[begin page 796] [begin page 797]

Roughing It was published in February 1872, a year and a half after Mark Twain began it. His second major book, it was also his second major success, comparable in some ways to his first, The Innocents Abroad (1869). Like Innocents, it was a thickly illustrated six-hundred-page volume published by Elisha Bliss and the American Publishing Company of Hartford and sold exclusively by subscription, at least in the United States. But because Innocents had been pirated in England and Canada, Clemens tried to make sure that Roughing It would not be, publishing it “simultaneously” in London through George Routledge and Sons. Just one year after publication, the combined English and American sales stood at ninety-three thousand copies, with royalties in excess of twenty thousand dollars. By Clemens’s own standards, these numbers represented an undeniable success—despite his worst fears to the contrary.

The success of Roughing It makes it easy to overlook, or discount, the author’s fears for his book, especially since his usual attitude toward it was fiercely upbeat: “We shall sell 90,000 copies the first 12 months,” he wrote typically to Bliss, “I haven’t even a shadow of a doubt of that.”1 But Clemens did have his doubts about Roughing It. Ten months before publication he was expecting it to be “a tolerable success—possibly an excellent success if the chief newspapers start it off well.” Three months before publication he was a little less confident: “If the subject were less hackneyed,” he told his wife, “it would be a great success.” Yet even that pessimistic forecast pales beside what he recalled telling his friend David Gray in early 1872, as the book was being issued: “You will remember, maybe, how I felt about ‘Roughing It’—that it would be considered pretty poor stuff, & that therefore I had better not let the press get a chance at it.”2

Acting on this belief, Clemens at first vetoed the distribution of review copies, even to the “chief newspapers.” As a result, only a handful of reviews ever appeared, and sales plummeted just six months after publication. He blamed this unexpected decline on the “engravings & paper,” and [begin page 798] on the “original lack of publicity,” admitting to Bliss that he had misjudged the importance of “early prompt notoriety,” which he had been “afraid of & didn’t want” until he was “dead sure of 50,000 subscriptions to R. I.”3

For a month Clemens made few exceptions to this ban, but among them were Charles Dudley Warner on the Hartford Courant, and William Dean Howells at the Atlantic Monthly. In late May 1872, when Clemens saw Howells’s comments, he thought the book might be a critical success after all:

Favorable though it was, this review failed to prompt any others: only three have been found later than May, and two of these Clemens himself solicited from the New York Tribune.5

Clemens’s relief at Howells’s praise is obvious. It is less obvious why he felt “afraid & shaky all along,” or why he thought of his book as a guiltily begotten “mulatto,” or even as “pretty poor stuff.” Clemens’s uneasiness was caused by two different but related concerns. First was the subject of the West itself. On the one hand, he realized that it had become suddenly “hackneyed,” even as he worked to complete his manuscript: too many journalists had already published books describing their recent tours of California and Nevada. For this reason he was prevented from writing a comic travelogue like Innocents, and was instead forced to remember and therefore transform his personal experiences “on the ground” between 1861 and 1866. On the other hand, despite his own delight in the “vigorous new vernacular of the occidental plains and mountains,” there were members of his audience who regarded slang and dialect, and the rough societies in which they flourished, as “coarse” and “low”—unfit subjects for literature, save perhaps in the refining hands of Bret Harte. Clemens therefore had at least some grounds for doubting whether, in Roughing It, he would, as he put it, “ ‘top’ Bret Harte again or bust.”6

Clemens’s second concern, which increased his uneasiness about the first, may be thought of simply as the multitude of problems he encountered in trying to write and publish Roughing It. These ranged from what [begin page 799] he referred to in 1882 as a “lapse of facility” in composing the early chapters,7 to the belated discovery that Bliss had been cutting corners with the “engravings & paper,” making books that might be regarded in more than one way as “pretty poor stuff.” Such problems ran counter to Clemens’s expectations: Roughing It proved much more difficult to write than Innocents, his only basis for comparison.

He had written Innocents in the spring and summer of 1868, published it within a year, and had since begun to receive an astonishing income from it—nearly $14,000 in the first year. “I mean to write another book during the summer,” he told Mary Mason Fairbanks on 6 January 1870. “This one has proven such a surprising success that I feel encouraged.” Producing the next book seemed at first a rather simple matter, in part because he believed he could always abandon the narrative form altogether, choosing instead to simply reprint his various sketches. On 22 January, for instance, he wrote Bliss that he was suing Charles Henry Webb to regain control of the Jumping Frog (1867), hoping to break up the plates “& prepare a new Vol. of Sketches, but on a different & more ‘taking’ model.” He was confident about the next book, whatever it turned out to be: “I can get a book ready for you any time you want it—but you can’t want one before this time next year—so I have plenty of time.”8

Clemens had less time than he thought, for during the next fifteen months he would be constantly interrupted by personal and family crises that eventually drove him to “a state of absolute frenzy,” and inevitably delayed the book.9 Three weeks after Clemens signed a contract for it, Olivia’s father died, leaving her deeply depressed. One month later a visiting schoolmate of hers (Emma Nye) contracted typhoid and, after weeks of feverish hallucinations, died in the Clemenses’ own bedroom. Five weeks after that, Olivia gave birth (prematurely) to her first child, Langdon, who was never strong and would not survive his second year. And when the child was just three months old, she herself contracted typhoid—from which she recovered slowly, once she was expected to recover at all.

But Clemens’s inability to finish Roughing It “during the summer,” or even within the five months his contract allowed him, is explained only in part by the lugubrious events of 1870–71. In particular, those events do not explain why he created half a dozen distractions during this same period. Just two months after deciding to write the book, he contracted to supply a monthly column for the New York Galaxy—a commitment so demanding that it alone could have stopped his progress on the book. In [begin page 800] addition, he undertook negotiations and preparations for three other books, long before he had completed Roughing It.

Before he had finished, Clemens admitted that for long periods during the book’s composition he was unable to be “thoroughly interested” in his subject.10 His penchant for self-interruption and this inability to be “interested” in his subject were closely related phenomena, although Clemens himself seems not to have recognized it at the time. They belong to a pattern of behavior which he eventually accepted as normal, adapting to it by learning to “pigeon-hole” manuscripts when they got “tired, along about the middle.”11 Encountering such resistance for the first time, however, must have been alarming—an open invitation to self-doubt, and to fault finding, since there were many people and things he could blame for his slow progress. But writing this book was Clemens’s first, naive experience with what proved to be his invariable pattern when writing fiction. This fact has not been noticed before, but it explains a good deal about how he wrote Roughing It.

By far the most important obstacle to such an explanation, however, has been the loss of Clemens’s manuscript printer’s copy, which was probably discarded once the book was in type. Without access to this document, it has been virtually impossible to understand, with any precision, Clemens’s letters describing his progress and recording (usually by page or chapter number) how far he had written.12 Now, for the first time, we propose a conjectural reconstruction of this missing document, which amounts to a map of what the printer’s copy initially contained, and how [begin page 801] it was revised before publication. The reconstruction was prepared by counting the words in each chapter of the first edition and arithmetically deriving from them the pagination of the printer’s copy, taking into account which passages were probably handwritten, and which were revised printings (usually newspaper clippings).13 Two charts are provided on pages 814–15: Figure 1 represents chapters 1–11, which were transmitted and revised together; Figure 2 represents chapter 12 through appendix C. Since these charts estimate pagination for the whole text by relying on word counts, they help to identify most of the pages and chapters to which Clemens himself referred during composition. And since Clemens’s references are independent of the estimates in the charts, they (and several other documents) constitute a significant check on the accuracy of the reconstruction.14

Choosing the subject of his second long book was, for Clemens, essentially a process of elimination: testing an old idea (writing about his experiences in the West) against a series of newer ideas. Until he signed the contract in July 1870, he toyed provocatively with the alternatives—publishers as well as subjects. He could write a “telling book” about England. He had heard from a subscription house in Philadelphia “offering unlimitedly.” He thought his “Noah’s Ark book” would be a “perfect lightning-striker.” The Appleton company wanted him to do a “humorous picture-book.” He gathered “material enough for a whole book” during a visit to Washington in early July (almost certainly the germ of The Gilded Age).15 And the idea of a new sketchbook persisted as an easy alternative to writing [begin page 802] something new. In fact, the first surviving indication that he had chosen the West as his topic came on 29 May 1870, in a letter to Mrs. Fairbanks: “Well, I guess we shall have to go with you to California in the Spring, for the publishers are getting right impatient to see another book on the stocks, & I doubt if I could do better than rub up old Pacific memories & put them between covers along with some eloquent pictures.”16 This remark implies his acceptance of a topic previously considered, and it assumes that the actual writing would be put off until the spring of 1871—a plan that makes sense only if he intended to satisfy Bliss’s needs with something else, presumably a sketchbook, which he could make ready “any time.”

Albert Bigelow Paine thought that it was Bliss who “proposed a book which should relate the author’s travels and experiences in the Far West.” If so, Paine’s assertion is now the sole evidence for it. More than a week after Clemens recorded his decision to “rub up old Pacific memories,” he turned aside a suggestion from Bliss: “I like your idea for a book, but the inspiration don’t come.”17 What Bliss proposed is not known, yet even if it was the western book, Clemens had anticipated him. Certainly a book based on his western experiences was not a new idea to him.

Six years earlier, Clemens had chosen the same subject (or one very like it) for what would have been his first book—a book he planned to write, and may have begun, before leaving California. All that is known of this project is contained in a letter he wrote from San Francisco on 28 September 1864 in response to a letter (now lost) from his brother and sister-in-law (Orion and Mollie) in Nevada, to whom he had earlier confided this literary ambition. “I would commence on my book,” he replied to their prodding, but he and Steve Gillis were “getting things ready for his wedding. . . . As soon as this wedding business is over, I believe I will send to you for the files, & begin on my book.”18

The “files” Clemens referred to in 1864 were probably the same as the “coffin of ‘Enterprise’ files” which, in early March 1870, he finally did ask Orion to send him from St. Louis, acknowledging their receipt in Buffalo on 26 March—several months before he signed a contract for his second book.19 These files were surely one reason that Roughing It at first seemed a relatively easy book to write. The files were scrapbooks filled with clippings of his work in the Virginia City Territorial Enterprise, but also of work by various other hands in the Enterprise and other Nevada and California newspapers. To judge from the surviving scrapbooks, the clippings documented Orion’s activities as well as his own, and included even multiple copies of work Clemens had published routinely in the West: everything [begin page 803] from local columns and letters in the Enterprise and the San Francisco Morning Call, to long, carefully crafted sketches in the Californian and the Golden Era.20

For Clemens, the scrapbooks held out the promise of two kinds of help. Some, perhaps most, of their clippings would simply remind him of the facts—incidents, people, and stories from his western years. But others, especially clippings of his own work, could surely be revised and reprinted as chapters in the book. Like Innocents, the new book would be a fundamentally factual account, a personal narrative of a real trip—with the trip providing an automatic source of coherence, as well as an excuse for humorous digression. But unlike Innocents, which Clemens had written within a year of the voyage it recounted and largely by cobbling together newspaper letters composed during that voyage, Roughing It had to be written five to ten years after the events it described, and without the help of any such contemporary account, save what could be salvaged from the scrapbooks—supplemented by whatever collateral material Orion was able to provide.

Between their arrival on 26 March 1870 and the signing of the contract in mid-July, the scrapbooks may well have prompted Clemens to write two long western sketches: “The Facts in the Great Land Slide Case,” published in the Buffalo Express on 2 April and later revised and reprinted in chapter 34 of Roughing It; and “A Couple of Sad Experiences,” about his notorious western hoaxes (“Petrified Man” and “A Bloody Massacre near Carson”), written no later than April and published in the June Galaxy, but not reprinted (or even mentioned) in the book. But as a source of ready-made chapters about Nevada and California, the scrapbooks turned out to be much less useful than Clemens had anticipated. He used scarcely anything of his own from the Enterprise, and nothing at all of what he had published in the San Francisco Morning Call, the Californian, and the Golden Era. Samuel C. Thompson, who was briefly Clemens’s [begin page 804] private secretary the year after he published Roughing It, recalled that he “seemed to have no great esteem for his newspaper contributions. They were too hurried and shallow. He told me that in preparing to write ‘Roughing It,’ he searched newspaper files covering years of his writings and got no help from them.”21

Although Roughing It may have been conceived as early as 1864, it was finally written at a time of greatly heightened interest in literature about the West—especially humorous short fiction and “dialect” poetry. Bret Harte and John Hay, both friends of Clemens’s, achieved instantaneous celebrity in 1870–71 through their dialect poems.22 Clemens himself was not much interested in poetry, but he was interested in dialect, and he was already well known for his western sketches. In a 27 December 1870 editorial in the New York Tribune, entitled “The Western School” (promptly reprinted in Clemens’s Buffalo Express), Hay announced that a “vigorous and full-flavored literature is growing up in the West. The period of echoes and imitations, of feeble reproductions of bad models . . . has gone by, and a school of writers is now coming up on the further side of the Alleghenies who have a message of their own to deliver, and who are uttering it in a way distinctly their own.” Hay cited George Horatio Derby (John Phoenix) as the “leader and founder” of this “field of eccentric fiction,” which had “since been so successfully worked by Mr. Francis Bret Harte and Mark Twain.”

Hay was in fact protesting a recent tendency to treat Bret Harte as the founder and sole legitimate member of this “Western school,” and to regard all others (including Hay himself) as Harte’s incompetent imitators.

For his part, Clemens felt personally indebted to Harte, telling Thomas Bailey Aldrich just a month after this editorial appeared that Harte had “trimmed & trained & schooled me patiently until he changed me from an awkward utterer of coarse grotesquenesses to a writer of paragraphs & [begin page 805] chapters that have found a certain favor in the eyes of even some of the very decentest people in the land.” Clemens was thinking especially of Harte’s most recent help, for Harte had “read all the MS of the ‘Innocents’ & told me what passages, paragraphs & chapters to leave out—& I followed orders strictly.”24 But Clemens also felt keenly competitive with Harte, who was rising meteorically in the eyes of the eastern literary establishment, even as Clemens struggled with the manuscript of Roughing It. During this period, Clemens reviewed Harte’s Poems in a way that suggests where he felt the competition between them was most intense:

Clemens’s own rank in American literature could not have seemed to him certain or permanent when he wrote this review—but he had written and published dozens of sketches about the West, especially California. Now he had to confront the problem of how to make a book-length narrative from such materials.

Clemens’s 1867–68 newspaper correspondence and even The Innocents Abroad contained many examples of his impulse to reminisce about his western experiences.26 But the true precursor of a narrative about the West arose from his longstanding interest in it as a lecture topic. As early as January 1867, he wrote Edward P. Hingston that he intended to speak in New York and other eastern states “on California & perhaps on other subjects.” Although Clemens did not write such a lecture that year, in May 1868 he again expressed his intention to “get up a lecture on California.”27 The following year he finally drafted his first narrative about the West, specific plans for which he described in a 10 May 1869 letter to James Redpath, his lecture agent. Clemens conceived the subject both as a lecture topic and as the subject of some newspaper letters he planned to write while revisiting California—and thus essentially in the mode of a tourist’s contemporary report:

Clemens never did return to California, nor did he ever deliver “Curiosities of California,” but he did write it. By 5 July he had “written more than enough for a lecture,” which he said “must be still added to & then cut down.” Only a single fragment of what he wrote has survived in manuscript (a description of Lake Tahoe),29 perhaps because he soon made use of the rest of it in the Buffalo Express—though not as a single narrative, but as a series of loosely connected articles. If we assume that he wrote “Curiosities of California” about as he described it to Redpath, then he probably published it between 16 October 1869 and 29 January 1870 as the so-called “Around the World” letters, five of which he subsequently reused in Roughing It.30 It may have been the potential of this protonarrative which led Clemens to ask Orion for the scrapbooks in early March 1870.

By the end of July Clemens had tentatively settled on the idea of a book about the West. According to Paine, Bliss came to Elmira in “early July” to negotiate the contract for Roughing It, but he seems not to have come before the middle of the month. On 4 July, Clemens wrote Bliss to say that he would be in Elmira “10 days or 2 weeks yet,” and to urge him to “Come—come either here or to Buf.”31 On the same day, however, Clemens took the train to Washington, D.C., returning to Elmira no sooner than 10 or 11 July. On 5 July, the Washington correspondent of the Sacramento Union had “quite a chat with him” in Washington:

This interview shows that Clemens was already committed to publishing his next book with Bliss, ten days before the formal contract was drawn up and signed. The original draft of that contract, in Bliss’s hand, indicates that it was “made this 15th day of July, AD 1870, at Elmira.” In it Clemens agreed to write a “manuscript for a book upon such subject as may be agreed upon,” and to deliver it “as soon as practicable, but as early as 1st of January next if they the sd company shall desire it. Said manuscript to contain matter sufficient for a book of about 600 pages octavo.” The contract forbade Clemens “to write or furnish manuscript for any other book unless for said company, during the time said manuscript & book are being prepared & sold.” And it stipulated that the American Publishing Company would “publish the said book in their best style—to commence operations at once upon receipt of manuscript & to push it through with all the despatch compatible with its being well done in text & illustrations.” The company also agreed to “a copyright on every copy sold of seven & one half per cent of the retail or subscription price.”33

In 1906 Clemens recalled in some detail—not all of it trustworthy—how he and Bliss had arrived at this percentage:

The contract did not name the subject of the new book, but the omission was deliberate, as Clemens explained to Orion on the day he signed it:

Orion did indeed have a memorandum book of the trip, which he promptly sent. He also agreed to write out some notes about their Nevada experiences, although these would not be ready until early November. Clemens did not, however, really begin writing on 15 July. Almost two weeks later he told his mother and sister that he was “going to write a 600-page 8vo. book (like the last) for my publishers (it is a secret for a few days yet.) It will be about Nevada & California & must be finished Jan 1. I shall begin it about a month from now. By request, Orion has sent me his notebook of the Plains trip.”36

Meanwhile, Bliss evidently voiced a suspicion that Clemens was unhappy with the terms of the contract, and was considering another publisher. On 2 August, Clemens reassured him in such candid terms that the letter still serves as a useful corrective of his 1906 recollections:

[begin page 809] Four days later, on 6 August, Olivia’s father died. Although Clemens had already decided not to begin on his book until late August, Jervis Langdon’s death ensured that he could not write anything else in August either. The funeral was held in Elmira on 8 August, the will probated on 12 August, and a memorial service conducted on 21 August.38 Clemens refrained from publishing anything in the Buffalo Express until 25 August, when “Domestic Missionaries Wanted” appeared.39 Presumably by then, or shortly thereafter, he also began on the book, for he wrote Orion on 2 September:

Orion’s journal served Clemens as an indispensable tool in writing the overland chapters (1–20). He relied on it for names, distances, times, landmarks, several particular incidents, and even its day-by-day account of the journey. He treated the original record like an outline, from which he could diverge at will without losing the essential thread of actual events. Especially at this early stage of composition, he seems also to have adhered closely to its chronology, and even its language. For example, when Orion noted on the fifth day, “Arrived at the ‘crossing’ of the South Platte, alias ‘Overland City,’ alias ‘Julesburg,’ at 11 A. M., 470 miles from St. Joseph,” Clemens reproduced Orion’s sentence almost verbatim, adding only a concluding phrase of his own: “At noon on the fifth day out, we arrived at the ‘Crossing of the South Platte,’ alias ‘Julesburg,’ alias ‘Overland City,’ four hundred and seventy miles from St. Joseph—the strangest, quaintest, funniest frontier town that our untraveled eyes had ever stared at and been astonished with.”41

Even with such help in hand, however, Clemens’s circumstances did not favor rapid composition. On the last day of August he wrote to his sister:

Within days Clemens reported that Emma Nye had become “right sick—she cannot go on to Detroit yet awhile, where she is to teach.” But he also reported progress on his manuscript: “I have written four chapters of my new book during the past few days, & I tell you it is going to be a mighty starchy book—will sell, too.”43 Two days later, on 4 September, he fired off a salvo of reassurance to Bliss:

His subject was now no longer a secret, for on 7 September the Elmira Advertiser reported that “Mark Twain’s new book, which is to be published next spring, is to be an account of travel at home, describing in a humorous and satirical way our cities and towns, and the people of different sections.”45

Clemens was soon obliged to confess that he had “no time to turn round” because Emma Nye was “dying in the house of typhoid fever (parents are in South Carolina) & the premises are full of nurses & doctors & we are all fagged out.”46 Still, he must have continued to write, for by 15 September he had completed two more chapters, which took him up through noon on the fifth day, at the end of chapter 6. By now he was already looking ahead to the ninth day, the subject of what became chapter 10, described in Orion’s journal as “Breakfast at Rocky Ridge Station, 24 miles from ‘Cold Spring,’ and 871 miles from St. Joseph.”47 Clemens had his own very particular memory of that day, which he had recalled the previous winter in his seventh “Around the World” letter:

But Clemens felt the need for more documentation of Slade’s history than he could find in the scrapbooks, or in Orion’s journal. He therefore wrote to the postmaster of Virginia City (Montana Territory), who was Hezekiah Hosmer, former chief justice of the territorial supreme court:

Hosmer’s reply to this appeal has not been found, but it was very likely he who referred Clemens to Thomas J. Dimsdale’s little book, The Vigilantes of Montana (1866), a compilation of newspaper stories Dimsdale had written and published in the Montana Post in 1865–66. Hosmer may even have realized that Dimsdale’s book would be hard to find in Buffalo, and therefore sent Clemens a copy of it—if not immediately, then probably within a month or two. The exact timing remains uncertain, because even though Dimsdale’s book became the explicit source for much of what Clemens said about Slade in chapters 10 and 11, he did not complete those chapters until mid-March 1871.

[begin page 812] Bliss had, meanwhile, asked Clemens to contribute to the trade newspaper he was starting up as an advertising medium for the American Publishing Company’s books. On or about 21 September, Clemens dashed off a note in reply: “Yes, will furnish article for paper. . . . Finished 7th or 8th chap. of book to-day, forget which—am up to page 180—only about 1,500 1500 more to write.”50 This statement, however, cannot be taken literally as a reference to chapters 7 and 8 as they were published in Roughing It—nor is it obvious why, having written “up to page 180,” Clemens then had “only about 1500 more to write.” Just where was he in the process of composition? To answer that question, it is first necessary to explain the reconstruction charts in greater detail.

Roughing It was, in most ways, a typical subscription book. Bliss published a salesman’s prospectus containing selected pages and illustrations from it. Both the prospectus and the book contained highly particular tables of contents and lists of illustrations, which can be compared with each other and with the texts and illustrations they refer to. Inconsistencies between the lists and the texts may, of course, be simple errors, but they may also reflect incomplete correction or revision. In this particular case, the bibliographical evidence of the prospectus is much richer than usual because Bliss began its production substantially before Clemens had completed his manuscript. Issued in November 1871, the first prospectus (Pra) was so filled with signs of incompleteness and inadvertence that Bliss took the equally unusual step of issuing a revised and corrected form of it (Prb) two months later. Much of what can now be pieced together about the history of composition and revision (especially the evolution of the early chapters, represented in Figure 1) depends on the evidence of “errors” in Pra—errors, that is, in the sense that they refer to or are part of an earlier form of the text than the one finally published in Roughing It.

The wording of some of the column headings in the charts also needs clarification: “clippings,” “real page nos,” and “equiv page nos” (for “equivalent page numbers”). It is not self-evident why Clemens said on 21 September 1870 that he had only “about 1500 more pages to write.” Several later statements show that he believed he needed a total of 1800 pages of printer’s copy to make a 600-page book—three pages of printer’s copy for each book page, an empirical average that took into account the variable [begin page 813] space needed for the illustrations.51 Yet when 1800 is reduced by 180 to 1620, that number does not seem a plausible approximation of “about 1500,” even if it is assumed that Clemens was rounding off to the nearest hundred.

This statement is, in fact, the earliest indication we have that Clemens sometimes counted the pages of the Roughing It printer’s copy, as he had earlier done with Innocents, in a way that took into account the larger number of words typically contained in a page made up wholly or in part from clippings. That Clemens used clippings to help make up the printer’s copy for both Innocents and Roughing It is now well established. One of the three surviving pages of the Roughing It manuscript (reproduced in facsimile on page 816) amply demonstrates how such clippings were mounted and then altered in the margins. (The page is extant because it was discarded after Clemens further revised it.) In June 1868, when the printer’s copy for Innocents was nearly complete, Clemens told Mrs. Fairbanks that he was “writing page No. 1,843. 2,343.” But the printer’s copy was never much longer than about 1300 pages: a manuscript chapter about Spain (very near the end of the book, and ultimately omitted from the text) was numbered 1289 through 1331. And in April 1869, when Bliss needed to cut the manuscript to fit within 650 book pages, Clemens wrote that he hoped “there won’t be a necessity to cut much, but when you say you are only to the 800 or 900th page you don’t comfort me entirely, because so much of the 400 or 500 pages still left are reprint, and so will string out a heap.”52 So it is clear that when he wrote Mrs. Fairbanks in June 1868, he added first 500, then 1000, pages to the real total of about 1343 pages. His hesitation between the two larger figures (1843 and 2343) shows that he was actually multiplying the same 500 pages of “reprint” (i.e., clippings from the San Francisco Alta California) first by two, then by three, so that the total equaled what he would have had if everything were in his handwriting. Clemens must have realized that he had incorrectly evaluated the length of his Innocents printer’s copy, since he ultimately had too much material. Now, with Roughing It, he seems to have decided that a clipping page contained four times the number of words on a holograph page—not merely two or three, as he had earlier calculated. So if a holograph page averaged 84 words (as assumed here), Clemens counted a clipping page as 336 words—the equivalent, in other words, of four holograph pages (4 × 84 = 336 words), or three more pages than were actually in the copy (real pages). It is this method of counting which explains his otherwise puzzling arithmetic here and in later statements about his

[begin page 816]

[begin page 817] progress—and which accounts for the headings on those columns of numbers in Figures 1 and 2.

[begin page 814] [begin page 815]

In the printed book, Figures 1 and 2 occur on pp. 814–15. The figures’ complex tables are available in PDF format.

How far had Clemens progressed with his book in September 1870, when he had written “up to page 180”? Figure 1 indicates that at the earliest “mappable” stage of composition, in March 1871, the last page in chapter 8 was 170, nine pages shy of what Clemens himself had said the previous September (assuming that “up to page 180” refers to the first page of chapter 9). In making this statement, however, Clemens may have considered that the clipping from the New York Times in chapter 4, which was slightly over one page long, was equivalent to five holograph pages. That would make the equivalent last page of chapter 8 number 174, just five pages shy of the number Clemens assigned it (179).53

If each clipping page that Clemens prepared was considered to contain the equivalent of four holograph pages, then he could reduce the total number of pages he needed by four times the number of clipping pages he used, or planned to use. If he estimated, for instance, that he would prepare 30 pages of clippings, then simple arithmetic shows that having written 180 pages of holograph, he could also count as “in hand” some 120 pages (4 × 30) which would result from these 30 clipping pages, for a total of 300 pages in hand—leaving him 1500 pages still to write.

Collation shows that by the time Clemens completed his book late in the fall of 1871, he had used more than 130 pages of clippings. But in September 1870, at the outset of composition, he expected to use many fewer, largely because he had not yet decided to reuse any of his 1866 letters to the Sacramento Union. Within a few weeks of his 21 September note to Bliss, he probably was able to count the clipping pages he planned to use from the Buffalo Express, and by December he had prepared nearly all of these by mounting and revising them for resetting. By that time, however, his plans for them had changed: he now intended to include them in a new sketchbook, rather than in Roughing It. On 22 December he told Bliss that he had “arranged” the principal sketches for this sketchbook, and by [begin page 818] sometime in January 1871 he had produced a working table of contents.54 When the sketchbook was again postponed, some of these mounted clippings found their way back into Roughing It: revised versions of “Around the World” letters 1, 3, 4, and 5, as well as “The Facts in the Great Land Slide Case,” all first published in the Express. If we count the number of clipping pages they represent—that is, if we count only the Express clippings Clemens planned at the outset to use in Roughing It—the total is 25. If he had used that figure (25) to compute the number of pages still left to write, the result would have been 1520, or “about 1500.”55

Although Clemens said on 21 September that he had reached the end of chapter 8, Figure 1 indicates that chapters 1–8 as first drafted differed in several ways from their published form. Chapter 6, for instance, was probably only about eleven manuscript pages, roughly half as long as the published version—lacking perhaps a long section in the middle. Chapter 8 ended with the pony-express passage, which now begins the chapter, and it began with the story of Bemis and the buffalo, which now fills all but the last three pages of chapter 7. The anecdote about Eckert and the cocoanut-eating cat, which now occupies the last three pages of chapter 7, was almost certainly not in the manuscript at all, for the references it makes to Siam have been shown to derive from Anna H. Leonowens’s English Governess at the Siamese Court, which Fields, Osgood and Company did not publish until December 1870, and which Clemens did not buy and read until sometime in 1871.56 Finally, chapter 7 was very likely given over to something that Clemens eventually left out of the book entirely—a description of Overland City (or Julesburg), “the strangest, quaintest, funniest frontier town that our untraveled eyes had ever stared at and been astonished with.”57

On or about 28 September (one week after finishing the “7th or 8th chap.”) Clemens may also have written to his former colleagues on the Enterprise for help similar to the kind he had asked of Hezekiah Hosmer. The Enterprise reported:

Two days’ journey “the other side” of Salt Lake City put Clemens on the ninth day, crossing the Continental Divide, which he described in chapter 12. He had not, presumably, heard yet from Hosmer on the subject of Slade, and he would leave chapters 10 and 11 incomplete until March.

The evidence of the next few months suggests that Clemens was having difficulty writing much beyond chapter 12. A dozen years later, he recalled that his problem was caused by his having given up smoking “during a year and a half”:

Although it shortens the time span, this recollection seems to be an accurate description of his difficulty, though perhaps not of its true or only cause. At any rate, Clemens did not resume his daily cigars until early 1871. (In mid-December 1870 he was still permitting himself to smoke only “from 3 till 5 on Sunday afternoons.”)60 The “lapse of facility” in composition actually lasted from September 1870 through most of March 1871.

On 29 September, at about ten o’clock in the morning, Emma Nye died in the Clemenses’ Buffalo bedroom. Exhausted by the ordeal, Clemens and his wife left for a week’s visit with his sister and mother (who had recently moved into their new home in Fredonia, New York), returning to Buffalo on 6 or 7 October. (His sister, Pamela, returned this visit shortly thereafter, staying for several weeks, at least in part because Olivia’s first pregnancy was then in its seventh month.) While in Fredonia, Clemens sent Bliss a contribution (not identified) for his trade newspaper—but work on the book had obviously been interrupted, and his attention began increasingly to be diverted to other projects.61 A week after returning to Buffalo, he wrote Bliss that he had “a notion to let the Galaxy publishers have a volume of old sketches for a ‘Mark Twain’s Annual—1871’—provided [begin page 820] they will pay me about 25 per cent. . . . What do you think? Write me at once—& don’t discourage me.” But he also confessed that he was “driveling along tolerably fairly on the book—getting off from 12 to 20 pages (MS.) a day. I am writing it so carefully that I’ll never have to alter a sentence, I guess, but it is very slow work. I like it well, as far as I have got. The people will read it.” On the same day he answered a question from Mrs. Fairbanks, implying that he had made less progress than he had expected: “My book is not named yet. Have to write it first—you wouldn’t make a garment for an animal till you had seen the animal, would you? I am getting along ever so slowly—so many things have hindered me.” He had not written to her earlier, he said, “because I am in such a terrible whirl with Galaxy & book work. . . . I never want to see a pen again till the task-hour strikes next day.”62

Clemens assumed that the burden (or lure) of writing for the magazine was affecting his capacity to sustain interest in the book. Five days later he told Francis Church that he would probably retire from the Galaxy in April “because the Galaxy work crowds book work so much.” But he did not resign immediately, he said, because he was “very fond of doing the Memoranda, & take a live interest in it always—& so I hang on & hang on & give no notice.” On 26 October, when he read Bliss’s inevitable objection to the book of Galaxy sketches, he replied that it was “too late now to get out the annual” anyway. But he took issue with Bliss’s notion that “writing for the Galaxy” hurt the sale of his books: “I cannot believe it. It is a good advertisement for me—as you show when you desire me to quit the Galaxy & go on your paper.” Still, if someone could prove that he was harming his reputation by writing for the Galaxy, he would “draw out of that & write for no periodical—for certainly I have chewed & drank & sworn, habitually, & have discarded them all, & am well aware that a bad thing should be killed entirely—tapering off is a foolish & dangerous business.”63

That Clemens was beginning to feel inhibited, if not yet “seriously obstructed,” in his efforts to write may also be inferred from his pursuing research into Slade and other matters. When he saw Bliss’s first issue of the Author’s Sketch Book at the end of October, he found several promising sources among the books advertised:

Bliss must have sent some of these books, but the request shows that Clemens had probably not drafted chapter 14 or 15, since passages in each suggest his familiarity with one of the books he asked for: Albert Deane Richardson’s Beyond the Mississippi. He had certainly not completed chapter 19, which explicitly refers to another of them, John George Wood’s The Uncivilized Races, or Natural History of Man. Clemens owned and annotated a copy of Wood’s book—and no doubt of Richardson’s as well, although his copy is not known to survive—in addition to a copy of (Colonel) Albert S. Evans’s Our Sister Republic: A Gala Trip through Tropical Mexico in 1869–70; all three books were published in 1869 or 1870 by the American Publishing Company or its subsidiaries.65

Soon there were additional distractions. During Pamela’s visit she had apparently suggested that if Orion were to succeed with the “machine” he was trying to invent, he needed a job that would give him more free time. On the last day of October, therefore, Clemens wrote Bliss, casually proposing quite another sort of favor.

In a final paragraph, Clemens made sure Bliss got the point: “When is your paper coming out? Did you ever receive the article I sent you for it from Fredonia? Tell me.”66 Two days later, Bliss cheerfully took the hint: “Yours recd Yes I got your article. ‘It is accepted’ (a. la. N.Y. Ledger) Thanks for same— . . . How would your Bro. do for an editor of it?” Clemens immediately sent this reply to Orion, urging him to “throw up that cursed night work & take this editorship.”

On 7 November, Olivia gave birth prematurely to Langdon. Four days later, Clemens told Orion that his wife was “very sick,” and that he did not believe the baby would “live five days.” At the same time he thanked Orion for sending “such full Nevada notes—though as they have just come & I am stealing a few minutes from the sick room to answer a pile of business letters, I haven’t read a sentence of them yet.” These notes have not been found, and Clemens may have done no more than glance at them; in early April 1871 he admitted to Orion that he had misplaced them.68

Olivia’s older sister, Susan Crane, had been visiting the Clemenses when Langdon arrived, and she remained to help care for Olivia and the baby until 12 November, when her place was taken briefly by Mrs. Fairbanks, visiting from Cleveland. On 19 November, after Fairbanks’s departure, Clemens wrote his mother-in-law that Olivia “lets me go up to the study & work, (which I ought not to do & yet I am so dreadfully behind hand that I get blue as soon as I am idle).” Later that month, he reported that he worked “in my particular den, from 11 AM till 3 P.M., rain or shine.”69 Even so, it is unlikely that during this month he advanced the narrative much past chapter 19, in which he could now make use of Wood’s Uncivilized Races. Most of his time was instead devoted to installments for the Galaxy—and part of it surely went to hatching plans for other projects, the first of which emerged in a 30 November letter that asked Bliss for an advance of $1500:

Demanding absolute secrecy and a royalty of 10 percent, Clemens went so far as to copy this letter (lest anyone steal the idea) before sending the original to Bliss. So completely did his new idea eclipse the western book that only in the next-to-last postscript of this long letter did he remember [begin page 823] to mention, somewhat lamely, that “Mrs Fairbanks (my best critic) likes my new book well, as far as I have got.” (Fairbanks had returned home more than a week earlier.) Clemens’s infatuation with his book-by-proxy scheme betrays the inventor’s pride in his invention, but the invention itself suggests how he may have diagnosed his problem with the western book—that is, as a lack of interest, due in part to the distance in time of his experiences. “I don’t care two cents whether there is a diamond in all Africa or not,” he wrote Bliss, “the adventurous narrative & its wild, new fascination is what I want.”70 The diamond-mine book (which was never really begun, let alone published) now seems a chimera, but it would be hard to find a more succinct expression of Clemens’s frustration with his western materials—especially if one considers that he obviously believed, for almost a month, that he could write about someone else’s recent adventures more easily than he could about his own relatively remote ones.

By Clemens’s own account, it would be several years before he learned that virtually any book he started to write was “pretty sure to get tired, along about the middle, and refuse to go on with its work until its powers and its interest should have been refreshed by a rest and its depleted stock of raw materials reinforced by lapse of time.” In 1906 Clemens recalled encountering such an interruption at “page 400” of the manuscript for Tom Sawyer, where “the story made a sudden and determined halt and refused to proceed another step.” That was probably in 1872, when he had written only about 100 pages, although he would eventually write another 400 before being interrupted by a second “halt.” What he could do about such interruptions became clear only after he began writing those 400 pages in 1874:

In the fall of 1870, Clemens knew that he was “obstructed” in the composition of his western book, but it is unlikely that he understood why or what to do about it. He seems to have alternated between pushing stubbornly ahead (trying in various ways to make himself write) and switching off, with more or less deliberation, to “other things.” Just three days after [begin page 824] announcing his diamond-mine scheme to Bliss, for instance, he declined an invitation to write for the Boston Saturday Evening Gazette on the grounds that

Yet “loaded” though he was, just one week later he had launched another project, quite unrelated to the western book, fully expecting to complete it before the month was out. On 7 or 8 December he telegraphed his Galaxy publishers, offering them a “pamphlet” that he wanted to publish in time for the holidays. On the morning of 9 December, Isaac Sheldon telegraphed his reply: “We will publish it & give you half of all profits.” In a letter written the same day, Sheldon promised to “do our very best as to getting it out in time &c &c,” but warned that it was “now of course late in the season to get out a book and there are always delays we can never calculate on.”73 This “book” became the fifty-page pamphlet known as Mark Twain’s (Burlesque) Autobiography and First Romance, which would not be published, as it turned out, until early March 1871.

The day after Sheldon’s telegram, 10 December, Clemens “shot off to New York to issue a pamphlet, & staid over 7 days,” not returning to Buffalo until 17 December. While in New York, he stayed at the Albemarle Hotel and (as he later confided to Twichell) “smoked a week, day & night.”74 Free (at least momentarily) from domestic concerns, and fully occupied with projects other than the western book, Clemens soon took the opportunity to postpone his original deadline. He spent the week only in part on the pamphlet, for which he sought to hire an illustrator, Edward F. Mullen, eventually settling for Henry Louis Stephens instead.75 He began his week in negotiations with Charles Henry Webb, from whom he succeeded in buying control of the Jumping Frog. On 11 December, he wrote an obituary letter about his old friend Reuel Gridley, whose “Famous Sanitary Flour Sack” he would treat at some length in chapter 45 of Roughing It. He gave this letter (signed “s. l. c.”) to the New York Tribune, which published it two days later.76 Perhaps while delivering this manuscript to the Tribune offices, Clemens called on several professional colleagues, including managing editor Whitelaw Reid and Reid’s first lieutenant, John [begin page 825] Rose Greene Hassard, as well as a recent addition to the Tribune editorial staff, John Milton Hay.

Clemens had become better acquainted with Hay through their mutual acquaintance in Buffalo, David Gray. Raised on the banks of the Mississippi, Hay shared Clemens’s midwestern roots and to some extent his literary tastes and ambitions. In December 1874, when Clemens published his first installment of “Old Times on the Mississippi” in the Atlantic, Hay declared it “perfect—no more nor less. I don’t see how you do it. I knew all that, every word of it—passed as much time on the levee as you ever did, knew the same crowd and saw the same scenes—but I could not have remembered one word of it all. You have the two greatest gifts of the writer, memory and imagination.”77 At this time in 1870, Hay was enduring a mixed reaction to “Little-Breeches,” which he had published (signed with his initials only) in the Tribune for 19 November, and which had since been widely reprinted. In a conversation that must have occurred during Clemens’s week-long visit to New York between 10 and 17 December (in 1905 Clemens remembered only that it had been “in 1870 or ’71”), they discussed this reaction. Clemens recalled that “Hay made reference to mention of the current notion that he was an imitator; he did not enlarge upon it, but he was not better pleased by it than you or I would be.” And Clemens also mentioned that his talk with Hay was “incidentally” the occasion of his “getting acquainted with Horace Greeley”:

[begin page 826] It is not too much to suppose that for Clemens, such gruff treatment demanded revenge—or at least poetic justice.

On 13 December, Clemens wrote Bliss again about the diamond-mine book, this time inviting him to come to New York to discuss it, which Bliss promptly did.79 And at about the same time, Clemens met with his intended proxy for the diamond-mine book, John Henry Riley, who had come up from Washington for that purpose. Later in the week Clemens received an invitation to join Whitelaw Reid, and “one or two friends, not more” over a “quiet bottle of wine” at the Union League Club, but by then he was too busy to accept.80

Well aware that he was progressing too slowly to meet his deadline for the western book, Clemens proposed to Bliss that he (rather than the Galaxy publishers) first publish a sketchbook in 1871, postponing the western book until later the same year—the very alternative he had had in mind since at least January 1870. Bliss tentatively accepted this change, in part because Clemens promised to discuss it further, in Hartford, as soon as his New York business was over. But instead of going to Hartford, Clemens returned to Buffalo, where on the afternoon of 17 December he telegraphed Bliss: “Got homesick. Will come shortly with sketches & manuscript.” Bliss replied that he was “most disappointed at your not coming here, & so was Twitchell— . . . Your Brother is here & we are getting at work in earnest. . . . Let me know about Mss. & also about the sketches & come on & have a talk if possible.”81 These references (and Bliss’s reply, quoted below) establish that Clemens expected to work on both books simultaneously, which meant submitting the western book in sections, as he completed them. But it would be mid-March before any manuscript for the western book was sent to Bliss, and early June before Clemens himself carried a later part of it to Hartford.

On 20 December, Clemens returned to Bliss the signed contract for the diamond-mine book: “Riley is my man—did I introduce him to you in New York? He sails Jan. 4 for Africa. . . . Riley is perfectly honorable & reliable in every possible way—his simple promise is as good as any man’s oath. I have roomed with him long, & have known him years. He has ‘roughed it’ in many savage countries & is as tough as a pine-knot.”82 Two days later he acknowledged the $1500 advance “for the foreign expedition,” [begin page 827] then launched into great detail about the sketchbook, trying to goad Bliss into immediate action:

Bliss replied cautiously, on 28 December: “Yours of 22nd rec’d. Glad to hear you are progressing with the Books— . . . Yes we will have Mullen illustrate the sketch book all right. . . . Are you coming on? Will canvass for Sketch book as soon as Prospectus is ready for it.”84 Clemens would need to be reminded more than once in 1871 that canvassing for a book could not begin until its prospectus was ready, which required a major portion of the manuscript. Oblivious to this hint, he replied on 3 January, trying to turn up the pressure on Bliss: “Name the Sketch book ‘Mark Twain’s Sketches’ & go on canvassing like mad. Because if you don’t hurry it will tread on the heels of the big book next August. In the course of a week I can have most of the matter ready for you I think. Am working like sin on it.”85 The next day he sent Bliss the manuscripts for two (possibly three) “new” sketches, specifying that “the one about the liar is to be first one in the book,” and insisting that Orion make security copies for all of them before they were sent to the artist for illustration. On 5 January he added a postscript: “The curious beasts & great contrasts in this Pre-duluge article offer a gorgeous chance for the artist’s fancy & ingenuity,” he wrote. “Send both sketches to Mullen—he is the man to do them, I guess. Launt Thompson, Albemarle Hotel, will find him when wanted.”86 The “Pre-duluge article” was an extract from the still incomplete “Noah’s Ark book.” The “one about the liar” was undoubtedly the Sandwich Islands sketch about Markiss, soon to be published in the Galaxy as “About a Remarkable Stranger” and eventually reprinted as chapter 77 of Roughing It. (The third “new” sketch, if any, has not been identified, although it may have been one that Clemens had listed as “Sailor Story,” and that may in turn have become chapter 50 in Roughing It.) Clemens presumably wrote [begin page 828] these “new sketches” only after preparing clippings of those he intended to reprint from the Express and the Galaxy, including “Around the World” letters 1, 3, 4, and 5, as well as “The Facts in the Great Land Slide Case”—all listed in the table of contents he drafted for the sketchbook, but all used ultimately in Roughing It.87

This “halt” in the composition of Roughing It gave Clemens’s “tank” time to refill. But it also gave him time to reflect on the hazards of his subject. The West, which in May 1869 had seemed to have “scope” and timeliness because of the recent “completion of the Pacific RR,” had by early 1871 been so often written about that Clemens could rightly call it hackneyed.88 In 1869–70 alone his fellow journalists had published half a dozen books on western travel, all based in part on sketches they had written for newspapers or magazines. Prominent among these were Richardson’s revised edition of Beyond the Mississippi, Samuel Bowles’s Our New West), J. Ross Browne’s Adventures in the Apache Country, and Fitz Hugh Ludlow’s The Heart of the Continent. Clemens must have been made increasingly aware that the West as described by passing visitors was no longer a novelty.

Clemens was also conscious of the risks he ran in renouncing journalism to write books—an intention he recorded as early as 26 October, and firmly embraced by 3 March 1871.89 John Hay was also pondering a career in literature. “The Western School” (his 27 December editorial in the Tribune) was his manifesto for a group in which he implicitly included himself.

[begin page 829] Even as Clemens worked to complete the printer’s copy for the sketchbook, Hay published the second of his Pike County ballads, “Jim Bludso (of the Prairie Belle),” in the 5 January Tribune, again signing it simply “J. H.” Clemens immediately reprinted the poem in the Buffalo Express and—in a letter not known to survive—wrote Hay praising it, but also suggesting a way to improve its authenticity. Hay replied on 9 January with a draft that adopted Clemens’s suggestion. He also alluded to their previous discussion of the hostile reaction to “Little-Breeches”: “The opposition is getting beaten out I think. Some of the heathen still rage furiously and the words of their mouths are ‘ribaldry’ ‘plagiarism B.H’ and ‘vulgar blasphemy.’ But there are compensations. ‘The Atlantic’ and Harpers and the Aldine have all asked me for some more foul vulgarity—and alack! I have not time to write it.”91 Clemens probably received this reply just before leaving for Cleveland, where he had to be by 11 January. On that day, he wrote Hay extending an invitation to join him (and probably David Gray) in a new “enterprise”—presumably a new journal. On 14 January, Hay declined this offer, but added:

Hay’s experience in publishing “low” verse (Pike County vernacular used to disguise a subversive point of view)93 doubtless warned Clemens that he, too, might be accused of “foul vulgarity” for what he had written, or planned to write, about the West. He had, in fact, just been accused by [begin page 830] Every Saturday of plagiarizing Bret Harte.94 And it was against Harte’s work that both he and Hay could expect to be judged. According to the Philadelphia Bulletin:

But Hay’s experience must also have been encouraging, because he clearly had won a large popular audience as well as the more enlightened one represented by Howells at the Atlantic Monthly. Two days after Hay published “The Western School,” he wrote to ask Howells:

It is a fair guess that Hay said much the same thing to Clemens about his book on the Far West (“I can’t do it—but you could”).97

Even before he received Hay’s reply, however, Clemens had decided to give up journalism. On 14 January he told Webb: “I dassent. I made up my mind solidly day before yesterday that I would draw out of the Galaxy with the April No. & write no more for any periodical—except, at long [begin page 831] intervals a screed that I happened to dearly want to write.”98 Meanwhile, Orion finished copying the two (or three) manuscript sketches within a week of their arrival in Hartford, and on 16 January Bliss set off with them to New York, intending to hire Mullen. He returned to Hartford a week later, on 23 January, without succeeding. Orion reported that Bliss had “hunted for Mullin and Lant Thompson, or whatever his name is, two days. He found the latter’s office in the hands of the plasterers. He is going back to-morrow and will find Mullin.” But Bliss clearly had his doubts about the publication strategy Clemens had proposed to him. Even before going to search for Mullen, therefore, he allowed Orion to tell his brother that he, Bliss, “hardly” knew “whether it is good judgment to throw the Sketch Book on the market & interfere with the Innocents.”99 On 25 January, Orion reiterated Bliss’s view: “About the sketch-book interfering with the Innocents—Bliss says he is going on with the sketch-book, and you will see which is right. The substance is that the new book will outsell the old one, and few people want to buy two books from the same author at the same time.”100 Even before this letter reached him, however, Clemens had conceded the point:

Three days later, he had developed this scheme into an elaborate but quite unrealistic plan to publish the western book as early as May—even though he had scarcely begun to write it. The inference is all but inescapable that Clemens was trying to force himself to write:

[begin page 832] No manuscript was sent, or presumably called for. On Tuesday, 31 January, Clemens telegraphed “E. Bliss or Frank” before taking the train from Buffalo to New York: “Have an appointment at Grand Hotel eleven tomorrow can you be there at noon Clemens.”103

That evening or early the next morning, Clemens checked into the Grand Hotel and, presumably on the morning of 1 February, met with Francis Church to negotiate a resolution of his Galaxy contract. Then he met with Bliss to discuss the western book.104 They evidently agreed to issue it in May, just as Clemens had proposed in his most recent letter, even though there was scarcely time to make such a deadline. Clemens also promised to contribute something from the book manuscript to the American Publisher, the first issue of which was to appear within a month.

On the evening of 2 February, Clemens took the train to Washington, where he planned, among other things, to lobby for legislation affecting the Langdon estate.105 On 6 February, Susan Crane, who was staying with Olivia and the baby, wrote him in Washington:

Olivia had typhoid fever, although no one yet realized it. Clemens was depressed by the news of his wife’s condition, but stayed in Washington to attend a dinner given in his honor by Samuel S. Cox (a Langdon family friend) on the evening of 7 February. The dinner was interrupted by a telegram from Buffalo, which summoned Clemens home. The danger of his wife’s illness was clear, and he sought frantically to cancel his “Memoranda” for the March Galaxy, lest its publication make him seem to jest in the presence of death.107 For the same reason, he telegraphed Bliss not to print the sketch he had sent in October for the Publisher. Throughout the next week Olivia’s condition threatened to prove fatal: “We cannot tell what the result is going to be. Sometimes I have hope for my wife,—so I have at this moment—but most of the time it seems to me impossible [begin page 833] that she can get well. I cannot go into particulars—the subject is too dreadful.”108

One week later, however, she was guardedly better, and Clemens was again trying to work on the western book. “Return to me, per express, the ‘Liars’ & the other 2 sketches—right away” he wrote Orion. “Livy is very, very slowly & slightly improving, but it is not possible to say whether she is out of danger or not—but we all consider that she is not.” The same day he gave an equally sober report to Whitelaw Reid: “My wife is still dangerously ill with typhoid fever, & we watch with her night & day hardly daring to prophecy what the result will be.”109

One week later still, Clemens had decided not only to leave journalism, but to leave Buffalo, ostensibly on his doctor’s advice.110 He put both the house (which had been a wedding gift from Olivia’s father) and his share in the Express up for sale. “I quit the Galaxy with the current number,” he told Riley on 3 March, “& shall write no more for any periodical. Am offered great prices, but it’s no go. Shall simply write books.”111 On 4 March he replied to Orion, who had returned the “Liars” sketch and one or two others, and who (at Bliss’s suggestion) had urged him to begin sending book manuscript—at least in part so the next issue of the Publisher could fulfill the promise of a contribution from Mark Twain.112 Clemens explained that he had wanted the “Liars” sketch back in order to “work it into the California book—which I shall do. But day before yesterday I concluded to go out of the Galaxy on the strength of it—& so I have turned it into the last Memoranda I shall ever write & published it as a ‘specimen chapter’ of my forthcoming book.” He made it clear that he had broken with the Galaxy, and would not write for it now under any circumstances. Somewhat casually, however, he also seemed to say that he was unable, or unwilling, to write anything for the Publisher, at least for the time being:

In short, Clemens acknowledged that he had written nothing in February (twenty-eight days), and did not expect to write anything in March until he had gone to Elmira and taken a week’s rest (another twenty-one days). Still, it is clear from the same letter that he remained sympathetic to Bliss’s needs for the Publisher:

Bliss wrote on 7 March, evidently before this letter had arrived, having “just returned from N.Y.” He identified two reasons for wanting the book manuscript as soon as possible:

As Clemens predicted, within five days of his 4 March letter he had finished his alteration of “the whole style” of one of his “characters,” at least up through what was then chapter 8. The character in need of alteration was almost certainly the narrator himself. New (and old) evidence discussed [begin page 835] below suggests that Clemens was pruning out some too-knowing remarks of a “tenderfoot” version of himself.115

At any rate, Clemens was by this time clearly reconciled to publishing one or more chapters in the Publisher:

The next day, however, Clemens reported having sent only “160 MS pages . . . to be copied,” expecting to “have it back next Tuesday,” when he would “ship it to Bliss & mark a chapter to be transferred to the Columns of the Publisher.”117 Clemens may simply have held back or destroyed eight pages. But if the reconstruction in Figure 1 is approximately correct, “160 MS pages” would not have taken him through the end of what he was then calling chapter 8 (later 7, and the start of 8), so there is at least one other possibility—that he mistakenly wrote “160” for “168.” Such an error seems somewhat more likely because the conjectured end of chapter 8, the pony-express incident, clearly was copied at this time along with what preceded it.118

Although Clemens had certainly drafted sections of his manuscript beyond chapter 9, he must have left chapter 10 unfinished, for in the same 10 March letter he begged Orion to

Orion promptly met this request on 11 March.119 Clemens incorporated most of his brother’s details almost immediately into chapter 10, which, [begin page 836] together with chapter 11 (presumably drafted earlier, upon receipt of Dimsdale’s book), completed the first batch of book manuscript he would send to Bliss—probably on 18 March.

Before Orion’s letter about Slade was written, however, Clemens received two letters from Hartford—one from Orion dated 8 March, and one from Bliss (now missing), probably written on the same day—making reply (but also clearly overreacting) to his casual refusal of 4 March to write for the Publisher. (Clemens had retracted this refusal almost immediately, and without prompting, in the two letters written on 9 and 10 March, before he received this double blast from Orion and Bliss on 10 March.) A sample of Orion’s badgering will explain, almost by itself, the sulfurous reaction Clemens had to it. Orion vowed, among other things, that he and Bliss would

Clemens wrote on the envelope of this missive, “Still urging MSS.” But he fully vented his unhappiness with it in a long letter, begun on 11 March and finished after “two days ‘to cool’ ”:

In spite of his irritation, Clemens added a postscript to the letter: “Shall ship some book MS. next Wednesday”121—that is, on 15 March—a deadline which he did not quite meet, postponing his shipment until 18 March, the day he and Olivia left Buffalo for Elmira.

On 17 March Clemens replied to Bliss, who had been trying to resolve their differences by rational explanation.122 His reply suggested that rational explanation was irrelevant. It also confirmed that his progress on the western book had been minimal since December.

It is scarcely surprising that under these circumstances Clemens’s opinion of his work was decidedly gloomy:

It is likely that the chapter Clemens threatened to “tear up” was eventually left out of the book. If so, no text for it has ever been found, and its contents remain unknown and to a large degree unknowable. And yet certain [begin page 838] things about it fall within the range of reasonable conjecture. It is clear, for instance, that some chapter originally preceded the Bemis incident, and that it was even set in type later that spring, as Bliss began work on the prospectus. The evidence for this conclusion is a small, bibliographical detail in the prospectus (Pra), which has long been known but never fully explained: two pages from the Bemis story, identical to the text as it appeared on pages numbered 62 and 63 in the first edition, are found in Pra on pages numbered 77 and 78—fifteen pages further along than where they ended up in the book.124 The most likely explanation for these higher page numbers is that at the time Pra was first set in type, there was enough manuscript somewhere before the Bemis story to make fifteen book pages (text and illustrations), which were removed before Pra was published in November. Such a reduction could have come piecemeal from several chapters, but probably included one large block (presumably a chapter), and must have been authorized by Clemens sometime in August 1871. These revisions are discussed in greater detail below.

Something can also be conjectured about the likely content of the chapter that preceded Bemis and the buffalo, and the reasons for Clemens’s dissatisfaction with it, if we assume that the reference to Overland City at the end of chapter 6 is a vestigial forecast of what originally followed it: “the strangest, quaintest, funniest frontier town that our untraveled eyes had ever stared at and been astonished with.” If Clemens had written a chapter about the quaint manners and odd characters of Overland City, its humor apparently struck him now as strained (“simply an attempt to be funny, & a failure”). It may even have been largely in this chapter that the condescending “style” of his narrator (“strangest, quaintest, funniest”) most required revision. If it seems implausible that he wrote such a chapter in the first place, it may help to recall the report in the 7 September 1870 Elmira Advertiser (published within days of his writing the first four chapters), which characterized his new book as “an account of travel at home, describing in a humorous and satirical way our cities and towns, and the people of different sections.”125

Despite the impulse to tear up this chapter, Clemens probably did not do so at this time—although his unhappiness on rereading it perhaps made him conclude, prematurely as it happened, that the first eight chapters contained nothing suitable for the Publisher. He and his family were in Elmira by 19 March, one day after he evidently sent Bliss all of the manuscript for what was then chapters 1–11 (some 258 pages)—even though only two-thirds of it (“168 pages”) had been copied for security.

After sending it, he may have looked through “the next chapters,” but [begin page 839] he found nothing there for the Publisher. On 20 March, however, he wrote Bliss that he had found something to contribute: it was in then chapter 8, which had in fact been copied along with the rest. He probably enclosed his security copy of it, referring Bliss and Orion to the original manuscript, which (as the postscript just below his signature shows) he assumed was by then already in Hartford:

It is clear that Clemens did not have his own manuscript of the pony-express passage before him, for otherwise he would simply have given its first page number, rather than saying it began “along about the 160th to 170th page of the MS.”127 (He seems to have remembered the approximate page, but could not check his memory against the security copy, which would not have preserved the original page numbers.) In fact, his ability to quote the first nineteen words of the passage, and his way of referring to it (“Here is my contribution”), would seem to make it all but manifest that he enclosed his security copy of the passage with the letter—both to make it easier to find it in his original manuscript, and to transmit the “marginal note” mentioned in the final postscript. It is also clear that the passage, [begin page 840] which he described for the Publisher’s audience as “a chapter,” did not begin a chapter, as it ultimately would in the book, but instead ended one (“Go on to end of chapter”), the beginning of which he also identified: “that bull story . . . that precedes the pony” can only refer to the Bemis tale, which ultimately began chapter 7 in the book, where it would be followed by the story of Eckert and the cat, which Clemens would supply in August to fill in when he turned the pony-express passage into the beginning of chapter 8.128

Having sent off the first eleven chapters of his manuscript,129 Clemens seems to have taken the “week’s rest” he earlier said he needed, while continuing to nurse Olivia through her convalescence. On 23 March, for instance, Charles Langdon, the Reverend Thomas K. Beecher, and Clemens attended an evening meeting of a local literary society, where Clemens spoke briefly: “Mr. Twain’s address was concluded by saying that he had a sick wife at home, and that he had duties which he could not delegate, and with the President’s permission he would retire, as his faith in doctors was rather limited.” The next day, 24 March, his old friend Joseph T. Goodman arrived in Elmira.130 Goodman’s visit would grow into several months, and provide Clemens with badly needed encouragement about [begin page 841] his western book. He and Goodman clearly discussed writing a novel together at this time (probably to be set in Washington, D.C.), but it is likely that shortly after arriving in Elmira, Goodman read Clemens’s manuscript for the western book—at least insofar as he could by relying on the incomplete security copy and any further manuscript chapters Clemens had drafted by that time. Clemens would tell Orion on 18 April that Goodman was “up here at the farm with me,” and that he was “going to read my MSS critically,” but the likelihood is strong that by then Goodman had already read substantial portions of the manuscript.131

In a 1910 interview, after Clemens’s death, Goodman recalled for a San Francisco reporter that it had frequently been his “privilege” to read Mark Twain’s “works in manuscript, before they were sent to the publishers.”

Paine, who also interviewed Goodman and heard this story directly from him, published a rather more detailed account, and possibly one that profited from Clemens’s own recollection of the facts. According to Paine, it

The first contemporary sign that Clemens’s interest had been rekindled came slightly more than a week after Goodman’s arrival. On 4 April, he wrote Orion:

Clemens had misplaced the notes that Orion had sent the previous November, and the specific details mentioned here indicate that on 4 April he was preparing to write what became chapter 25. Indeed, four days later, on 8 April, he recorded being up “to the 6 570th page,” conjecturally the next-to-last page of chapter 24, about the Mexican plug.135 These facts would seem to confirm that during October, November, and December, at least, Clemens had in fact made some progress beyond chapter 12, if only because it seems unlikely that he wrote all of chapters 12–24 (more than 300 pages) in the two weeks between 25 March and 8 April.

On the other hand, much of what Clemens had written seems to have been still in draft form: he clearly continued to revise chapters 12 and following during March and April. One of the three surviving manuscript pages (reproduced in facsimile on page 843) shows, in fact, that he must have revised chapter 12 sometime after 13 March, when he received Orion’s letter about Slade. Chapter 12 covers the final part of the journey to Salt Lake City and is, in general, notably dependent on Orion’s journal for names, distances, and particular incidents, as well as for occasional phrasing.136 But the journal barely mentions another landmark that Clemens treated rather fully in chapter 12—the east-west stream on the Continental Divide—stating merely, “Near this Station are the Pacific Springs, which issue in a branch, taking up its march for the Pacific Ocean.” Clemens’s cue for the incident in chapter 12 was therefore almost certainly not

[begin page 843]

[begin page 844] Orion’s journal, which he had had since August, but Orion’s letter about Slade, which he received in March. That letter began by describing South Pass, “where the clouds looked so low, where we saw the first snow, and where a spring with waters destined for the Atlantic stood within a man’s length (or within sight) of another spring whose waters were about to commence a voyage to the Pacific.”137 If Clemens wrote (or elaborated) his own description of this stream only after reading Orion’s 11 March letter, then the unnumbered manuscript page—clearly a draft, not even quite filling the page—was written no earlier than mid-March. It contains a version of the “snapper” used to conclude this incident, which, by the time it appeared in the prospectus, had been condensed into a two-sentence paragraph: “I freighted a leaf with a mental message for the friends at home, and dropped it in the stream. But I put no stamp on it and it was held for postage somewhere.”138 Gone from this final version is any hint of the narrator’s sweetheart (“she”) in New Orleans (“the old city”)—perhaps because such details were not compatible with his having “never” been “away from home” before.

An early version of the first page of chapter 18, about crossing “an ‘alkali’ desert,” also survives in manuscript. The page (reproduced in facsimile on page 845) is numbered “423,” and the chapter is numbered “20,” altered from “19,” indicating that the number of the “desert” chapter was at least one higher than its final number in the first edition. A higher chapter number is confirmed by the evidence of Pra, which included several pages from chapters 15, 16, 21, 23, and 24, anachronistically identified as chapters 16, 17, 22, 24, and 25—that is, one chapter number higher than in the book. Where, prior to chapter 15, was a full chapter deleted? The answer is suggested by the author’s footnote at the end of chapter 14, which refers the reader to the appendixes for “a brief sketch of Mormon history, and the noted Mountain Meadows massacre.” That reference, together with some minor discrepancies in chapter 13, makes it all but certain that the material in the appendixes was originally part of these chapters on Mormonism. The “Brief Sketch of Mormon History” (appendix A) may have belonged originally in chapter 13—perhaps following Clemens’s reference to having “picked up a great deal of useful information and entertaining nonsense,” not otherwise mentioned. If so, that might explain why references

[begin page 845]

[begin page 846] in the next paragraph to visiting “the king,” and to the king’s “high-handed attitude toward Congress,” occur without the king or his offenses against Congress being clearly identified—something the last paragraph in what became appendix A had done, noting that

Likewise, the material that became appendix B, “The Mountain Meadows Massacre,” was very likely a separate chapter, perhaps following chapter 14, and therefore preceding the one remaining reference to the subject in chapter 17. But however these chapters were in fact configured, the likelihood is great that they originally included the material on Mormonism preserved in the appendixes. When the words in the appendixes are counted as if they were part of the manuscript for chapters 13–15, the calculated page number for the start of chapter 18 is 424—quite close to the actual number on the surviving draft page, 423.140

Although Clemens sent Bliss only eleven chapters on 18 March, he knew perfectly well that the contract obliged Bliss “to commence operations at once upon receipt of manuscript.” He therefore added a postscript to his 4 April note to Orion which expressed his impatience for Bliss to begin illustrating and typesetting what he had sent: “Is Bliss doing anything with the MS I sent? Is he thinking of beginning on it shortly?”141 His anxiety to see illustrations and typesetting begin is the more remarkable because, having sent those chapters rather in haste than otherwise, he soon found that they needed more work. On 8 April he replied to another letter (now lost) from Orion, who apparently asked whether he and Bliss should use the manuscript of the Bemis episode for the Publisher and for the book, or wait for the revision promised in Clemens’s first postscript of 20 March. Clemens repeated his original instructions, with one condition:

But on reading his security copy that evening, Clemens was unable to resist the temptation to revise it—and not just in the Bemis chapter. He therefore did add a postscript, which referred explicitly to changes in chapter 6, in the Bemis incident, and in other chapters not specified:

By the time Clemens finished and mailed this letter on 10 April (Monday), he had written as far as page 610, which probably fell near the start of chapter 27, about the trip to Humboldt, the virtual beginning of his “silver fever.” And although the original letter no longer has its enclosures, he probably did send with it (or separately, soon thereafter) both “the bull story, altered the way I want it,” and an additional “20 or 30 pages” of “corrections marked in purple ink,” which he asked that Bliss follow in setting up the eleven chapters already in hand. Both the revised “bull story” and these “corrections” were presumably in the form of altered and augmented pages from Clemens’s security copy, which thereby grew even more incomplete. (Because Clemens used his typical purple ink, the changes were visibly distinct from the ink of the copy, evidently a different color.) Although none of the enclosures survives, it is still possible to make some informed guesses about what they contained.

The “Jack & ‘Moses’ ” yarn, which Clemens here identified as beginning on “about” page 117 of the manuscript, was ultimately included in chapter 6 despite this instruction to omit it (the revised closing words for the chapter were also rejected). If, as Figure 1 indicates, chapter 6 began on page 114, then the “Jack & ‘Moses’ ” story would have begun on page 126, not 117. The discrepancy indicates that among the revisions Clemens may have made at this time was the addition of up to nine pages toward the beginning of chapter 6, thus forcing the “Jack & ‘Moses’ ” episode back [begin page 848] to its present position.143 The new material may have been designed to spell out more clearly, in preparation for chapters 10 and 11 on Slade, exactly how the “division-agent” fit into the hierarchy of stagecoach employees. Some evidence for this conjecture is provided by Orion’s Slade letter (11 March), which implied that Clemens may have mistakenly referred to Slade as a conductor: “My impression is that he was a division agent, from Overland City to Salt Lake—having several conductors under him. . . . Slade was not a conductor. He had the conductors and drivers under him.”144 The somewhat disjointed style of chapter 6 may result from a revision designed to correct this error by introducing the division-agent to a narrative that originally described the conductor. Clemens’s addition of nine pages (nearly doubling the chapter’s original eleven pages) may even have suggested to him that he could now leave out “Jack & ‘Moses’ ” entirely. Moreover, it seems necessary to speculate that then chapter 7, which Clemens had found so unsatisfactory even as he sent it to Bliss in March, was now submitted to further revision—chiefly by the addition of material. These 10 April “corrections” totaled, conjecturally, twenty-five pages: nine pages added to chapter 6, and sixteen to then chapter 7. There certainly were, in addition, changes to the Bemis story, which was now “the way I want it”—conjecturally, ten pages longer than before (see Figure 1).

Given the limitations of the evidence, the precise size and location of these changes must remain in doubt. But whatever “corrections” were actually made in April, their net effect was to increase the length of the manuscript between chapter 6 and then chapter 8, without changing the original pagination. (Clemens could not have renumbered his pages because the printer’s copy was in Hartford.) These added pages might well have been distributed in some other way, but their total could not have been very different. Without the bulk they added, there would be no way to explain why pages 62–63 of the Bemis story were forced back to pages 77–78 in the initial Pra typesetting, for if all the material, including then chapter 7 (later removed in proof), were present in the original manuscript, thus affecting its pagination, the conjectural first page of the pony-express passage would be forced well beyond the range of pages 160–70, to which Clemens assigned it.

Two days after sending this letter, Clemens was in New York.145 Probably [begin page 849] by the time he returned to Elmira, Orion’s reply was waiting for him. The letter itself has not survived, but it evidently reported Bliss’s reluctance to begin on the prospectus without more manuscript in hand, or at least without a more representative selection from it. Orion seems also to have requested permission to put “Jack & ‘Moses’ ” in the Publisher, and even to have suggested further revision of the Bemis episode for the Publisher in order to distinguish it from an ostensibly similar account by another one of Bliss’s authors, Thomas W. Knox. On 18 April, Clemens responded coolly to such meddling:

On the same day, Clemens left Elmira for Buffalo “to finish the sale of my ‘Express’ interest,” as he told Mrs. Fairbanks. The next day he completed this errand, picked up his mail at the Express office, and apparently met briefly with his friend David Gray.147 The Buffalo Courier reported on 21 April:

Clemens did not immediately acknowledge this gift from his recent famous acquaintance. When he did, probably in early May, it must have been after Greeley’s joke at his expense had received some coverage in the press (reprinting the Buffalo Courier), for he evidently protested Greeley’s [begin page 850] “public” criticism of him. Clemens was, at any rate, in a foul mood during the last week of April. Orion’s meddling had merely annoyed him, but shortly after returning from Buffalo, he received a truly disheartening letter from Bliss, probably on Monday, 24 April. Bliss wrote that he was about to begin work on the prospectus, and that he planned to “get it out very quickly,” but felt the need to clarify what Orion had already communicated:

Bliss also mentioned that he had already “made selections from Mss. here for Pros.,” suggesting that if Clemens had “any choice cuts further along in the book for it” that he should “send them on” and Bliss would “heave them in.” Bliss continued, “Your brother says he wrote you Knox had written up something similar to the Bull story—I never saw it & do not know anything about it. It Yours struck me as a good thing, every way. Your first chap. is splendid—smacks of the old style—”149

These afterthoughts scarcely blunted Bliss’s point, which was that Clemens’s delay in finishing his manuscript had cost him ground with his audience and, by implication, that he now needed Bliss’s help to reverse that loss.150 Clemens was clearly depressed by Bliss’s testimony, and perhaps by other evidence, telling Mrs. Fairbanks on 26 April: “I am pegging away at my book, but it will have no success. The papers have found at last the courage to pull me down off my pedestal & cast slurs at me—& that is simply a popular author’s death rattle. Though he wrote an inspired book after that, it would not save him.”151 It is not known which newspapers he thought were casting “slurs” at him, or just what occasioned their criticism. But on Sunday, 30 April, a week after reading Bliss’s remarks, he reproached Orion for writing him “discouraging letters”: “Yours stopped [begin page 851] my pen for two days—Bliss’s stopped it for three. Hereafter my wife will read my Hartford letters & if they are of the same nature, keep them out of my hands. The idea of a newspaper editor & a publisher plying with dismal letters a man who is under contract to write humorous books for them!” Yet Clemens went on to make clear that he was steadily at work on his book. His inability to write had lasted no more than a week, in clear contrast with the longer intervals of indifference he had experienced the previous fall and winter.

If on 30 April Clemens was “away up to page 750 of the MS.,” Figure 2 indicates that he was probably at the end of chapter 34 (a revision of “The Facts in the Great Land Slide Case”) or the start of chapter 35 (the trip to Esmeralda). Neither chapter was selected for the prospectus, which in fact included nothing from chapters 25–37; of the two, the tale of the landslide case seems more likely to have inspired Clemens’s enthusiasm.

The 100 pages of manuscript Clemens mentioned to Orion as having been sent “yesterday” (29 April) are identical with what, three days later, he told Bliss he had sent “yesterday” (2 May): “I mailed you the 12th, 13th, 14th & 15th chapters yesterday, & before that I had sent you the previous 11 chapters.”153 Chapters 12–15 describe arriving in Salt Lake City and encountering Mormons. As Figure 2 shows, the page numbers of chapters 12–15 ran from 259 to 358 (exactly 100 pages), but only if it is assumed that they included material later removed and, later still, relegated to the appendixes. Clemens also told Bliss that his book was by then “half done,” so he had accumulated many more pages than he sent—probably some 900 equivalent pages, placing him near the end of chapter 40 (the first of two chapters on the blind lead) by 3 May.154

[begin page 852] On 8 or 9 May, Clemens received Greeley’s somewhat testy and defensive reply to his letter of thanks:

Greeley’s notoriously difficult handwriting was the subject of innumerable jokes and anecdotes in the press, but this letter was the first Clemens himself had ever received from him. Their brief exchange, part of which was even then being reported in the newspapers, may well have suggested to Clemens the basic comic story of chapter 70, which concerns someone driven mad while corresponding with Greeley “about a trifle of some kind,” only to discover that their correspondence had become “the talk of the world.”156 Clemens’s first letter from Greeley became, in any case, the physical model for the only slightly exaggerated imitation of it, reproduced as a pseudo-facsimile in chapter 70.157 Greeley’s letter arrived during the most intensive period of book composition that Clemens had yet experienced, but the Sandwich Islands setting of chapter 70 suggests that it was probably not written until July, or even August, when Clemens again wrote Greeley (on another matter), and when the need for additional manuscript had become more apparent.

On Monday, 15 May, less than two weeks after he declared himself “half done,” Clemens counted himself well past the halfway point:

By this time Clemens had clearly shifted to counting by equivalent pages. So when he said that he had “1200 pages of MS,” he was probably near the end of what became chapter 51, with 1202 equivalent pages (but only 1100 real pages). Twelve hundred equivalent pages were “two-thirds” of the eighteen hundred pages he thought he needed: enough, by his reckoning, “to make (allowing for engravings) about 400 pages of the book,” which was to be 600 pages long. His arithmetic makes sense in other ways as well: if he had 900 equivalent pages on 3 May, and had written more than 30 pages a day “nearly every day for the last ten,” he would have added approximately 300 pages. His increased rate of progress may have had something to do with his subject: in chapter 42 he began to describe his experiences as a newspaper reporter in Virginia City. But his new-found speed was also due to his increased use of clippings in this section of the book (chapters 34, 37, 38, 46, and probably 51 all relied extensively on clippings), which now made it virtually necessary to count by equivalent pages.

Since Clemens wrote this letter on a Monday (15 May), the extraordinary string of daily stints “during the past week” can be plausibly assigned to Monday through Saturday, 8–13 May. It is therefore likely that he began on 8 May with 23 pages, a stint that corresponds to chapter 44 (puffing stocks), estimated at 25 pages (see Figure 2). The second day’s stint, on 9 May, was 30 pages, which corresponds to chapter 45 (Reuel Gridley and [begin page 854] the flour sack), estimated at 24 pages (only a fair approximation of 30). The third stint, on 10 May, was 33 pages, which corresponds to chapter 46 (Nevada nabobs), estimated at 31 equivalent pages. The fourth stint, on 11 May, was 35 pages, which corresponds to chapter 47 (Scotty Briggs and Buck Fanshaw’s funeral), estimated at 36 pages. The fifth stint, on 12 May, was 52 pages, which corresponds to chapters 48 (western desperadoes and the vagaries of jury trials) and 50 (Captain Ned Blakely): 25 plus 24 pages, or 49 pages, a good approximation of 52.159 And the sixth and last stint, on 13 May, was 65 pages, “nearly half” being “a re-print sketch,” which is to say, clippings. If this stint does indeed correspond to what became chapter 51 (the Weekly Occidental, with the poem called “The Aged Pilot Man”), estimated at 62 pages, the only part of it that could have been “re-print” was the poem. This suggests that the poem was in fact set from a clipping of the lost Occidental or even the Enterprise—although no such printing has been found.160

Having just written chapter 48 (on the failings of the jury system) on 12 May, Clemens concluded his 15 May letter somewhat giddily, with a dedication for the book which was not, in the end, adopted:

Nor could he resist a gloating postscript about his reputation, which had recuperated, without Bliss’s help, in less than three weeks’ time: “The reaction is beginning & my stock is looking up. I am getting the bulliest offers for books & almanacs, am flooded with lecture invitations, & one periodical offers me $6,000 cash for 12 articles, of any length & on any subject, treated humorously or otherwise.”161 Bliss replied almost immediately, on 17 May:

Nearly four weeks had passed since Bliss had said he was beginning on the prospectus. Clemens’s reply to this latest request for his “best stuff” is not known, but he probably did not send any more manuscript in May. On 1 or 2 June, however, he carried his third batch of manuscript to Hartford. This third batch probably extended from chapter 15 (his 16) through 51 (his 50), omitting chapters 36 and 49, which were probably added later. He may also have submitted some “selected” chapters” from later in the narrative which he thought especially suitable for the prospectus.163

Shortly before Clemens made this delivery, David Gray published an announcement of the book which was obviously informed by Clemens himself: “Mark Twain’s new book, same size and style as the ‘Innocents Abroad,’ and as copiously illustrated, will be published in the fall, and will appear simultaneously in England and America. Dealing as it does with certain hitherto unrecorded phases of western life, it will be of historic value as well as aboundingly humorous.”164 This is the earliest known public reference to Clemens’s plan to publish his western book “simultaneously” in England with George Routledge and Company—a step that would establish a valid English copyright, provided English publication preceded American by a day or so. Clemens had asked Bliss to arrange the matter with Routledge, probably soon after learning in March 1870 that only simultaneous publication could protect the book in England.165

[begin page 856] Clemens did not stay in Hartford long enough to read proof, or even to revise chapters 1–15, which were already in Bliss’s hands. On 5 or 6 June he returned to Elmira, where he may have begun one more chapter. He did not, at this time, write what became chapter 52 (Virginia City mining statistics, probably added in August or even later), nor did he write chapter 53 (Jim Blaine’s ram story), which he probably finished on 10 August, when he told Olivia of having written “a splendid chapter today, for the middle of the book.”166 But he may well have begun chapter 54 (the Chinese in the West) in early June, for its first paragraph reads, in part: “As I write, news comes that in broad daylight in San Francisco, some boys have stoned an inoffensive Chinaman to death, and that although a large crowd witnessed the shameful deed, no one interfered.” This sentence could not have been written before 3 June, when the news report that Clemens saw appeared in the New York Tribune:

But if Clemens began chapter 54 after returning home on 5 or 6 June, by 9 June he apparently felt so relieved at having delivered some fifty chapters to the publisher that he stopped work on the book altogether. The next day he wrote his lecture agents, Redpath and Fall, that “without really intending to go into the lecture field,” he had written “a lecture yesterday just for amusement & to see how the subject would work up,” but having “read it over,” he had now decided to “deliver it.”168 Clemens spent the rest of the month working on this lecture, and, when he grew dissatisfied with it, on two others. He also managed to write, or at least revise for publication, four magazine articles, three of which he sent, on 21 June, to Bliss:

[begin page 857] It is likely that Clemens did not again work on the book manuscript until the end of the month. On 27 June he told Orion: “I wrote a third lecture to-day—& tomorrow I go back on the book again.” On 29 June he reported writing “2 chapters of the book to-day—shall write chapter 53 to-morrow.”170 The two chapters were probably his chapters 51 and 52, which became chapter 54 (the account of the Chinese community, begun earlier in June) and chapter 55 (the “Prodigal,” and Clemens’s brief stint as chief editor of the Enterprise). His chapter 53, probably completed on 30 June, would eventually become chapter 56 (about tiring of Virginia City and leaving for San Francisco).

Back in Hartford, Bliss and Orion had begun to have the illustrations drawn, engraved, and electrotyped—at least for those chapters selected for the prospectus. Bliss had said in April (with only eleven chapters to choose from) that he had already made a selection—and indeed, out of the first eleven chapters, all but three (6, 9, and 11) were represented in the prospectus. If he followed normal procedure, Bliss would have ordered and received engravings for most of the illustrations in these chapters before ordering any type set, so that the first proofs could show the illustrations in place. Some evidence indicates that he was, to a degree, bypassing normal procedure for this book,171 but it was still desirable to give the manuscript to the illustrators as soon as possible.

Bliss hired at least three artists: Edward F. Mullen, whom Clemens had wanted to hire for the Burlesque Autobiography, and then for the sketch-book, in December and January; Roswell Morse Shurtleff (1838–1915), later a successful landscape painter; and True W. Williams.172 All three artists had earlier drawn illustrations for Richardson’s Beyond the Mississippi, and Shurtleff and Williams had both provided drawings for The Innocents Abroad. All of the new drawings would be engraved by the firm of Fay and Cox in New York, which had also done the work for Beyond the Mississippi and Innocents, among others.

Williams was the principal artist for the book, signing fifty-four drawings (usually with his distinctive monogram) and contributing at least as many more that were not signed.173 He received the first, as well as the [begin page 858] largest, assignment of manuscript—possibly in May, when Bliss had chapters 1–15, or in June, when he also had chapters 16–50 (15–51, as later numbered in the book). His hand is evident starting with the clear likeness of the author in chapter 1, in which Mark Twain sports the same checked trousers that he wears in Williams’s drawings for Innocents. Over half of Williams’s signed drawings appear in chapters 2–37, but many of the later drawings are his as well. (In chapter 62, he may even have managed to sneak in a portrait of himself as “Williams,” the captain’s nemesis: see page 427.) Shurtleff signed the frontispiece and the full-page drawing of South Pass for chapter 12. Fourteen additional illustrations bear his initials or monogram, the earliest in chapter 42 and the last in chapter 79, for a total of sixteen.174 Mullen signed only four drawings—one in chapter 38, one in chapter 43, and two in chapter 67175—although he likewise contributed several more that are unsigned.

In addition to commissioning the drawings for Roughing It, Bliss followed the common practice of reusing engravings from earlier books, especially those published by his own company and its subsidiaries. He took engravings from Thomas W. Knox’s Overland through Asia, Nelson Winch Green’s Mormonism, Junius Henri Browne’s Sights and Sensations in Europe, as well as from Richardson’s Beyond the Mississippi and Evans’s Our Sister Republic. He also used seven illustrations from J. Ross Browne’s Adventures in the Apache Country, presumably by purchasing a set of electrotypes from Browne’s publisher, Harper and Brothers. In all, Bliss reused thirty-three illustrations from known sources.176 In addition, [begin page 859] Roughing It included some twenty-five illustrations for which no previous publication has yet been found, but which are likely to have been taken from standard or stock sources, rather than drawn and engraved specifically for the book.

In early June, with the first fifty-one chapters now in hand, Orion presumably read the manuscript through for the first time, and in mid-June wrote his brother praising it. Clemens replied on 21 June: “Am very glad, indeed, you think so well of the book. I mean to make it a good one in spite of everything—then the illustrations will do the rest. When the prospectus is out I believe Bliss will sell 50,000 copies before the book need be actually issued.”177 Orion’s comments are lost, but it was almost certainly he who wrote the following squib in the July American Publisher: “A New Book by Mark Twain.—The American Publishing Company, of Hartford, has in hand manuscript for a new book by Mark Twain. It is a lively account of travel and a stirring recital of adventures in the Far West. He was miner and reporter, and had some experiences, which, though sober reality, will tax credulity.”178 This last reference suggests that Orion had read at least as far as chapters 40 and 41, about the blind lead. David Gray, on the other hand, probably did not have a chance to read the manuscript, although on 20 July he published a description that must reflect Clemens’s private remarks about it—albeit inaccurately: “Mark Twain’s new book will be published by the American Publishing Company, of Hartford, Connecticut. It will describe a journey to California in the ‘flush times’ of 1849, or scenes in the early history of the Golden State, with probably much piquant personal matter. It is intended to serve as a companion volume to ‘Innocents Abroad.’ “179

Despite this talk about publication, Clemens was well aware that he had still not completed his manuscript. On 27 June, therefore, he had Olivia write to Mrs. Fairbanks, postponing their planned July visit to Cleveland: “Mr. Clemens feels that it will be a month or six weeks before his book will be finished.” Clemens himself added an apologetic postscript: “I have lost so much time that I am obliged to give it up. This book has been dragging along just 12 months, now, & I am so sick & tired of it. If I were to chance another break or another move before I finish it I fear I never should get it done.”180 Orion was conscious of the same problem, [begin page 860] and in late June wrote his brother asking when to expect the next installment of printer’s copy. Clemens replied on 2 July: “My MSS? Shall bring it there myself before long. Say 2 to 4 weeks hence. Am just finishing Chapter 56. Have already nearly MS enough, but am still writing—intend to cut & cull liberally.”181 Chapter 56 would become chapter 59, about his adventure with a San Francisco beggar. If he was still writing the chapters in sequence, he must have written chapter 57 (San Francisco pioneers), 58 (the earthquake in San Francisco), and 59 between 30 June and 2 July.182 Even so, he did not have “nearly MS enough” to complete his book: he was still some 450 pages short of the 1800 he thought he needed. Not for the last time did his eagerness to be done mislead him about where he stood. Orion replied on 4 July:

The drawings Orion mentioned here were for chapter 24 (the Mexican plug), which was eventually included almost in its entirety in the prospectus. Frank Bliss, who was presumably helping his father prepare the book, followed Orion’s advice, despite his concern that eastern readers would not understand the auctioneer’s being seated on a horse. No tree appears in any of the pictures of Carson; one illustration shows the horse bucking; and another shows him taking his rider over a post-and-rail fence.184 Only the drawing of the horse “going through a gate, raking his rider off with the top beam” was discarded. It is apparent that the Blisses and Orion were passing judgment on preliminary drawings, before they were sent off to be engraved.

Three days later (on 7 July), Bliss replied to Clemens’s letter of 21 June:

This letter confirms that some large part of chapters 1–23 was already in production, since with the “engravings mill driving,” Bliss would soon be able to begin the typesetting. He was eager to have the subjects for full-page illustrations chosen and the plates completed, since the time required for that, rather than for typesetting, would determine when the prospectus could be done (“Shall have prospectus ready early as possible to get the cuts ready”). Clemens replied in his turn on 10 July:

It remains uncertain just how many more chapters, if any, Clemens did “fix up & forward” after writing this letter. But this fourth batch probably contained only what became chapters 54–57, plus chapters 76 (Maui’s Iao Valley and Haleakala) and 78 (Clemens’s first lecture in San Francisco). [begin page 862] Chapters 76 and 78 must have been submitted at this time (if not sooner), out of sequence, for both were well represented in the prospectus, albeit as chapters 73 and 75.187 The presence of a shared typographical feature also strongly suggests that chapters 1–57, 76, and 78 were in Bliss’s hands substantially before the remaining chapters (58–75, 77, and 79): they all appear in the book with headings in type size A, while all the remaining chapter headings are in type size B.

This difference was clearly inadvertent, the result of using different typesetters for the two sets of chapters, or of using the same typesetters, but only after a long interval, so that they forgot (or were unable for other reasons) to use the same size type in the second set as they had in the first.

Clemens’s letter of 10 July also makes it apparent that his intention even then was to include only a relatively small number of chapters on the Sandwich Islands—“4 or 5” at the end of a narrative principally devoted to “flush times” in Nevada. Back in December he had indicated that he still thought of the Sandwich Islands as the subject for a book he intended to write: “I am going to do up the Islands & Harris. They have ‘kept’ 4 years, & I guess they will keep 2 or 3 longer.”188 Presumably over the next three weeks in July, he culled his own notebooks and scrapbooks for possible material. He almost certainly made use of the printer’s copy he had already prepared in 1866–67 from his Sacramento Union letters for a book on the Sandwich Islands, which he had submitted—unsuccessfully—for publication. And he ultimately included not “4 or 5” but fifteen chapters on the Sandwich Islands—all but four of them based on his original letters to the Union, clippings of which he pasted up and revised to serve as printer’s copy.189 Only about one-fifth of the material published in the twenty-five Union letters was used in the book: twelve letters were not used at all, while portions ranging from one-sixth to four-fifths were extracted from the remaining thirteen. Clemens probably wrote very little new material for the Sandwich Islands chapters, relying instead on the work he had done in 1866–67. The unusual number of simple errors in these chapters in the first edition also suggests that they were rather hurriedly prepared. By the time Clemens had chosen the material for them, it [begin page 863] was clearer than it had been on 10 July that “Flush Times in the Silver Mines” was no longer an appropriate title. On 3 August, just days before Clemens arrived in Hartford, Bliss copyrighted another possibility, “The Innocents at Home,” which served as the interim title for several weeks.190

On 2 or 3 August, Clemens left Elmira for New York, where he stayed briefly before taking the train to Hartford on about 5 August.191 He probably brought with him all the remaining printer’s copy—some 460 equivalent pages—except for those few chapters he would add in August, or even later. He remained in Hartford for almost the entire month, working on proofs for some portion of the early chapters (possibly only those chosen for the prospectus), and revising the printer’s copy for chapters not yet sent to the illustrators or typesetters. On 10 August, five days after setting to work, he summarized his progress for Olivia:

The “splendid chapter” has been conjecturally identified as chapter 53 (Jim Blaine’s ram story).193 The calculated pages in Figure 2 correspond [begin page 864] exactly to Clemens’s own statement: when he arrived in Hartford with his final chapters he had a total of 1806 equivalent pages (1488 real pages); his new chapter added 24 pages, bringing the total to 1830 equivalent pages (1512 real pages). This amount of material would still not, as it turned out, be enough for a 600-page book—partly because of revisions Clemens would soon make in the first fifteen chapters, but chiefly because his estimates of the equivalence between printer’s copy and book pages were too high (overcorrected from his estimate for Innocents). The need for more copy, however, would become apparent to him, and to Bliss, only as he worked through the proofs during August, and even somewhat later that fall.

It was probably during August that Clemens read proofs for chapters 1–25—or at least the chapters in this section that were needed for the prospectus. His intention to “cull & cut down the MS,” especially in the “Overland trip,” was not entirely relinquished, for he evidently decided at last to remove the chapter he had wanted to tear up back in March (then chapter 7). Figure 1 illustrates the revisions in chapters 7–9 which were set in motion by this deletion, made after the early chapters were in type. Clemens probably wrote and inserted the anecdote about Eckert and the cocoanut-eating cat after the Bemis episode, which was now chapter 7 (rather than chapter 8), at least in part so that the pony-express passage could begin (instead of end) what was now chapter 8. But in order to make chapter 8 a normal length (five, rather than three, book pages), he divided chapter 9 so that the opening passage, about Scott’s Bluffs Pass, now ended chapter 8.194 Chapter 9 was thereby reduced to about three-fourths its original size, while the chapter numbering remained intact. That these changes were made only after the first twelve or so chapters were typeset is indicated by at least two things: (a) the absence of the Eckert episode from the analytical Contents, even though the other changes in chapter division were incorporated there, and (b) the typesetters’ failure to change the Pra folio numbers (77–78) to their correct book pages (62–63) in the Bemis incident, even though the Eckert episode (68–69), which now followed Bemis, was correctly paged.195 Also in August, presumably, Clemens [begin page 865] went on to revise chapters 13–15, about the Mormons, reducing the amount of “dry history” in chapter 13, and eliminating chapter 15 altogether, although not at this point with the intention of printing the excised material in appendixes. This last change did affect the chapter numbering in a way that was not corrected in Pra: it caused parts of chapters 15, 16, 21, 23, and 24 to be anachronistically identified there as chapters 16, 17, 22, 24, and 25.

Clemens no doubt made other changes in the text of the early chapters. The running headline on the last page of chapter 1 (“hermaphrodite steamer”) may refer to a remark that was cut out at the last moment (this chapter exactly fills its last page in the first edition). Similarly, the last subheading in the Contents for chapter 3 (“Warning to Experimenters”) may refer to a passage that is no longer present, suggesting that the chapter was revised in proof without a corresponding revision in the Contents.

It may also have been in August that Clemens added chapter 36 (Nevada quartz mills) to his manuscript—in this case, before the surrounding chapters had been typeset.196 Clemens must have added roughly twenty pages to his manuscript somewhere between chapters 15 and 46, for if he did not, the calculated pages for chapter 46 would be too high to include the page number (“968”) on the draft manuscript page from that chapter.197 That the added pages came in chapter 36, however, is only a reasonable guess. No illustrations were specifically drawn for it: three of its four engravings were demonstrably reused from other works, and the fourth, a tailpiece, appears to be a stock engraving.

Late in August, or sometime in the next few months, Clemens must also have added three more chapters to his manuscript, somewhere between chapter 46 and chapter 76, since chapters 73 and 75 in the prospectus became chapters 76 and 78 in the book. The most likely candidates are chapter 49 (shooting affrays in Virginia City), which is made up largely of Enterprise clippings, and which lacks any illustration except for a stock tailpiece; chapter 52 (mining statistics and techniques), which is also based on clippings, and which appears to be illustrated exclusively from borrowed engravings (the source for only one of its three engravings has actually been found); and chapter 75 (the exploration of Kilauea crater), whose four engravings include at least two stock drawings, and possibly three. Chapter 75, which is not based on a Sandwich Islands letter to the Union, may have been the sketch originally called “Fearful Adventure” in Clemens’s draft table of contents for the sketchbook, written no later than January 1871. If so, then the piece was available in January, but presumably [begin page 866] not included until, like the other added chapters, there was little time to commission drawings for it.198

By 17 August, Clemens had probably reached the proof for chapter 20 (Horace Greeley and Hank Monk), for on that day he wrote to Greeley:

Clemens had made the statement he referred to in a footnote at the end of chapter 20—a footnote he had probably supplied at the behest of Joseph Goodman, when Goodman had read his manuscript back in March and April.200 It is not known how, or even whether, Greeley replied to Clemens’s question.

On 19 August Clemens heard from his friend Adolph Sutro, who was planning to leave for Europe later in the month, but would be in New York for some days prior to departure. Clemens wrote him from Hartford:

[begin page 867] Clemens may have just drafted chapter 52, on mining statistics and techniques, which prompted him to ask Sutro for information about his tunnel. Sutro must have telegraphed Clemens shortly after arriving at the Gilsey House on 22 August, for two days later Clemens telegraphed him there: “When do you sail? how long shall you remain in NY when leave & whither?” Sutro evidently replied that he sailed on 30 August, for on 25 August Clemens again telegraphed: “All right will see you in New York before you sail.” Clemens almost certainly met with Sutro in New York between 26 and 28 August, and returned to Hartford almost immediately, telegraphing from there on 29 August, “How long will tunnell be when finished,” a question that assumes a previous conversation.202

At any rate, it was from Hartford on 30 August that Clemens found himself “peremptorily called home by sickness” in his family. The next day he wrote Orion from Elmira: “We have scarcely any hope of the baby’s recovery,” although his fears soon proved groundless. On 2 September, the Hartford Courant reported that Clemens had just spent “a month” in Hartford, and that he had left “a few days ago” after “attending to matters connected with the publication of his new book, ‘The Innocents at Home,’ which is to be brought out by the American Publishing company of this city.”203

It remains uncertain exactly how much of the book Clemens saw in proof by the end of August, but it could not have been much more than chapters 1–25. Three months later, on 6 December, Bliss reported that type had been set only up through page 300 (the second page of chapter 43), but since portions of chapters 46, 51, 76, and 78 had by then already appeared in Pra, they might have been ready for proofreading as early as August. Bliss’s statement indicates, however, that while Clemens may have added chapter 52 in August, he did not supply the concluding footnote to it until sometime in December, for it reflects a communication from Sutro received after Clemens departed Hartford: “Since the above was in type, I learn from an official source that the above figure is too high. . . . The tunnel will be some eight miles long, and will develop astonishing riches.”204 This note, which Clemens must have supplied on [begin page 868] proofs sent to him on his lecture tour, may be usefully contrasted with a revision in chapter 27, where he inserted a parenthetical addition rather than a footnote, presumably before the passage was typeset. On the second page of chapter 27, William Clagett is identified parenthetically as “now member of Congress from Montana,” a fact that could not have been added before 9 August, when the election results were first published in the New York Tribune.205

Langdon soon recovered from the immediate threat to his health, and much of September was given over to other matters, including Clemens’s two visits to Washington to patent his elastic garment strap, which he had invented in August while working on his manuscript. Clemens was in Washington on 7–9 September, returned to Hartford a day or two later, and then went on to Elmira on the afternoon of 13 September.206 On 15 September he asked Redpath to send his lecture schedule to Hartford, because he and his family planned to “take up our permanent residence” there on “the last day of this month,” and because he had “to read proof half the winter.”207 The letter implies that he did not expect to complete the proofreading before he went on tour in mid-October.

On 17 September, Clemens and Olivia went to Buffalo “to pack up for Hartford.”208 From Buffalo, on 19 or 20 September, he again went to Washington in pursuit of his patent, returning to Buffalo by 22 September. Interviewed by the Washington National Republican just outside the patent office, Clemens seized the occasion to remind a recent correspondent of their memorable first meeting:

Back in Buffalo, Clemens reported that he and Olivia were still “packing our furniture & shipping it to Hartford & we are in a mess—house upside [begin page 869] down—my wife sick—can’t leave her bed for perhaps a week yet—& yet we must take possession of our house in Hartford Oct. 1.”210 These various activities suggest that little proofreading was possible during most of September, even if proofs were available.

By 1 October Clemens was in Hartford and, though preoccupied with other matters, in a better position to examine proofs when and if they were ready. On 3 October, Orion wrote to Mollie that Clemens was extremely busy, and would “be over in a day or two”:

Orion’s remark about the “book just going into the printer’s hands” must refer to some portion of the text following chapter 24 which had been illustrated and was now going to the typesetters.

On 9 October, Clemens asked Redpath for details about his lecture schedule: “Send along the first end of my list & let me see where I am to talk. Please send a copy to my publisher, E. Bliss Jr. 149 Asylum st Hartford—for I must read proof for the next month or so.”212 Clemens’s estimate—that he would finish proofreading by mid-November at the latest—was optimistic, for there were certainly problems, even before he left Hartford on 13 October to begin his tour. Before leaving, he probably settled the question of the title, since that was necessary for the prospectus, and he may well have seen some further proof.213 On 19 October, after only six days (and three lectures), Clemens wrote Bliss from Wilkes-Barre, Pennsylvania:

[begin page 870] Clemens referred not to chapter 20 (with its allusion to the Great American Desert),215 but to chapter 18, which describes crossing “an ‘alkali’ desert” west of Salt Lake City. It is apparent that the illustrators or the typesetters had lost part of the manuscript for this chapter, and that Bliss had asked Clemens to rewrite it, relying on what had survived.216 (Clemens’s suggestion that Bliss leave out the “whole chapter” if the attempt to rewrite it failed shows that not all of it had been lost.) Despite his temptation to give up, Clemens soon did rewrite the text. Since in the opening paragraph he correctly recalled at least one detail from Orion’s journal—that the stage station was “forty-five miles from the beginning of the desert, and twenty-three from the end of it”—it seems likely that this passage was in the portion Clemens “brought . . . away” with him. The rest of the chapter could have been recreated from memory. Toward the end Clemens even offered a jocular excuse for a “narrative” that might “seem broken and disjointed, in places.”217 And there were other signs of strain: the chapter contains only one illustration, a tailpiece that is a stock engraving, suggesting that the text was lost before it could be illustrated. Finally, the chapter heading was clearly set by a typesetter who had either forgotten, or never knew, how to match the letterspacing of the other chapter headings, even though he used the correct size of type (“size a”). Such details imply that the chapter was typeset much later than the surrounding chapters. Compare the letterspacing in the headings for chapters 17, 18, and 19:

Two weeks into his lecture tour, on 31 October, Clemens wrote Olivia from Milford, Massachusetts, that he had received no mail at the previous day’s lecture site, Brattleboro, Vermont: “I got no letters at Brattleboro. None had come. None in the post office, either. No proofs from Bliss. Brattleboro is unreliable, I guess.” This statement may mean that he had received no proof at all since leaving on tour. On 8 November, he lectured in Hartford, and could have picked up (or delivered) proofs, although no indication has been found that he did either. He passed through Hartford [begin page 871] again on 19 November, three days before the first copies of the prospectus came from the bindery on 22 November.218 By 27 November, he had seen a copy, for he wrote Olivia:

About ten days later, probably on 8 December, Clemens received a package from Bliss, who sent printed signatures instead of proofs:

Bliss had printed no more than eighteen signatures (since signature nineteen comprised pages 289–304), and probably fewer. He had typeset, but not yet printed, page 300 (the second page of chapter 43). His reason for not sending proofs seems somewhat specious: why was his own set the only one available? Perhaps by this late date he wanted to prevent Clemens from making further revisions, since that would interfere with keeping the presses “running.”

Bliss’s letter, and other evidence, have led to the inference that Clemens did not read proof for the last third of his book. The other evidence is the appearance in the first edition of some sixteen cases of terminal punctuation followed by a dash (“last word.—First word”) between page 426 (in chapter 58) and page 532 (in chapter 74). The absence of this form of punctuation from any earlier part of the text has led to the conclusion that Clemens did not read or correct proof for any text beyond chapter 57—for if he had, he would surely have removed the superfluous [begin page 872] dashes.221 Terminal dashes, however, occur frequently in The Innocents Abroad, even in chapters for which we are certain that Clemens read proof, simply because the proofreader ignored his corrections. And in Roughing It the dashes occur in the same late chapters that have now been identified as a group by their anomalous “type size b” chapter headings (58–75, 77, 79). Therefore the terminal dashes are, in all likelihood, more evidence that the compositors for these late chapters were not instructed as carefully as the compositors for the early ones had been. In addition, Clemens’s comments about two of the illustrations suggest that he did read at least some proof for chapters in the final third of his text.

In chapter 44 (on page 309 of the first edition) Clemens almost certainly added the following sentence in proof: “[My revenge will be found in the accompanying portrait.]” He referred to the “portrait of mr. stewart” that appeared on page 310 (page 289 of the present edition), which he probably first saw in page proof. Likewise, in chapter 54, he presumably added (or modified) the following sentence on first-edition page 392: “They always send a bill, like this below, pinned to the clothes.” Here he referred to the untitled illustration of four Chinese characters (page 370 of this edition), which he probably also first saw in proof.222 Since both references occurred on pages later than 300, and since Bliss said that he had set only up to page 300 by 6 December, it follows that Clemens saw and corrected proofs after that date, including at least one page that fell in the final third of the book (page 392 of the first edition).223

By early January Bliss had realized that the printer’s copy was still not long enough to make what the first prospectus promised the reader, “Between 600 and 700 Octavo Pages.”224 The last page of the final chapter was only 570, and since Clemens was still on tour, there was no likelihood of his writing anything new. No documents have been found to tell us whether it was Bliss or Clemens who thought of adding appendixes to fill out the missing thirty pages. Either one could have suggested using the material that in August Clemens had removed from chapters 13 and 15 on [begin page 873] the Mormons, which was now reset in small (extract) type. The footnote referring the reader to appendixes A and B must also have been added at this time, in the blank space on the last page of chapter 14. For appendix C Clemens almost certainly supplied material he had already written, but then set aside. He may have drafted it while he was writing Roughing It, or while preparing the sketchbook, but he clearly did not include it in the original printer’s copy. Appendix C is devoted to a relatively recent incident in Virginia City, and reprints a long article that first appeared in January 1870. Its relevance to the subject and period of Roughing It is at best remote, suggesting that it too was added solely to make the book longer: appendix C made the final page number 591. Accordingly, the revised prospectus (Prb), which came from the binders in late January, claimed that the volume would consist of “Nearly 600 Octavo Pages.”225

The earliest known public announcement of the book with its correct title, probably based on the information in the recently published prospectus (Pra), appeared in the Buffalo Courier for 30 November:

On 8 December, Clemens telegraphed Redpath that he had decided to “talk nothing but selections from my forth-coming book Roughing it, tried it last night suits me tip top.”227 Among the first to hear this evolving selection was David Gray, who described it on 9 December in the Buffalo Courier:

[begin page 874] If Roughing It may be said to have evolved from the lecture called “Curiosities of California,” it now gradually evolved back into a lecture much like what we know about its ancestor. On 10 December, Clemens wrote Mrs. Fairbanks:

Bliss was no doubt eager to take advantage of the publicity that Clemens’s new topic provided, and with prospectuses finally available, he was ready to enlist agents to sell the book.230 On 11 December he began to advertise:

By 19 December, a more specific appeal for agents began to appear:

These figures show that canvassing began in Cincinnati as early as 7 December, the day after Bliss formally copyrighted the book. The orders were placed, of course, without any actual books in hand: the first bound books were not delivered until February.

Meanwhile, Clemens continued his lecture tour in various towns and cities of Michigan, Illinois, Indiana, Ohio, West Virginia, Pennsylvania, and Maryland. On 24 January 1872, just a day after the revised prospectus (Prb) issued from the bindery, he gave his Roughing It lecture at Steinway Hall in New York. The next day, his performance was described by a knowing but anonymous hand in the New York Tribune:

Clemens had no difficulty identifying the author of such praise:

[begin page 876]

The first copies of Roughing It came from the bindery on 30 January 1872, but were probably not released to agents for several weeks. On 8 February, the Publishers’ and Stationers’ Weekly Trade Circular reported the book “almost ready for publication.” On the same day the Hartford Courant said it had “already reached a circulation of twenty thousand copies,” meaning that so many subscriptions had been taken (none had yet been filled). On 13 February, Clemens gritted his teeth for Mrs. Fairbanks: “I killed a man this morning. He asked me when my book was coming out.” A copy was deposited in the Library of Congress on 19 February, and publication was formally announced on 29 February.235

These final weeks of delay were probably a mistake, or at least unnecessary—the result of Bliss’s failure to fully understand the British copyright law. The transatlantic arrangements for Roughing It had probably been worked out with Routledge as far back as July 1871, and Clemens’s (and Bliss’s) working relationship with Routledge was of even longer standing. In 1870, Bliss had cooperated with Routledge in publishing “the only authorized and unabridged edition” in the United States of Wood’s Uncivilized Races, which Routledge had first issued in London in 1868–70. And Routledge had begun to court Clemens as early as 1868, hoping eventually to be named his “only authorized London publishers.” The Routledge edition of Roughing It was to be the first result of that courtship, and it was soon followed by a series of Routledge “authorized” editions, all published in 1872, which included both A Curious Dream; and Other Sketches and Mark Twain’s Sketches, as well as a revised edition of The Innocents Abroad. Roughing It would prove reasonably profitable for Routledge, but its importance to Clemens, who accepted a flat fee of only £37 ($185) for it, was chiefly as an answer to the English “piracy” of his books by John Camden Hotten.236

Collation shows that the Routledge edition of Roughing It, which appeared in two volumes, was typeset from proofs (or printed signatures) of the American edition (state Aa). The proofs were sent to London in at least two batches: the first in December 1871, and the last in January 1872. No effort was made to include the illustrations (although these were certainly present in the American proofs), and the Routledge text omitted the dedication to Calvin Higbie, as well as the three appendixes—further evidence that Clemens agreed to add the appendixes in January, too late even for the final batch of proofs. Since the dedication had appeared as early as November, in the first prospectus (Pra), Routledge may have omitted [begin page 877] it because the proofs lacked some of the front matter, or because it would have made the first volume asymmetrical with the second. Collation establishes that Clemens did not alter the text, or provide any new material for the English edition—both things that he did soon do for the Routledge edition of Innocents, which he prepared in June and July 1872. Apart from the missing dedication and appendixes, there are fewer than forty variant substantives in the Routledge text—none of them authorial. Even the substitution of “d——d” for “bloody” (at 334.17) must be attributed to the English proofreader, since he was more likely than Clemens to know that “bloody” was more offensive than “damned,” at least to a British reader. Although Clemens certainly gave his blessing to the English edition, its text is wholly derivative.

Chapters 1–45 appeared in volume 1, which was titled Roughing It, and chapters 46–79 appeared in volume 2, titled The Innocents at Home, perhaps at Bliss’s or Clemens’s suggestion. Volume 2 also included Routledge’s text of the Burlesque Autobiography (tacked on at the end), presumably to help equalize the size of the two volumes—a sign that Routledge did not omit the appendixes, but instead set from proofs that did not contain them, and that he determined the size of volume 1 before knowing the total length of the work. Six thousand copies of the first volume were bound on 6 February. It appeared in the London Athenaeum’s “List of New Books” on 10 February, and a copy was deposited in the British Museum on 15 February. The second volume was announced in the Athenaeum’s “List of New Books” on 17 February, but its first 6,000 copies were not bound until 28 February. Bliss seems to have held back on announcing publication of the American edition until both volumes had been published in London, although he probably would have been safe as early as 10 February, when the first volume issued.237

As soon as publication in Hartford was formally declared, Orion abruptly accused Bliss of fraud in manufacturing the American edition, and was promptly fired (or resigned) for his trouble. Clemens surely understood why, under the circumstances, Bliss could no longer trust Orion, even as he knew that Orion had long nursed a resentment of Bliss’s condescension toward him. He therefore did not intervene to save his brother’s job, writing him on 7 March, in part:

Clemens was indeed thankful for Orion’s warning. Five days later, on 12 March, he met with Bliss to try to resolve the questions that had been raised by Orion’s “indefensible” attack. And shortly before (or shortly after) that meeting with Bliss, he consulted Hartford lawyer Charles E. Perkins. Clemens recalled in 1875:

A partial, and indirect, record of Clemens’s 12 March conversation with Bliss is preserved in a letter he drafted to Bliss on 20 March—a document he carefully preserved, sending a fair copy to Bliss. This draft shows that Orion’s charges led Clemens to suspect Bliss of overstating his manufacturing costs, which in turn led to the suspicion that “half profits” were not, as Bliss had maintained in July 1870, equal to a 7.5 percent royalty. Written just three weeks after Roughing It was published in the United States, and just one day after Clemens’s second child, Susy, was born, this remarkable document shows Clemens adopting his lawyer’s advice and maneuvering to get “half profits” written into his contract:

For his part, Bliss could hardly regard the result of any revision as “the same,” and he evidently declined to amend or replace the original contract, probably soon after receiving this letter. Despite his lawyer’s earlier advice, Clemens decided to sue. He recalled in 1880 that eventually “Bliss went into the accounts & details & satisfied Perkins & his expert that 7½ per cent did represent half profits up to a sale of 50,000, & that after that the publisher had a mere trifling advantage of the author. So we dropped the matter.”241 The precise timing of these events remains unclear, but it seems likely that Perkins and his (unidentified) expert interviewed Bliss sometime in late May, and that Clemens was persuaded to give up the suit by the end of June or, at the latest, by mid-July.242

Orion’s charges against Bliss are not known to survive in their original, presumably written, form, but their general import is clear. As mentioned earlier, Bliss had used probably more than fifty borrowed engravings in Roughing It—that is, engravings originally made for other books, but reused (at substantial savings) in this one. Since this common practice could hardly have been a secret, Orion probably accused Bliss not of borrowing illustrations, but of overstating their cost. Yet Orion must have pointed to other abuses as well. On 15 May, Clemens urged him to “Ask Chas. Perkins if he wants you to give him points in my lawsuit. But give none otherwise.”243 On 17 May, Orion replied: “I didn’t know you had commenced a law suit. My plan did not contemplate a law suit by you. I suppose it is a suit for damages. I did not think there was any chance for enough to be made that way to justify such a proceeding.” Orion’s plan was to force Bliss to sue Clemens, to prevent his publishing with someone else.

These remarks suggest that Orion had accused Bliss of misrepresenting the cost not only of the illustrations (“borrowed engravings”), but of the paper and binding as well. Orion’s point was that his brother could hardly recover significant damages for the way Roughing It had been manufactured, but that “the fraud you can prove concerning the printing of Roughing It” could be used to free him from his multiple contracts with Bliss by frightening the publisher with the prospect of seeing subscription-book manufacturing standards held up to ridicule by “newspapers favorable to the trade,” which is to say, publishers who sold their books in bookstores, rather than by subscription.244

Whether or not Bliss really cheated Clemens remains an unanswered question.245 It is clear, however, that Orion’s charges inadvertently provided [begin page 881] Clemens with yet another reason to fear what newspaper critics would say about Roughing It. He did not believe that he was receiving “half profits,” even when he dropped the lawsuit in June or July, but he was convinced that Roughing It had been printed on substandard paper and with badly executed engravings. It is important to recognize that his complaint was principally with the crude way the illustrations had been engraved, electroplated, and printed—and only incidentally, if at all, with the quality of the drawings themselves. Eventually he recognized that these drawbacks were to some extent inherent in subscription publishing. Two years later, after True Williams had contributed drawings to his next book (The Gilded Age), Clemens told Thomas Bailey Aldrich: “There is one discomfort which I fear a man must put up with when he publishes by subscription, & that is wretched paper & vile engravings.”246

Although Clemens was less concerned about the quality of the actual drawings, he was ambitious to improve them as well. In December 1872, he wrote Thomas Nast: “I do hope my publishers can make it pay you to illustrate my English book. Then I should have good pictures. They’ve got to improve on ‘Roughing It.’ ”247 But Nast never did illustrate a book by Mark Twain—at least in part because of the high cost of his services. In [begin page 882] 1905, Clemens put the blame for this parsimony entirely on Bliss, who “always had an eye for the pennies”:

Williams continued to help illustrate Mark Twain’s books, up through A Tramp Abroad (1880).

It was presumably with Orion’s accusations fresh in his mind that Clemens told David Gray he was afraid Roughing It “would be considered pretty poor stuff,” and that therefore he “had better not let the press get a chance at it.”249 But did Clemens actually withhold review copies, as this remark implies? To answer this question, more than sixty American, English, and Hawaiian newspapers and magazines were searched for periods ranging from six to twelve months (as available) in 1872–73.250 The search showed that while extracts from the text of Roughing It were frequently published,251 the total number of reviews, even very brief ones, was only fourteen. Reviews that Clemens or Bliss might have solicited soon after publication are certainly few: ten of those found can be safely excluded from this category, either because they were published in England and reviewed the Routledge edition, or because they appeared well after Clemens [begin page 883] had changed his mind, and was actively trying to solicit reviews. On the whole, the reviews that did appear failed to confirm his worst fears, but they contained enough negative comment to suggest that his fears might have been more fully realized if he had not restricted review copies.

In March 1880, Clemens said that David Gray, Charles Dudley Warner, and William Dean Howells were the only men he could “trust to say the good thing if it could be honestly said” about his books, or, if it could not, to “be & remain charitably silent.”252 His trust in their discretion may well have begun with the publication of Roughing It. Clemens knew that Gray had reviewed The Innocents Abroad favorably, and intelligently, on 19 March 1870, and it seems inconceivable that he did not send him a copy of Roughing It in March 1872. But a search of the files of the Buffalo Courier shows that Gray did not review the book. On 9 March, in fact, he had this to say about Clarence King’s Mountaineering in the Sierra Nevada:

The closest Gray came to publishing his opinion of Roughing It was to print half a dozen extracts from it (including parts of chapters 26, 30, 44, 51, 59, and 67), which he introduced as follows:

The inference must be that Gray was obliged by his real opinion of Roughing It to “be & remain charitably silent.”255

On 18 March (a week after his meeting with Bliss) Clemens wrote to Howells, thanking him for an autographed copy of Their Wedding Journey: “I would like to send you a copy of my book, but I can’t get a copy myself, yet, because 30,000 people who have bought & paid for it have to have preference over the author. But how is that for 2 months’ sale? But [begin page 884] I’m going to send you one when I get a chance.”256 Howells’s copy was almost certainly among the twenty-three “½ moroccos” that Clemens ordered sent to “friends of mine” in his 20 March letter to Bliss. In the final paragraph of that troubled letter, he professed to be

This statement may well be part of Clemens’s strategy in dealing with Bliss, so just how “easy & comfortable” Clemens actually was at this time is in doubt. The passage does suggest, however, that he was relying chiefly on the word of friends, not on book reviews. It is also not known just how Clemens learned of Anna Dickinson’s opinion, but since she was a friend of the Langdons’, her words may have come by private letter to his wife or mother-in-law. Yet Dickinson’s real feelings, less than a year later, were quite different from those she had expressed to the Langdons (and Clemenses). On 14 March 1873 she wrote her mother:

Because of Clemens’s cautious strategy with review copies, opinions like Dickinson’s were only rarely expressed in print. On 6 March 1872, however, the Manchester (England) Guardian reviewed the Routledge edition, taking a more critical tack than has been found in any American review:

And on 6 April, a reviewer in the London Examiner expressed similar, albeit slightly more tolerant, views. Mark Twain’s

Whether or not Clemens saw either of these reviews is not known, but is almost irrelevant: although British, they represent precisely the kind of reaction Clemens had predicted to David Gray (“pretty poor stuff”).

It is likely, but not certain, that Clemens also sent a copy of Roughing It to the Hartford Courant, edited by his friend Charles Dudley Warner—soon to become co-author of The Gilded Age. The Courant’s review appeared on 18 March, the same day Clemens wrote to Howells, and is remarkable less for its deliberate praise than for an undercurrent of alarm, which emerges as a kind of defensiveness about the book. Warner began by reassuring the reader of Mark Twain’s moral purpose, and the genuine clarity of his style:

Warner then took up one of the more troublesome issues, the proper uses of “slang”:

And he included a defense of the book which, except for its final note, chimed very close to Clemens’s own professed feelings about it:

Similarly defensive, though briefer, was a notice on 25 March in the Cincinnati Gazette, which observed that Mark Twain’s adventures were “related with a liveliness, and, we must add, with more than occasional imaginativeness peculiar to himself. One may not always approve the taste of what is said, yet he can rarely record his dissent without a smile. The many illustrations are not the least amusing feature of the book, which is published by subscription exclusively.”262

These conservative views did not go wholly unanswered. On 11 April, the New York Independent remarked that among Roughing It’s

By the time this notice appeared, Clemens had evidently begun to regret his cautious policy toward reviews. On 19 April he asked Frank Bliss to send a copy of the book to William C. Smythe, “a splendid old friend of mine,” who had written him requesting a chance to review it: “He is city editor of the principal Pittsburg paper—a city where I drew the largest audience ever assembled in Pittsburg to hear a lecture. Send him a book. I want a big sale in Pittsburg.” Smythe’s review (if any) has not been found, but Clemens’s letter shows that his own attitude toward reviews had changed.264

The next day, 20 April, Clemens wrote to James Redpath in Boston, weeping a few crocodile tears about the press’s neglect of his book:

No such item has been found in the Boston Advertiser, but on 1 May, a generous review did appear in the Boston Evening Transcript:

Redpath’s reply to Clemens has not been found. He presumably told him what, if anything, he had accomplished with the Boston press, and may also have said something about the book itself. On 15 May Clemens replied in his turn: “Thank you with all my heart. I want to send a copy to the Boston literary correspondent of the N.Y. Tribune—Louise Chandler Moulton, isn’t it? I will have it sent to you. Will you give it to her with my compliments?”267 Redpath presumably did as requested, for on 10 June, Moulton noticed the book in her regular letter from Boston to the New York Tribune:

Shortly before Moulton’s letter appeared, probably sometime between 25 and 30 May, Howells published his review (yet another voice from Boston) in the Atlantic Monthly:

This last sentence shows that even Howells could not avoid mentioning the book’s occasional lack of “all the literary virtues.” It is worth recalling that John Hay had identified Howells himself as a promising member of the “Western school.” And Clemens therefore acknowledged this praise, as Howells recalled in 1910, in a way that “stamped his gratitude into my memory with a story wonderfully allegorizing the situation, which the mock modesty of print forbids my repeating here.”270

Clemens’s response to Howells is also worth comparing with what he said just three weeks later to Louise Chandler Moulton, writing her on 18 June:

These remarks were disingenuous in more than one way, especially since Clemens had probably not yet dropped his lawsuit against Bliss. It is clear that the responsibility for not sending copies “to any newspaper to be reviewed,” at least until April, belonged squarely to Clemens, not to Bliss.

Given the controversial nature of the West as a literary subject, Clemens could reasonably expect to be treated kindly by western journals, but no evidence has been found that he sent copies to them. The book seems to have reached western journals at about the same time copies were being delivered to agents in the West, which implies that Clemens did not do what he had previously done for Innocents, order it sent to the Overland Monthly so that Bret Harte could review it before anyone else. In late May an unidentified critic reviewed Roughing It in that journal with understandable enthusiasm:

On 28 April, Clemens’s old San Francisco employer, the Morning Call, published a long, rambling notice that was basically positive:

Western reviewers were certainly not all so kind. On 18 May, a critic for another old employer, the Sacramento Union, reviewed the book with some asperity, sharply drawing attention to its “padding”:

Sales were, in fact, excellent. On 21 March Clemens had chided Bliss for getting “caught in a close place with a short edition.” And by the end of May the American Publisher announced, “The book is having an unparalleled sale. About 50,000 copies in a little over three months. . . . We have been unable to fill orders at sight, for this book; 10,000 copies are ordered ahead now.” In mid-June Clemens again boasted to Howells of “62,000 copies of Roughing It sold & delivered in 4 months.”275 The bindery records of the American Publishing Company corroborate these impressive figures. It is clear that Bliss was indeed unable to meet the demand for copies: his general agents for Chicago reported on 24 March that they were “away behind in filling their orders,” and that their “five hundred agents, throughout the North and Southwest,” were “taking from five to ten orders each, daily.” By 31 March the bindery had produced 23,695 copies, and it was apparently unable to meet all back orders until sometime in June: by the end of that month it had shipped a total of 67,395 copies.276

But sales declined sharply and unexpectedly in July, when only 2,645 copies were bound, less than a quarter of the number bound the previous month. The total for July through December was only 7,773, a mere one- [begin page 891] eighth of the number for the first half of the year. Total sales for the year were thus 75,168, with 90 percent occurring in the first six months. If the rate of sale for January through June had continued for the rest of the year, Clemens’s prediction in September 1870 that he would sell 90,000 copies during the first year would have been easily met, even exceeded.277

To someone with Clemens’s high expectations, the sales of Roughing It were ultimately disappointing. A total of only 7,831 copies of the American edition were sold in 1873, followed by 5,132 in 1874. In comparison, fewer copies of Innocents—only 67,680—were sold during its first year, but its sales did not decline so precipitously: during the second year of publication, a total of 21,822 copies were sold. On 4 March 1873 Clemens remarked to Bliss that Roughing It was now selling “less than twice as many in a quarter as Innocents, a book which is getting gray with age.”278 But he was willing to take at least part of the blame:

Perhaps in response to this decline, Clemens engineered one more review in the New York Tribune. On 12 January 1873 he wrote to John Hay:

But Whitelaw Reid did not assign Hay to review it, much to Clemens’s annoyance. The reviewer was instead George Ripley, “the profound old stick who has done all the Tribune reviews for the last 90 years. The idea of setting such an oyster as that to prating about Humor!”281 Ripley’s comments were hardly unfavorable, concluding that Roughing It “may be regarded as one of the most racy specimens of Mark Twain’s savory pleasantries, and their effect is aggravated by the pictorial illustrations which swarm on every page, many of which are no less comical than the letterpress.”282 But they certainly did not exhibit the sort of sympathetic understanding Clemens expected (and had earlier received) from John Hay. Just one year before, Hay had reviewed the Roughing It lecture in the Tribune, describing Mark Twain as “a true humorist, endowed with that indefinable power to make men laugh which is worth, in current funds, more than the highest genius or the greatest learning.”283

Sales of the English edition of Roughing It followed the same pattern as the American, but on a smaller scale. In 1872 Routledge printed 18,000 copies of each volume, and followed in 1873 with another 8,000 of each. From 1872 through 1899, Routledge sold a total of 63,750 copies of volume 1, and 68,000 copies of volume 2. In 1882, Routledge also issued a new, one-volume illustrated edition, printing 4,000 copies that year. The plates and unsold sheets of this edition were sold in 1885 to Chatto and Windus, who printed an additional 5,000 copies before 1900.284 Routledge was well satisfied with the sales of the original two-volume edition: in [begin page 893] 1892, when Chatto and Windus offered £100 for the “stereos & copy-rights” of Roughing It, The Gilded Age, and Mark Twain’s Sketches, Edmund Routledge responded, “You can have as I offered before the two 2/. vols of Twain for £60, but Roughing it we don’t wish to sell at all.”285

Although Clemens did not alter the text of any edition after the first American, he did sanction republication by Routledge, and later by Chatto and Windus, in England, as well as by Bernhard Tauchnitz in Germany, and (indirectly, through Routledge) by George Robertson in Australia. Clemens also sanctioned, but did not revise, later editions in the United States, both by the American Publishing Company and by Harper and Brothers. Despite Routledge’s English copyright, however, John Camden Hotten reprinted several extracts from the text, without Routledge’s or Clemens’s permission. And beginning in 1880, Belford and Company in Toronto (later Rose-Belford Publishing, and later still, Rose Publishing Company) reprinted the whole text of Roughing It, lacking only chapters 22, 36, 45, 49, 52, 71, 72, and 77, again without Routledge’s or Clemens’s approval.

Hotten had written to Clemens in February 1872, shortly before Roughing It was published: “Will you oblige by mailing to me on receipt of this some of the proofs—a few chapters. You may depend upon my dealing honourably with you & I will place to your credit whatever is fair & equitable.”286 But Clemens had long since agreed to sell the English rights to Routledge, precisely in order to foil Hotten, who had reprinted both of his earlier books, the Jumping Frog and The Innocents Abroad, in England, and had also issued various collections and anthologies of American humor, much of it Mark Twain’s. In 1870 Hotten had published The Piccadilly Annual of Entertaining Literature, which included five Mark Twain sketches from the Galaxy; and in 1871 he had issued two smaller volumes, Eye Openers and Screamers, containing several more sketches from the Galaxy and a few from the Express, the Chicago Republican, and the San Francisco Alta California.287

Hotten was made more cautious by the Routledge copyright on Roughing It, but in August 1872 he issued yet another collection of humor, Practical Jokes with Artemus Ward, Including the Story of the Man Who Fought Cats, which included fourteen Mark Twain sketches taken from the Express, the Galaxy, and other newspapers, as well as three extracts from the English edition of Roughing It. Whether or not Hotten thought Routledge’s copyright was valid, he was apparently confident that these [begin page 894] Roughing It anecdotes would not be detected. He supplied the titles “Editorial Skits,” “Sending Them Through,” and “The Union—Right or Wrong?” And he edited the texts rather more heavily than those from other sources that were clearly not protected by copyright.288

In March 1873 Hotten published The Choice Humorous Works of Mark Twain (HWa), reprinting virtually everything by Mark Twain which he had included in the earlier books. In London during the fall of 1872, Clemens had told Hotten of his willingness to revise and correct this volume before it appeared. Hotten welcomed the offer, but Clemens left for home before he could follow through on it. HWa included eight sketches corresponding to passages in Roughing It. Two of them were the anecdotes that Hotten had previously extracted from Roughing It (“Editorial Skits” was excluded). Also reprinted was the Practical Jokes text of “Mark Twain’s Remarkable Stranger,” derived from the April 1871 Galaxy, which Clemens had independently used as the basis for chapter 77. Hotten’s 1870 Jumping Frog provided two extracts from the 1866 Sacramento Union which Clemens had independently incorporated into Roughing It from Union clippings: “Honoured as a Curiosity in Honolulu” and “The Steed Oahu.” Hotten also reprinted “Mark in Mormonland” from Roughing It, “A Nabob’s Visit to New York” from the American Publisher of January 1872, and “Baker’s Cat” from his own Screamers, which had in turn come from the Buffalo Express of 18 December 1869.289

When Hotten died only a few months after HWa issued, his assistant, Andrew Chatto, took over the business in partnership with W. E. Windus. Chatto was eager to follow up on Clemens’s 1872 offer to revise HWa, and so wrote to him again when he was in London: “I am sincerely anxious to establish more cordial relations as between Author & Publisher, than have hitherto existed, between you and our firm, and I beg to submit to you a set of the sheets of a volume of your writings, in order that you may (as I understand you expressed a desire to do) correct certain portions of the contents.”290

An agreement was reached, and Clemens set to work revising a set of HWa folded and gathered sheets, which became the basis for a new edition, HWb, issued in April 1874. Collation of HWa with HWb shows that the original HWa plates were altered to incorporate Clemens’s corrections and, in some cases, his deletion of entire sketches. Some of the texts—such as those from the Union and the Express—had undergone a parallel [begin page 895] evolution: they had been subjected to Hotten’s cutting and editing, while Clemens had independently supplied his own revisions of the same sketches for Roughing It. Thus when he came to revise Hotten’s texts, he was annoyed and sometimes puzzled, since their history was not clear to him.291 Clemens’s demonstrable changes to these eight sketches for HWb, however, were not revisions of Roughing It but of independent sketches, and therefore have not been adopted in the present text.292

The second English edition of Roughing It and The Innocents at Home (illustrated by F. A. Fraser; Routledge, 1882) was set from the first English edition, but contained the appendixes of the first American edition, and the revision of “thirteenth” to “sixteenth” (136n.6) introduced in the fifth state of the first American edition (Ae, 1874). Chatto and Windus reissued the second Routledge edition in 1885, 1889, and 1897. Clemens made no revisions in any of these reimpressions.

Bernhard Tauchnitz of Leipzig, Germany, published a Continental edition of Roughing It in 1880–81, which was based on the 1872 Routledge edition. In September 1880 he sent Clemens 300 gold marks (about $75) as a voluntary payment for his use of the text. Clemens responded on 7 October, expressing his “distinguished appreciation of a publisher who puts moral rights above legal ones, to his own disadvantage.” He also evidently explained that the Routledge version of the book comprised two volumes, of which Tauchnitz had published only one. Tauchnitz promptly replied: “As to ‘Roughing it’ you are quite right. I have published in my edition only the Tale which bears this title in the ‘Routledge edition.’ In consequence of your kind explanation I will now publish in a further separate volume ‘The Innocents at Home,’ and beg to offer you, as for ‘Roughing it,’ Three Hundred Mark Gold for this volume.”293 Clemens provided no revisions for either volume in this edition.

From 1899 until 1903 the American Publishing Company issued several impressions of a two-volume edition of Roughing It (A2), typeset from the last state of the first American edition (Ag, 1892). Later impressions of A2 (e.g., the “Japan” and “DeLuxe”) incorporate corrections suggested by the company’s proofreader, Forrest Morgan, who compared a copy of the “Royal edition” (an early impression of A2) with the first edition. Although in at least one instance a query from Morgan was referred to Clemens, he offered no revisions of his own. The American Publishing Company [begin page 896] retained the copyright on Roughing It until 1903, when it sold the plates of A2 to Harper and Brothers. Harper continued to issue impressions of A2 from these plates until about 1914, when it published a new, more compact two-volume edition. Numerous impressions of this last typesetting appeared with the Harper imprint over several decades.294

No royalty statements (at least none itemized by title) are known to survive for the years 1880–98, although there were lump-sum payments for the years 1890–91 on the seven Mark Twain titles for which the American Publishing Company owned the rights.295 On 31 December 1896 the company signed a new contract for these books, whereby it agreed to pay a royalty to Olivia Clemens of 12.5 percent of the cover price, or one-half of the net profits—whichever was greater—on the seven titles and on the forth-coming Following the Equator. Between January 1899 and June 1903, extant royalty statements (covering all but six months of that period) show 2,280 copies of Roughing It sold, representing payments of at least $997.296 For the years 1903–7, the records of Harper and Brothers show that Mark Twain’s three most popular books were The Innocents Abroad (46,125 copies), The Adventures of Tom Sawyer (40,962 copies), and Roughing It (40,334 copies).297

The printer’s copy used by the American typesetters and illustrators is not known to survive, but the conjectural reconstruction of that copy in Figures 1 and 2 (pages 814–15), together with the history of composition and production just described, affords a reasonably clear idea of it. To recapitulate, three pages of manuscript draft for Roughing It do survive, and are reproduced in facsimile above: they contain passages revised so completely that Clemens simply replaced them in the printer’s copy.298 The page numbered 423 is typical of the many chapters that were entirely in [begin page 897] Clemens’s holograph—those listed in Figures 1–2 as having no “clipping” words: they consisted of leaves inscribed on a single side, in purple ink, with the usual cancellations, interlineations, and insertions. The page numbered 968 is typical of the chapters that included “clipping” words: the printer’s copy for them was partly holograph, and partly in printed form. Such printings were either passages cut from the Book of Mormon or Catharine V. Waite’s Mormon Prophet,299 or clippings from the New York Times, the Buffalo Express, the Galaxy, the Sacramento Union, the Virginia City Territorial Enterprise (and possibly even from proofs of the Weekly Occidental, also of Virginia City), which Clemens mounted on the same-sized paper that he used for holograph sections. Draft page 968, which contains a clipping from the Express, shows that such clippings could be revised both on the newsprint itself and in the margins. Among the sketches Clemens prepared in 1870–71 were two additional Express articles: “A Ghost Story” and “Adventures in Hayti” (CU-MARK). The draft page and these sketches, totaling sixteen pages, demonstrate that Clemens attended chiefly to changes in his own diction, while revising or correcting the spelling, capitalization, and punctuation much less often. On the draft page he changed “can’t” to the more formal “cannot”; he altered “letter” to “chapter” (for obvious reasons); and he substituted “shifted” for the slangy “swapped.” He also changed “newspaper article” to “little sketch” (in pencil, on the clipping itself), and then to “brief sketch” (in ink, in the margin). In “A Ghost Story” he changed “a tiny” to “an,” “What” to “Now what,” “in” to “under,” and “infamous” to “mean.” He also corrected a typographical error (“of” to “off”) and changed “foot print” to “footprint,” so that it would be spelled consistently as one word throughout the sketch, as was the analogous “footstep.” In “Adventures in Hayti” he deleted some extreme phrasing, like “carrying their bleeding hearts outside their shirt-bosoms.” He also deleted “just” and “soulless,” corrected “dollars” to “dollar,” “volcanos” to “volcano,” and “I” to “It,” but passed over several other errors: “what what” (dittography), “halucination,” and “decripit.” Collation of the first edition of Roughing It against the original printings of articles reused in it shows that Clemens made many similar changes (and some similar oversights) in them, presumably on the printer’s copy, but possibly also in proof.

There were some likely exceptions to these two basic kinds of copy. In some cases, Clemens’s revisions were so extensive that he must have copied out the passage, as he decided to do with the first uncanceled sentence in the clipping on draft page 968, preserving his change from [begin page 898] “swapped” to “shifted,” but making further deletions and revisions as well. Judging from collation, and other evidence, the quotations in chapters 10 and 11 from Dimsdale’s Vigilantes of Montana were probably copied out, rather than supplied in torn pages from the book, which may have been loaned, rather than given, to Clemens by Hezekiah Hosmer. In addition, some small part of Clemens’s manuscript in the first seven chapters may have been replaced before typesetting began with the “20 or 30 pages” of corrections that he sent to his publisher on or about 10 April 1871. Although these corrections were in Clemens’s hand, some of them must have been inscribed on pages from the amanuensis security copy of his first 168 pages (up through the pony-express passage). It also seems likely that, when it was decided to put the material removed from chapters 13 and (then) 15 into appendixes, the typesetters availed themselves of the early proofs for those chapters, rather than digging out and setting from Clemens’s printer’s copy. Finally, the table of contents and list of illustrations were not drafted by Clemens, but by Bliss or his associates (Orion or Frank Bliss are both likely candidates), and therefore copy for them was also not in Clemens’s hand.

It is clear that Clemens did not write or revise all of the chapters in sequence. Nor were they all typeset in sequence—in fact, they were probably not set by the same compositors throughout. It will be recalled that in October 1871, Clemens was forced to rewrite parts of chapter 18, combining newly written material with what had been preserved in proof, or possibly in manuscript. Chapters 58–75, 77, and 79 were clearly set in type well after the earlier chapters, probably by different typesetters, but at least by those who had not been instructed (among other things) to ignore Clemens’s habitual end-line dashes. In addition to the revisions Clemens submitted by mail in April 1871, he revised virtually all of the copy during August 1871 (“I cut & slash & lick & trim & revamp it”), when he also read and drastically revised parts of the first fifteen (possibly twenty-five) chapters in proof, removing two chapters (conjecturally, one about Overland City and one about the Mountain Meadows massacre), as well as sharply reducing the length of a third (chapter 13); he also added new material (Eckert and the cat) and redivided chapters 7–9. Even so, the continuing problem of the book’s overall length forestalled his wish to “scratch out half of the chapters of the Overland narrative.”300 It is likely, but not demonstrable, that Clemens read and revised proof for most of the book, occasionally adding footnotes (chapters 14 and 52), deleting anecdotes or episodes like “Waking Up the Weary Passenger” (chapter 48), adding two specific references to the illustrations (chapters 44 and 54),301 and writing or adding several additional chapters (conjecturally 36, 49, 52, 53, and 75), as well as the appendixes, long after he thought the book was complete.

[begin page 899] Orion (and both Blisses) saw preliminary sketches for the illustrations, but no evidence has been found that Clemens himself did. He nevertheless saw a high percentage of the illustrations in proofs of the prospectus and of the book itself, and therefore had a chance to reject any that he did not approve. And although he was disappointed in the technical execution of at least some of the wood engravings, no edition of Roughing It can claim to fulfill his intentions without including them. There can be no doubt that, even before beginning to write, he expected his book to be “well & profusely illustrated” with “snappy pictures—especially pictures worked in with the letter-press.”302 They are therefore reproduced photographically from the first edition, occasionally reduced or enlarged slightly to accommodate the present page design.

Each chapter of the text has its own peculiar history of authorial revision and printing-house transmission, but the chapters with clippings of Clemens’s own work were the most complex. In those cases, only the original newspaper (or magazine) typesetter worked exclusively from Clemens’s holograph. The book typesetters set partly from his holograph, and partly from printed copy already one remove from holograph, despite Clemens’s efforts to revise and correct it. This circumstance complicated the task of the book typesetters and proofreaders, insofar as their goal was to maintain uniformity in punctuation, spelling (especially of compound words), and capitalization, as well as in various nontextual matters, such as whether to retain certain manuscript abbreviations, how to set extract quotations, what typographical treatment to give titles and foreign words, and whether or not to spell out numbers.

In the absence of the manuscript printer’s copy for Roughing It, the first American edition, which was set directly from that copy, would ordinarily be chosen as copy-text throughout. Indeed, for the present edition, the first edition (A) is copy-text for all parts of the text that reprint no earlier text (or at least none that is now available).303 But wherever Clemens reprinted a text of his own, the earlier printing was chosen as copy-text, since it is the text nearest to the author’s original holograph and is therefore the most likely to preserve his punctuation, spelling, capitalization, and paragraphing. And wherever Clemens quoted from a text by another writer, that text too was chosen as copy-text, to ensure that any errors in transmission would be removed, on the assumption that Clemens intended to quote accurately.304 These several copy-texts were then emended [begin page 900] to incorporate all changes, whether to substantives or accidentals, which Clemens is thought to have made on the printer’s copy (or on the proofs) of the first American edition.

This procedure is somewhat more cumbersome than relying exclusively on the first edition as copy-text, but it has several advantages. It increases the likelihood that the author’s spelling and punctuation will be preserved throughout, while it forces the editor to recapitulate the process of revision and correction which Clemens himself performed on the clippings. And since the apparatus must record the editor’s decisions about which differences between the copy-texts and the first edition are authorial and which are not, the list of emendations provides as full a record as we are ever likely to have of the author’s revisions on his printer’s copy and proofs.

But while the choice of first printings as copy-texts has these positive results, it does not automatically yield a text that is free of errors. We know from Clemens’s statements over a long period of time that he always welcomed the correction of mistakes in his texts. In 1881, for example, he told the publisher of The Prince and the Pauper: “For corrections turning my ‘sprang’ into ‘sprung’ I am thankful; also for corrections of my grammar, for grammar is a science that was always too many for yours truly.” And in 1897 he made it clear that while the printer’s proofreader was not to concern himself with punctuation, he was expected to correct misspelling, “which is in his degraded line.”305 Since errors are, by definition, not intended, the following have been emended, whether they occur in authorial or in quoted copy-texts: (a) typographical errors such as “carrried” and “welome”; (b) missing or incorrect quotation marks (single where double are needed, and vice versa); (c) misspelled words (that is, spellings not sanctioned by any known authority since 1800, such as “logarythm”); (d) misspelled proper names of real people, places, things, and institutions such as “Holliday,” “Fort Kearney,” and “Roscicrucian”; (e) subject-verb disagreement (except in quoted speech), as in “statement were”; (f) “end-line dashes,” which were never intended as punctuation; and (g) manifestly defective punctuation (at 183.18, 233.12, 304.31, 406.32, 430.15, 431.5, 487.18, and 523.9).

The largest potential for error in the copy-texts, however, lies in their inconsistent punctuation, spelling (particularly of compound words), and capitalization. In the first edition, the punctuation of parenthetical phrases, for example, follows no particular pattern: sometimes a comma [begin page 901] precedes the opening parenthesis, sometimes the closing one, and sometimes both. Collation reveals that the typesetters occasionally “corrected” the punctuation of the earlier printings, while in others they followed their copy exactly. The first edition also contains dozens of words spelled in more than one way, largely because it followed spellings in the earlier newspaper and magazine typesettings. The book typesetters attempted, albeit unsuccessfully, to spell words consistently, often succeeding within a chapter while overlooking variants in more distant chapters. While uniformity in spelling was clearly a goal, there was evidently no clear house style—at least none which was known by all the compositors, and which would have allowed them to resolve all such inconsistencies. It is clear, for example, from a detailed survey of holograph letters written between 1853 and 1870, as well as of numerous literary manuscripts from the same period, that Clemens spelled “summer” and “winter” without initial capitals. In some twenty occurrences of each of these words in manuscript, none was capitalized, including the two which appear on draft page 423 (where they are part of a sentence that Clemens canceled). And yet in both the Buffalo Express and Sacramento Union printings of Clemens’s work, these words were consistently rendered with capitals, no doubt because the typesetters followed their own house style. In setting passages from holograph, the book compositors adopted Clemens’s invariable lowercase form. But when setting passages based on Express and Union clippings, they followed the capitalized spellings in their copy, and even supplied capitals nearby, in an attempt to achieve uniformity. The result is that both authorial and nonauthorial spellings of these words occur almost randomly in the first edition, not because Clemens ever intended to spell them both with and without initial capitals, but because the typesetters tried and failed to create a uniform texture of spelling, defeated by printer’s copy that contained a bewildering mixture of spellings.

We know quite a lot about Clemens’s attitude toward compositorially or editorially imposed consistency in punctuation. His strong view, often and vigorously expressed, was that punctuation was his province and his alone, and that the compositor or proofreader ought simply to “have no opinion whatever regarding the punctuation, that he was simply to make himself into a machine and follow the copy.”306 In August 1876, while seeing Tom Sawyer into print, Clemens explicitly turned down an offer from Bliss to make the punctuation “uniform . . . here & hereafter.” The narrow question was whether the single word “No” (in dialog) should always be followed by a period, or always by an exclamation point. Clemens [begin page 902] replied on the proof for page 75 (chapter 7): “No, not uniform; follow copy; sometimes it is a quiet negative, & sometimes an exclamatorily vigorous one. The copy isn’t always the way I want it, though. The thing takes a different look in print from what I thought it would.” To which Bliss replied: “Of course! alter wherever it dont look right. We will follow copy & make your alterations afterwards.” Clemens responded: “Very well, what better way is there, than that? Do I give you one-fiftieth the trouble that Richardson did?” To which Bliss in turn replied, “Richardson made more trouble over every page than you do in a whole book. Your model Ms is my standard to gauge others by, & must not be much better & cant be really.”307 This exchange shows that Clemens did not tolerate compositorial regularizing of his punctuation; that despite his care with preparing printer’s copy, he did sometimes change his mind about punctuation when he saw it in type; and that in his own eyes and in Bliss’s, his printer’s copy manuscript was typically a model of clarity and accuracy.

Variant patterns of punctuation which were not deemed erroneous before the twentieth century, and which may be intentional, have therefore not been emended. Clemens’s occasional use of a comma before an opening or closing parenthesis, or just before the sentence verb, or in combination with a dash, are all well documented in holograph manuscripts. Their occurrence in the printed copy-texts does not necessarily reflect the author’s inscription, but it might. Variants of this kind either do not require uniformity (as Clemens firmly told Bliss in 1876), or cannot be made uniform without significant risk of obscuring the writer’s intention.

Clemens’s attitude about the limits of compositorial responsibility, however, did not extend to misspelled, or inconsistently spelled, words. For instance, in April 1869, while proofreading The Innocents Abroad, he wrote Bliss:

This letter makes clear that two years before he proofread Roughing It, Clemens thought there was one and only one way to “properly” spell “anywhere” and “everything,” even though it must be conceded that he occasionally wrote them with what looks for all the world like a space between the two halves of the compound. He was chastising the proofreader [begin page 903] not simply for misspelling the words, nor just for failing to correct them, but also for spelling them in more than one way (“& sometimes even compound words”).

The printing-house goal of uniformity in spelling can be traced at least as far back as 1808, to an essay by Joseph Nightingale in Caleb Stower’s Printer’s Grammar; or, Introduction to the Art of Printing, which included a paragraph subsequently reprinted in dozens of printer’s manuals.309 It appeared, for instance, in a manual Clemens himself is likely to have used as a printer’s apprentice—Thomas F. Adams’s Typographia; or, The Printer’s Instructor (first published in 1845):

By the time Clemens wrote Roughing It, the demand for uniformity, both from readers and from the typesetters themselves, had significantly increased. Writing in 1901, Theodore Low De Vinne noted pointedly that

To emend the copy-texts of a published work by Mark Twain so that in its spelling, and in nontextual questions of typography, the text is as authorial [begin page 904] as possible, as well as uniform in spellings for which the author had no distinct or discoverable preference (except the tacit one that uniformity be preserved in print), is therefore not to “modernize” the work but to restore part of its authenticity.312 The lack of uniformity in the first edition cannot by itself show that uniformity was not a goal, both of the author and of the typesetters. There is no reason to suppose that Clemens intended his spelling to be pointlessly varied. It is no more likely that he would intentionally spell words in more than one way—without some reason for doing so—than it is that he would willingly allow misspellings in the text (always excepting the representation of speech, especially dialect). Clemens agreed with what his friend Joseph T. Goodman said to him in 1881, when he questioned Clemens’s habitual (and no doubt unintentional) misspelling of “champagne” as “champaign”:

Apart from the correction of error, emendation of the copy-text accidentals was carried out with three priorities in mind. The first was to restore authorial spellings wherever possible, not just in resolving ambiguous end-line hyphenation, and to record each change as an emendation.314 Taking that step alone brings the text very close to uniformity in spelling throughout. The second priority was to achieve uniformity in spelling the relatively few words for which the author’s preference is either unknown or nonexistent. And the third was to complete the task begun by the book [begin page 905] typesetters, emending typographical or nontextual variants that they tried, but failed, to make consistent.315

In the absence of the printer’s copy and the proofs for Roughing It, authorial preference in spelling must be determined from independent, or collateral, evidence.316 Such evidence was gathered by a computer-assisted search of holograph manuscript letters for 1867–70 and a reading search of manuscript letters for 1853–66 and 1871–72, plus the following literary manuscripts (as well as miscellaneous later manuscripts as needed): “Sarrozay Letter from ‘the Unreliable’ ” (1864); “The Mysterious Chinaman” (1864–65); “Angel’s Camp Constable” (1865); “The Brummel-Arabella Play Fragment” (1865, 1870–71); “The Only Reliable Account of the Celebrated Jumping Frog of Calaveras County” (1865); “Burlesque ‘Il Trovatore’ ” (1866); “How, for Instance?” (1866); Sandwich Islands lecture notes (1866–67); “Interview with Gen. Grant” (1867); “A New Cabinet ‘Regulator’ ” (1867); “A Plea for Old Jokes” (1867); an unpublished letter to the San Francisco Alta California (1867); “Mr. Brown, the Sergeant-at-Arms of the Senate” (1867–68); draft of “The American Vandal Abroad” lecture (1868); “Assaying in Nevada” (1868); “Boy’s Manuscript” (1868); “Colloquy between a Slum Child and a Moral Mentor” (1868); “The Frozen Truth” (1868); “The Legend of Rev. Dr Stone” (1868); “Remarkable Sagacity of a Cat” (1868); “I Rise to a Question of Privilege” (1868); manuscript fragments C, D, E, F, G, H, J, K, L, M (1868), discarded from The Innocents Abroad (NPV A7, A11, A20, A22, A24, A25, A26, A27); “L’Homme Qui Rit” (1869); “Remarkable Idiot” (1869); “Scenery,” draft of the “Curiosities of California” lecture (1869); untitled burlesque letter from Lord Byron to Mark Twain (1869); “Chinese Labor &c” (1870); “Housekeeping No. 1” and “Housekeeping No. 2” (1870); “Interviewing the Interviewer” (1870); “A Protest” (1870); “The Reception at the President’s” (1870); “The Tennessee Land” (1870); “A Wail” (1870); draft of the Artemus Ward lecture (1871); notes for the “Roughing It” lecture (1871–72); “An Appeal from One That Is Persecuted” (1872); “The ‘Blind Letter’ Department, London P.O.” (1873); “Foster’s Case” (1873); play fragment, a dramatization of the “Arkansaw Incident” (1873?); “Samuel Langhorne Clemens” (1873); “A True Story, Repeated Word for Word As I Heard It” (1874); “The Experience of the McWilliamses with Membranous Croup” [begin page 906] (1875); the partial manuscripts of The Gilded Age (1874) and A Tramp Abroad (1880); and the complete manuscripts of The Adventures of Tom Sawyer (1876); The Prince and the Pauper (1882); Adventures of Huckleberry Finn (1885); and A Connecticut Yankee in King Arthur’s Court (1889). For each word or word pair (like “a while”) spelled in more than one way in the Roughing It copy-texts, a context-sensitive examination was made of every occurrence found in these manuscripts, determining whether or not the manuscript usage was reliably comparable to the usage in Roughing It. More weight was given to manuscripts written before 1885, and especially to those written in the period 1869–71. A count was made of all collected usages, and the results were judged to belong to one of three levels of evidence. It should be noted that a number of the adopted spellings are not more, but less, modern than the spellings probably imposed by the typesetters.

Level 1. At least ten occurrences of a word or word pair were found in holograph manuscript, all spelled alike.

Level 2. At least four occurrences were found in holograph manuscript, and at least three-fourths of them were spelled alike—that is, one form predominated over the other (or others) in a ratio of 3 to 1.

Level 3. Three or fewer occurrences were found in holograph manuscript—or more were found, but no single form predominated in a ratio of 3 to 1.

Level 1 evidence was deemed strong enough to justify emending invariant spellings of two words in the copy-texts: the noun “envelope” to “envelop” (twice), and “Saviour” to “Savior” (once). Four spellings that predominated in the copy-texts were confirmed by level 1 evidence: “anybody,” “anyhow,” “whisky,” and “winter.”

Level 2 evidence was deemed strong enough to establish clear authorial preference, and was used to overturn copy-text predominance, even when the predominant form was found to occur occasionally in manuscript. Thus for eight words the predominant spelling in the copy-texts was ignored, and the author’s preference uniformly adopted: “a while,” “backwards,” “cañon,” “County,” “offense,” “pretense,” “river,” and “Sanitary.” In one case, “barkeeper,” the spelling found three out of four times in manuscript was adopted even though it never occurred in the copy-texts, which had either “bar keeper” or “bar-keeper.” For ten words in which no form predominated in the copy-texts, level 2 evidence established a preferred spelling: for example, “enterprise,” “half past,” “mould,” and “practice.” In twenty-seven cases a spelling that predominated in the copy-texts was confirmed by level 2 evidence: for example, “afterward,” “by and by,” “cannot,” “centre,” “ecstasy,” “King,” and “worshipped.”

Level 3 evidence was regarded as significant, but less conclusive about the author’s preference than levels 1 or 2. When variant spellings fell into this category, the spelling predominant in the copy-texts was adopted, provided that it was also found at least once in manuscript. This pattern [begin page 907] occurred in sixty-two cases: for example, “boulder,” “height,” “Islands,” “lawsuit,” “sage-brush,” “shoveled,” “skurrying,” and “tattooed.” In six cases no manuscript usage was found that confirmed the predominant copy-text spelling, and so the manuscript form that was found (if only once) was adopted: “brim full,” “enclosure,” “log house,” “pocket-mining,” “station,” and “wild-cat” (n.). When the copy-text variants occurred in equal numbers, any manuscript usage found was relied on to resolve the variation: level 3 evidence established the preferred spelling in twenty-six such cases: for example, “caulk,” “foot-notes,” “Garden of Eden,” “per cent,” and “St. Joe.” In one case, “shotgun,” the spelling found in manuscript was adopted even though it never occurred in the copy-texts, which had either “shot gun” or “shot-gun.”

For forty-three variant spellings in the copy-texts, no exact manuscript evidence at all was found. When one spelling predominated in the copy-texts (as happened in twenty-seven cases), it was necessarily adopted. When no single spelling predominated (sixteen cases), holograph spelling of similar words was considered, as well as the preferred form in Webster’s 1870 Unabridged Dictionary.

Resolution case by case (“spring,” “summer,” and “winter,” all individually considered) rather than by category (the seasons) was attempted for most spelling variants, partly because that method was most likely to recover genuine authorial practice, but also because the result turned out to be categorically consistent as well. In a few cases, however, the categorical method was effective both in recovering Clemens’s practice and in maintaining overall uniformity where the manuscript and copy-text evidence was otherwise scarce.317 Emendation to retrieve authorial spelling and achieve uniformity was applied in the following spelling categories by using the same criteria as case-by-case emendation, except that both copy-text and holograph occurrences were counted in groups. (a) The spelling “county” was emended to “County” when it was part of a name, as in “Carson County.” (b) “River” and “Station” were emended to “river” and “station” when part of a name, as in “Green river station.” (c) The hyphen between “a” and a participle was emended to a space, as in “a hunting” rather than “a-hunting.” (d) Word space between the numerator and denominator of fractions was emended to a hyphen: “two-thirds” rather than “two thirds.” (e) Two-em dashes to show interrupted speech were [begin page 908] emended to one-em dashes: “and—” rather than “and——.” (f) The possessive of “Jarves” (a proper name ending in “s”) was emended with an additional “s”: “Jarves’s” rather than “Jarves’.” (g) Compound modifiers that included numbers but no hyphen, or only one hyphen, were emended: “ten-cent piece” rather than “ten cent piece,” and “three-thousand-foot precipice” rather than “three-thousand foot precipice.” (h) All references to “the Lake” (Mono or Tahoe) were emended to “the lake.” (i) All spelled-out forms of “Captain,” “Colonel,” “General,” “Governor” (as parts of names), and “United States” (as an adjective) were emended to their abbreviated forms: “Capt.,” “Col.,” “Gen.,” “Gov.,” and “U.S.” (j) All closing quotation marks preceding other punctuation were emended to follow it instead. (k) Parentheses to signal interpolations or asides were emended to square brackets: “[falling inflection]” rather than “(falling inflection).”

Variant spellings that denote a difference in meaning were not emended. The distinctions between “Heaven” and “heaven,” “Overland” and “overland,” “Plains” and “plains,” and “Nature” and “nature” are difficult to articulate, but nevertheless real. Likewise variable are Clemens’s neologisms or slang set off, or not, by quotation marks (“flush times,” “square meal,” and “feet”) in virtually identical contexts.

Despite Clemens’s jealous control of his punctuation, there certainly were aspects of typesetting that he willingly if tacitly ceded to the printing house. In other words, Clemens’s practice in manuscript is not invariably a sound guide to his intention for a literary work—the ampersand, for example, which he used throughout his manuscripts, always expecting it to be expanded to “and.” In the use of numbers versus figures, or in the treatment of long quotations, Clemens accepted the participation of the printshop in devising and applying a uniform standard. A comparison of the first-edition text with the author’s manuscript usage, as well as with the several copy-texts, provided a good deal of information about what kind of styling, textual and nontextual, the book typesetters applied. For example, they usually (but not invariably) spelled out numbers, even though the newspaper compositors (and Clemens himself, at least with informal manuscripts like letters) used numerals for convenience and speed. In spelling out numbers the book compositors may well have had the warrant of Clemens’s holograph manuscript chapters: draft page 968 spells out both “six thousand” and “twelve thousand dollars.” But whether or not Clemens’s manuscript was consistent in this matter, the book compositors tried to be (albeit with imperfect success), deliberately printing numerals only for hours of the day followed by “a.m.” or “p.m.” Collation reveals that they achieved nearly consistent results in all categories except for dollars. They tended to leave the dollar figures they encountered in printed texts, such as the Buffalo Express (where newspaper style made figures conventional), perhaps because a clipping gave the impression of greater finish, having already been styled by the previous compositors. Emendation (always recorded) was therefore undertaken to complete the [begin page 909] pattern begun by the book compositors in spelling out numbers—including dollars, except in two instances where numerals were deemed acceptable (both discussed in textual notes).318

Similarly, the book compositors set all extended quotations in smaller type, sometimes enclosing them in quotation marks, sometimes not. In a few places they set double quotation marks within an extract already enclosed in quotation marks, which suggests that they added the surrounding double quotes without attending to the internal ones. Examination of Clemens’s practice in marking quoted material in his literary manuscripts of the 1870s and 1880s reveals that all three possible treatments occur with seemingly equal frequency: quotation marks with, and without, a request for small type, and small type alone. It is also clear that the compositors did not always follow their copy, but instead adopted a house style—with varying degrees of consistency. When the author wrote a specific directive in the manuscript, such as “Put this in small type,” they usually followed it, but sometimes they removed quotation marks, and sometimes supplied them. While early-nineteenth-century printer’s manuals make no mention of how to treat extracts, the later ones explicitly advise compositors not to use quotation marks when using small type; from this we can infer that redundancy was becoming less acceptable.319 This edition perfects the apparent intention of the book compositors by following their prevailing pattern: quotation marks enclosing extracts in small type were removed by emendation.

Several additional minor points of typographical style required emendation for uniformity. The book compositors were nearly, but not entirely, consistent in styling names of ships in roman type and names of newspapers in italic. Three emendations in these two categories were made to achieve uniformity. The compositors also failed to achieve a consistent treatment of foreign words, and in this instance the evidence does not point to a categorical solution. Three foreign words were invariably set in italic type: “riata,” “tapidaros” (a misspelling, emended to “tapaderas”), and “awa.” Numerous other foreign words occurred only in roman type: “cañon” and “poi,” for example. Five words and one abbreviation appeared in both italic and roman type: “adobes,” “hula-hula,” “i. e.,” “kahilis,” “tabu,” and “taro.” No distinct pattern of normalizing was discernible in the evidence of collation. The italicized words appeared in material set from the author’s holograph as well as from clippings from the Sacramento Union; in some cases the italic form was changed to roman, in others [begin page 910] the reverse was true. The fact that some italic forms also represent a word used as a word (“adobes, the Spanish call them”; “Tabu . . . means prohibition”) added a complication. It seems probable that the first-edition typesetting reflects Clemens’s own inconsistency, overlaid by incomplete compositorial attempts, at one or more stages, to extend his sporadic markings for italics to all like instances. A word occurring in a single invariant form was deemed likely to reflect at least some authorial impulse to render it in that form; all invariant forms were therefore left unemended. Furthermore, it seems likely that any word occurring in both roman and italic forms reflected some authorial impulse, albeit sporadic, to render it in italics; all variant forms were therefore consistently italicized.

Italic punctuation following italic words (an accepted convention in American typesetting since at least the early nineteenth century) was adopted in the first edition in all but three instances, which have been emended here. Footnotes in the first edition were typeset with a paragraph indention except for two instances, in which the notes were centered: both have been emended here. Mark Twain’s initials at the end of a footnote were emended to delete brackets in one instance.

The front matter in the first edition (the List of Illustrations and the Contents), the illustration captions, and the headlines (running heads) on each page were all undoubtedly supplied by Elisha or Frank Bliss, perhaps with Orion’s assistance. All usages in these sections, being nonauthorial, were excluded from the first-edition tally of spelling variants and the like, but they were emended for uniformity according to the same policy as the rest of the text. Whenever the titles in the List of Illustrations, the wording of subjects in the Contents list, and the illustration captions themselves correspond to wording in the text, emendation was applied to make them accord exactly, and the quoted phrases were enclosed in quotation marks. Whenever the title of an illustration, as given in the List of the first edition, differed from the wording of the caption, the caption was deemed more accurate, and the title in the List emended to match it.320 The running heads, which cannot be adapted to the paging of the present edition, have been silently omitted. In addition, the following alterations were carried out without being individually reported in Emendations of the [begin page 911] Copy-Text and Rejected Substantives. (a) All publication information no longer applicable to the present edition (such as “issued by subscription only” and “AMERICAN PUBLISHING COMPANY. . . . 1872”) was removed from the title page. (b) The seven full-page illustrations were printed in the first edition as separate plates, which were not included in the pagination. For these seven illustrations the reading in the List—“(Full Page,) (Face Page)”—was altered to read merely “(Full Page),” followed by the appropriate page number in this edition. (c) All first-edition front matter was styled in accordance with the design of this edition: periods were removed from the title page, and illustration titles on the List and the subjects in Contents were altered so that nouns, adjectives, and adverbs begin with initial capital letters, prepositions and conjunctions with lowercase letters. (d) Arabic chapter numbers, with no punctuation, were substituted for roman numerals followed by periods, and the opening of each chapter and appendix was styled in capitals and small capitals and set flush left with no paragraph indent.

H.E.S.