APPENDIX A

Editing The Grail Mass Manuscripts

In his notes to The Roman Quarry, Grisewood described his edition as simply one version of many that could have been constructed from Jones’s manuscripts. The Grail Mass presents one of those alternative versions, and the editors wish to acknowledge a debt to Hague and Grisewood.

Our intention in constructing the present edition has been to provide the reader with an edition of Jones’s project in what we consider its most complete and reader-accessible form. The structure presented here is that of the original project on which Jones was working before he undertook The Anathemata, and to which he returned after its publication: first considering it a continuation, then a separate work and finally a series of related sequences. We treat Jones’s previously published ‘fragments’ and The Grail Mass as independent texts. This new edition makes no claim to establishing a definitive edition that reflects Jones’s final intentions for this body of material. In fact, we would argue that Jones had no stable ‘final’ intention for his project, save trying to complete it in some form.1

Structurally, Jones undertook two types of additions: in a number of places he re-arranged already written sections to fill in or to minimize narrative gaps caused by the earlier extractions; he also introduced at least three insertions, two into ‘On the Traverse of the Wall’, and at least one into ‘The Agent’ where he extracted material from the Crucifixion scene and his dinner conversation. Actual changes in ‘content’, what Jones referred to as ‘matter’, tend to be minimal, most often consisting of altering or adding a single word, changing a name or amending a phrase. When there are substantive changes between ‘matter’ used in a published ‘fragment’ and that same ‘matter’ as found in the unpublished project, they tend to be one of three types. First, some changes indicate Jones’s apparent attempt to obscure the original narrative unity amongst the published Roman ‘fragments’. Second, Jones made some substantive changes to older material during the post-Anathemata re-workings of it as a result of new historical evidence he had subsequently discovered, which he felt invalidated the original content.2 Third, after 1960 and with the publication of ‘The Tutelar of the Place’, Jones’s published Celtic ‘fragments’ increasingly employed the Welsh language.3 Based on the evidence in the National Library of Wales, it appears that, with the two notable exceptions of the Hall Priest’s prayer in ‘The Sleeping Lord’ and a long and previously unpublished passage found in the 1958 fair-copy of ‘The Hunt’ that he subsequently excised from the published ‘fragment’, the only entirely new material that Jones composed after The Anathemata was ‘Under Arcturus’ and ‘Cailleach’. ‘Under Arcturus’ was, in all probability, intended as an insertion to or a derivation from ‘The Book of Balaam’s Ass’. We have chosen to publish ‘Under Arcturus’ in a slightly different version to that found in The Roman Quarry. Given the limited poetic interest of the text and the fact it has been published in The Roman Quarry, we have not included ‘Cailleach’ here.

THE GRAIL MASS MANUSCRIPTS

Given the project’s complex compositional history, the publication history of his published ‘fragments’ and the chaotic state of Jones’s manuscripts at the time of his death, we do not have an entire copy-text, a complete fair-copy text or a full manuscript for The Grail Mass. The reconstruction is based on the body of manuscripts in the David Jones Papers at the National Library of Wales and involves extensive use of those manuscripts rejected by Hague and Grisewood in their edition. Most of the sheets of the missing first MS B, the section the puzzle sheet denoted as 1 to 57, were recovered from those discarded sheets (DJP, LR 8/6). The most prominent are the sheets comprising the seven-sheet opening to MS B printed here as ‘The Alternative Mass Opening’; with sheets three to seven located among the manuscripts to The Anathemata (DJP, LA 1/3), and sheets one and two discovered in the discard file to The Roman Quarry (DJP, LR 8/6).

Like Hague and Grisewood, we used the first puzzle sheet as a base from which to proceed in the reconstruction. Of the 143 sheets from MS B and MS C noted in Hague’s ‘puzzle page’, only five sheets were not found in manuscript form. The matter of four of those sheets, however, already existed in print.4 Of the original 143-sheet version, then, only one sheet is lost, although it is clear that there are some linkages that Jones had neither completed nor incorporated into his project. Of the principal insertions and developments made during his 1960–1962 revision of that 143-sheet manuscript, approximately three sheets are missing. For the most part, Jones composed on white foolscap 13 × 8, the two major exceptions being his use of a pale blue-green foolscap which he used for many of the MS sheets and their drafts that comprised the three Celtic insertions written circa 1943–1946; and his use of some official stationary taken from the Italian Consulate in 1939, which he used to ‘draft pages of The Anathemata concerning utilitarian civilization as exemplified by Roman battering rams’ (Dilworth, 2017: 209). Interestingly, the Celtic material used in the construction of The Anathemata is also written on the same pale blue-green foolscap and was composed during the same time period. As a general rule, Jones wrote only on one side of the sheet. The exception to this rule concerns his notes. Jones wrote his notes on the reverse of the page that preceded the material to which the note referred. For example, a note to sheet 67 would be written on the back of sheet 66. Given Jones’s method of composition by the insertion of long passages of new material into already written matter, aligning the footnote to the material to which it referred proved crucial in the reconstruction process.5

For the most part, Jones wrote in a blue-black ink; however, he sometimes composed in pencil, and many sheets are a mixture of pencil and ink. In composing, particularly in the places where long insertions or new avenues of development were to be introduced, Jones would take a draft version of a sheet, re-copy the lines in ink up to the point where he intended to develop the line of thought or to place the insertion, and then leave the remainder of the sheet blank (or leave a large break between the point of insertion and the next already-written section to follow the insertion) before continuing. Jones did not compose using a biro until after he submitted the 1949 typescript of The Anathemata. The only two works that he composed entirely in biro are ‘Under Arcturus’ and ‘Cailleach’ and for both of these he used a fine-point biro.6 In the final years of the project, he also made minor alterations and annotated already published sections in fine-point biro.

The following outlines both the principal cruces we faced in editing the current edition and the chief editorial decisions that we made in The Grail Mass.

THE OPENING MOVEMENT: ‘THE MASS’

The first editorial decision concerned the opening. There were three choices. We could open with a version of the text published as ‘The Grail Mass’ in The Roman Quarry and from which we have taken the title of this edition; we could open with what we have titled ‘The Alternative Mass Opening’; or we could begin with what Jones noted was the opening to his continuation, the bugle call. All three had merit and all three were used by Jones at various stages in the life of the project.

Our first decision was to open our text with one of the two versions of ‘The Mass’, rather than beginning with the bugle call. In a letter to Grisewood written shortly before the publication of The Anathemata, Jones wrote that ‘I’m half thinking about getting out my “new book.” It starts with the word: Taratantara!’ (Quoted in IN, 41n). Shortly after the publication, Jones wrote a letter to Chute on a sheet that was once used for a draft of that opening, noting it as such. The lines that follow the trumpet blast read ‘You can break the hearts of god and men/but you won’t break Caesar’s/ division of time/nor the august routine’ (86), lines that are found in Section VIII. In rejecting this bugle call as the opening, we are keeping in mind three things. First, this is a variation of the opening to the MS C in the first puzzle sheet, and the opening lines of MS B in the second. Manuscript evidence shows that Jones laboured over the location of those specific lines and considered placing them as the introduction to what became the watch on the walls. We have kept them in their original position in Section VIII. These lines do not contain the word ‘Taratantara’ itself; Jones eventually used the word as part of the inscription used as the frontispiece for ‘The Tribune’s Visitation’. Additionally, at the time that he wrote his letters to Grisewood and Chute he still considered the ‘new book’ to be a continuation of The Anathemata which opened with a version of ‘The Mass’ that ultimately became the text that frames The Anathemata. As a result, and based on subsequent development of the project, including the Judas and Caiaphas material, we decided to open with one of the two versions of ‘The Mass’ that acted as the introduction at other times in the evolution of the project.

This became our second editorial decision. As we noted in our introduction, Jones began with a version of ‘The Mass’ that came directly after ‘The Book of Balaam’s Ass’, published here as ‘The Absalom, Mass’and in The Roman Quarry as the beginning of ‘The Old Quarry, Part I’. From this original, Jones developed two other openings. As with his placement of the lines ‘Taratantara’ and the lines following, Jones showed considerable concern as regards which of the two versions of ‘The Mass’ to use. At different points in the project’s development, both texts served as the introduction. In fact, at different times, both versions of ‘The Mass’ moved into the Judas material and the watch on the walls. At one point, it appears that Jones even attempted to incorporate both versions into the full text. The earlier of the two subsequent texts, and the one that opens the current edition, became the text of ‘The Mass’ recorded for the BBC; it is published as ‘The Grail Mass’ in The Roman Quarry. Composed circa 1944 but, according to Jones, subsequently lost and recovered only later, it served as the opening to MS A, where it introduced the earliest version of the Judas material. In later drafts, this opening moved directly into the material that became ‘The Wall’. This text for a now-lost BBC production evolved into ‘The Kensington Mass’. The second text, which here is published as the ‘The Alternative Mass Opening’, was originally the opening to MS B of the first puzzle sheet, as seen in the lines ‘We already’ (214) and later was the opening to the MS A of the second puzzle sheet, as seen in the line ‘These, at the sagging end’ (215). This text also moved from the Cenacle directly to the soldiers on the walls in some drafts and into the Judas material in others. This is the version of ‘The Mass’ that Jones would eventually employ as the opening frame to The Anathemata.

We have chosen as our opening the text associated with the BBC recording for three reasons. First, manuscript evidence shows that originally Jones considered using a version of it as the opening to his entire project, even to the point of splitting it to act as a frame for the work; a version that at its most ambitious subsumed even The Anathemata. Second, it is the predecessor of his final piece ‘The Kensington Mass’, even though his project at that time was ‘very differently gescepenne from the thing I made in the ’40s’ (DGC, 262). Third, while it is in many ways the more poetically and structurally appropriate opening, ‘The Alternative Mass Opening’ is nevertheless very close to the opening of The Anathemata.

In choosing to publish ‘The Alternative Mass Opening’ as it is, we made our third editorial decision. In publishing The Anathemata, Jones began on what was initially sheet 3 of the original version of the ‘The Alternative Mass Opening’. We have chosen to include sheets one and two of the original text which were found in the file of discarded material associated with the manuscripts used by Hague and Grisewood in constructing The Roman Quarry. In addition – and as we noted earlier – there is strong manuscript evidence that ‘The Alternative Mass Opening’ was initially preceded by ‘The Book of Balaam’s Ass’. Additionally, it is clear that Jones eliminated the opening two sheets only after he began the reconstruction leading to The Anathemata. Finally, presenting the original two sheets offers added insight into Jones’s shifting focus. Where The Anathemata begins with the act of consecration, ‘The Alternative Mass Opening’ in its focus on the detail of the priest’s vestments begins with the act of making, giving added emphasis to de la Taille’s comment that in the Cenacle, Christ ‘placed Himself in the order of signs’.

SECTION I TO IV: ‘THE AGENT’

The second set of editorial choices concerned the material that Hague published as ‘The Agent’. The question became whether to present it on its own as a sequence or whether it was to be included as part of The Grail Mass. While, like Hague and Grisewood, we consider the four sections to be capable of standing alone as a complete sequence, we present the movement as part of The Grail Mass. There are, however, significant differences between the version presented here and the version presented in The Roman Quarry. The manuscripts used in constructing Sections I to IV are located in the National Library of Wales in file LR6/1.

The material was first developed as part of MS A of the first puzzle page and in its earliest form can be found in ‘The Old Quarry, Part I’ after the opening version of ‘The Mass’. Following development and refinement, it was then extracted from that text and incorporated into the B and C text of the first puzzle sheet after the 143-sheet MSS B and C was constructed. According to the first puzzle sheet, the insertion of the Judas material took place at sheet seven of MS B; however, in its final stage and after subsequent development and minor re-pagination, this material began at sheet nine following a version of ‘The Mass’, opening with the lines ‘Why’s he elect’ (30) and ending with sheet 21 and the line ‘before the light of Venus-Day’ (52). Initially 13 sheets when Jones first incorporated the material into MSS B and C, he also made a number of insertions that increased the length of the material to approximately 23 sheets. As other material was later developed, the four sections eventually totalled 38 sheets. At the end of the Judas material, Jones then introduced what became ‘The Wall’. In a subsequent re-organization, Jones took the Judas material out and then, in a still-later version, re-introduced the material. Finally, in his 1962 letter to Grisewood he speaks of revising the material for separate publication. Jones either never finished revising the material for publication or, after it reached a final state, decided against publication. Given that final decision by Jones, the editors restored the material to its original position where it is followed by the earliest version of ‘The Wall’.

It should be noted that based on the manuscript evidence the 1960–1962 revisions were minimal. The poetry itself is written in either pencil or with the black-blue ink consistent with the material composed in the 1940s and in a hand indicative of that time period. In contrast, Jones has a number of annotations to the text regarding who is speaking and concerning passages that he wants to keep, to revise or to re-arrange with such annotations written in fine-point biro and in a hand consistent with the mid-to-late 1960s.

Section I

This section was originally sheets nine to 15 and while there are a number of drafts, this section shows little revision after it was first introduced.

Where Hague and Grisewood begin with Judas already involved in his internal debate after Satan has entered him (John 13:27), this edition begins with ‘Why’s he elect’ (John 13:23, 30) and continues to the line ‘vatic crow’ (30). Jones has a number of drafts that point out that ‘Why’s he elect’ is the opening to Section I, including those that come after both versions of the opening. In fact, he marked at least one draft ‘N.B. KEEP’ in red crayon.

1.Hague and Grisewood delete the passages from ‘I see the devil knows how to rhyme’ (31) to ‘And do it quickly’ (31). We have restored those passages.

2.The passages from ‘O truth, O fact’ (31) to ‘some give and take’ (32) vary considerably and this edition uses an expanded version.

Section II

This section was originally sheets 16 and 17 and is noted in Hague and Grisewood’s puzzle sheet. It was subsequently expanded with an insertion at 17 labelled 17A to 17E.

1.Hague and Grisewood have taken the passages that our edition uses as the beginning of Section I from ‘why’s he elect’ (30) to ‘in the house of friends’ (30) and moved them to the beginning of Section II.

2.The insertion introduced by Jones comes with the line ‘Remembers what on hilly Thabor/ once he saw’ (35) and continues to the end of the section.

Section III

This section began as sheet 18 and initially concerned Judas and Caiaphas determining where Judas would meet the soldiers going to arrest Christ. Jones made a major insertion into the text, labelled 18A through 18P.

1.This is the section that Jones called ‘The Agent’ and to which he returned in his 1960–1962 revision attempt. In an attempt to maintain both historical accuracy and to develop a colloquial form of speech, Jones revised the opening as to Judas’s reflections on what path Christ would choose on his way to Gethsemane. There are minor differences between the drafts chosen by Hague and Grisewood and the drafts we have chosen.

2.The main insertion begins with the lines ‘And soon, maybe, his beauties’ (39) and continues to the end of the section. There was also a minor insertion, 18F1 to 18F4, which begins with the lines ‘But were we the last to wear the horned bonnet’ (40) and continues to ‘Why’s here all best part leveled’ (43).

Section IV

This section was originally sheets 19 to 22 and while Jones has a number of drafts, he made no significant changes after he introduced Section IV into his text. The version that we have published is substantially the same as the version in The Roman Quarry.

ON THE TRAVERSE OF THE WALL AND THE CELTIC INSERTIONS

The Unpublished Manuscripts and Published ‘Fragments’

One of the concerns in editing Jones’s manuscripts that form ‘On the Traverse of the Wall’ and the three Celtic insertions is that the material is a mixture of work he published as fragments and work that he never published, and the text moves from one to the other throughout. Jones’s five Roman fragments (‘The Wall’,‘The Dream of Private Clitus’,‘The Fatigue’ and ‘The Tribune’s Visitation’ from The Sleeping Lord and ‘The Narrows’ published in the Anglo-Welsh Review and Agenda) and the three Celtic fragments (‘The Tutelar of the Place’,‘The Hunt’ and ‘The Sleeping Lord’ from The Sleeping Lord) exist both as published fragments and as unpublished parts of his project. In addition, the material published as fragments and the manuscripts to that same material as part of the larger project differ at times, sometimes in significant ways. In reconstructing Jones’s project, we have used those manuscript sheets that maintain the dramatic and unified nature of the original work, rather than those manuscripts that present a particular ‘fragment’ as an autonomous work. Of the four Roman fragments found in both The Sleeping Lord and the current edition, ‘The Wall’ is presented in an expanded version in this edition, while ‘The Dream of Private Clitus’ evidences two principal substantive changes between the published and unpublished versions: the change of the characters’ names and the late addition of Welsh material, introduced just before publication. The material published in 1965 as ‘The Fatigue’ has undergone the greatest change. As with ‘The Wall’, the version we present includes material Jones excluded from the published fragment. In his notes to ‘The Fatigue’, Jones writes that the work falls into three ‘movements’ but that together they form a continuous piece. In The Grail Mass, those same movements form what Jones designated as Section IX and Section X. In the case of ‘The Narrows’, Jones changed the characters’ names as he did with ‘The Dream of Private Clitus’ and such changes appear to have been undertaken to allow the fragment to stand apart from the other fragments.7 ‘The Tribune’s Visitation’ is substantially the same in both The Sleeping Lord and in the current edition. Regarding the Celtic material, ‘The Tutelar of the Place’ is, like ‘The Tribune’s Visitation’, very close to the published version while ‘The Hunt’ is similar to ‘The Dream of Private Clitus’ in that it shows the introduction of Welsh in the stages before publication. The work evidencing the greatest change is ‘The Sleeping Lord’ where the original third Celtic insertion is replaced by the passages on Arthur’s foot-holder and candle-bearer and the hall-priest’s prayer.

The manuscripts to ‘On the Traverse of the Walls’, as mentioned, are written in blue-black ink and pencil, and there is only one short passage written in biro, strongly suggesting that the manuscripts were composed before Jones began writing in biro. Some minor annotations, however, are in biro8 and, as we noted in the introduction, only five sheets were not eventually recovered in manuscript. Of these five sheets, two exist in print form in the published version of ‘The Wall’ and two exist in the published version of ‘The Tutelar of the Place’. The one sheet missing, unfortunately, comes when Crixus is first introduced. At the same time even where no sheets are missing, it appears that Jones had still not completed all transitions. This is particularly evident with the lines, ‘the countersign, my man’ (73) to ‘well, the cowson’s gone’ (75), when Jones introduces the Tribune for the first time in an abrupt fashion.

Dismantling, Recovering, Editing

In The Roman Quarry, Hague notes that what they published as ‘The Roman Quarry’ ‘purports to be “Sections VIII–XII”’; however, as to ‘what sections I to VII were we can only guess’ (RQ, 216). The recovery and editing of the missing sheets to what Jones called the MS B of the first puzzle sheet and the subsequent reconstruction of his project represents the most significant difference between The Roman Quarry and the present edition. When the current editor first undertook an examination of the manuscripts in the National Library of Wales, little work had been done regarding Jones’s method of composition. Likewise, save for the work done by Hague and Grisewood and the subsequent ordering and cataloguing of Jones’s papers by the staff at the National Library of Wales, there had been no other work done on the manuscripts themselves. While Jones made numerous comments regarding how his work was shaped, such comments – without the original manuscripts at hand – appeared either cryptic or deliberately vague.9 As such, the process of constructing this edition involved three stages: a dismantling of The Roman Quarry in order to ascertain Jones’s compositional process, the recovery of the missing MS B and subsequent establishment of the complete narrative on the walls of the Antonia, and the subsequent ordering and editing of the project.

From the beginning of the editorial process, there was the suspicion that that there had been an Ur-Anathemata that actually formed a work in and of itself. It was not simply that Jones had noted in his preface to The Anathemata that he hoped to make a ‘continuation, or Part II’ (A, 15) from what remained of his ‘experiments’ or that in a number of private letters, as in the letter he wrote to Lewis, that there were ‘wodges of stuff … largely concerned with conversations of grousing Roman soldiers (Celts) doing duty on the Wall of Jerusalem’ that he wished to make available. The first indication can be found in the four Roman poems of Jones’s collection, The Sleeping Lord. While Jones’s comments, including his introductions to the poems, noted that the works were drawn from his experiments and they formed a sequence insofar as they were concerned with Romans in the first century CE serving in Palestine, he gave no indication that they once were part of a unified, single work. When taken collectively, however, beginning with ‘The Wall’ and following through to ‘The Tribune’s Visitation’, the four Roman works and ‘The Narrows’, suggested a clear narrative that covers a watch from its opening to its conclusion. The second was the evidence of The Roman Quarry itself which contained versions of ‘The Fatigue’,‘The Narrows’ and ‘The Tribune’s Visitation’ that were surprisingly close to the versions published in Jones’s collection. What made reading The Roman Quarry disconcerting, though, was that the Celtic material, later to become ‘The Hunt’ and ‘The Sleeping Lord’, was embedded into the Roman material to which it appeared narratively unrelated. On examining the manuscripts that comprised The Roman Quarry, however, two other elements stood out. First was the proliferation of numbers on the individual sheets of manuscripts, a complex combination of numbers and letters which, as we have noted, suggested both a code and multiple versions of the material with different paginations. Second was the absence of any material relating to ‘The Wall’, ‘The Dream of Private Clitus’ and ‘The Tutelar of the Place’.10

The first step in ordering the manuscripts with a view to determining whether such an Ur-Anathemata ever existed involved dismantling The Roman Quarry to try to understand Jones’s compositional method and to make sense of the proliferation of numbers. This first involved extricating the Celtic insertions from the mass of material in Hague and Grisewood’s version. In doing so, three things became apparent. First, there were three separate major insertions, what we have labelled as the first, second and third Celtic insertions, and that all three were composed independently of each other and then inserted one inside of the other. Second, this Celtic material was all inserted between two lines11 of an earlier text and that Jones’s constructional method consisted of developing long passages, ‘fragments’ in a rather different sense, and inserting them into a body of previously composed material. Third, while the three insertions were introduced on different dates, all three originated no later than 1946 and all were inserted into this earlier text before the publication of The Anathemata. In fact, all evidence suggests that the composition of the first and second Celtic insertions began in mid–1943 and the third Celtic insertion was composed in late 1945 and early 1946.

After extricating the Celtic material, the Roman material was dismantled by extricating what were fragments again inserted into an already existing Roman narrative. Primarily, this entailed separating the material in what became the second and third movements of ‘The Fatigue’. The first insertion – the four-sheet meditation on the working of the world empire (SL, 38–40) – was introduced at sheet 68 of this earlier manuscript and was labelled 68A to D. The second – the meditation on the Cross (SL, 31–38) – was introduced at sheet 69, that sheet was also labelled 125, 68 and 45. This second six-sheet insertion was then labelled 125A to 125F, and was inserted by Jones after the construction of the 143-sheet manuscript of the first puzzle page. As with the Celtic insertions, the material was introduced in its entirety between two lines of the earlier text that preceded the insertion. Again, evidence establishes that the material was initially composed beginning in 1943 and revised and expanded intermittently through 1946.

After dismantling The Roman Quarry, extracting the Celtic insertions and the insertions into the Roman narrative, and deciphering Jones’s code system, three things started to come into focus. First, there had been an earlier and continuous Roman narrative that, at least as regards the material in The Roman Quarry, began with the blowing of the classicum signalling the end of the middle watch, contained a version of both ‘The Fatigue’ and ‘The Narrows’ and continued until the end of what became ‘The Tribune’s Visitation’. In The Roman Quarry, this earliest version was approximately 19 manuscript sheets and went from MS sheet 58 to MS sheet 76/143/85. This 19-sheet work was both the concluding part of an original MS B that pre-dated the first puzzle sheet and the concluding sheet of the MS B of the second puzzle sheet. The second was that although this original version had undergone a number of expansions between 1943 and 1946, Jones always used the same original linear narrative as a kind of spine into which he made his insertions. The original version was first expanded by its Celtic and Roman insertions so that sheets 58 to 76 became sheets 58 to 143, the MS C of the first puzzle sheet. The material was then re-paginated so that 58 to 143 became sheets 1 to 85, the MS B of the second puzzle sheet.

The third development, though, was most puzzling. After dismantling The Roman Quarry, there were versions of the material that eventually became ‘The Fatigue’,‘The Tribune’s Visitation’ and ‘The Narrows’ from the Roman material, and versions of ‘The Hunt’ and ‘The Sleeping Lord’ from the Celtic strand; however, there was a complete absence of any material relating to ‘The Wall’,‘The Dream of Private Clitus’ and ‘The Tutelar of the Place’ with the manuscripts Hague and Grisewood used for their version of The Roman Quarry. Equally surprising was that there was very little material regarding ‘The Wall’ deposited with the papers on The Sleeping Lord. While at the National Library of Wales there was both a manuscript draft entitled ‘The Dream of Private Crixus’ with the papers on ‘The Dream of Private Clitus’ located in file LS3/3 and a manuscript draft of ‘The Tutelar of the Place’ in LS 6/1, there were few draft versions to either.

Crucially, though, there was a folder of material, file LR8/6, consisting of approximately 397 sheets of manuscripts that were not used by Hague and Grisewood. When the editor of the current edition first started to examine these sheets, they were in disarray. He was given permission by the Keeper of Manuscripts to organize the material, and they were subsequently organized into 11 subfolders. In this folder were alternative drafts to material found in The Roman Quarry, some stray sheets belonging to The Anathemata manuscripts and, finally, manuscript sheets that were related to the material of The Sleeping Lord, including ‘The Wall’,‘The Dream of Private Clitus’ and ‘The Tutelar of the Place’. It eventually turned out that this folder also contained most of the missing sheets from the MS B of the first puzzle page.

The recovery process began by first separating the material associated with the works in The Sleeping Lord from the rest of the sheets. It was here that manuscript sheets to versions of ‘The Wall’,‘The Dream of Private Clitus’ and ‘The Tutelar of the Place’ were uncovered. Additional material to ‘The Fatigue’ that Hague and Grisewood chose not to incorporate into The Roman Quarry was also found. Next, all draft versions of material from The Roman Quarry not identified with The Sleeping Lord were organized. Finally, all Roman material not included in The Roman Quarry was separated and organized. Here what proved to be still another insertion into Jones’s MS C of the second puzzle sheet made at sheet 135 and denoted Section XI – an insertion eliminated by Hague and Grisewood – was discovered.

Even on a first examination of the versions of ‘The Wall’,‘The Dream of Private Clitus’,‘The Tutelar of the Place’ alongside the Roman material not published with The Roman Quarry, it was evident that the material formed a once-complete text that went from sheet 8 to sheet 57, although there were still at that time lacunae, most notably sheet 7 of MS B of the first puzzle sheet. That this body of material was connected to the material of The Roman Quarry was borne out by the notation on sheet 56 that reads ‘End of Section VI … Section VII in Folder B’. There was then the issue of the notations that needed to be deciphered. Save for three exceptions – sheets 25A, 44A and 47A – there were no obvious insertions and no proliferation of letters that marked The Roman Quarry. Most sheets, however, had two numbers, one in the centre and circled, and one to its immediate right and crossed out. Some sheets between 9 and 25 had three numbers, two being the same; to illustrate, a sheet that had a circled 9 in the top centre would also have next to it a 32 crossed out and on the far right side another 9 that was also crossed out. For the first 24 sheets, there was always a differential of 23 between the higher and lower numbers so that sheet 21, for example, was also sheet 44. Beginning with sheet 26, and following on 25A, however, there was a differential of 24 so that a sheet labelled with a circled 26, 27 or 28, also generally had a crossed out 50, 51 or 52 to the immediate right. As the editor eventually discovered, Jones had written a 76-sheet narrative of soldiers on the wall, introduced a 23-sheet insertion, and then later removed that insertion. This insertion later proved to be a 23-sheet version of ‘The Agent’.

While the recovery and initial reconstruction of the missing MS B of the first puzzle sheet proved that there had existed a full manuscript, the text recovered began at sheet eight with the phrase ‘the full clear call degenerates’ (53). It was at this point that sheets six and seven were discovered in a file to The Anathemata (LA/1–3, f6 and f23), sheet seven containing the phrase ‘on night-gust’ and the introduction to the original version of ‘The Wall’. Those sheets conclusively linked the watch on the walls to the version of ‘The Mass’ that ultimately became the frame for The Anathemata and which we have published here as ‘The Alternative Mass Opening’. An examination of the manuscript sheets to the final page of The Anathemata further established that the original version of ‘The Mass’ that opens and closes The Anathemata had first opened the watch on the walls. Finally, those two sheets contained a version of the notations found on the puzzle sheet. With the discovery of sheets six and seven, the final organization and editing of the material became possible.

The discovery of the missing 57 sheets of MS B raised two issues. The first was why Hague and Grisewood had never recovered them. One possible answer is that in editing The Roman Quarry, Hague eliminated the lines ‘They say he is himself both sign/and thing signified’ to ‘fish in urbes/throughout orbis’ (85) because he felt that the lines were ‘dauntingly theological’ and the phrase ‘the sign and the thing signified’ was a phrase ‘misused by expounders of DJ’ (RQ, 218). The lines, though, were on sheet 58 which contained the conclusion to MS B of the first puzzle sheet and were followed by the opening to the second half of the watch on the walls, MS C. In cutting out those lines, Hague inadvertently broke one of the links necessary to reconstructing the work. This problem was compounded because sheet seven of the first puzzle sheet, the beginning of the watch on the walls, included on it the markings noted on the puzzle sheet. This sheet, though, was filed with the manuscripts to The Anathemata, which were deposited in 1978 at the National Library of Wales. As such, both the first and last sheet were eliminated. Without either the opening sheet or the closing sheet, Hague and Grisewood were unable to recover the missing sections. Again, this problem was exacerbated when Hague raided the material in the unused file, extracting some of the sheets from what proved to be the conclusion of MS B12 to fill in textual gaps, a decision that further fragmented the original dramatic narrative.

The second issue concerns the extent and nature of Jones’s post-Anathemata poetic production. All of the poetry that Jones brought to publication after The Anathemata originated in his earlier ‘experiments’ as the author explained in his notes to The Sleeping Lord. Given the similiarities between the works published in The Sleeping Lord and the earlier versions, however, it appears that almost all of Jones’s poetic activity after the publication of The Anathemata consisted in extracting fragments and preparing them for publication, with this concentrated on the layout of the work on the page. It also appears that save for the central section of The Sleeping Lord’, Jones initiated no new works that resulted in publication after 1952.13 Instead, he appears to have spent the remainder of his life extracting sections of his vast project for separate publication, while continuing to refine, revise and re-arrange a vast project he was simultaneously dismantling into fragements.


1 The structural changes Jones wrought on his final poem ‘The Kensington Mass owed much to his having been separated from many of his manuscripts while living in Calvary Nursing Home and the fact that these manuscripts themselves were in a state of utter disarray.

2 The clearest statement of Jones’s insistence on fidelity to historical accuracy in art is found in ‘An Introduction to “The Rime of the Ancient Mariner”’. After discussing Coleridge’s detail and accuracy, Jones poses a rhetorical question: ‘You may say: But what does any of this matter?’ His response is revealing:

It is precisely because of the greatness of the poetry, and the imagination which informs it that these things do matter. In poetry everything matters, and the greater the poetry so much the more is this true.

E & A, 207

One important change of ‘matter’ that caused him serious difficulties came about when he discovered that the troops stationed at the Antonia in Jerusalem at the time of Crucifixion were not Roman legionaries of the Legio X Fretensis, as he believed when he wrote ‘The Dandy Xth are my regiment/who diced/Crown and Mud Hook/under the tree’ in In Parenthesis (IP, 83). When he found out that they were, in fact, local auxilia troops, he had already completed a first draft of the long Roman narrative of the soldiers on the walls of the Antonia.

3 The earliest suggestion of this shift occurs in the 1958 fair-copy of ‘The Hunt’ that Jones used for the now-lost BBC recording. The section of the poem where the influence of the Welsh language is most evident is in a previously unpublished section. Although it is concerned with Welsh matter, the original section from The Grail Mass employs little Welsh. The previously unpublished fair copy introduces long passages concerning Welsh matter that also employs the Welsh language. The ‘fragment’ he published in Agenda in 1965 as ‘The Hunt’ rejects the Welsh matter he first introduced in the fair copy of 1958, but it employs the Welsh language and, at the suggestion of Watkins, includes a list of the Welsh words Jones employed and their English translations.

4 Of these four sheets, the material from two were from the section published as ‘The Wall’ and two were from ‘The Tutelar of the Place’. In both instances, we employed the published ‘fragments’ as the means of restoring the material, and in both cases, the proof was conclusive that the content was identical to the original.

5 This was particularly important in uncovering the original Roman narrative and the subsequent insertion of approximately 45 sheets of Celtic material, introduced in three separate insertions, that divided the original colloquy between Crixus and Oenomaus occurring on the walls of the Antonia.

6 Both texts were written post-1966, with ‘Under Arcturus’ almost definitely written in 1971. While ‘Under Arcturus’ has the date 1971 written on one of its sheets, both manuscripts used an altered form of marking insertions; a colour-coded notational system involving red and green circles that Jones first employed in 1968 while trying to write an essay on Gerard Manley Hopkins. The drafts for this essay are on file in the David Jones Papers, Burns Library, Boston College, BC18032. The essay is published in David Jones on Religion, Politics, and Culture, 101–268.

7 Some alterations also reflect minor changes in detail in the interests of historical accuracy that resulted from Jones’s ongoing research.

8 This occurs in the phrase ‘mightn’t be me – mightn’t be you, might be that cissy Greek/from Attica’ (78). However, it appears that Jones wrote in biro over those original lines which had faded.

9 Two examples from his preface to The Anathemata should serve as illustrations. Jones writes that ‘what is now sheet 166 of my written MS has at different times been sheet 75 and sheet 7’ (A, 15). Jones fails to note that the work was constructed by way of an insertional method, and this sheet 7, 75, 166 was always the final sheet of the manuscript. Similarly, Jones writes:

If it has a unity it is that what goes before conditions what comes after and vice versa. Rather as in a longish conversation between two friends, where one thing leads to another; but should a third party hear fragments of it, he might not know how the talk had passed from the cultivation of cabbages to Melchizedek, King of Salem.

A, 33

Here Jones has embedded in the preface not only the general outline of his Roman narrative, the ‘longish conversation between two friends’ but also part of the conversation itself, a part of the conversation that Hague eliminated:

They say he is both sign and

thing signified.

They say he can raise a gale o’ wind and

still one.

They say no tree can hang him but

a cabbage tree. (85)

10 There was one major exception as to location and the material published separately as ‘The Dream of Private Clitus’, where the sheets were located with the drafts to the published version in file LS3/3.

11 The earliest draft places the insertion between the lines

So mate, their very signa we fetch for

them — let history weave but long

enough (96).

and

When calibans of Logia Sinus

swear by Bron that tree-tops walk

the spume because the green troughs

hide all but the top-trees of our

cruising biremes (124).

MS 66. Compare: RQ, 12 & 39

This is the sheet employed by Hague and Grisewood in their edition. As we discovered, in developing this material and introducing his first Celtic insertion Jones took the first part of this passage ending with ‘let history weave’ and moved it into the first Celtic insertion (96).

12 This is most evident with the passage in The Roman Quarry beginning with ‘Should ever the men of rule’ and ending with ‘Back to the womb of Tellus’ (RQ, 43). This passage was sheet 56 of the original and continued on with sheet 57 beginning with ‘When the young hero’ and concluding with ‘They say he cuts signs in his skin’ (85), all passages that Hague eliminated, but which, in turn, moved directly into what became the opening to The Roman Quarry.

13 In his biography of Jones, Dilworth examines the extent to which the medications prescribed after Jones’s second nervous breakdown had an adverse influence on his artistic productivity. The manuscript evidence regarding his post-Anathemata activity not only supports Dilworth’s findings, but also suggests that the impact might have been even more severe.