University of Virginia Library

The Surviving Galley Proofs of Lawrence's Sons and Lovers
by
Helen V. Baron

When D. H Lawrence's final manuscript of Sons and Lovers is compared with Duckworth's first edition there are found to be thousands of variants. They are hard to quantify because substantive revisions can be one word or several sentences, but when they are listed according to the points in the text at which they occur, the list has approximately seven thousand entries, of which approximately four and a half thousand are punctuation. Discovering who was responsible for each item is not now possible but some bearings can be taken from the small number of revised galley proofs which has come to light.[1]


232

Page 232

An unmarked set of page proofs has been in the possession of the Harry Ransom Humanities Research Center at Austin for many years, but James Boulton's more recent location of the galley proofs is both the rewarding by-product of scholarly endeavor and an invaluable discovery for editors and others who ponder the many changes that took place between Lawrence's submission of his manuscript to Duckworth in November 1912 and publication of the first edition in May 1913. For although the galleys that have survived are unfortunately incomplete, amounting to only about a ninth of the novel, they belonged to the set corrected by Lawrence.

The galley proofs are of interest in the pure bibliographical sense of providing evidence about a sample of printing practice. They also help to make sense of the compositors' marks on the Sons and Lovers manuscript, which number more than eighty and may well puzzle the many scholars who are currently consulting the photographic facsimile of the manuscript published in 1977.[2] Most importantly, they provide the only firm evidence of how accurate the compositors were; what changes and suggestions were entered on the proofs before they were sent to Lawrence; how well Lawrence corrected substantive errors; how he responded to the massive quantity of punctuation changes; and how much and in what way he revised.

Selective Chronology

On 1 July 1912, while William Heinemann was writing to Lawrence to announce his rejection of 'Paul Morel', Walter de la Mare, Heinemann's Reader, was writing to Edward Garnett, Reader for Duckworth, to inform him of the firm's decision. As a result of these exchanges between publishers in London, Lawrence, in Gargnano beside lake Garda in Italy, received Heinemann's rejection of his novel one day and Garnett's offer to take an interest in it for Duckworth the next. To state, as Mark Sexton does, that: "Undaunted, Lawrence arranged for Edward Garnett . . . to read the work" (see note 1) not only misrepresents the chronology of incident as revealed by Lawrence's letters, but (more significantly, from my editorial point of view) implies that Lawrence was more in control of events than was evidently the case.[3]

Garnett had in the same way secured publication by Duckworth for Lawrence's previous novel, The Trespasser, when Heinemann, to whom it was contracted, was unenthusiastic. Garnett had sent Lawrence detailed notes on the manuscript of The Trespasser which stimulated Lawrence to tackle it again and produce a final version; and he now did the same with the rejected Sons and Lovers manuscript. The same Guildford printers, Billing and Sons, who had set The Trespasser set the new novel, in both instances using Lawrence's


233

Page 233
manuscript as copy. In the case of The Trespasser only page proofs were produced, an unmarked set of which has survived. There the parallels appear to end, but it would be interesting to read a parallel study of the manuscript and proofs of The Trespasser, to see for example whether the compositors were the same men and how their house-styling compared with that in Sons and Lovers.

As is well known, Lawrence triumphantly sent off his final manuscript of Sons and Lovers on 19 November 1912 but Garnett could not accept that it was so much longer than The Trespasser and wrote him a withering letter, before settling down to reduce it by a tenth. He then sent it directly to the printers. It was an ultimatum, and Lawrence had no option but to acquiesce.[4]

Having been asked to check only one proof for each of his earlier novels, Lawrence assumed that the proofs of Sons and Lovers which began arriving in batches on 5 February 1913 were all he would have to correct. He does not mention whether the relevant pages of his manuscript were returned to him with each batch of proofs, although this was the usual practice. However, the proofs themselves were unusual. They were in the style of the French placard proofs of the period: four columns printed side-by-side on large sheets of paper. The columns were numbered (but not the sheets) and each contained eighty-eight lines of type, which was double page-length. Lawrence had corrected them by 3 March 1913 and expected his novel to be on sale in three weeks (Letters 1: 524).

He knew that Duckworth's terms included a downpayment of one hundred pounds on the day of publication and awaited the day with intense eagerness not only because he believed the book would make an impact but because his financial anxieties were acute. He had earned very little since resigning from his teaching job and he now had Frieda to support as well as himself—and indeed expected to be fined as a result of the imminent divorce proceedings. So he was aghast when he learned that Duckworth, who evidently wanted the novel to be censored further, had decided on another round of proofs, and he wrote back in desperation: "I don't mind if Duckworth crosses out a hundred shady pages in Sons and Lovers. It's got to sell, I've got to live" (Letters 1: 526).

The page proofs were in signatures of sixteen pages with forty-four lines of type to a page. They appear to have begun reaching Lawrence on 22 March 1913 and he seems to have finished correcting them by 11 April. This time the revised galley proofs were sent back to Lawrence with the batches of page proofs. His revisions had caused some realignment of text and therefore the galleys had been trimmed so that the text began and ended at the same point as each signature of page proofs. As Lawrence finished working on each batch he posted the page proofs to Duckworth and the corresponding galley proofs to Ernest Collings, an illustrator.


234

Page 234

In his letter to Collings he referred to them, perhaps confusingly for modern readers, as 'the first batch' (Letters 1: 529). However, it is certain that what Collings received was a complete set of corrected galleys, for it was in his son Guy's possession that the few surviving galleys came to light when James Boulton wrote to enquire whether any of Lawrence's letters to Collings were still in existence.[5]

At this time, March 1913, Lawrence and Collings had not met, but they had been exchanging letters since November 1912, when Collings sent Lawrence a book of his illustrations in appreciation of Lawrence's first two novels. Lawrence's reason for sending Sons and Lovers proofs to Collings was that Garnett had asked Lawrence to produce some drawings of collieries for the novel's dust-cover, but Lawrence felt unable and decided to pass the request on to Collings. The proofs were a gift, chiefly to give Collings some idea of what the novel was about and also as a kind of return favour.

Only six of the original fifty-two sheets of galley proof survive and they are in fact three whole and three part galleys, containing eighteen whole and three part slips. They are in two separate sections and have been cut to tally with page-proof signatures: 23 (pages 359-368) and 26-27 (pages 401-423, i.e. to the end of the novel). The clicker's marks in the manuscript indicate that the first section was set by Moore and the second by Knowles and L. Bristow.

The Clicker's Marks in the Manuscript

As Lawrence had sent it to Edward Garnett, the finished manuscript was 533 pages long but numbered 1-540. Garnett had then pruned the text, largely by deleting passages but also by scissors-and-pasting a little, so that it was 530 pages long when it reached the printers and still numbered 1-540. Evidence that Garnett had calculated the length of the manuscript, to determine format and page-length, in preparation for printing, is found in two sets of 100-word enumeration marks by him in the manuscript.

Lawrence had written on one side of the paper only, but most of the novel, that is the 436 pages he wrote out at the last revision stage onto continental paper,[6] consisted of gatherings usually of about five but sometimes as many as ten sheets folded over together. These were now torn along the folds so that the manuscript stood in a pile of single leaves which could be handed round to the compositors.

The annotations they then made on the manuscript show that it was set by a team ("companionship", or "ship") of nine compositors, working "upon lines", a method or organisation described by Philip Gaskell as "by the later nineteenth century . . . the normal arrangement in English houses for all classes of work." The team-leader or clicker "received the copy, passed on the overseer's instructions about letter for distribution and typographical


235

Page 235
style, and handed out takings of manuscript."[7] The publishers had normally "calculated the length of the whole manuscript, and it was no longer customary to cast it off [measure it in terms of format and page-lengths] in the composing room. . . . The clickers . . . divided the batches of copy into individual takings for the members of the ship, the length of the taking varying but being typically from one to four leaves of a manuscript written on one side of a 'quarto' paper. . . . The pattern of composition of a companionship job, as revealed by surviving copy marked up both by the individual compositors and by the clickers, was complex, not surprisingly since the six or eight members of a ship worked simultaneously on takes of different length, and probably had three or four different books on hand at a time" (Gaskell, p. 293).

There are in fact two sets of clicker's marks[8] in the Sons and Lovers manuscript: entries in pencil recording the casting off of copy into estimated, numbered galley-lengths, and the distribution of these to the compositors; and entries mainly in ink, overlapping with the former, which indicate the actual galley proofs pulled. The relationship between these two sets of marks is instructive, but frequently baffling; and the pages below are an attempt to understand them and to gain a clearer picture of the working routine by which the novel was printed.

Because Billing and Sons were not working towards page proofs, which Gaskell indicates were the norm at least until the 1880s (p. 293), but galley proofs consisting of four columns of eighty-eight lines of type, the manuscript appears to have been cast off in passages estimated at the equivalent of 352 lines of type. The notation entered in the manuscript, referring to each proof as a "galley" and each eighty-eight-line column as a "slip", ran counter to the normal use of the terminology at the time, which referred to the long trays of type as galleys and the proofs printed from them as slips. This suggests that this style of proofing was sufficiently recent for the terminology not yet to have settled down.[9] The clicker's annotations to the manuscript, marking the start of "galley one", "galley two" etc. (although the printed galleys were not numbered) occur on average every ten pages. Not only is this longer than the usual takings gauged in terms of page-lengths, but in practice the takings of this manuscript often consisted of more than one such cast-off galley.

The Rate of Work

The surviving galley proofs were annotated with the date 21 February 1913 as they were checked and despatched to the author; and since they all belong to the latter part of the novel, this provides a clue as to how long the manuscript was in the printing shop. For the clicker entered another


236

Page 236
date, "27.1.", i.e. Monday 27 January 1913, in the manuscript in three places, pages 196, 205 and 324. As this note first occurs in the body of a 23-page taking set by one compositor, it does not in context appear to mark the start of setting the book, but a fresh large-scale distribution of copy at the beginning of a new week. If so, the first 195 manuscript pages had been started if not completed the week before. Thus, the period from sometime before Monday 27 January to Friday 21 February being four to five working weeks, it is possible that the compositors set the novel in three or four weeks and the last batch of proofs was checked and despatched within another week.

The book when printed was in the region of just under a million (c. 975,000) "ens" in length. The average rate of work for an English compositor at the time has been calculated at about a thousand ens per hour including the distribution of type back into the cases, but this type was left standing and so the setting-rate would probably average more like 1,500 ens per hour at a minimum. The job therefore represented about 650 hours' work, and as the week was normally 50 or 55 hours, a companionship of five or six compositors could have set the novel in only two weeks.[10] Thus the nine compositors named in the manuscript cannot have been working solely on this book.

Since the total work of 975,000 ens was printed in 52 galley proofs, each proof was about 18,750 ens long. At an average rate of 1,500 ens per hour a compositor may have set enough matter for a galley proof in about twelve and a half hours, or one and a quarter working days, and so in a 55-hour week could have set four and a half galley proofs. This consideration may have influenced the clicker in his distribution of the manuscript in takings to the compositors, for he gave them anything from one to six cast-off galleys at a time, so that the takings vary between ten and sixty manuscript pages.

Casting Off

Edward Garnett had prepared the manuscript for publication and calculated its length. The task of casting-off enough text to fill the galley-trays was complicated by the fact that the amount of text on each MS page ranged from sixty to nine hundred words because of variation in paper-size, changes in Lawrence's handwriting (in the amount of interlinear revision and in the degree to which he crammed lines together on unlined paper), and deletions of irregular lengths of text by Garnett.

The clicker's entries in pencil are typically as follows. The cast-off galleys were marked for example "Gal 27" or just with a number. The numbers ran roughly from one to fifty-two and some have a compositor's name pencilled beside them. It seems that when the clicker gave a compositor a taking of one or more cast-off galleys he wrote the man's name beside the first number; but it is not uncommon to find the same name beside more than one of the cast-off galleys in a long taking.

The marks in ink are fuller, and they were evidently entered as the lines


237

Page 237
that had been set were imposed in galley proofs. But in fact they indicate only where every alternate proof began. A typical entry marks with a large square bracket the precise point in the text where the first slip of a new proof begins, even mid-word in some cases, and records the galley number, the slip number, the compositor's name, and the chapter number. The code for denoting the chapter number referred back to the last chapter-heading which had been set, as "l.c." or, once, "Last chapter", plus a roman numeral.

For example on MS p. 489 (in chapter XIII) such an entry reads: "Moore Gal 45 Slip 177 Lc XIII", and there is a large square bracket in the eleventh line of the page around the first word of "too flagrantly give herself away before the other girls" which coincides precisely with the start of slip 177 on the only one of the surviving galley proofs that is recorded in the manuscript. The few others are not so recorded because of the clicker's system of marking only alternate proofs, and because even within this sequence he omitted to note proofs at the very end (and indeed the beginning) of the novel.

Thus fifty-two galleys were cast off in pencil and fifty-two appear to have been pulled, but the two sets of numbers in pencil and ink fluctuate in relation to each other as they progress through the manuscript. Evidently the clicker hoped that the length of text he estimated when entering his pencil marks would turn out to fit the 352 lines of type wanted for each galley proof, but in fact his cast-off galleys varied a great deal and, when set, ranged between approximately one hundred and five hundred lines of type.

Twenty-one of the fifty-two cast-off galleys started with a new page in the manuscript, and all except two began with a new paragraph. The two exceptions, cast-off galleys 4 and 20, started mid-sentence with a square bracket marked in the text, and this suggests that here casting off was made in terms of a point already reached in the setting. On the other hand, the clicker's pencil can also be seen at work making improvised adjustments within a sequence already cast off.

He pencilled-in a galley-number 39 towards the end of "galley" 38 as originally cast off, evidently because he wanted the compositor of 38 (which would have been 470 type-lines long) to make a new start. He then had to change his original cast-off galley 39 a few pages later to 40, and the one after that is numbered not 41 but 40A. This suggests that here the later 39 at least, if not also the later 40, had been cast off in advance (though the 40A may have been a subdivision of a cast-off galley that proved too long—40 and 40A together totalled nearly 460 lines of type). Presumably the reason why the clicker did not extend the new short "galley 39" into the next, and tell the compositor to continue setting, was that he had already given the next to someone else; and the reason that he renumbered the next "40" instead of "39A" must be that he used that tactic only within a taking.

Distribution of Takings

The clicker did not distribute this manuscript equally among the compositors named in it. He must have had a number of considerations in mind. He had to give copy to the different compositors according to their rate of


238

Page 238
work and the need to keep the setting-rate in proportion with the rate of imposition and proofing; and he had to dovetail the setting and imposing of these proofs with those of the other books the compositors were working on.

Whatever further principles governed his method of distribution, it was surely intentional that, apart from the distinct break between Parts I and II of the novel, the chapter divisions were never used as breaks between compositors, even though they were sometimes used to start a new cast-off galley within a taking. Although in the page-proofs and first edition a new chapter would start on a new page, in the galley-proofs there was only a gap of ten lines between chapters, and therefore it was probably convenient for one man to set the juncture.

The clicker generally ensured that a new taking began with a new paragraph (not merely each new galley as noted above). The only exception occurred very near the beginning of the work and was probably untypical for that reason. In fact there are only twenty-three transitions between compositors, and in fourteen of them the compositor started his new taking with a new paragraph on the first line of a manuscript page, which must have minimised the extent to which compositors needed to pass copy to those setting adjacent text.

The striking feature of the takings is not simply that they were rather large, but that some of them appear to have been set by compositors working in pairs, simultaneously setting alternate stints. The names of the nine compositors, Cook, Rendell, L. Bristow, Notley, Moore, Deacon, Wright, Knowles and F. Bristow, seem to occur, recur and disappear in random succession in the manuscript, but when it is surveyed as a whole, the names group together to indicate only ten fundamental divisions of the text among them.

The general pattern is not immediately evident because the marks are scant at the beginning of the work. Cook's name is the first to appear, on p. 29, and thereafter it occurs only once again, deleted. But page 29 was not necessarily the start of Cook's take (his name may have been written also on the edge of p. 8) and unless a specimen was first set and submitted to the publishers, it seems likely that Cook set the first 41 pages and then worked on something else.

The next compositor, Rendell, began his taking on the second line of p. 42 and it was marked unusually as beginning mid-sentence, with a square bracket in pencil around the first word of "an hour of joy", a phrase which certainly starts a new line in the page-proofs, on the first page of chapter II. This may therefore be the point at which the general distribution of Part I began.

After the obscure start, the total distribution of takings fell into the following shape: Rendell and L. Bristow cast-off galleys 4-10, Notley 11-16, Moore 17-18, Deacon and Wright 19-30, Knowles 31-34, L. Bristow and Rendell (again) 35-39, F. Bristow and Moore 40-49, Knowles 50-51, and L. Bristow 52.


239

Page 239

Compositors Working in Pairs

In the four shared takings the compositors set one or more cast-off galleys in a series of stints. In the first, Rendell set 4-6, L. Bristow 7-9 and Rendell 10. But in their next taking they worked even more closely together. L. Bristow set 35, Rendell 36-37 and then L. Bristow set 38 (MS pp. 404-413), although one page of it, which started with a new paragraph on the first line and ended with a complete paragraph on the last, was set by Rendell (p. 407). Then in the middle of p. 413, L. Bristow continued setting an unusually short section of only three and a half small pages which the clicker numbered "39" in the apparently improvised manner described above. The fact that this forced him to renumber the next cast-off galley appears to reveal that he needed to make some minor adjustments in his casting off at this point, but since the cast-off galley he cut short was one that these two compositors shared and set in unusually rapid overlapping stints, it may reveal instead that here the rate of imposition had caught up with the rate of setting, or they were rounding off a week's work.

Of the other two pairs of compositors who worked together, Deacon and Wright set in consistently longer stints: Deacon 19-23, Wright 24-25, Deacon 26-27, and Wright 28-30, tackling in this way nearly 120 pages of the manuscript, which was their total share of the book, amounting to about 3,770 lines of type or 10.7 galley proofs. It probably took them little more than a week. Similarly F. Bristow and Moore set a hundred pages in alternating stints, but their shares were unequal: F. Bristow 40-41, Moore 42, F. Bristow 43, 43A, 43B, Moore 44-46, 46A and F. Bristow 47-49. Moore had already set two cast-off galleys and in the final totals of lines he and F. Bristow were among the three who set the greatest number. The fact that he had not worked in conjunction with F. Bristow in his first taking suggests that the overlapping arrangements were ad hoc.

This interpretation, that the compositors' alternating stints were large takings distributed to pairs working simultaneously, is supported by the way the date, "27.1", was pencilled in by the clicker. It first occurs on p. 196, on the fifteenth page of Moore's twenty-three-page solo taking: he must therefore have set the first fourteen pages the week before. His taking ended at the end of Part I on MS p. 204, and the date then occurs a second time on p. 205, the start of Part II and of the 119-page taking set simultaneously by Deacon and Wright. Immediately after their last stint it occurs again, on p. 324 where Knowles the next compositor began. Thus it seems that on Monday 27 January the clicker began distributing Part II of the novel, the first twelve cast-off galleys to Deacon and Wright, the next four to Knowles.

Imposition

Meanwhile the work of imposition continued concurrently, and the clicker made ink annotations in the manuscript as the lines were imposed. He recorded alternate galleys, plus slip numbers and (as explained above) chapter numbers and compositors' names. It is evident that these entries do


240

Page 240
not simply refer to the proofs that had been imposed but form a bridge between the casting-off and the imposition. They are, indeed, so cryptic that it is only possible to speculate on the answers to the questions that arise. How did the cast-off galleys he had estimated tally with the precise lengths of text imposed for each proof? Did the compositors set portions of text according to the pencil record or the ink record of their names in the manuscript? How did they hand on copy or spare lines between takings?

As the compositors delivered the lines of type to the clicker, he probably made them up into the correct lengths for the galley-trays ready for imposition then headed and numbered each column or slip. But he entered in ink in the manuscript both slip and galley numbers. These ink entries contain irregularities which make the sequence hard to follow simply by turning the pages of the manuscript, so they are summarised in the next three paragraphs.

Since there are four slips to a galley proof, the number of the first slip of each proof is one more than four times the previous proof-number, e.g. galley proof 3 should start with slip 9. The first entry, however, is "Slip 9 Gal 2", but the "2" may simply be an error for "3". The next entry records only the slip number, 17, and omits a galley number, presumably 5. Thereafter the entries in ink consistently note alternate galleys, such that galley 7 starts at slip 25 and so on, until galley 19, at which point some confusion occurs.

On reaching galley 19, which started with slip 73, the clicker omitted the galley number from his entry. He next entered galley 22, instead of 21, as starting with slip 81; and then galley 24 also as starting with slip 81. This latter must have been a simple error for 89, and the net effect was the loss of galley number 21. Then from galley 26 nearly to the end of the book, all the even-numbered galleys are consistently recorded, but the slip number for the start of each is now lower than it should be by a difference of four, such that the first slip of galley 26 is numbered not 101 but 97, and so on. This was apparently rectified at galley 43 by the insertion of subdivisions 43A and 43B; for thereafter odd-numbered galleys are recorded and the slip numbers are as one would expect at four to a galley.

What in fact happened at galleys 21 and 43 was that the casting off turned out, as the lines of type were imposed, to be significantly inaccurate. When the lines are counted[11] it is clear that too little copy had been cast off up to galley 21 and 22, and too much for galleys 38-43. Therefore the clicker first omitted the number 21 from his ink marks and then he rearranged cast-off galley 39 and subdivided 40 and 43 in his pencil marks, and finally as the matter was imposed he noted the new galley numbers in his ink marks.

From this it becomes clear that the clicker's ink marks actually combine two separate records: the slip number refers to the proofs, while the galley number refers back to the cast-off galleys as pencilled and adjusted in the


241

Page 241
manuscript. Thus the ink entries form a link that enabled him to relate the manuscript copy to the first proofs. The original casting off was in places widely divergent from the imposition, but once these adjustments had been made, it is generally the case that (apart from the opening six galleys which are untypical) the ink entries occur on average five MS pages after the pencil entries referring to the same galley number, though in one place the gap is only four lines. To take an illustrative example: the pencil entry on p. 75 (Lawrence's numbering), "7 L. Bristow", marks the start of L. Bristow's first stint in the large taking of cast-off galleys 4-10 set by him and Rendell. The next entry by the clicker, p. 79 (Lawrence's numbering), is in ink and records: "L. Bris. Gal 7 Slip 25 lc III" with a large square bracket in the text before: "'Well our mother!' he answered."

Two observations can be made about this pair of entries. First, the pencil entry had occurred at a new paragraph five lines from the top of p. 75, and therefore the most likely procedure would be that when L. Bristow had set from there to the bottom of the page, he handed the leaf of MS back to Rendell, so that Rendell could complete the remaining four lines of the preceding paragraph when he reached the end of his stint. Second, although L. Bristow had begun setting at p. 75, it was not until p. 79 that the clicker was able to mark up the first proof imposed by him. The intervening text that he had set amounts to about eighty-six lines of type, nearly a whole slip. These lines were needed to complete slip 24 of the previous proof, set by Rendell, and probably imposed by him. Therefore L. Bristow must have passed these lines of type set by himself back to Rendell, or to the clicker. This suggests that the inclusion of names in the ink entries may indicate the compositor who imposed the proofs, most of which had been set by him, but not all.

Thus in general the compositors began setting the takings given them at the new paragraph where their name was pencilled, and had set on average as many as 177 lines of type before the part of their taking they actually imposed began. Then, because only alternate proofs were noted, the next pair of proofs appears to be marked in ink as imposed by them, whether or not they had in fact set both. Thus, on MS p. 109 "L. Bris (9) slip 33." indicates that L. Bristow was responsible for galley 9 slip 33, and it is followed on p. 130 by "Notley (11) slip 41 l.c. IV", so that there is no record in these ink annotations that Rendell set cast-off galley 10, allocated to him, of course, in pencil.

If therefore the clicker kept a debit and profit account of the lines passed backwards in this way between compositors so as to pay them fairly, he must have done so in a line-book, for his pencil and ink marks in the copy are widely discrepant from the evidence of the proofs. If one supposes that the compositors set all the text allocated to them in pencil, and compares that with the lines of type which that amounted to in the page proofs, the following figures emerge.[12] Wright's total work of c. 1865 lines set (62 MS pages) is


242

Page 242
represented in pencil as 5 cast-off galleys and in ink as 6 in proof; Rendell's c. 1900 lines (68 MS pages) as 6 cast-off and 4 in proof; Deacon's c. 1910 lines (57 MS pages) as 7 cast-off and 6 in proof; Knowles's c. 1910 lines (56 MS pages) as 6 cast-off and 4 in proof (but his name is omitted at proof 51 and his total should clearly be 6 proofs); Notley's c. 2045 lines (60 MS pages) as 6 cast-off and 6 in proof; Moore's c. 2345 lines (53 MS pages) as 7 cast-off and 4 in proof (but again his name was omitted in error at proof 17 and his total should be 6 proofs); L. Bristow's c. 2380 lines (74 MS pages) as 7 cast-off and 6 in proof; and finally F. Bristow's largest number of lines, c. 2560 (67 MS pages) is represented as 9 cast-off galleys and 8 galley proofs.

The ink entries may have had one further purpose. The fact that the clicker noted only alternate galley proofs indicates that he was thinking in terms of the 16-page signatures of the next round of proofs. His ink annotations may indicate which compositors would be responsible for which pairs of galley proofs when it came to incorporating the author's corrections and revisions, and imposing the revised type into page-proofs.

The Surviving Galley Proofs: How Accurate Were the Compositors?

The primary question which presents itself at this stage, is whether the compositors made a good job of printing Lawrence's manuscript.

The document itself was not a severe challenge to a compositor. Lawrence's handwriting is characteristically neat, and even though there is heavy revision in a number of places, the result is always legible. From time to time it is difficult to distinguish his "h" from his "k", his small from his capital "c" and "s", or a new line from an indentation for a new paragraph; but these are not matters that would lead to a serious corruption of the text. Garnett had been systematically through the manuscript, marking in pencil and confirming in ink passages for deletion, and censoring by removal or minor alteration isolated words and sentences; and at the same time he had occasionally clarified or changed the paragraphing. There remained Lawrence's use of dialect in speech, but in fact this only rarely led the compositors into error. The thing that did, however, present them with a major stumbling block, was his punctuation, which differed in a variety of ways from their house-styling practice.

Omitting from the total such categories of house-styling as regularisation of the length of dashes, the insertion of hyphens in words like "to-day", and various normalisations of spelling, the average number of changes in punctuation, paragraphing, and hyphenation introduced by the compositors amounts to approximately ten per page or one nearly every four lines, in a book totalling c. 18,050 lines of type.

Finally, the evidence of the few extant galley proofs is that the compositors made a small number of substantive errors not all of which were corrected in the first edition. The main problem is that because so few


243

Page 243
galley proofs survive, other similar errors cannot easily be distinguished from revisions which Lawrence might have made on the lost galley proofs and the lost set of revised page proofs.

The surviving galley proofs were set, on the evidence of the clicker's marks in the manuscript, by Moore, Knowles and L. Bristow.

Moore

The fragment of galley proofs set by Moore comes from his stint of cast-off galleys 44-46A (MS pp. 478-500) in the long taking of MS pp. 417-516 shared by him and F. Bristow which covered the end of chapter XII, chapter XIII, and nearly half of chapter XIV. This specimen of Moore's work represents his setting of a passage from chapter XIII beginning midword in a description of Clara's passionate preoccupation with Paul during working hours at the factory (MS p. 486; E1 p. 353): "[Every second she ex]-pected him to come through the door," and ending with Paul's return home to his mother after his fight with Baxter Dawes (MS p. 498; E1 p. 368): "She was there, he was in her hands."

Moore set this passage, the longest specimen of the three compositors' work, in 704 lines of type and introduced a total of 232 punctuation changes (counting pairs of brackets and inverted commas as one). This is approximately one every three lines, a higher rate than the other two compositors, but Moore made less serious substantive errors. These were "marram" for MS "marrain" and "train" for MS "tram", both of which Lawrence corrected on the galley proofs; and "every" for MS "ever", which had been cleared up by the first edition. Moore also corrected four substantive errors in the manuscript: "always" for MS "alway", "One evening as Paul" for MS "One as Paul", and "one" for MS error "on", all of which were probably pen-slips. Finally, he set a correct form, "different from", instead of Lawrence's "different to", which, being spoken by a character (Clara) rather than the narrator, may possibly have been a deliberate grammatical error (MS p. 495; E1 p. 364).

In addition to his many alterations of punctuation, Moore corrected two small punctuation slips in the manuscript, and introduced six changes of paragraph. Finally, he omitted one meaning-bearing comma, which Lawrence reinserted on the galley-proofs, after "Behind" in: "Paul glanced round. Behind, the houses stood on the brim of the dip, black against the sky, looking like wild beasts . . ." (MS p. 497; E1 p. 366).

Knowles

Knowles and L. Bristow finished the last twenty-four pages of the MS. Knowles's cast-off galleys 50 and 51 covered nearly all the second half of chapter XIV, and L. Bristow's cast-off galley 52 started on the last page of the chapter and continued to the end of chapter XV.

Knowles began in the middle of Mrs Morel's illness but the surviving specimen of his work begins after her death, with Walter Morel's words: "'No!' he said. 'Why—has she gone?'" (MS p. 523; E1 p. 401) and it ends


244

Page 244
with the end of his taking and the start of L. Bristow's (MS p. 531; E1 p. 410), in the middle of the final scene between Paul, Clara and Baxter Dawes where, at Clara's decision not to go back to Nottingham with Paul, Baxter "jerked suddenly, as if he had been held on a strain. He looked out over the sea, but he saw nothing." This fragment is 406 lines of type, the smallest of the three specimens, and in it Knowles made the fewest punctuation changes, only 76, which is about one every five and a third lines, and he introduced only one new paragraph. His substantive errors were also fewer, being only two, but they were more serious than Moore's and neither of them was picked up by Lawrence or the proof-correctors.

In the first, Knowles omitted the word "now", which seems a trivial slip, but it may be significant that L. Bristow also omitted this word. Here (MS p. 524; E1 p. 402), Lawrence was describing Walter Morel's inability to look at his dead wife: ". . . he got out of the room again and left her. He never looked at her again. He had not seen her now for months, because he had not dared to look." Knowles's second error occurred during the initial awkward moments after Clara's arrival in the Skegness lodgings shared briefly by Paul and Baxter, when Paul, covering his unease by busily making arrangements, urged Clara to change out of her wet boots and offered her his slippers. Knowles set "felt" for "left" in the concluding sentence (MS p. 529; E1 p. 408): "He put the slippers near her feet. She left them there." The error was not corrected on the proofs, and "felt" continued to be printed until the 1960s.

L. Bristow

The third compositor, L. Bristow, who set the second largest share of the book, c. 2380 lines or thirteen percent of the whole, made in this extract of 520 lines (or twenty percent of his total share) ten substantive errors, four of which remained unnoticed, and 129 punctuation changes, an average of one every four lines. He made no changes of paragraphing but inserted one dubious substantive correction of MS and omitted one significant comma.

L. Bristow began at a new paragraph on the second lines of MS p. 531 (E1 p. 410), beginning "'There are one or two books in the corner,' said Morel." When he had set the page, which ended the chapter, he probably handed the leaf back to Knowles, so that he could set the last words of the preceding paragraph, which were the top line on this leaf: "over the sea, but he saw nothing." On this first page, where, at the very end of the chapter (MS p. 531), Clara asked Baxter to take her back, and ". . . put her fingers through his fine, thin dark hair . . .", L. Bristow made his first error, setting "her fingers through her . . . hair". Since this reading makes sense, only Lawrence would be likely to notice the error. He overlooked it on the galley proofs but the correct reading was restored in the first edition (E1 p. 411). Similarly, Lawrence did not notice at once that his "saw, far away," had been printed as "saw far, far away," but by the first edition the reduplication had been removed and the phrase reduced simply to "saw" (MS p. 533; E1 p. 413).


245

Page 245

The next error may have been set correctly but while the lines of type were being moved, the word "his" was dropped from the beginning of one line: "his own glass on the slopped, mahogany board. There was", and replaced at the end. The resulting nonsense was spotted by the printer's reader who wrote the correction onto the proofs before they were sent to Lawrence (MS p. 534; E1 p. 514). L. Bristow then understandably misread Lawrence's "k" as "h" and set "Colwich" for "Colwick" twice in one paragraph (MS p. 535; E1 p. 416). Only Lawrence could decide these spellings were wrong, and he corrected only one of them, the other remaining in a number of subsequent editions despite the obvious inconsistency. Another slip, "quiet" for "quite", was self-evident in context and had been corrected by the first edition (MS p. 536; E1 p. 417).

L. Bristow's last four errors are more serious and three of them have never been corrected. In the last meeting between Paul and Miriam, Lawrence inserted an extension of Miriam's inner monologue in very small handwriting at the bottom of MS p. 538, beginning with the sentences: "Oh, why did not he take her! Her very soul belonged to him. Why would not he take what was his!" The first of the inversions "not he" L. Bristow set correctly but the second he reversed to the more common "he not" (E1 p. 420). A few lines later he omitted the word "now", written clearly if small, and not obscured by interlinear revision: "'Stop now all this restlessness and beating against death . . .'". The word has never been restored (MS p. 539; E1 p. 421). But when, a few lines further on, L. Bristow again reversed Lawrence's word order, the error was more easily identified: on the galley proofs Lawrence corrected "By her tone she knew he was despising him" back to "he knew she" (MS p. 539; E1 p. 421).

Finally, on the last, now damaged, page of the manuscript Lawrence had first written, of Paul's tilt towards despair: "On every side the immense dark silences seemed pressing him into extinction, and yet, tiny speck, he could not be extinct" (MS p. 540). Then he changed the middle of the sentence to: "pressing him, so tiny a speck, into extinction, and yet, almost nothing, he could not be extinct" (E1 p. 422). Lawrence wrote his revision "so tiny a speck" very small, right at the edge of the page where it is difficult to read, but there is no doubt from the context that this is the reading he intended. However, L. Bristow set "so tiny a spark", which has never been corrected.

L. Bristow also "corrected" Lawrence's prose by inserting a conjunction into the disjointed perceptions of Paul's "semi-intoxicated trance" at the opening of "Derelict". I quote the passage in full because it exemplifies both the deliberate disjointedness of the prose and some typical effects of the compositors' punctuation changes:

Suddenly, the electric light went out, there was a bruising thud in the penny-in-the-slot meter. He did not stir, sat gazing in front of him. Only, the mice had scuttled and the fire glowed red in the dark room. (MS p. 533).

Suddenly the electric light went out; there was a bruising thud in the penny-in-the-slot meter. He did not stir, but sat gazing in front of him. Only the mice had scuttled, and the fire glowed red in the dark room. (E1 p. 414)


246

Page 246

The lack of conjunction in the second sentence has a parallel in the first, where the comma separating "light went out, there was a bruising thud" in MS suggests Paul's disjointed perception of the near-simultaneous events. In the third sentence, Lawrence's "Only," meant "But", and this was a common usage of his which frequently depended on the comma; but L. Bristow's removal of the comma changes the meaning to "The mice alone . . ." .

Finally, in the near-famous last sentence of the novel (MS p. 540): "He walked towards the faintly humming, glowing town, quickly.", L. Bristow removed the comma before "quickly", but Lawrence restored it on the galley proof (E1 p. 423). In the manuscript the comma had originally been written over a full stop, and "quickly" was added after the sentence had been completed without it. Therefore Lawrence's reinstatement of the comma indicates that its rhetorical value was important to him.

The Punctuation Changes

Not all the changes to Lawrence's MS punctuation made by the compositors altered the specific meaning of a word, as in the case of "Only," cited above. But Lawrence used commas to introduce pauses in the narratorial voice which frequently carry a subtle innuendo, and he indicated tones of voice in speech by unorthodox use of punctuation, particularly question marks, exclamation marks and dashes, a large number of which were altered throughout the novel by the printers' house-styling. Some typical effects are illustrated below by a series of parallel quotations from the galley proofs.

(1) Removal of commas

The house was empty, except for her. (MS p. 523)

The house was empty except for her. (Knowles E1 p. 401)

(In this case, the "her" is the corpse of Mrs. Morel: hence the pause indicated by the comma.)

"Have you seen her?" Annie asked of him, sharply, after breakfast. (MS p. 524)

"Have you seen her?" Annie asked of him sharply after breakfast. (Knowles E1 p. 402)

He never thought of her, personally. Everything deep in him he denied. (MS p. 525)

He never thought of her personally. . . . (Knowles E1 p. 403)

(In his preoccupation with Walter Morel's deeper side, Lawrence somehow makes the comma lead the reader to take "personally" to indicate both Mrs and Mr Morel, to include the meaning as it affected himself; whereas Knowles's removal of the comma leads the sentence to suggest something like her as a person.)

Still the curls were fine and free, but her face was much older. . . . (MS p. 537)

Still, the curls were fine and free, but her face was much older. . . . (L. Bristow, E1 p. 418)

(Here "Still" picks up references earlier in the novel to Miriam's curls; the compositor suggests a concessive "Still" meaning "Even so", or "At least".)


247

Page 247

(2) Exclamation mark changed to question mark

"But perhaps even then you were a bigger fool," said Dawes.

There was a touch of triumph and malice in it.

"Do you think so!" said Paul.

They were silent for some time. (MS p. 527)

. . ."Do you think so?" said Paul. . . . (Knowles E1 p. 406)

(3) The introduction of exclamation marks

"Sit down, Jack", he said.

"I don't want that chair," said Dawes.

"Sit down," Morel commanded. (MS p. 528)

. . ."Sit down!" Morel commanded. (Knowles E1 p. 407)

(4) Removal of dashes

"When are you going back?" she asked.

"Well—the rooms are taken until tomorrow, so he wants me to stop. He's going back tonight." (MS p. 529)

. . . "Well, the rooms. . . . (Knowles E1 p. 408)

(Dawes appears to hesitate more with Lawrence's dash, and to be more composed with Knowles's comma.)

He was a mean fellow after all—to take what he wanted and then give her back. (MS p. 530)

He was a mean fellow, after all, to take what he wanted and then give her back. (Knowles E1 p. 409)

(Lawrence's dash suggests: "So, in the end he turns out to be a mean fellow!—on the evidence of the fact that he takes what he wants and then gives her back." Knowles's commas suggest: "You must admit that he was a mean fellow to take what he wanted and then give her back.")

"Oh, I don't think it won't be a great deal. Only you'll find earning your own living isn't everything."

"No," she said, swallowing with difficulty. "I don't suppose it is—"

"I suppose work can be nearly everything to a man," he said. . . . (MS p. 537)

. . . "No," she said, swallowing with difficulty; "I don't suppose it is." (L. Bristow E1 p. 418)

(Lawrence's dash tantalisingly suggests Miriam might have said more but was interrupted by Paul; L. Bristow takes her statement as final.)

(5) Removal of exclamation marks

"Goodbye!" he said to Clara.

"Goodbye," she said, giving him her hand. (MS p. 531)

"Good-bye," he said to Clara.

"Good-bye," she said, giving him her hand. (L. Bristow E1 p. 410)

"Must you go straight home?"

She looked at him, then hid her face under her hat brim.

"No," she said. "No! It's not necessary." (MS p. 535)

. . . "No," she said "—no; it's not necessary." (L. Bristow E1 p. 416)


248

Page 248

(In both these instances Lawrence's exclamation mark suggests a sort of rousing out of lethargy, a rising to the occasion, on the part of the speaker; L. Bristow's punctuation makes them sound more neutral.)

(6) Question mark added

"Hm!" he said, as she paused at a sketch. "I'd forgotten that! It's not bad, is it."

"No!" she said. . . . (MS p. 536)

"H'm!" he said, as she paused at a sketch. "I'd forgotten that. It's not bad, is it?"

"No," she said. . . . (L. Bristow E1 p. 417)

(All the tones of voice are affected here by the removal of exclamation marks and the change from fullstop to question mark.)

The Printer's Reader's Revisions

Before the proofs were despatched to London they were marked up at Billing and Sons with corrections of typographical errors and suggestions for revision. In the first sequence of surviving galleys the queried readings were accompanied by "qy." in the margin, in the later sequence by "?". This suggests that more than one person was responsible for marking up the proofs. Lawrence responded to the majority of their promptings.

He evidently did not wish to reconsider the words "shapen" and "realest" which were underlined and queried in the margin, for he merely deleted the query (MS pp. 486, 532; E1 pp. 353, 413). But where three words were underlined and queried in his description of Dawes as "trying to get by unnoticed past every person he met", Lawrence deleted "by" and the query (MS p. 495; E1 p. 363). A similar query in the margin, without specific underlining, against "his experience had been It and not Clara" led Lawrence to replace "It" with "impersonal" (MS p. 489; E1 p. 356). In a tidy-minded vein the proof-reader underscored two sentences in a paragraph describing how Miriam sat with her "hands . . . clasped over her knee," and then, five sentences later, "suddenly took her finger from her mouth . . ." (MS pp. 57-58; E1 p. 419). Lawrence obligingly inserted a sentence: "She put her finger between her lips."

The final query concerned a repetition in the fourth paragraph from the end of the novel: "a level fume of lights. Beyond the town the country, little fuming spots for more towns . . .". Lawrence changed "fuming" to "smouldering", and then made three more changes to the paragraph, which he might not perhaps have focussed upon, had his attention not been drawn to "fume-fuming" (MS p. 540; E1 p. 422). In addition he deleted "low fuming of the" four sentences from the end of the novel (MS p. 540; E1 p. 423): "Turning sharply, he walked towards the low fuming of the city's gold phosphorescence."

In some instances the proof-readers did not merely enter a query and leave Lawrence to revise or not, they made specific suggestions. In chapter XIII, the reader wrote "qy. trs." and circled the words "nights" and "evenings" in the sentence: "The days were often a misery to her, but the


249

Page 249
nights and the evenings were usually a bliss to them both." Lawrence's word order had prioritised "nights" as the time most significant of the physical bond between Paul and Clara: the evenings were of secondary importance as leading up to the nights. By contrast the reduction to mere chronological order is banal; but only the query is deleted, and the words were reversed in the page proofs (MS p. 486; E1 p. 353). Towards the end of chapter XV, where Miriam admired the flowers on Paul's table, L. Bristow had printed correctly one of Lawrence's MS sentences lacking a conjunction: "'Have them!' he said, and he took them out of the jar, dripping as they were, went quickly into the kitchen." The reader entered a caret mark after "were," and wrote: "? and /" in the margin, and Lawrence deleted the query (MS p. 539; E1 p. 421). The other intervention of this kind was to a sentence rendered ambiguous by L. Bristow's omission of a comma: "He wanted everything to stand still so he could be with her again." The reader wrote "? that" in the margin, and Lawrence both reinserted a comma after "still" and deleted the query (MS p. 532; E1 p. 413).

Finally, the punctuation was altered in two places. Both instances occur in the passages following the eighth paragraph of chapter XV. Towards the end of Paul's semi-intoxicated inner dialogue Lawrence had written:

"Painting is not living."
"Then live."
"Marry whom?" came the sulky question.
"As best you can."
"Miriam."
But he did not trust that.

What he did not trust was a statement, but the proof-reader deleted the fullstop and wrote "?/" in the margin. It was not altered, and '"Miriam?"' appears in the page proofs (MS p. 533-534; E1 p. 414).

In the other case, fifteen lines later, Lawrence's punctuation was entirely adequate to his meaning: "Sometimes he ran down the streets as if he were mad. Sometimes he was mad: things weren't there, things were there." (MS p. 534). But the compositor, L. Bristow, had revised it to: "Sometimes he ran down the streets as if he were mad: sometimes he was mad, things weren't there, things were there" (E1 p. 415). The proof-reader deleted the comma after "mad" and wrote ";/" in the margin.

The evidence which makes it possible to distinguish the handwriting of punctuation revisions is given in the section below describing Lawrence's response to the house-styling.

Edward Garnett's Revisions to the Galley Proofs

When Edward Garnett checked through these proofs before forwarding them to Lawrence, he directed his attention to the early morning swim taken by Clara and Paul in chapter XIII. Lawrence had written two similar sentences a few paragraphs apart:

"OO" she said, hugging her breasts between her arms, "it's cold!"

. . . She gathered her breasts between her arms, cringing, laughing.


250

Page 250

"OO, it will be so cold!" she said.

He bent forward and kissed the two white glistening globes she cradled. (MS p. 491; E1 p. 358).

Garnett had already censored the passage when preparing the manuscript for the printer by changing the last sentence to "He kissed her neck." Now, on the proofs, he deleted the first of Clara's exclamations, and changed "gathered" to "hugged". He may have done this to remove a repetition, but it seems likely that his motivation was also censorship. Given how slight were the revisions that constituted censorship in The White Peacock, such as the removal of "have a real bellyful of me" in favour of "be really glutted with me",[13] it is quite possible that Garnett judged "gathered" to suggest an image of Clara presenting her breasts to Paul and "hugged" of covering them up.

Lawrence's Response to the House-styling

The surviving galley proofs amount to 1630 lines of type and in them there are 437 differences of punctuation from manuscript (counting pairs of brackets and inverted commas as one). Lawrence did not change any punctuation, but he added a little, nearly all of which had been in his MS. This provides an insight into the nature of his attention to punctuation. He made eight interventions. Three have been discussed above.[14] The others are as follows.

In chapter XIII Lawrence had made Paul reflect on Clara as they dried themselves after their swim (MS p. 492; E1 p. 360): "'But she is magnificent, and even bigger than the morning and the sea.—Is she?—is she?'" The compositor, Moore, had set this without the dashes and Lawrence reinserted dashes but in a different place: ". . . Is she—? is she—?" It seems highly improbable that anyone but Lawrence should have entered this change on the proofs, and since the notation used is a long slash which bends upwards at the end so that it looks like a tick, this provides the evidence by which it is possible to identify Lawrence's punctuation revisions. When he entered verbal revisions he used a straight slash in the margin, as did the proofreaders.

Later in the same chapter, after the fight, Baxter Dawes returned to Nottingham (MS p. 498; E1 p. 368): "And dimly in his consciousness as he went, he felt on his foot the place where his boot had knocked against on[e] of the lad's bones." Moore omitted the comma, as did Knowles in chapter XIV: after Mrs. Morel's death, Walter Morel waited up late for Paul to return home (MS p. 524; E1 p. 402): "Paul realised with a start that he had been afraid to go to bed, alone in the house with his dead."

The fact that Lawrence reinstated these commas suggests that he heard a pause in these sentences. In the second example the comma is essential to the meaning: "alone in the house"; Moore's omission of it introduces a possible ambiguity: "to bed alone."

Finally, L. Bristow had set precisely Lawrence's MS version of a sentence


251

Page 251
near the start of chapter XV which ideally required a comma (after "so"), and Lawrence provided it on the proofs (MS p. 532; E1 p. 412): "It hurt him so that things had lost their reality."

Statistically, Lawrence's alterations to the compositors' punctuation changes were minimal, but they underline the rhetorical element in his own practice. There is some evidence from The White Peacock proofs[15] that at the outset of his career he had tried to reverse more of the printer's house-styling, but he probably gave up the unequal struggle. Since this novel had been cut and censored, he had more urgent matters to attend to, and indeed, as was his habitual practice, he entered a larger number of substantive revisions.

Lawrence's Substantive Revisions on the Proofs

There is insufficient space here to quote or discuss Lawrence's revisions, but the statistics are revealing. On the surviving galley proofs, he rewrote words and sentences in twenty-nine places, only two of them in response to a query in the margin. He deleted material in twelve places. But his additions were infrequent: he added the sentence about Miriam putting her finger between her lips, and a further three short phrases of three, five and seven words.

Not only did Lawrence delete and rewrite or simply delete far more than he added, he also did so far more than the proofreaders and Garnett. Although the "interference" provided by them, which we now know about, is sometimes dramatic, we may perhaps be reassured that it does not loom larger. It is regrettable that more of the author's proofs have not survived, but we must be grateful for the information supplied by this small sample.

Notes

 
[1]

It is puzzling that Mark Sexton should assert (in his "Lawrence, Garnett, and Sons and Lovers: An Exploration of Author-Editor Relationship", Studies in Bibliography, 43 [1990], 209) that "neither the galley nor page proofs survive", since both page and galley proofs are listed in Warren Roberts's A Bibliography of D. H. Lawrence, 2nd ed. (1982), p. 524.

[2]

Mark Schorer, ed., Sons and Lovers, a facsimile of the manuscript (1977).

[3]

See James T. Boulton, ed., The Letters of D. H. Lawrence, vol. 1 (1979), 421-424 and footnotes; hereafter Letters.

[4]

Letters 1: 476-477, 481-482; George Jefferson, Edward Garnett, A Life in Literature (1982), p. 150.

[5]

These galley proofs are deposited at Nottingham University Library, U.K. Mr. Guy Collings has no theory as to what happened to the rest; they are not among the materials which Ernest Collings left to the Victoria and Albert Museum, London.

[6]

See my "Sons and Lovers The Surviving Manuscripts from Three Drafts Dated by Paper Analysis", in Studies in Bibliography, 38 (1985), 289-328.

[7]

Philip Gaskell, A New Introduction to Bibliography (1972), p. 193.

[8]

The annotations are in more than one hand, but as it is now impossible to identify the writers, it is simpler for the purposes of general exposition to refer to the annotations as if they were all by the clicker.

[9]

The fact that the galleys as cast off did not tally precisely with the galleys as printed means that simple reference to "galleys" is confusing. The following account will therefore, refer to "galley-trays" "cast-off galleys" and "galley proofs".

[10]

I am grateful to Dr. Gaskell for help with these calculations.

[11]

At least as they are represented by the page proofs, where the differences in the text caused by Lawrence's revisions would at most alter one or two lines per page and would not for present purposes materially affect quantities across whole galley proofs.

[12]

Despite Lawrence's revisions on the galley proofs which altered the exact number of lines in the same sections of page proofs, these still are more accurate figures than can be calculated by reference to the manuscript.

[13]

See The White Peacock, ed. Andrew Robertson (1982), pp. xxxv, 150, 380.

[14]

They were "Behind, the houses . . ."; "to stand still, so he could . . ."; and "town, quickly."

[15]

See Robertson's edition p. xxxii; he dealt with this matter more fully in his unpublished Proposal submitted to the editorial board of the Cambridge Lawrence edition.