Joyce's 'habit of composing his material in blocks with only arbitrary attempts at transition or connection. This practice is basic to the entire copybook: the eight scenes constitute large blocks, and Joyce gives no indication of how he planned to connect them, even if he knew at the time'.[101] Groden's description applies to the various text blocks on this manuscript as well, and I would argue that this practice is a basic and fundamental aspect of Joyce's compositional method throughout his career. If Joyce had a plan for a continuous narrative for 'Cyclops' at this stage (and this seems unlikely), it was certainly different from the episode's later form. Therefore, this manuscript stage cannot be called an early 'draft' or even a 'proto-draft' without significant qualifications to what those terms usually denote. In these 'Cyclops' manuscripts we have important examples of the storehouse of material that allowed Joyce to work like an assemblagist before he wrote continuous narrative drafts of episodes.

Even with the discovery of this manuscript, much of the documentation of the evolution of 'Cyclops' is still missing. There were most likely other documents at this draft stage (like MS V.A.8 and this manuscript that were either other copybooks or loose pages). The next extant draft stage is Buffalo MS V.A.6,[102] which is also composed of non-sequential fragments, only some of which were copied from MS V.A.8 and this manuscript. Since the texts in MSS V.A.8 and V.A.6 are often significantly revised (at times based on BL Notesheet entries), at least one intermediary draft stage between these two extant stages must have intervened. Furthermore, there were probably ancillary documents of a similar kind as the later manuscript (MS V.A.6), though they too are now missing. Furthermore, since MS V.A.6 covers only a small portion of the episode, at least one draft stage intervened between it and the next extant draft stage, the Rosenbach faircopy manuscript.[103] The discovery of whatever draft stage Joyce wrote to produce the faircopy manuscript would shed a great deal of light on the evolution of this episode. The Rosenbach manuscript in turn was used to prepare the typescript for 'Cyclops' from which both The Little Review and Ulysses were set.

This copybook begins in mid sentence with the continuation of the 'eighth' fragmentary scene,[104] from Buffalo MS V.A.8, p. [24r], and both sections are in pencil.[105] In fact, Joyce numbered the conclusion of the scene '8' in blue crayon on p. [1r] here, just as he had the last-written page of the Buffalo manuscript. The main body text is heavily revised and there is much additional text in the margins. Neither the layout of the text on the page nor the BL Notesheet draft usage makes it possible to determine whether Joyce was copying the scene from an earlier version or whether it is a first draft, though Joyce's handwriting and the fact that he wrote in pencil suggest that the latter is more likely.[106] This scene ends about two-thirds of the way down the page with blank space below it.

Joyce then wrote a new scene at the top of p. [1v], again in pencil, that he intended to follow another fragmentary scene in Buffalo MS V.A.8, p. [22v], which he correspondingly tagged here '7) | b)'. He relied heavily on BL 'Cyclops' Notesheet entries to construct this short scene (to an even greater degree than he did for the rest of the scene on MS V.A.8), including the poem.[107] Joyce transcribed the entire scene from MS V.A.8, with its continuation here, on a missing intermediary document, which accounts for Joyce having crossed through the text in red crayon. He then transcribed it in the later draft of the episode (Buffalo MS V.A.6), but he did not include it in any subsequent extant draft or in Ulysses.

The next fragment on the page is even shorter, also in pencil, and in the form of a statement of dialogue attributed to an anonymous speaker, simply tagged '–',[108] about syphilis in the British Army. Joyce did not use this material as such again, though the theme persists (see U 12.1197), and it was incorporated with other material from this copybook in a later draft.

At the bottom of p. [1v] is the first bit of exchange between Joyce and Nora and the dialogue is continued that same day on p. [4v]. He playfully demands that she loan him '10 ^(or 5 frs)^ francs' to pay for the whiskey he has 'just stood' her: 'If not, be damned!' Joyce's note is imbued with the style of 'Cyclops' and the draft-like nature of the banter is evident in the way he revised the details as he normally would his creative work. His handwriting is noticeably different from the one he used to write either the pencil or ink drafts in this manuscript and in other Ulysses draft manuscripts. It resembles Joyce's note-taking hand, particularly as we know it from the many extant Buffalo Finnegans Wake notebooks.

On p. [2r], Joyce began a new, integral, and well-developed scene that is continued directly on pp. [3r] and then [4r] (see U 12.1675–735). I have been unable to locate any BL Notesheet entries or prior textual state of this elaborate parody of a church scene, but as Joyce wrote it in ink and only on the recto pages, this suggests that he may have written an earlier draft of this fragmentary scene elsewhere and transcribed and then further revised it here. Though this is the first unnumbered fragment in these early 'Cyclops' copybooks, it is crossed out in red crayon and so it seems most likely that Joyce transferred it to at least one missing document before the scene next appears in the Rosenbach manuscript version of the episode.

Joyce had left pp. [2v] and [3v] blank and then he wrote the uppermost block of text on p. [3v] first. Like the text block on p. [1v], it is in pencil and Joyce tagged it as an additional text for the 'sixth' fragment on MS V.A.8, p. [22r]. In fact, Joyce had marked the end of the scene there with instructions to himself: '(v. p. 28)'; that is, see p. [3v] here, which is yet another clue before 2002 that there must have been this companion manuscript. This is a scene about Molly, 'the ravenhaired daughter of Tweedy', which Joyce recopied on Buffalo MS V.A.6, pp. [4r]–[5r], with the rest of the 'sixth' scene (see U 12.1003–7).

Below it (and therefore almost certainly later), Joyce wrote another new scene, this time in ink, and continued it on the previous verso, p. [2v] (see U 12.1593–620). It is yet another parody scene: the arrival of Martin Cunningham, Jack Powers, and the variously-named Orangeman is told in a nineteenth-century style of the 'travellers' arriving at the 'rustic hostelry'.[109] Though both pages are again crossed out with red crayon, this scene does not appear in MS V.A.6. Nonetheless, it is virtually identical to the Rosenbach version and Ulysses, except for the adjective Joyce used to portray Bloom's countenance. It is twice described as 'olive' here but appears both times as 'pleasant' in subsequent versions.[110]

On p. [4v], the dialogue between Joyce and Nora is continued with her initial response to his demand for a loan on p. [1v]. Her reply begins 'In answer to yours of todays [sic] date' and raises an interesting question: how much of the texts on the intervening pages had Joyce written between his note and hers? It is impossible to know for certain, but a likely scenario is that Nora read Joyce's note on p. [1v] and replied on the first blank page she found. If this were the case, it would mean that Joyce had written at least two new, significant scenes as well as some other fragments all on the same day between their first exchange of notes. For whatever reason, she too adopts a parodic, business-like tone that suits 'Cyclops' and states that she 'regret[s]' being unable to 'advance' him 'the maximum sum' but is willing to provide the lesser amount he had suggested (that is, 5 rather than 10 Francs).

In one of the strangest cases of intertextual transference in the Joyce archive anywhere, he then replies to her with a line directly from his notes for Ulysses that begins 'The curse of a lopsided God light sideways [...]' and he continued the note, jokingly insulting her, but then signed himself 'yours affectionately | J J'.[111] At some stage, Joyce added the entire Ulysses text, with yet another line, in the left margin of MS V.A.8, p. [10r] in pencil. It is impossible to determine the exact priority of the Notesheet entries, the notes to Nora, and the addition to the draft, but this text next appears in the context of the 'syphilisation' theme from p. [1v] here in the Rosenbach manuscript and then in Ulysses at 12.1197–9. Nora has the last word in the matter when she echoes something Joyce had in fact written to Sir Horace Rumbold a year earlier. In Rumbold's name, she presents her 'compliments' to Mr Joyce and 'suggests that he shall go to Hell'.[112]

Finally, two years later (in the summer of 1921 as an addition to the printers' typescript), consciously or not, Joyce has the citizen in turn echo Nora's curse in this precise context:


–Their syphilisation, you mean, says the citizen. ^To hell with them!^ The curse of a goodfornothing God light sideways [...] (U 12.1197–8).


Finally, complaining of his ongoing work on 'Ithaca' in late June 1921, Joyce quoted this curse again to Frank Budgen: 'In the words of the Cyclops narrator the curse of my deaf and dumb arse light sideways on Bloom and all his blooms and blossoms. I'll break the back of Ithaca tomorrow so 'elp me fucking Chroist' (LIII 46).

Returning to this manuscript, Joyce wrote another parody on pp. [5r] and [6r], leaving p. [5v] blank. This is a mock Irish revivalist heroic epic about a 'sinewyarmed hero' and his 'savage animal of the dog tribe' that became Ulysses 12.151–205. The draft here is very similar to the text on the Rosenbach manuscript (though Joyce continued to amplify his description of the giant's appearance, especially on the proofs).

Without access to this manuscript, Groden wrote that, 'Joyce created the parodies first, the barroom scene soon after, and the narrative voice developed last'.[113] In fact, now we know that Joyce discovered both the first-person narrator's voice and developed the method of integrating it with the expansive parodic commentaries that define the episode's style when he wrote a new opening for the episode on pp. [7r]–[10r] of this manuscript. Although the entire scene has most of the elements that readers know from the published texts, this is very much a draft-in-process. Joyce revised it fluidly and heavily, at various times in ink and pencil, with his usual further additions in the left margin. As in Ulysses, the scene here begins with the unnamed 'I' meeting Joe Hynes and recounting the story about the near miss between the street sweeper's gear and his 'eye'. The episode's Homeric parallel with Polyphemus is now clearly established. In this draft, Joyce seamlessly combined the opening's anti-Semitic dialogue and what became the episode's first parody.[114] Naturally enough, Joyce had also already determined that Hynes and the narrator would visit Barney Kiernan's pub to see 'the citizen' and tell him about the meeting in the City Arms Hotel. Notably, this portion of the draft is only lightly revised.

The new opening of the episode is continuous in this draft up to Ulysses 12.67 at the top of p. [10r],[115] but then jumps directly to the scene where the pair enter Barney Kiernan's and meet the citizen and Garryowen already in situ (this scene begins at U 12.118). In the Rosenbach manuscript and Ulysses these narrative actions are interrupted by the 'land of holy Michan' digression. Joyce had already written this scene on Buffalo MS V.A.8, pp. [1r]–[2r] and probably intended it as the original opening of the episode.[116] There the mythic and parodic description of the area around Barney Kiernan's also served to introduce 'O'Bloom' ('the noble hero, ^the son of Rudolph^' [...]) as he walked on his errand to the pub. Joyce's solution in integrating the older and newer texts was simple enough: in a typically utilitarian fashion, he just broke up the older scene. He merely combined the first and third descriptive paragraphs (and they appear basically the same, though much expanded, on the Rosenbach version and then as U 12.68–99 and 12.102–17). On the other hand, he also reserved what he had already written about Bloom and later inserted it in what would be lines 12.215–7, about a hundred lines later in Ulysses.

For now, Joyce continued the narrative on pp. [10r]–[11r] in this manuscript. Here the 'Stand and deliver' exchange between the citizen and Hynes is virtually identical to the published versions, including the way the citizen rubs 'his hand in his eye'. Joyce also fluidly wrote Hynes's dialogue about his 'opinion of the times' with the citizen, and it is already remarkably similar to its appearance in Ulysses.[117] Joyce had also by now already determined the narrator's catchword, 'begob'; it appears for the first time on p. [11r]. Similarly, the dialogue as the men order drinks (some of it based on BL 'Cyclops' notes) and the reason Hynes has the money to pay for them is already established. So far, the narrative as we know it in Ulysses has proceeded in these pages uninterruptedly from U 12.118 to 12.147. Then, as noted above, by the time Joyce wrote the Rosenbach version of the episode, he had seamlessly integrated the mock-heroic description of the citizen and his faithful companion that he had written on pp. [5r]–[6r] in this copybook into this scene, with just some additional transitional material before and after the parody (see U 12.151–205).

The placement of the various elements of the text in these pages sheds light on Joyce's later deliberations about how to introduce Bloom in 'Cyclops'. On p. [11r] he wrote:


And begob he [Hynes] ^<outs> lands out^ with golden sovereign.

― Were you robbing a poorbox, Joe? says I.

― Sweat of my brow, says Joe. (U 12.206–11)


Then, as an addition on the facing page ([10v]), referring to an earlier scene in 'Aeolus', Joyce has Hynes say:


'Twas the prudent member gave me the wheeze.[118]

(Description of LB)


Presumably, the second line was Joyce's note to himself to add the unused material from the older description of Bloom in Buffalo MS V.A.8, p. [1r]. Most likely, Joyce first returned to the earlier text and revised it in pencil, but this did not suit him either. So he also rewrote Bloom's entrance in full (just as it appears in Ulysses) right below the note and also added the narrator's insulting account of his first sighting of Bloom (see U 12.213–14) right above, all on p. [11v]. Joyce pieced all the elements together by the time he wrote the Rosenbach manuscript and it remained virtually unchanged from then on.

Hynes's revelation that he had been paid by 'the old woman of Prince's street' sets the citizen off on his rant about the Irish Independent on pp. [11r] and [12r] (see U 12.218–43). The entire scene is quite similar to the subsequent versions, including the list of names read aloud. Although Joyce did not use any of the extant BL Notesheet entries for this list, it is based on the Irish Daily Independent for 16 June 1904.

The episode's new opening ends towards the bottom of p. [12r]. Joyce later drew a short centred horizontal line separating it from yet another parodic text, the execution scene (see U 12.525–678). The sketch continues directly on pp. [13r]–[15r], with additions in the margins and on pp. [13v] and [14v].[119] This relatively long scene here is followed by two unrelated further fragments and the rest of the copybook is blank, except for yet another list of incidents on its final page that accounts for the arrival of some of the characters at Barney Kiernan's.


[back to top]


To be continued.





Manuscript Collections:




British Library, London


Poetry Collection, University at Buffalo, State University of New York


National Library of Ireland, Dublin


Rosenbach Museum & Library, Philadelphia, Pennsylvania



Works by Joyce:


James Joyce Archive, edited by Michael Groden et al. 63 volumes (New York: Garland, 1977–8); cited by volume and page number.


Letters of James Joyce. Vol. I, edited by Stuart Gilbert (New York: Viking, 1957; reissued with corrections 1966). Vols. II and III, edited by Richard Ellmann (New York: Viking, 1966).


Selected Letters of James Joyce, edited by Richard Ellmann (New York: Viking, 1975; London: Faber, 1975).


Ulysses, edited by Hans Walter Gabler et al. (New York and London: Garland, 1984, 1986), also published by Random House, Bodley Head, and Penguin; cited by episode and line number.



Secondary Sources:


Genetic Joyce Studies


Joyce's Notes and Early Drafts for 'Ulysses': Selections from the Buffalo Collection, edited by Phillip F. Herring, (Charlottesville: University Press of Virginia, 1977).


Richard Ellmann, James Joyce, revised edition (New York: Oxford University Press, 1982).


Rodney Wilson Owen, James Joyce and the Beginnings of 'Ulysses' (Ann Arbor: UMI Research Press, 1983).


James Joyce Quarterly


Joyce's 'Ulysses Notesheets in the British Museum, edited by Phillip F. Herring, (Charlottesville: University Press of Virginia, 1972).


The Workshop of Daedalus, edited by Robert Scholes and Richard M. Kain, (Evanston, Illinois: Northwestern University Press, 1965 (see


[back to top]



Census of the Extant Ulysses Holograph Manuscripts:


Dates & Place of Composition:


'Telemachus' (1):

Faircopy for Typescript

September–October 1917: Zurich


'Nestor' (2):

Faircopy for Typescript

October–early December 1917: Zurich


'Proteus' (3):

Earlier Proto-draft

Fall 1917: Zurich

NLI 36,639/07/A

Later Draft

Fall 1917: Zurich

Buffalo V.A.3

Faircopy for Typescript

December 1917: Zurich


'Calypso' (4):

Faircopy for Typescript

February 1918: Zurich


'Lotus Eaters' (5):

Collateral Faircopy

January–May 1918: Zurich


'Hades' (6):

Collateral Faircopy

January–July 1918: Zurich


'Aeolus' (7):

Collateral Faircopy

January–August 1918: Zurich


'Lestrygonians' (8):

Collateral Faircopy

1918: Zurich


'Scylla and Charybdis' (9):

Earliest Draft

Early to mid 1918: Zurich

NLI 36,639/08/A &

Earliest Draft


NLI 36,639/08/B &

Earliest Draft


NLI 36,639/08/C

Collateral Faircopy

Late 1918: Zurich


'Wandering Rocks' (10):

Faircopy for Typescript

January–February 1919: Zurich


'Sirens' (11):

Partial Early Draft & Fragments

Late 1917–early 1918: Zurich

NLI 36,639/07/B

Later Draft

January–May 1919: Zurich

NLI 36,639/09 &

Later Draft


Buffalo V.A.5

Collateral Faircopy

June 1919: Zurich


'Cyclops' (12):

Earlier Fragmentary Texts

June 1919: Zurich

Buffalo V.A.8 &

Earlier Fragmentary Texts


NLI 36,639/10

Later Draft of Texts

June–July 1919: Zurich???????????????????

Buffalo V.A.6

Additional Text


Buffalo V.A.7

Faircopy for Typescript

August–September 1919: Zurich


Additional Manuscript for Proofs

October 1921

Buffalo V.A.9

'Nausicaa' (13):

Early Draft

November 1919–January 1920: Trieste

Buffalo V.A.10 &

Early Draft


Cornell 56A &

Early Draft


Cornell 56B

Mixed Faircopy for Typescript

January–February 1920: Trieste


'Oxen of the Sun' (14):

Earlier Draft

February–March 1920: Trieste

Buffalo V.A.11 &


Earlier Draft


Buffalo V.A.12 &

Earlier Draft


NLI 36,639/11/A &

Earlier Draft


NLI 36,639/11/B

Later Draft

February–March 1920: Trieste

Buffalo V.A.13 &

Later Draft


Buffalo V.A.14 &

Later Draft


NLI 36,639/11/C &

Later Draft


Buffalo V.A.15 &

Later Draft


NLI 36,639/11/D &

Later Draft


Buffalo V.A.16 &

Later Draft


Buffalo V.A.17 &

Later Draft


Buffalo V.A.18 &

Later Draft


NLI 36,639/11/E &

Later Draft


NLI 36,639/11/F

Pre-faircopy Fragment

March–April 1920: Trieste

Cornell (Uncatalogued)

Mixed Faircopy for Typescript

May 1920: Trieste


'Circe' (15):

Earlier Draft

January–July 1920: Trieste

Buffalo V.A.19

Intermediary 'Quinn' Draft

July–December 1920: Paris

NLI 35,958

Later Draft

July–December 1920: Paris: Paris

NLI 36,639/12

Faircopy for Typescript

December 1920–January 1921: Paris


Additional Manuscript for Proofs

September 1921: Paris

Buffalo V.A.20

'Eumaeus' (16):

'Eumeo' Manuscript: Earlier Draft

January–February 1921: Paris

Private Collection

Later Draft

Mid February 1921: Paris

Buffalo V.A.21

Faircopy for Typescript

Mid to Late February 1921: Paris


'Ithaca' (17):

Early Proto-draft

March–August 1921: Paris

NLI 36,639/13

Faircopy for Typescript

November 1921: Paris


'Penelope' (18):

Early Draft

Early summer 1921: Paris

NLI 36,639/14

Faircopy for Typescript

July–September 1921

Rosenbach &

Faircopy for Typescript


Buffalo V.A.21


[back to top]

[1] I would like to thank Dirk Van Hulle, Geert Lernout, Hans Walter Gabler, Ronan Crowley, Terence Killeen, and Daniel Ferrer for their insightful comments and suggestions.

[2] I wrote a previous version of this survey to accompany the 'Ulysses in Process' installation in the 'James Joyce and Ulysses at the National Library of Ireland' exhibition (Dublin, June 2004–March 2006), which I curated with Catherine Fahy and Katherine McSharry.

[3] The 'Joyce Papers 2002' are NLI MSS 36,639/01–19 and this study covers MSS 36,639/01–10. Peter Kenny prepared an initial catalogue of all the 'Joyce Papers 2002' and it is available online at I am preparing a revised, complete catalogue of the Joyce manuscripts at the NLI.

[4] See Michael Groden, 'Ulysses' in Progress, (Princeton: Princeton University Press, 1977), pp. 115–65.

[5] I completed the catalogue of the Buffalo Joyce Collection in September 2010 and it is available online at

[6] See Jean-Michel's Rabat?'s discussion of the 'genreader' in James Joyce and the Politics of Egoism (Cambridge: Cambridge University Press, 2001).

[7] See Daniel Ferrer's recently-published rich and provocative analysis of the potential of genetic criticism: Logiques du brouillon: model?s pour une critique g?n?tique (Paris: ?ditions du Seuil, 2011).

[8] Hans Walter Gabler, with Wolfhard Steppe and Claus Melchior, Ulysses: A Critical and Synoptic Edition (New York and London: Garland Publishing Inc., 1948, 1986), vol. 1, p. 152.

[9] This complex topic requires a monograph-length study, so this section is only meant as a simplified pr?cis; hence the descriptions here are deliberately generic and the number of citations have been kept to a minimum.

[10] See the appendix, a Census of extant Ulysses holograph manuscripts for further details about all the extant manuscripts. I have catalogued all of the Ulysses manuscripts at Buffalo; see

[11] See Christie's [Chris Coover et al.], 'James Joyce's Ulysses: The John Quinn Draft Manuscript of the ?Circe? episode.' Christie's, New York, Thursday, 14 December 2000: this is now NLI MS 35,958.

[12] See Sotheby's [Peter Selley et al.], 'The Lost 'Eumaeus' Notebook: James Joyce, Autograph Manuscript of the 'Eumaeus' Episode of Ulysses.' Sotheby's London, July 10, 2001.

[13] The Early Commonplace Book (NLI MS 36,639/02; see pp. [19r]–[21v], for example) is a very early example of Joyce's method of working from prior notes. At the other end of Joyce's career we have the fifty-odd 'Work in Progress'/Finnegans Wake Notebooks, virtually all of which are at Buffalo; see

[14] Frank Budgen, James Joyce and the Making of 'Ulysses' and Other Writings, with an introduction by Clive Hart (London: Oxford University Press, 1972), p. 176.

[15] Budgen, Making of 'Ulysses', pp. 176–7.

[16] With the notable exception of 'Proteus', before the discovery of the new NLI manuscripts, all of the then known manuscripts cluster around the middle of the book from 'Sirens' to 'Eumaeus'. Given the fact that Joyce had been working on Ulysses since at least 16 June 1915 (about five and half years before Ulysses was published), the manuscripts that were then known to survive are all relatively late, dating from 1917 to early 1921. In 1999, there was no documentary evidence of the early genesis of episodes 4 to 9, 'Calypso' through 'Scylla and Charybdis'. Since Joyce began to hold on to his working drafts relatively late Joyce began to hold on to his working drafts relatively late, most likely what has survived for these episodes is just based on chance. More often than not he simply disposed of the manuscripts once they had been recopied because he no longer needed them.

[17] Various typists used the Rosenbach faircopy manuscripts for only ten episodes: 'Telemachus' through 'Calypso', 'Wandering Rocks', 'Cyclops', and 'Circe' through 'Penelope'.

[18] Harriet Shaw Weaver marked up her own copies of The Little Review for the printers to set up the few sections of Ulysses that appeared in The Egoist.

[19] There are numerous and sometimes substantial variations in the text between the Rosenbach manuscripts and the typescripts for certain episodes, but there is not enough textual or contextual information to determine the precise relationships between the extant individual 'faircopy' manuscripts and the typescripts. It is possible that certain now lost manuscripts produced both the Rosenbach manuscript versions as well as the typescripts at different times; or else, one or more subsequent documents intervened between the Rosenbach manuscripts and typescripts, but the situation may be different with different episodes.

[20] Budgen, Making of 'Ulysses', p. 22.

[21] Ronan Crowley and I prepared a census of all the Ulysses proofs that was published in GJS (Issue 8), 2008, as 'The Ulysses Proof^finder'; see Proofs by Episode. Also, see Crowley's introduction to the Placards for further, specific information on these important documents in the genesis of Ulysses.

[22] Sylvia Beach, Shakespeare and Company (New York: Harcourt, Brace, 1959), p. 47.

[23] Almost six months after Ulysses was published, Beach recorded that the printing of Ulysses had cost 42,492 Francs (twice as much as Darantiere had originally estimated) and the postage for shipping this bulky tome had cost 3,200 Francs so far. By then 39,505 Francs had been paid to Joyce, which is almost thirty percent of the net receipts of the book's sales from 19 May 1921 to 27 July 1922, whereas Beach had only received 13,978.80, or just about ten percent. (See Beach's 'ULYSSES | Account Rendered' memorandum: Buffalo MS XVIII: Miscellaneous Material Related to Joyce's Works, E.1, folder 21.)

[24] See WD 3; The Critical Writings, edited by Ellsworth Mason and Richard Ellmann (New York: Viking, 1968), pp. 15–16.

[25] Dirk Van Hulle, 'Joyce and Beckett Discovering Dante', Joyce Studies 2004, Number 7, edited by Luca Crispi and Catherine Fahy (Dublin: The National Library of Ireland, 2004). It is not known which printing of this Divina Commedia Joyce used but the pagination between them remained consistent.

[28] Wim Van Mierlo's The Subject Notebook: A Nexus in the Composition History of Ulysses-A Preliminary Analysis is a pivotal essay on this manuscript. It was published in GJS Issue 7 (Spring 2007) and the reader is referred there for information on the sources of these notes.

[29] The manuscript has been reproduced in colour photo-facsimile on JJA 7.109–56. It has been transcribed and annotated in WD 92–105.

[30] This manuscript was previously catalogued as Buffalo MS VIII.A.5.

[31] See The manuscript has been reproduced in colour photo-facsimile on JJA 12.129–66. It was transcribed and annotated in JNEDU, pp. 3–33.

[32] The vast majority of the Buffalo Finnegans Wake notebooks are 'first-order' compilations; see

[33] There was at least one further early Ulysses notebook (that was also compiled in 1918 and in tandem with Buffalo MS V.A.2.a), but it survives only in the form of a partial amanuensis's transcription in Buffalo Finnegans Wake MS VI.C.16, pp. [232]–[274]. It is also catalogued as a so-called 'Missing Notebook', Buffalo MS VI.D.7; see; and JJA 42.348–59. Danis Rose and John O'Hanlon prepared an edited version of Buffalo MS VI.C.16 as The Lost Notebook: New Evidence on the Genesis of 'Ulysses' (Edinburgh: Split Pea Press, 1989).

[34] These manuscripts have been reproduced in colour photo-facsimile on JJA 12.02–95. The manuscripts were transcribed and annotated in UNBM, see p. 526.

[35] See The manuscript has been reproduced in colour photo-facsimile on JJA 12.97–125. It was transcribed and annotated by Phillip Herring in JNEDU, pp. 37–118

[36] For an invaluable discussion of this little-known period of Joyce's writing, see Rodney Owen's JJBU.

[37] For example, the text-blocks that constitute the proto-draft of 'Proteus' (NLI MS 36,639/07/A, pp. [1r]–[5r]) as well as the fragments at the end of the earlier 'Sirens' draft (NLI MS 36,639/07/B, pp. [10r]–[14r]) may be indicative of the earlier state of these transitional texts.

[38] Significantly, although Joyce had already settled on the character of Molly Bloom (or 'Mollie' as she is named here on p. [2r] under the heading 'Leopold'), she is virtually absent from these notes.

[39] Joyce's most consistent method of heading a notebook can be seen in NLI MS 36,639/05/B. There he usually headed only the recto page and then used the facing verso pages for further notes when the recto pages were full.

[40] Also, the notes under the heading 'Theosophy' on p. [7r] are, again unusually, continued on p. [7v] rather than the blank p. [6v].

[41] There are similar catchall heading at the end of Buffalo Finnegans Wake Notebook VI.A (the so-called 'Scribbledehobble' Finnegans Wake notebook): 'Books' (twice), 'Words' and 'Names'. See The inscribed pages of this manuscript have been reproduced in black and white photo-facsimile in JJA 28.01–253 and some pages were also reproduced in colour on 28.255–86. An edited preliminary version of the manuscript has been published as James Joyce's Scribbledehobble: The Ur-Workbook for 'Finnegans Wake' by Thomas Connolly (Northwestern University Press: 1961).

[43] The NLI's manuscript numeration (as '04', '05/A', and '05/B') was assigned by Peter Kenny, but was based on earlier and only partial analyses by Michael Groden and others. (See Michael Groden's 'The National Library of Ireland's New Joyce Manuscripts: A Statement and Document Descriptions', JJQ Vol. 39, Number 1 (Fall 2001 [February 2003]), pp. 29–51, as well as 'The Archive in Transition: The National Library of Ireland's New Joyce Manuscripts' in Michael Groden, 'Ulysses' in Focus: Genetic, Textual, and Personal Views (Gainesville: University Press of Florida, 2010), pp. 14–31.) These designations are misleading in several respects, but cannot be altered at this stage, though my catalogue will address this issue in a more comprehensive manner. Most importantly, obviously, these are three distinct manuscripts and so they should have been listed as MSS '04', '05', and '06' (with all subsequent manuscripts being assigned a +1 number). Based on my current research, I would tentatively order the manuscripts in the following order: MS 36,639/05/B first, probably 5/A second, and therefore 4 last.

[44] See Geert Lernout's recently published essay on the use of search engines to track down note sources as well as inter-textual references in Joyce's writings: 'Joyce World-Wide Intertext', JJQ Vol. 47, Number 2 (Winter 2010), pp. 247–53.

[45] I am indebted to Ronan Crowley for sharing his original work on 'Circe', most of which is still unpublished. He was the first to determine the complex relationship between the NLI notebooks and the BL 'Circe' Notesheets. See "His Dark Materials": Joyce's "Scribblings" and the Notes for 'Circe' in the National Library of Ireland' in GJS Issue 6 (Spring 2006).

[46] This particularly important letter should be read in its entirety to appreciate Joyce's state of mind as he set about the arduous task of finishing Ulysses.

[47] This is Ellmann's translation; see SL 275–7.

[48] Further work on the printed sources of these notes may indicate how many of them are based on Joyce's reading in Paris in 1920–1.

[49] In fact, there are only five other pages of notes for 'Penelope' all together in the other two NLI notebooks, though, of course, there are seven BL Notesheets for the episode (according to Herring's numbering; see UNBM 490–517), as well at least seven-and-a-half pages in Buffalo MS V.A.2.b, pp. [1r]–[3v], [4v], and [19v] (see JNEDU 55–73, 77–9, and 116), which further reinforces the connection between these two late notebooks.

[50] Along with all the other NLI Ulysses notebooks, Joyce passed this notebook on to his amanuensis, Mme France Raphael, who transcribed it (as Buffalo MS VI.C.7, pp. [235]–[254]) in early 1935 while he was writing 'Work in Progress'/Finnegans Wake. See; JJA 41.431–6. It is odd that the Buffalo notebook (MS V.A.2.b) was not similarly transcribed with these other Ulysses notebooks. On the other hand, the fact that this notebook was transcribed after the earlier Ulysses notebook (NLI MS 36,639/03) and before the other late notebooks (NLI MSS 36,639/05A and 5/B) seems simply to have been a coincidence.

[51] Along with all the other NLI Ulysses notebooks, fourteen years after he had compiled it, Joyce also passed on this notebook to his amanuensis, Mme France Raphael, who transcribed it (as Buffalo MS VI.C.7, pp. [136]–[198]) in early 1935 as he worked towards finishing 'Work in Progress'/Finnegans Wake. See; JJA 41.406–22.

[52] That is, Joyce compiled and sorted it from previous notes as well, along with the BL Notesheets (for 'Cyclops' through 'Penelope' only), the NLI MSS 36,639/04 and 5/A, and the even later Buffalo MS V.A.2.b.

[53] Since only some of the manuscripts for those episodes survive, Joyce's various uses of these notebooks can also be inferred by collation.

[54] See Sotheby's [Peter Selley et al.], 'The Lost 'Eumaeus' Notebook: James Joyce, Autograph Manuscript of the 'Eumaeus' Episode of Ulysses.' Sotheby's London, July 10, 2001.

[56] The same pattern holds for those parts of the Rosenbach manuscript of the episode that are not covered by Buffalo MS V.A.21.

[57] Along with all the other NLI Ulysses notebooks, Joyce also passed on this notebook to his amanuensis, Mme France Raphael, who transcribed it (as Buffalo MS VI.C.7, pp. [202]–[234]) in early 1935 while Joyce was writing 'Work in Progress'/Finnegans Wake. See; JJA 41.423–31.

[60] The discovery of this manuscript was the impetus for my current project, Becoming the Blooms.

[61] Fragments 6 and 7 are not separated by Xs, presumably because there was no room at the bottom of p. [2r], but clearly they are distinct textual fragments. The number of Xs Joyce used varies but that does not seem to be significant.

[62] For example, see the visit with Uncle Richie (U 3.70–103 on p. [2v]) as well as the midwives scene (U 3.29–44 on [3v]).

[63] See U 3.271–81 and 3.286–9 (on p. [1r]), 3.107–24 and 3.48–52 (on p. [2r]), 3.303–9 (on p. [4r]), 3.313 & 316–30 (on p. [4v]) and 3.406–18 on (pp. [4v–5r]).

[64] For example, see Stephen's recollections of Paris waking and of ^<Joe> Kevin^ Egan (U 3.209–57) that is composed of two distinct fragments (on pp. [5r], [3r] and [3v] that are separated by two pages, the latter fragment preceding the first text in this manuscript).

[65] 'What Song the Sirens Sang ... Is No Longer Beyond All Conjecture: A Preliminary Description of the New 'Proteus' and 'Sirens' Manuscripts', JJQ Vol. 39, No. 1 (Fall 2001), pp. 53–67. See also Sam Slote's essay in GJS Issue 5 (Spring 2005), Epiphanic 'Proteus'.

[66] In this respect they resemble all of the extant 'Cyclops' manuscripts (Buffalo MS V.A.8 and NLI 36,639/10), the later portion of the early 'Sirens' draft (MS 36,639/07/B), as well as the early 'Ithaca' manuscript (NLI MS 36,639/13), for example.

[67] Such as the fragmentary narrative style and typographical layout of the 'Wandering Rocks' episode. Joyce claimed he composed the Rosenbach draft of this episode from 'notes' that may even have resembled these fragments. But Joyce wrote that episode at least one year later and after Ulysses had undergone significant elaboration.

[68] David Hayman's description of Joyce's 'piecemeal or mosaic' method of composition on 'Work in Progress'/Finnegans Wake seems particularly applicable to this draft of 'Proteus'. He writes that this method 'was useful when the passage depended more heavily upon ornament and logic than upon plot development. Working from some sort of rough plan or at least from a coherent concept, Joyce wrote a series of unintegrated passages which, when their number was significant, he organized into a unit'. David Hayman, A First Draft Version of 'Finnegans Wake', (Austin: University of Texas, 1963), p. 13.

[69] Buffalo MSS I.A; see The rectos and versos of this manuscript have been reproduced in black and white photo-facsimile on JJA 7.01–44.The manuscripts were published first in Epiphanies, edited by Oscar A. Silverman, with introduction and notes (Buffalo, New York: University of Buffalo, 1956; and reprinted in 1979 [Snyder, New York: Richard West]).

[70] Although this manuscript is composed of at least two distinct parts, Joyce nonetheless numbered the entire 'Sirens' section of the copybook continuously as pp. '1'–'18'; the final page (with the Homeric notes) is unnumbered; and the loose leaf is foliated as p. '20'. These facts indicate that Joyce had probably already compiled the last page of notes before he started this draft, and that another page of the draft (what must have been p. '19') is still missing. It is also possible that further loose pages of this draft stage followed p. '20'.

[71] Based on the way in which Joyce used the pages in other, later copybooks, he usually started by only using the rectos when writing the main text of first or earlier drafts, reserving the versos for additional texts. Generally, it is only when he thought a text had reached a relatively stable state that he would use the recto-verso-recto pages sequentially, like here. Of course, this manuscript proves that Joyce can be quite mistaken about the relative stability of his work.

[72] Owen also wrote that, 'Because the 'Alphabetical Notebook' and Giacomo Joyce were used in Ulysses prior to Joyce's return to Trieste in 1919, and because they evidently remained in Trieste since 1915, episodes which used them were likely sketched before 1915' (JJBU 65). This evidence is less convincing than the developed nature of the text in this manuscript for the presumption that at least one version of 'Sirens' pre-dates this draft (and so was written before 1917). In fact, only one note from the so-called 'Alphabetical Notebook' (Cornell MS 25) appears in this manuscript (on p. [8r]): 'Mooney sur mer' (Joyce had noted it under the heading 'Devin'; see WD, p. 96), but Joyce could easily have transferred that note to an intermediary notebook or notesheet (or simply recalled the colloquial name of the pub) and so did not need direct access to that notebook.

[73] As part of the stylistic overhaul the episode had undergone, on the next draft Joyce echoed the phrase (see MS 36,639/09, p. [3r]) just as it appears in Ulysses (11.214) and then on the Rosenbach manuscript Joyce repeated it a third time (now as a noun: U 11.217). Interestingly, none of these repetitions are indicated as additions in either of the extant drafts.

[74] A more detailed and expanded version of this argument is forthcoming.

[75] See the colour reproduction of the Rosenbach 'Scylla and Charybdis' manuscript, f. '37' in Ulysses: A Facsimile of the Manuscript, volume I (New York: Octagon Books, 1975).

[76] Although all of their early drafts are missing, Joyce certainly had written parts of 'Telemachus', 'Nestor', and 'Hades' by 1917.

[77] The 'Monsieur de la Palice' and the '[...] brave medicals [...]' entries appear on pp. [18v] and [24v].

[78] See Hans Walter Gabler, 'The Rocky Road to Ulysses', Joyce Studies 2004, Number 15, edited by Luca Crispi and Catherine Fahy (Dublin: The National Library of Ireland, 2005), particularly pp. 26–9.

[79] Sometime between 1916 and 1918, Joyce compiled a notebook (Buffalo MS V.A.4; see, which he labelled 'Shakespeare | Dates'. As the title indicates, it is composed of notes on the life of William Shakespeare from 1593 to 1616; each of the twenty four pages covers one year (except for 1612) and was compiled from Sidney Lee's works on Shakespeare. This notebook is not textually related to this draft. The manuscript has been reproduced in black and white photo-facsimile on JJA 12.323–348 and Richard M. Kain transcribed it in 'James Joyce's Shakespeare Chronology',"Massachusetts Review Vol. 5, No. 2 (1964), pp. 342–55. Also, see Rodney Wilson Owen's incisive treatment of this manuscript and its relation to A Portrait and the 'Scylla and Charybdis' episode in JJBU 91–2.

[80] It is possible that by 1917 Joyce had already transferred it to another note-repository and its draft usage here does not come directly from Cornell MS 25; see WD 95.

[81] The notes come from pp. [7r], [7v], and p. [15v], the final page in the notebook.

[82] In 2009 Ronan Crowley discovered that these notes came from Alexander Schmidt, Shakespeare-Lexicon: A Complete Dictionary of All the English Words, Phrases and Constructions in the Works of the Poet. 3rd. ed. rev. and enl. by Gregor Sarrazin. Vol. II. M-Z (Berlin: Reimer, 1902).

[83] See also Peter Spielberg (compiler), James Joyce's Manuscripts and Letters at the University at Buffalo (Buffalo: University at Buffalo, 1962), p. vii.

[84] Bernard Gheerbrant, James Joyce: Sa Vie, Son ?uvre, Son Rayonnement (Paris: Librairie La Hune, 1949), n.p., item 254. See also John J. Slocum and Herbert Cahoon, A Bibliography of James Joyce 1882-1941 (1953; rpt. Westport, CT: Greenwood, 1971), Item 5.b.iii, p. 140. Slocum and Cahoon's description is based on 'hasty notes made by one of the compilers during a private examination of the library in the spring of 1949'.

[85] The use of two differently coloured crayons usually indicates that Joyce returned to the manuscript at various times.

[86] The missing texts include the following: U 9.96–9, 9.137–8, 9.158, 9.163, 9.183, 9.221–4, 9.308–13, 9.381–5, 9.506, 9.515, 9.517–21, 9.651–4, 9.674–80, 9.729–30, 9.754–7, 9.889–91, 9.999–1006, and 9.1072–80.

[87] See Hans Walter Gabler, with Wolfhard Steppe and Claus Melchior, Ulysses: A Critical and Synoptic Edition (New York and London: Garland Publishing Inc., 1948, 1986), 3 volumes; especially vol. 3, pp. 1876–82.

[88] It is likely that the Rosenbach manuscript of the 'Nausicaa' and of 'Oxen of the Sun' episodes are also a mixed manuscripts, which makes them analogous to the textual states of this 'Scylla and Charybdis' draft and its Rosenbach version.

[89] There may have been more than one draft and these may have been manuscripts or typescripts or both.

[91] The episode's textual history is further complicated by the fact that this draft level (NLI MSS 36,639/09 and Buffalo MS V.A.5) is not the direct source of the next extant draft, the Rosenbach manuscript, and that the Rosenbach manuscript was not the source text for the typescript, which in turn was used for The Little Review, and then after further revisions for Ulysses as well.

[92] Another difference between the draft stages is that, for whatever reason, Joyce named one of the barmaids both Miss 'Douse' and 'Douce' on the earlier version but in this draft she is consistently 'Miss Douce'.

[93] Although Joyce knew the circuit of Bloom's itinerary and errands, he added some of his contrapuntal thoughts in the margins. Similarly, Boylan's journey to the Ormond and then on towards Eccles Street is also well developed in this draft. Joyce had also determined that Bloom would follow Boylan to the Ormond before this draft. Here it reads: '[...] Can't see me there. [...] Still be near and hear: At four.' (p. [5r]).

[94] Presumably, this happened after Joyce had found the factual detail he needed among his notes or elsewhere: the shop was actually located at 31 Wellington Quay.

[95] The following are typical examples of Joyce's method: '^<That fellow is most aggravating> Most aggravating that fellow is^.' and '^<His dark eyes went by by Bassi's blessed virgins,> by Bassi's blessed virgins, his dark eyes went by^' (both on p. [2r]). Joyce's rearrangement of the same phrases in different orders was already conspicuous on the earlier draft, but became more intense here and continued at every subsequent level.

[96] Interestingly, whether purposefully or not, the order in which Bloom would have passed the shop fronts is inverted here and remains so in each subsequent version and in Ulysses. See Thom's 1904, pp. 1615–6; as well Ian Gunn and Clive Hart with Harald Beck, James Joyce's Dublin: A Topographical Guide to the Dublin of 'Ulysses' (London: Thames and Hudson, 2004), p. 120.

[97] For further information on these notes, see Susan Brown's essay in GJS Issue 7 (Spring 2007), The Mystery of the Fuga per Canonem Solved and Michelle Witten's further contribution to the debate in GJS Issue 10 (Spring 2010), The Mystery of the Fuga per Canonem Reopened?.

[98] Michael Groden, 'Ulysses' in Progress, (Princeton: Princeton University Press, 1977), p. 119, n. 6. The reader is referred to Groden's works for a comprehensive analysis of the entire evolution of the episode and of Joyce's activities in 1919, see 'Ulysses' in Progress, pp. 115–65.

[99] See also Michael Groden, 'Cyclops in Progress, 1919', JJQ Vol.12 Nos. 1/2 (Fall 1974/Winter 1975), pp. 123–68 and 'Joyce at Work on 'Cyclops': Toward a Biography of Ulysses', JJQ Vol. 44, No. 2 (Winter 2007), pp. 217–45.

[100] According to Groden, Joyce 'began the episode [in Buffalo MS V.A.8] without a clear idea of the technique he would use; he even planned briefly to continue the monologue method' ('Ulysses' in Progress, p. [115]).

[101] Groden, 'Ulysses' in Progress, p. 131. Buffalo MS V.A.8 has been reproduced in black and white photo-facsimile on JJA 13.83–132; for further information about this draft stage of 'Cyclops'; see the catalogue description of this manuscript:

[102] See the catalogue description of this manuscript for further information about this later draft stage of 'Cyclops':

[103] Also, Buffalo MS V.A.7 contains an addition to the text on V.A.6; see

[104] This scene is the courtroom parody; see Ulysses 12.1111–40.

[105] The first four scenes in MS V.A.8 are in ink and the later four are in pencil, but in this NLI copybook ink and pencil fragments alternate in a more random manner.

[106] Like the rest of the scene, Joyce relied heavily on the BL 'Cyclops' Notesheets to write this fragmentary scene; here specifically Notesheets 1, 3–5, and 10.

[107] See BL 'Cyclops' Notesheet 3.36–9.

[108] See Groden's discussion of Joyce's use of symbols to tag dialogue that he only later assigned to characters: 'Ulysses' in Progress, p. 117.

[109] This scene must be the one Joyce refers to in his list of incidents (Buffalo MS V.A.7) simply as 'Arrival Martin'.

[110] Like so much of the evolution of 'Cyclops', this change happened on one or more missing documents.

[111] Joyce noted the curse twice: 'lopsided God' appears on BL 'Cyclops' Notesheet 1.43, and the fuller text quoted above (that is continued 'light sideways on the bloody thicklugged sons of whore's gets') is found on BL 'Cyclops' Notesheet 7.47–8.

[112] Rumbold was the British Minister to Bern, Switzerland, and Joyce and he had several unpleasant encounters; see JJII 447, 459, and elsewhere for Ellmann's version of the account.

[113] Groden, 'Ulysses' in Progress, p. 124.

[114] Although Joyce relied on a few BL Notesheet entries for the narrator's part of the dialogue, no notes have been located for the pseudo-legalistic parody that he wrote directly as part of the opening (see U 12.32–51). Nonetheless, he must have relied on Thom's to write it since Joyce noted both correct addresses for Herzog and Geraghty.

[115] It is also remarkably similar, as subsequently revised here, to the published version.

[116] Joyce tagged it with a '1)' in blue crayon and he noted it as number '1' in the other list of incidents at the end of this copybook (p. [22v]).

[117] This exchange is based on BL 'Cyclops' 1.73–9. In 1966, Hugh Staples identified the source of dialogue here as Alexander M. Sullivan's New Ireland: Political Sketches. Also see UNBM 19–20.

[118] See U 12.211–2.

[119] In early October 1921, Joyce wrote the 'Dublin Metropolitan police' addition (Buffalo MSS V.A.2 and then V.A.9) that the printers added to the first setting of 'Cyclops' in proofs (that is, the first placards).