University of Virginia Library


Page 241

James E. May

Edward Young's prose satire and homily The Centaur Not Fabulous was published in 1755 and reprinted twice in revised editions that year. It was a sufficiently popular work—often enough reprinted and translated—to deserve bibliographical and textual scrutiny.[1] The work is also of interest be-


Page 242
cause of the collaborative involvement of its printer, Samuel Richardson. Richardson's friendship with Young led to an extraordinarily patient and on-going pre- publication revision of the edition, encouraging bibliographical complexities and affording us the correspondence that partly records the work's growth. Last minute revisions to the first edition led to an errata list (on the back of the title-page) with nine corrections, a manuscript correction to all copies, and the cancellation of at least ten leaves.

Letters between Young and Richardson record many major developments in the revision and printing of The Centaur. On 14 July 1754, Young wrote Richardson that he had something “about the length of five sermons” that he wished to publish with Richardson's assistance.[2] He sent Richardson his draft of its satirical dedication on 21 July 1754, remarking “Blot, add, alter, as you please; and, if then you approve it, print it” (Pettit, 405). On 28 July Young, estimating the manuscript's length as that of “five or six sermons,” proposed sending “not all the copy at once, but letter after letter, being four in all [five would be published],” and on 1 August he promised to send “the first letter” four days later (Pettit, 407). Richardson responded on 5 August with thoughts about the difficulty of pricing individual pamphlets, persuading Young to publish the Letters as a single unit and, perhaps, encouraging him to expand the text (Pettit, 408-409). Richardson then assured Young that, although he would “attend, as you desire, to the copy.... You shall see every proof,” and Richardson added, “I will send you a proof of the dedication as soon as possible” (Pettit, 409-410). After Young wrote 12 August to request that Richardson find an artist to cut a frontispiece illustrating the modern centaur described in the dedication, the printing began, with Young sending copy in installments from late summer 1754 through January 1755 and Richardson returning proof sheets for Young's correction. Young worked slowly, requiring multiple proofs to revise, not just correct, his text. In apologizing to Richardson for “slovenly copy” on 1 August, Young remarked, “my eyes are bad; and I had rather compose two letters than write one” (Pettit 408; see also 412 on Young's needing his housekeeper Mary Hallows to read for him). The conclusion to the fifth and final letter is dated 29 November 1754, and the undated postscript defending the work's “mixture of Levity with Solemnity” (375) was only written after Richardson's next extant discussion of The Centaur, in a letter of 21 January 1755.

On 21 January, when the press had printed to sheet Z and probably to Aa, but certainly not the last five leaves with the end of “The Conclusion” and the five-page postscript, Richardson proposed substantial revision to the dedication, requiring “a few pages [be] cancelled... after some such manner as I have presumed to offer in a waste sheet of the print, which I enclose” (Pettit, 416). In addition, he offered detailed corrections, mostly verbal, on


Page 243
sixteen pages from pages 24 to 325 (D4v to Z3r) of the first edition (Pettit, 416-417). Most of these corrections involve grammatical and phrasing errors that Richardson should have caught if he had carefully read proof earlier. Few of these changes were thought significant enough to require insertion into the first edition. Two corrections (to pages 24 and 92) made their way into the errata, and another resulted in the words “for their” being added in manuscript on page 131 of all copies. Richardson's corrections to pages 92, 158, 172, 207, 291, 307, and 325 were inserted into the second edition. In addition, Richardson's question whether too many points were not “crowded” into Letter V led to its division into Letters V and VI for subsequent editions. More significantly, Richardson remarked:

In another impression methinks it were to be wished that all from, If this is a man of pleasure, p. 161, to, from a higher hand, p. 163, were omitted, as it interrupts, by ludicrous images, emotions that were nobly excited.

For the same reason, suppose, in p. 163, were omitted the words, Fain would I bury &c. to the end of the paragraph, real men, p. 164?

(Pettit, 416-417)

This remark led to the cancellation of leaves N1-N2 (pp. 161-164) and their replacement with an altered text.

In apologizing for the liberty taken in printing a revised draft of the dedication, Richardson explained that, having “seen the whole together,” he was “apt to think that the reader is not sufficiently prepared in that dedication for the solemn and elevated subjects of the following letters” (Pettit, 416). Young, apparently the next day (“Wednesday”), approved these changes and asked that Richardson “alter, as you think good; and let not delay or expense be any objection to any thing now practicable.” And he particularly asked for help with “the close now sent” (Pettit, 418). Although the final paragraphs of “The Conclusion” might be termed a “close” (“now, my Friend, Farewell,” Cc1v/370), Young probably refers to a manuscript version of the Postscript (Cc3v-Dd1v), which directly addresses Richardson's objections on 21 January regarding excessive levity. Closing, he added, “I desire this last sheet [presumably sheet Bb] again before worked off” (Pettit, 418). When Young next wrote (“Sunday,” Pettit supposed the very next, 26 January), he returned Richardson's revised printing “for ye Dedication part,” requesting the cancellations required for Richardson's “most judicious Insertions.” He added that “Tomorrow I expect ye last Sheet, & when I return yt, I shall... consult with you if the d[eletion]s p. 161 &c: can be made in this Edition. Expence, if yt is all shall not hinder it” (Pettit, 419). Young approved cancellations in the dedication during the very week that the final five leaves of the book were sent to be set in type, and the contents of leaves N1-N2 were soon to be rewritten. Thus, some of the cancellations were likely to be impressed on the same sheet or sheets with the final five leaves of the book. During the next fortnight, prior to the book's being bound for its sale in the first days of March, Young had to rewrite at least leaves N1 and N2 and read proof for them and the final five leaves, and the press had to print these leaves along with sheet Bb and the cancellantia within the dedication. On 13


Page 244
February (Pettit's conjecture for a “Th” date), six days before Robert Dodsley purchased for himself and Andrew Millar the copyrights on 19 February 1755 (document printed in Pettit, 599), Young had finished work on the first edition and turned to revising for the second (Pettit, 420-421).

Young and Richardson's correspondence prepares us for finding that some leaves in the B sheet (containing the dedication) as well as N1 and N2 are cancels. But nothing in the correspondence refers to three other canceled leaves, R1, R2, and U1. T. C. Duncan Eaves and Ben D. Kimpel rightly perceived that these, in addition to N1 and N2, were cancels, and they correctly identified two of the five cancellantia in gathering B when they claimed, “Three pages of the dedication (iii, viii, xi) were certainly reset since they are short by one or two lines.”[3] However, as will be shown, although leaves B2 and B6 (carrying pages iii and xi) were canceled, leaf B4 (with page viii) was not. The text cancels are easily detected since leaves N1.N2 and R1.R2 are folds, N2 and R2 are unsigned, N1.N2 have a different paper-stock from N3-N8, and the stubs of the cancellanda are more often than not visible. Eaves and Kimpel checked the press figures in a few copies and found the stubs of the obvious cancels, but they did not collate editions or scrutinize bibliographical features such as the paper-stock (nor as biographers would they have been expected to do so). They have left the task of identifying the cancels in the B gathering and of interrelating the cancels, singletons, and partial sheets in the edition. With evidence drawn from 58 copies (listed in Appendix 2, where abbreviated codes are noted), including an uncut copy at St. John's College, Cambridge, I will reconstruct the three final printed units of the book, placing 16 leaves that remained to be printed on 26 January on a whole sheet and on the halves of two other sheets. Scholars working on comparable cases of cancellation may find applicable the arguments drawn from press figures, point holes and dimensions of an uncut copy, watermarks and tranchefile patterns, and the position of chain-lines in many extant copies. The multiplicity of approaches and the many extant copies examined allow us to test the value of conclusions about imposition patterns drawn from chain-line positions.

The edition has many partial gatherings and cancellatia: the frontispiece, a confusing nine- leaf B gathering, two cancellans folds in the text (N1.N2 and R1.R2, with material from Letter III and from the beginning of Letter V), half-sheet Cc4, and four singletons (title-leaf A1, final leaf Dd1, and two cancels: B3 [with the third leaf of the preface] and U1 [with material from the “Thoughts for Age” section of Letter V]). As I shall argue, leaves A1, B3, U1, and Dd1 completed a sheet containing Cc4; cancellans folds in the B gathering that appear as if they are B2.B7 + 1 and B6.B7 were printed adjoined on half of a sheet; and cancellans folds N1.N2 and R1.R2 were both parts of a separate whole sheet. Thus, the octavo has the collational formula:

8°: (frt +) A1 B8 (-B2, B3, B6, B7; B1 + χ4, with χ2-4 between B5 and B8 [appearing as B6, B7, B7 + 1]; χ1 + 2χ1 [`B3'] C-M8 N8 (-N1,N2; + `N1'.N2 [=3χ8 -6 leaves in-


Page 245
cluding `R1'.R2]) O-Q8 R8 (-R1,R2; + `R1'.R2 [=3χ8 -6 leaves including `N1'.N2]) S-T8 U8 (± `U1') X-Bb8 Cc4 Dd1; $1-4 (-N2, R2) signed; cancellans χ1 and 2χ1 missigned “B2” and “B3”]; 200 leaves; pp. [4 (frontispiece and title-page)] i-xvi [2 (section title)] I 2-53 54-56 57-221 222-224 225-378. [Note. Although for clarity I have not embedded the information in the collational formula, the reconstruction proposed in this paper would create a Cc8 sheet containing leaves A1, B3, U1, and Dd1 positioned as follows: A1=Cc3, B3=Cc7, U1=Cc2, and Dd1=Cc6, with the leaf signed `Cc2' occurring in the position of Cc4.]

First, I need offer some prefatory observations on the paper-stocks and their watermarks. The edition was printed in common octavo imposition, with the watermark appearing principally on either leaves $1 and $4 or on $2 and $3, and tranchefiles (which produce narrower chain- widths at the shorter sides of the sheet) sometimes on the first four leaves of the gathering but never the last four. All four paper-stocks employed in The Centaur (each produced from a pair of nearly identical moulds) have watermarks, and three have tranchefiles. The dominant paper-stock (also the sole paper-stock in the second edition) has a single “NH” watermark (no sheet of any copy has been seen with double marks). The 14 mm-tall letters are divided by a chain-line; the “N” has single-lined verticals and a double-lined diagonal; the “H” is entirely double-lined. This paper-stock has fairly distinct chain-lines slightly wider than those in other paper-stocks in the book, four chain-widths usually measuring 104-107 mm. It has one tranchefile space of 12-14 mm at both short sides of the sheet (seen regularly in the uncut Csj copy and in many copies only slightly trimmed). The three other paper-stocks all have fleur-de-lys watermarks centered on chain- lines, “IV” countermarks, and more distinct, or easily visible, chain-lines 25-25.5 mm wide. The three are distinguishable from each other by differences in the shape and size of their fleur-de-lys watermarks. One fleur-de-lys paper-stock is found in sheets E and N of all copies; a second is found in sheet Q and the frontispiece of all copies. Of concern to us is the third fleur-de-lys paper-stock, found in folds B1.B8 and B4.B5. Unlike the other paper- stocks, it lacks a tranchefile. The watermark has a broad tail the size of that in the first paper-stock (23-24 mm across), but a broader head (17.5-19 mm across), sometimes asymmetrically larger on one side of the chain-line. Both the “NH” and fleur-de-lys watermarks are placed so close to the short exterior edge that they almost never appear on the four internal leaves of the sheet, and the “NH” is usually seen on only one of the exterior leaves of the short axis.

That said, let us consider the book's preliminaries, correcting Eaves and Kimpel's remarks on the cancellantia in gathering B. The frontispiece, with a fleur-de-lys paper-stock, and title-leaf (A1), with the “NH” stock, are disjunct. The frontispiece was almost certainly printed by itself and not with an earlier, canceled title- page without an errata list since watermarks appear in the frontispiece sometimes at the top outside corner (as in C, CtY, FU, MH1, MiU, OTH) and sometimes at the top inside (as in MH2, NcU, Ose, and ViWCF). All copies seen have an errata list on A1v. As originally printed, the title-page was followed by a dedication occupying leaves B1-B7, with a sec-


Page 246
tion title to Letter I completing the B sheet, but, as revised to accommodate Richardson's suggestions of 21 January, the dedication required an additional leaf after B5v. As published, the nine-leaf unit with dedication and section title contains four folds, two enveloping (B1 and section title B8, and B2.B7 + 1 [χ1.χ4]) and two back to back (B4.B5 and B6.B7 [χ2.χ3]), and a disjunct leaf, `B3' (2χ1), glued in. Two original folds, B1.B8 and B4.B5, are distinguished by a fleur-de-lys paper-stock. Fold B4.B5 is regularly seen conjugate, and the section title to Letter I is folded improperly so as to precede B1 and seen conjugate to it in ten copies (CaOTU, CLU-C, MB, MH2, O, OAU, Ose, PSt, TnU and TxU). Leaves B1 and B4 frequently have complementary halves of the watermark (as in C, CoU, InU, L, LU, MdBJ, MR1, MR3, O, OCU, Ose, and WU) or of the countermark (as in CaAEU, CaOTU, DLC, LdU-B, May, MH2, NRU, Occ, OTH, Owo, TnU, and Wns). Furthermore, in no copies does the same portion of the watermark appear in both B1 and B4. Thus, it is evident that the folds of which they are parts were formerly adjoined at the top as one half of the original B sheet. Within the B cancellantia, enveloping fold B2.B7 + 1 is seen conjugate in LdU-B, May, OCiU, and O-HF copies, and back-to-back fold B6.B7 is regularly seen conjugate. All five cancellantia are printed on the paper-stock with “NH” watermark, also found in the cancellantia within the text. Whereas the type- settings for the pages with fleur-de-lys stock are all reset for the second edition, those on the cancellantia with the “NH” paper-stock (B2-B3 and B6-B7 + 1) are not reset for that edition but reappear on B1-B2 and B5-B7, again on the “NH” paper-stock.[4] From this fact, the signatures, and the errata change for page ii (B1r), we can be certain the two folds on the fleur-de-lys stock are not the cancellantia but part of the original gathering.

Since, in addition to the title-leaf and the B cancellantia, the “NH” paper- stock is also found in the cancels within the text (N1.N2, R1.R2, and U1) and in the five final leaves (Cc4 + Dd1), we have sixteen leaves to interrelate, the sixteen likely to be those last printed (if R1.R2 and U1 were also late alterations). For the sake of efficiency, Richardson would presumably have wished to print these sixteen leaves on as few units as possible, ideally on two sheets.

The press figures offer some guidance in hypothesizing what leaves were printed with what others.[5] For the 16 leaves of cancellantia and partial units,


Page 247
press figures appear in all copies on B7+1r (figure 7), R2r (figure 2), and both sides of leaf Cc3 (Cc3r-1 and Cc3v-4). Note that four different figures, potentially presses or teams of pressmen, were involved in printing the final units of the book. This may suggest that multiple presses were at once printing off the final leaves during early February, as the work was hurriedly prepared for the binders. In the preceding sheets of the edition (C-Bb), the same figure sometimes recurs in sequential sheets, as in sheets H-I (3), O-P (3), and X-Z (1), and figure 7 appears on the inner forme of sheet Aa (with figure 2 on its outer forme) and as the sole figure in Bb. Since cancellans B7+1r also has figure 7, we can presume that the cancellans folds in B were not printed until February, for on 26 January Young was expecting the return of sheet Bb the next day for his approval (Pettit, 419).

The four press figures exclude certain possibilities. The two figures in Cc4 suggest that it was not produced with its eight page-settings locked in the chase and the sheet perfected from the same forme but that Cc4 was printed with four other leaves. Half-sheet Cc's two figures support the hypothesis that Cc4 was printed with the singletons, for were the two cancellans folds in B or were R1.R2 printed with it, the sheet would be figured three times. Also indicated is that pages B7+1r and R2r were not printed on the same side of the same sheet (as cancels they need not be both in the inner forme as the second and eighth rectos normally would be). There is no need to assume they were on different sides of the same sheet, for most sheets in this edition are not figured on both sides. Indeed, one might reason that, since the vast majority of whole sheets have a single figure, and since both R2r and B7+1r are figured, they were probably on separate sheets; however, that conjecture must be qualified since the sheet with Cc4, printed close to publication, has two figures. (Watermark evidence below will indicate that the B cancels were not printed with cancels N1.N2 and R1.R2.) Finally (although, atypically, sheet E and, after cancellation, sheet U are never found with a press figure), we can suppose it likely from the press's general figuring of sheets that the lack of a figure on the group of leaves A1, B3, U1, and Dd1 indicates that they were not printed by half-sheet imposition but with other leaves on a sheet.

St. John's College, Cambridge, has an untrimmed copy in a contemporary binding providing both point-hole and diverse paper evidence about the


Page 248
printing of the sixteen leaves of cancellantia and other partial gatherings.[6] The leaves of the book are typically 212-213 mm tall (such as Cc1 and Cc2). dividing a sheet width of 425 mm, but individual leaves can vary in height from 208 (C5) to 217 (C8); leaves are typically 130 mm wide, though individuals vary from 124 to 138 mm.

In The Centaur, the point holes (from pins in the tympan that pierced the paper along the short axis at the center of the forme) appear between 110-165 mm from the top of the leaf, with one 20-30 mm higher than the other. What Philip Gaskell records as a general practice is true of this and other octavos of Young's works printed by Richardson during the 1750s: the “near-side point was placed further in towards the middle of the tympan (and of the sheet).”[7] For instance, in the Csj copy, leaves C5 and C6 share a point hole at 148 mm down leaves 208 mm tall; C7 and C8 share a hole at 120 mm down leaves 217 mm tall. Of the cancellantia in gathering B, only the leaf appearing as if B7+1 (χ4) has a point hole. It appears as a hemispherical bite along the edge, 114 mm down from the top edge. Within the singletons, point holes are not found on cancellans B3 and Dd1. However, half-sheet Cc4, whose press figures suggests it may have been printed with these leaves, has point holes on Cc3 down 110 mm and on Cc4, down 153 mm. As for folds N1.N2 and R1.R2, R2 has a large hole at the edge, down 165-170 mm on a leaf 210


Page 249
mm tall at the edge: a bite marking out 270 degrees of a circle is missing. On N2, along the edge at 165-170 mm down a leaf 210 mm tall at the outer edge, there is an outward flap that constitutes the contents or surplus of the point hole on the edge of R2. The bite and surplus are alike in size, shape and position. (If N2's flap were the surplus from a point hole on a leaf other than R2, it could not be from the other leaf in the same sheet with a point hole, for that hole would be closer to 110 mm down from the top of the leaf.) Furthermore, at 21 mm above the start of the bite on R2 and the surplus on N2, we find a small 1-mm nick in R2 and a compatible 1-mm knob on N2. It is evident that N2 lay along the edge of R2 at the center of a whole sheet, in a row of four leaves along the long axis of the sheet, with the recto of N2 adjacent to the verso of R2, as N1v.N2r.R2v.R1r. Since the point hole in leaf B7+1 is only 114 mm down from the top edge and that on N2 and R2 at 165 mm down could be the complementary hole, we might be tempted to suppose them a pair, but that inference is blocked by watermark evidence (soon to be offered) that B2.B7+1 and B6.B7 were adjoining at their top edges.

The untrimmed margins of the Csj copy provide other useful evidence. First, there are matching uneven edges produced during cutting and tearing apart of formerly adjacent paper into separate leaves. As with outer edges of leaves N2 and R2, there are deficits and surpluses in the ragged upper edges of B2 and B6 that indicate these leaves were formerly joined at the top. If we measure in from the outer edge on B2, we find that the leaf is short of the uniform top edge until 15 mm in toward the fold and then again from roughly 25-35 mm in, but B6 has a surplus from 0-15 and 25-35 mm in, making it taller than the uniform height. And from 55-70 mm in on that top edge occur two dips where B2 has a deficit below the uniform edge and B6 a surplus above it (the valleys in B2 and the hills in B6 match exactly when they are xeroxed and cut along the edge, and these images are laid against one another with one image up and the other down). In addition, matching hills and valleys along the top edge are also shared by Dd1 and B3, providing further support for their formerly adjoining at their top edges. At 1-3 mm in from their outer edges, there is a deficit on B3 and a surplus on Dd1; from 50 to 66 mm in, there is a rectilinear deficit on B3 and a comparable surplus on Dd1; and, from 66 to 90, there is a rectilinear surplus on B3 and a comparable deficit on Dd1.

In addition, since in the Csj copy the cancels and partials were not trimmed but only cut apart (as the point holes and watermarks testify), the leaves' dimensions provide further evidence of the leaves' positions in the sheets before division. The dimensions of the singletons support the arrangement hitherto argued: leaf A1 is between 213-215 x 137 mm, and U1, 212-213 x 137 mm; Dd1 is 215 x 132 mm, and B3, 212-213 x 132 mm, providing normal dimensions for half of a sheet in this book, 425-430 x 269 mm. The shared widths of 137 mm for leaves A1 and U1 and of 132 mm for leaves B3 and Dd1 suggest a cutting along the long axis of the half-sheet down a fold


Page 250
between two unequal halves. The shared heights of leaves A1 and Dd1 and of leaves U1 and B3 suggest these leaves had received the same fold along their top edges before they were separated.

Both the watermarks and the tranchefiles of the uncut Csj copy offer telling evidence. In this copy we find leaf A1 with a tranchefile and a portion of the “NH” watermark that complement those on U1 (8-9 mm of the top of “N” and one vertical of “H” on A1 and 4.5 mm of bottom of “N” and one vertical of “H” on U1); in addition, the chain- lines match up at the top of the page: at 12-13, 41, and 66 mm out from the spine on A1 and 14, 42, and 68 mm on U1. (Since they are singletons glued to stubs, we cannot expect the chain-lines to appear at exactly the same distances out from the spine, but they should be close and the distances between the chain-lines should be proportional.) Similarly, we find comparable chain-line locations in B3 and Dd1 as if these leaves without watermark or tranchefile were adjoined at the top (10, 37, 64, 91 mm out from the spine on B3 vs. 13, 39, 66, 92 mm on Dd1). In line with our hypothesis that Cc4 was printed on a whole sheet with the singletons, we note that Cc1 and Cc2 do not share a watermark (as they should not since it is on A1 and U1), but they do share a tranchefile. It is also reassuring to note that the combined height of Cc1 over Cc2 measures 425 mm, the usual sheet width. Leaves B2 and B6, which we would like to show were printed together adjoining at the top, lack watermarks but both do have a tranchefile, 12-13 mm, and their chain-lines are a comparable distance from the spine (B2: 10-11, 36-37, 64-65 mm; B6: 9, 36, 62 mm). The chain-line distances on leaves B7 and B7+1 are also compatible.

Paper evidence involving the leaf sizes and watermark of folds N1.N2 and R1.R2 supports the point-hole evidence in placing the folds on the same sheet's long axis: N1.N2.R2.R1. There are no surpluses or deficits along the top edges of these folds that suggest they were adjoined at their top edges; in fact, the reverse is true: the top edge of N1, from the outer edge to 18 mm inward, is raised a full one wire-line taller (1-2 mm) above the general height of the leaf, and there is no comparable dip near the outside along the top of R1. Moreover, leaf R1 has 9 mm of the watermark's “N” showing, but none of the watermark shows on N1. This being an untrimmed copy, we would expect at least several mm of watermark, perhaps as many as 5 mm, to be visible there. It might be objected that a cancellans fold like N1.N2 was the product of scissored cutting, not the opening of a folded whole sheet, and so some trimming could have occurred. However, all the other cancellantia and partial gatherings have not been trimmed but only cut, and there is no reason for the binder to have done additional trimming for leaves N1.N2 since these leaves are shorter on the bottom edge and have their running-titles slightly above those adjacent. Furthermore, the dimensions of leaves N1.N2 and R2.R1 suggest they were all cut from the whole sheet with a single cutting along the long axis. Leaf N1 is only 208 mm tall for the most part; N2 is 208 close to the fold and grows to 210 at the edge; R2 is 210 near its outer edge and grows to 211-212 near the fold; R1 is 210-211 near the


Page 251
fold and 213 at the outer edge. Thus, the heights of all four leaves progress with a single movement from 208 on N1 to 213 at the outer edge of R1. The same slanting cut is evident if one considers the upper margin above the running titles: the space from the running title to the top edge on N1 is 20.5 mm, on N2 21.5, on R2 23.1, and on R1 (above the bracketed page number) 23.5. In addition, the combined distances between the running titles and the top edges on N1 and R1 total 44 mm, a total white space between type-settings that is comparable to the figures for B3 and Dd1 (45) and for B2 and B6 (45). If N1 and R1 were joined at the top edge, and if the 5 mm of missing watermark were cut from N1, then the running titles would have had roughly 4 mm more space between them than we find elsewhere in the book. All the evidence suggests N1 was untrimmed and lacks a watermark and so was not printed above R1.

Watermarks, tranchefiles, and chain-lines in all the trimmed copies seen confirm as a general pattern what is true of the uncut St. John's copy and consonant with press figure evidence: the leaf-pairs A1/U1 and Dd1/B3 were joined at their top edges in the same half-sheet; the singletons' half-sheet was printed on a whole sheet with Cc4; folds B2.B7+1 and B6.B7 were printed on the same half of a sheet; and N1.N2 and R1.R2 were not joined at the top to form a half-sheet but ran laterally across a full sheet. In addition, this evidence from the paper of all copies indicates that leaves A1 and U1 were printed at the outside of the sheet and that the inside edge of A1 was conjugate with Dd1, just as that of U1 was with B3. Also, the watermarks and tranchefiles apparent in many copies and the chain-line positions in most copies will confirm what is true of the Csj copy: disjunct leaves and partial gatherings printed on the same individual sheets of paper ended up in the same copy of The Centaur. Indeed, demonstrating the imposition pattern for the sheets last printed also involves demonstrating this uniform procedure in the binding shop. It allows the evidence from the uncut Csj copy to be applied safely to other copies of the book. Finally, confirming the imposition pattern with evidence from most extant copies rules out the possibility of multiple imposition patterns caused by interrupting the press run to reimpose on a single sheet the four-leaf units with the cancellantia in the dedication and the text.

Given that the vast majority of extant copies have been examined, we can generalize with some confidence regarding which problematic leaves may have tranchefiles and watermarks and which may not. Tranchefiles have been seen in some copies on A1, B2, B6, N1, R1, and U1, and five of these leaves (all but B6) have tranchefiles in the TnU copy. The six leaves with tranchefiles have been seen with watermarks also. In no copy does a tranchefile appear in leaves B3, Dd1, N2, and R2. In no copy does a watermark appear in leaves B3, Dd1, N2, and R2 without it also appearing more fully in leaves A1, U1, N1, and R1—leaves that are or were formerly adjoining and that commonly have tranchefiles and so were at the exterior of the sheet's


Page 252
short side. Usually with this paper-stock, the small watermark appears only in the two leaves at the same exterior side and is not shared by the next leaves in. However, in the LdU-B and WU copies, the double-lined base of one of the verticals of the “H” appears on leaf B3, and in these copies we find the “N” on leaf U1, suggesting U1 and B3 were conjugate and share the same watermark. Also, in the FU and NcU copies, the double-lined base of the “H” appears on Dd1, and in these copies we find the “N” visible on A1, suggesting A1 and Dd1 were conjugate and share the same watermark.

Ample watermark evidence testifies that at their upper edges leaf A1 adjoined U1 and Cc1 adjoined Cc2. Eleven copies examined have watermarks divided in a complementary manner of tops and bottoms between A1 and U1: Csj, DLC, FU, LdU-B, LU, MnU, MR1, MR2, NBiU, NSbSU, and TxU. In no copy does the extent of the watermark on one leaf rule out the possibility that the mark is part of the same one showing in the other leaf. Also, in no copy do we find the same portion of the watermark on both leaves A1 and U1. Also, as will be noted below, when a copy has a watermark only on either A1 and U1, the watermark often fails to complement watermarks found on other partials and cancels. Similarly, Cc4 appears to be a regular half-sheet from complementary watermark halves shared by leaves Cc1 and Cc2 in at least ten copies (CaOTU, CSmH, MH[2], MiU, Occ, Ose, OTH, Owo, and PSt).

Additional evidence on the placement of A1, U1, and all the disjunct leaves and partial gatherings can be drawn from chain-line position. Above I found it useful to note the matching chain-line positions on A1 and U1, as on B3 and Dd1, of the uncut Csj copy. Similar evidence confirming my hypotheses can be drawn from all the copies examined. Moreover, since in this edition we have other strong evidence from watermark and tranchefile patterns and from the uncut Csj copy for the imposition patterns within three sheets, we can test the value of evidence from chain-line position. Although there is nothing novel about employing chain-line position to situate conjugate folds and disjunct leaves on a reconstructed sheet, The Centaur offers an ideal case for applying the method systematically. For here we have good confirmation from other more reliable evidence, and measurements can and have been obtained from a large number of extant copies. Furthermore, every case is unique: here we are working in octavo on disjunct leaves both at the exterior edge and within the sheet and also on conjugate folds; also, we have a paper stock that has regular spacings between chain-lines and comparable spacings between the tranchefile and closest chain-line at both ends of the sheet. Because of the regularity of the chain-widths, when examining copies I often neglected to record chain-line positions across the entire leaf. Yet I believe the position of the closest chain-line to the spine will be shown useful in situating disjuncts and partial gatherings in editions without watermarks or tranchefiles and for which there are no uncut copies or copies with offset.

In common octavo imposition, one notes how the chain-line positions on leaves adjoining at their top edges ($1 and $4, $5 and $8, $6 and $7, and $2


Page 253
and $3) are very similar relative to the spine fold (or to the outer edge). Accordingly, with octavos (as with folios) the distance of chain-lines out from the spine can be used to argue that unmarked leaves were contiguous at their top edge or along the inner or outer edge before the sheet was cut apart. Note that this also allows us to situate on an octavo sheet leaves that were never part of a conjugate fold but were adjacent at their outer edges at the center of the sheet's long axis where the point holes are located (where I am hypothesizing N2 and R2 were adjoined). Although chain-line position would be more useful if the “NH” paper-stock had chain-widths widely varying, nonetheless it is worth considering. The evidence will be more significant for the second leaf in from the exterior of the sheet, since, the further in, the more the variations of chain-widths will accrue. The value of chain-line position is stronger for matching two folds (as B2.B7+1 with B6.B7) than for matching singletons glued to stubs (as A1 with U1)—and not just because twice as many chain-lines are involved. The distance out for the first chain-line on two singletons will vary with the fashion in which the two leaves are attached, whether folded to form a stub that can be sewn or glued on to an old stub or adjacent leaf. Also, one must expect misinformation and anomalies to creep in even if one takes measurements with the same strategy and keeps careful records.[8] Finally, it would be best to record the position of all chain-lines on the leaves so that irregular or varying chain- widths can aid identification.

As Appendix 1 with chain-line positions within 42 copies indicates, the distances from the spine of the closest chain-line to it in leaves A1 and U1 of the same copy differ on average 1.88 mm (see the third last measurements in the tables of Appendix 1; these differences total 79 mm for the 42 copies). In all but three copies the distance of the chain-lines from the spine on leaves A1 and U1 is within 4 mm of each other (in May the first chain-line out from the spine on A1 is 12-13 mm away, but on U1 it is 9; in Occ, A1: 11-12 vs. U1: 5-6; and in OCU, A1: 10-11 vs. U1: 17 [only OCU is rebound]). The first chain-lines from the spine on B3 and Dd1 at the top edge differ on aver-


Page 254
age 1.77 mm in their distances to the spine (note the second last measurements in the tables; these differences total 74.5 mm in the 42 copies). In all but six copies the distance of the chain-lines from the spine on B3 and Dd1 is within 4 mm of each other (in May, B3: 11 vs. Dd1: 18; in MdBJ, B3: 6-7 vs. Dd1: 11; in OAU, B3: 11-12 vs. Dd1: 16; in Occ, B3: 20 vs. Dd1: 14; in OCU, B3: 12-13 vs. Dd1: 6-7; and in PSt, B3: 15 vs. Dd1: 10 [only OCU is rebound]). It is noteworthy that the four copies with greatly differing chain-line positions on A1 and U1 are among these six. Although these anomalies may result from some copies' having A1 and U1, or B3 and Dd1, from different sheets, they are more likely caused by vagaries in the size of the leaf as cut, the depth glued inward, and the slanting or bending of the chain-lines. For instance, in my own copy of the edition, when I measure the distance of first chain-lines from the spine to the outer edges on leaves B3 and Dd1, I find the distance to be, respectively, 114 and 106 mm (of a copy trimmed 125 mm wide for B3 and 124 for Dd1), indicating that there is 11 mm to the spine on B3 and 18 on Dd1. However, this apparent disparity is actually caused by the insertion of B3 with a stub wrapping around and appearing 7 mm out from the fold after leaf B5; leaf Dd1, on the other hand, is glued to leaf Cc4. If B3 were glued in even with the fold, and the leaf trimmed accordingly, the first chain-lines out from the spine on B3 and Dd1 would be 107 and 106 from the outer edge, and the distances from the spine to the closest chain-lines would be 17 for both. Some of the other copies with chain-line positions breaking the pattern have irregularities in their binding, such as the rebound OCU copy, where A1 is attached to a conservation stub, and perhaps these copies should be excluded from consideration. Also, one might turn to the distances between chain-line across the leaf when disparities occur in the position of the first chain- line. Thus, in the PSt copy where B3 and Dd1 are 5 mm apart, one might find evidence that these leaves were adjoined at the top edge by noting that chain-lines appear out from the spine at regular distances on both: B3: 15, 42, 68, 95, and 120 mm; Dd1: 10, 38, 63-64, 90, and 115 mm.

In addition, if, as hypothesized here, leaves A1.Dd1 and U1.B3 were conjugate pairs prior to the separation of a half-sheet, then the distances from the spine to the closest chain-lines in each pair should add up to roughly 25.5-27 mm, the general range of the spaces between chain-lines. In fact, we find that the first chain-lines out from the spine usually total approximately 26 mm (see respectively fifth and fourth last measurements in the tables). Since the chain-widths can vary a millimeter from 26 mm and since my own measurements are subject to one mm of error for each measurement, we might consider any total within 4 mm of 26 an instance arguing the singletons were once conjugate folds. All the totals for the first chain-lines from the fold on A1 and Dd1 are within 4 of 26 mm except the May copy by.5 mm (21.5 mm) and the rebound OCU copy (17). The totals of the first chain- lines out on B3 and U1 are within 4 mm of 26 except for MdBJ (17.5) and, off by 1 mm, rebacked OTH (20). (With B3's stub factored in, the May is no longer


Page 255
an exception.) Totals a little short of the usual chain-width are easily explained" by trimming or folding the disjunct leaf before gluing or sewing it into place. If the chain-lines in this paper-stock were not as straight and evenly spaced as they are, we might argue from chain-line position alone that leaf Dd1 was attached to A1 and B3 was attached to U1. Because the chain-lines in this paper-stock are so regularly close to 26 mm apart, this argument that A1 was attached to Dd1 and U1 was attached to B3 is all the stronger—perhaps strong enough on its own to situate them on a half-sheet. Of course, in determining the imposition pattern of The Centaur we only look for confirmation from chain-line position for we have considerable watermark evidence as well, such as from four copies with watermarks on Dd1 and B3 that the Dd1 leaf was conjugate to A1 and B3 was to U1. But the evidence from watermarks and tranchefiles, and from the Csj copy, corroborates the evidence of chain-line position and attests its potential value.

There is substantial watermark evidence that leaves B2 and B6 were adjoining at the top edge. First, there is the converse evidence that in none of 58 copies examined does the same portion of the watermark appear on both B2 and B6. More importantly, complementary halves of the watermark appear on B2 and B6 in eleven copies: CaAEU, CaOTU, CoU, CSmH, MdBJ, NjP, NRU, O, Occ, Owo, and TnU. Additional support for supposing B2 and B6 were printed adjoined at their top edges comes from the matching position of chain-lines at the top of these leaves and of their conjugates, B7 and B7+1. In 41 of 42 copies measured for Appendix 1, the measurements of the first chain-lines out from the spine on B2 and B6 (as on B7 and B7+1) fall within 2 mm of each other (they are 3 mm off in CoU); the average distance is under 1 mm (41 mm total difference divided by 42 copies).

Although the uneven outer edges of leaves N2 and R2 in the Csj copy indicate these edges were formerly conjugate in that copy, let us still test other paper evidence on the placement of these folds. Since the tranchefiles in the “NH” stock are comparable in width at both sides of the sheet, those on N1 and R1 match in the Csj copy and 13 other copies (C, CaOTU, CtY, FU, MR1, NBiU, NcU, NIC, NRU, O, Oa, TnU, and Wns). As with the tranchefile evidence, some watermark evidence could mislead us into hypothesizing that the two folds were adjoining at their top edges. For instance, not a single copy has the same portion of the watermark on N1 and R1, which would rule out the hypothesis that they were imposed with top edges adjoining. However, the absence of any watermarks shared by leaves N1 and R1 supports paper-tear and dimension evidence that N1.N2 and R1.R2 were adjoined, not at their upper edges, but at the outer edges of N2 and R2. The absence of any matches is striking. Since one-fifth of the copies seen show watermarks shared by leaves A1 and U1 (11 copies), by leaves B2 and B6 (11), and by leaves Cc1 and Cc2 (10), we might expect to see at least one copy with watermarks shared by N1.N2 and R1.R2 if those folds had been adjoining at the top. We should particularly expect to see watermarks on both N1 and R1 in large copies like that at O, where 7-8 mm of the “NH” watermark often


Page 256
appears, divided between two leaves sharing a watermark; yet, at O, R1 has only 4 mm of watermark, but none appears on N1 (nor can we suppose the watermark is lost in the fold on N1 since their chain-lines are only a few mm off from matching at the top).

In addition, since, whenever either N1 or R1 has a watermark, the other does not, and since more than half the copies have a watermark on either leaf N1 or R1, we can infer that the two leaves were probably printed at opposite ends of the same sheet. In the 58 copies examined, I recorded watermarks on N1 in 27 or 28 copies (CaOHM, CoU, ICN1, L, LdU-B, LU, May, MH2, MoU, MR1, MR3, N, NcU, NIC, NjP, NRU, NSbSU, Oa, Ose, OTH, Owo, PSt, RPB, TnU, ViWCF, Wns, WU; possibly CaOTU) and on R1 in 15 or 16 copies (CLU-C, Csj, CSmH, CSt, DLC, The Hague, MdBJ, MH1, MiU, NBiU, O, OAU, Occ, O-HF, TxU and possibly MB). I still found 42 to 44 of 58 copies examined to have the watermark on either N1 or R1 but never on both. Watermarks, thus, appear on N1 or R1 in a high percentage of the total copies seen: roughly 75%. That is a very high percentage when one considers that I probably failed to identify one or two and that watermarks were commonly trimmed off. There are copies cut short enough that no watermark appears in the sheet with the singletons and Cc4, such as CSt (194), CtY (195), NeU (194), and OCU (194)—indeed, no watermark appears in any cancellantia or partial gathering of CtY. Since three quarters of the copies seen have a watermark in N1 or R1, it is likely that N1 and R1 were at the two opposite ends of one sheet, with one receiving the watermark when the other did not. That may be inferred from how much greater the total 42-44 is than the total number of copies with watermarks in the singletons half-sheet (27 copies, possibly 28), the half-sheet formed by folds B2.B7+1 and B6.B7 (26 copies), or the half-sheet Cc4 (24 copies).[9] Indeed, the total number of copies with watermarks on leaves N1 or R1 is comparable to the total with watermarks on the sheet comprising the singletons and Cc1-Cc4, 51 (possibly 52).

Since N1.N2 and R1.R2 were not positioned head to head but side by side, N1 and R1 on different ends of the sheet, chain-line positions in those leaves provide a good check against the value of chain-line position as used above for evidence that two cancellans leaves or folds were formerly adjoined at the top edge. As the tables indicate, there is less evidence from chain-line positions that N1.N2 and R1.R2 were adjoined at the top than there was


Page 257
for the singletons and the B cancellantia. There is greater than 4 mm difference in the positions of the first chain-lines from the fold on N1 and R1 in 9 of 42 copies: InU (7-8 mm), May (5), MiU (11-12), MR1 (4-5), NSbSU (6-7), OAU (4-5), Ose (7), PSt (5), and WU (5), and the difference is 4 mm in three others (CtY, LdU-B, and MH2). Only three copies had 4 mm of disparity between chain-line positions on AI and U1, and none did for B2 and B6. On average the positions of the first chain-lines from the fold on N1 and R1 are 2.54 mm apart (107 mm total for 42 copies); whereas, the averages for disjuncts A1 and U1 and for disjuncts B3 and Dd1 are under 2 mm (1.88 and 1.77 respectively). More to the point, however, is comparison with the folds B2.B7+1 and B6.B7, since as conjugate folds they, like N1.N2 and R1.R2, would have been sewn in: the average difference between chain-line positions on B2 and B6 was under 1 mm. Still, the position of chain-lines on N1 and R1 has not offered valuable evidence that N1 and R1 were not adjoining at the top, but nothing dramatic could be expected since this paper-stock has regular chain-widths and also tranchefiles spaced within 1 mm of 13 mm at both ends of the sheet.

Besides trying to match portions of watermarks to determine what leaves were printed beside each other, we can also use the record of watermarks to establish that three sheets were required to print the 16 leaves with cancellantia and partial gatherings. In no copies does either the top or bottom half of the watermark appear more than three times in cancellantia and the partial gatherings. Thus, there is no watermark evidence that more than three sheets were needed to print these 16 leaves or that all leaves printed together on a sheet were not bound in the same copy. But that three sheets were needed to print these 16 leaves, as was shown by the presence on them of five leaves with tranchefiles within the TnU copy, is also evident from the appearance of three separate watermarks in the sixteen leaves within at least six copies: CSmH (complementary halves on B2 and B6, Cc1 and Cc2, and bottom half on R1), LdU-B (complementary halves on A1, and U1 and bottom on B6 and N1), MiU (complementary halves on Cc1 and Cc2 and bottom on B6 and R6), O (complementary halves on B2 and B6 and also the bottom on R1 and U1), Owo (complementary halves on B2 and B6 and on Cc1 and Cc2 and bottom on N1), and PSt (complementary halves on Cc1 and Cc2 and bottom halves on B6 and N1). Furthermore, if we accept the evidence for placing on the same sheet the singletons and Cc4, or for placing on the same sheet N1.N2 and R1.R2, then other copies evince the need for three sheets. For instance, if either reconstructed unit is presupposed, three printed units are indicated by CoU, for it has the bottom of the watermark on AI and B6 and top on B2 and N1, and three printed units are indicated by L, for it has the bottom half on B6 and N1 and the top on Cc1.

Assuming that the sixteen leaves derive from three sheets and that partials cut from the same sheet were distributed to the same copy, the distribution of watermarks in these copies indicate what units were not printed with what other units (and by implication what units were printed with what


Page 258
others). Thus, watermarks described above for the CoU copy (bottom halves on A1 and B6 and tops on B2 and N1) indicate that the B cancellantia was printed on separate sheets from A1 and N1, and those described above for the O copy (complementary halves on B2 and B6 and also bottom halves on R1 and U1) indicate the singleton half-sheet, the B cancellantia half-sheet, and R1.R2 were printed on separate sheets. Watermark patterns rule out the possibility that the singletons could have been printed with any leaves other than Cc4 (not the half-sheet with B2 nor N1.N2, for in LdU-B leaves A1 and U1 as well as B2 and B6 have complementary halves and N1 has a watermark, and not R1.R2, for in Csj U1 and R1 have bottom halves). With evidence drawn from multiple copies, the same logic can show that the half-sheet with B2.B7+1 and B6.B7 could not have been printed with the singletons, N1.N2, R1.R2, or Cc4; or that N1.N2 could have been printed with the singletons, the B cancels, or Cc4; and so forth. Finally, the general watermark patterns show that leaves A1 and U1 were printed with Cc4. When watermarks appear on leaves A1 or U1, they also—in at least some copies—appear on B2, B6, N1, and R1, but none of 27 (possibly 28) copies with watermarks on A1 and/or U1 also has them in Cc4. Furthermore, none of the 24 copies known to have a watermark in Cc4 has a watermark in A1 or U1. Yet copies with a watermark in Cc4 fairly often also have a watermark in the canceled folds in B (as C CaOTU CSmH L MdBJ MiU NIC NjP NRU Occ OCU Ose Owo PSt TnU), in N1 (as ICN1 L May MH2 MiU MoU NIC NjP NRU Ose OTH Owo PSt TnU and possibly CaOTU), and in R1 (as CSmH, MdBJ, MH1 MiU Occ). These watermark patterns indicate that A1 and U1, with their likely conjugates B3 and Dd1, formed a whole sheet with Cc4. This seems all the more certain because leaf Dd1 continues the postscript text on Cc4v.

We know that B2.B7+1 and B6.B7 formed a half-sheet printed separately from the other cancels and partial gatherings in the edition. We might now ask whether this half-sheet of B cancellantia was imposed for work-and-turn and printed by itself or whether it occupied a sheet with four leaves of other text. Half-sheet imposition might have been preferable if the rest of the book had been printed, if cancellantia N1.N2 and R1.R2 were not yet composed, or if Richardson wished the B cancellantia solely to occupy a chase to facilitate Young's revision and correction. Then too half-sheets offered the simplicity of printing separately Young's text on the paper-stock billed to him. (Were a paper-stock different from the dominant “NH” stock found in the cancellans folds of B, we might suppose they were printed with other jobs.) Since only half the sheet is watermarked and roughly half the copies have a watermark in the B cancellantia half-sheet (26), we are unable to conclude whether it was produced through work-and-turn or with other material in whole sheets. If either considerably more than or fewer than half of the copies had watermarks on B2 or B6, then only full-sheet imposition could account for that pattern. We could suppose that half of the forme with the B cancellantia was regularly draped with the watermarked or the non-


Page 259
watermarked half of the sheet. It is certain that all the B cancellans folds for all copies were not printed off with the watermark end of the sheet, for uncut Csj and a number of large copies (FU, MR1, MR3, and TxU all 200 mm tall, MH2 at 201 mm, and O at 203) lack watermarks on B2 and B6. We can assume that the roughly one-half figure is representative and that it was never considerably above fifty-percent, for the singletons and Cc half-sheets yielded the comparable figures of 25 (or 26) and 24 respectively.

In conclusion, the 16 leaves of cancellantia and partial gatherings to be accounted for were printed on three units: the singletons and Cc4 on one whole sheet, folds N1.N2 and R1.R2 on another whole sheet (with unidentified material), and the cancellans folds B2.B7+1 and B6.B7 separately through either whole-sheet or half-sheet imposition. The pattern of watermark distributions supports this conclusion and argues that almost always the leaves printed together went into the same copy. We know that the cancellation of material on leaves R1-R2 occurred at the same time as that of N1-N2, days after Young's letter of 26 January 1755. Also after 26 January, leaf U1 was canceled, for cancellans U1 was printed with cancellans B3, and Young returned with approval the proofs for the revised B gathering only on 26 January, responding to Richardson's proposed changes penned 21 January 1755. The whole sheet with B3 and the half- or whole-sheet with the other B cancellantia must have been printed at nearly the same time since the texts on B3 had to begin and end in sequence with that on B2v. Perhaps the printing of cancels N1.N2 was further delayed by changes, unrecorded in the correspondence, that appear on cancels R1-R2.

APPENDIX 1 Chain-Line Distances from the Spine

These tables provide, for 42 copies (cited with abbreviations glossed in Appendix 2) the distance of the first chain-line from the spine at the top of disjunct and canceled leaves. Leaves in gathering B are numbered as they appear; when the first chain-line is too close to the fold to allow measurement, 26 or 27 has been subtracted from the second chain-line to place the first one.

CaOHM  CaOTU  CoU  Csj  CtY  DLC 
A 10  12-13  12  13  16  13 
B2  8-9  7-8  14  10-11  7-8 
B3  17  15  16  14-15  10  6-7  12 
B6  8-9  10  11  11 
B7  16  17  19-20  15  15  12  13-14 
B7+1  16  17-18  18  11  15  12  13-14 
N1  6-7  5-6  5-6  4-5 
R1  5-6  5-6 
U1  7-8  10  14  16  13 
Dd1  17  13  19  16  13  6-7  12-13 


Page 260

A1+Dd1  27  25-26  27  28  26  22-23  25-26 
B3+U1  24-25  24  26  23-24  24  22-23  25 
A1 vs U1  2-3  3-4 
B3 vs Dd1  1-2  0-1 
N1 vs R1  0-1  2-3 

FU  InU  KU  LdU-B  LU  May 
A 10-11  10  10  13-14  3-4 
B2  15-16  11  12-13 
B3  14  13  14  17  23  15  11 
B6  16  10-11  10  14-15 
B7  14-15  10-11  17  11  17  17  16 
B7+1  14-15  13-14  17  11  17  17  15 
N1  4-5  5-6  10 
R1  6-7  24 
U1  9-10  10  11-12  7-8  15 
Dd1  12  13  17  16-17  21-22  13-14  18 
A1+Dd1  22-23  23  27  25-26  26-27  26-28  21-22 
B3+U1  23-24  23  25-26  24-25  28  30  20 
A1 vs U1  1-2  1-2  1-2  5-6 
B3 vs Dd1  0-1  1-2  1-2 
N1 vs R1  7-8 

MdBJ  MH1  MH2  MiU  MoU  MR1  MR2 
A 15  10-11  15  10-11  11  5-6  14 
B2  6-7  14  16  13-14  17 
B3  6-7  13  11-12  18  14  15  14 
B6  9-10  8-9  12  17  15  15 
B7  16  19  12  13  15  20  11 
B7+1  17  19  12  12-13  15  18  12 
N1  6-7  7-8  8-9 
R1  3-4  20  2-3 
U1  18  13-14  13-14  11  13-14  13 
Dd1  11  15  12  18  15  17-18  13 
A1+Dd1  26  25-26  27  28-29  26  22-24  27 
B3+U1  24-25  26-27  24-26  29  27-28  22  27 
A1 vs U1  1-2  0-1  2-3  1-2 
B3 vs Dd1  4-5  0-1  2-3 
N1 vs R1  0-1  11-12  4-5 

MR3  NBiU  NcU  NeU  NIC  NjP  NRU 
A 12  8-9  10  11  10 
B2  12  14  5-6  13  13  10 
B3  17  13  16-17  14  16  14 
B6  10  14  5-6  13  13  5-6  10 
B7  17  11-12  16-17  12-13  13  19  16-17 
B7+1  15-16  10-11  16-17  12-13  13  18  15 
N1  5-6  2-3  4-5  5-6 
R1  3-4  4-5 
U1  11  12-13  8-9  12-13  10-11  11-12 
Dd1  17  15  18  14-15  13-14  16  13 
A1+Dd1  29  23-24  26  24-25  24-25  25  23 
B3+U1  28  25-26  24-26  26-27  26-27  23  26-27 
A1 vs U1  0-1  2-3  .5  1-2 


Page 261

B3 vs Dd1  1-2  0-1  2-3 
N1 vs R1  0-1  1-2  1-2  0-1 

NSbSU  Oa  OAU  Occ  OCU  Ose 
A 10  11  13  11-12  10-11  11 
B2  8-9  12-13  11 
B3  16.5  15  17-18  11-12  20  12-13  15 
B6  7-8  11  6-7  13  12 
B7  18-19  15-16  16  16-17  13  22  12 
B7+1  18-19  15  16  16-17  12  22  12 
N1  8.5  1-2  2-3  10-11 
R1  3-4  3-4  5-6  3-4 
U1  13  10  11  11  5-6  17  8-9 
Dd1  16  16  16  16  14  6-7  16-17 
A1+Dd1  25  26  27  29  25-26  16-18  27-28 
B3+U1  29-30  25  28-29  22-23  25-26  29-30  23-24 
A1 vs U1  6-7  2-3 
B3 vs Dd1  0.5  1-2  4-5  1-2 
N1 vs R1  6-7  2-3  4-5  2-3 

OTH  Owo  PSt  TnU  TxU  Wns  WU 
A 12-13  15-16  12-13  12  6-7 
B2  13  12-13  8-9  11 
B3  11-12  10  15  19  17  18-19  21-22 
B6  8-9  12  13-14  11  11 
B7  20-21  11  11-12  20  13  16  22 
B7+1  18-19  10  12-13  20  13  15  22 
N1  8-9  7-8  6-7 
R1  3-4  5-6 
U1  8-9  15-16  8-9  11 
Dd1  12  10  10  19  17  15  21-22 
A1+Dd1  24-25  25-26  22-23  28  29  23  27-29 
B3+U1  19-21  25-26  23-24  27  28  27-28  29-30 
A1 vs U1  1-2 
B3 vs Dd1  0-1  3-4 
N1 vs R1  0-1 

APPENDIX 2 Copies Examined

The frontispiece is present unless noted otherwise.[10]


Page 262

C (Cambridge) N.17.25; 199 x 128; rebound

CaAEU (U. of Alberta) PR3782.C39 1755; 194 x 125; cont[emporary] calf

CaOHM (McMaster U.) B-623; 194 x 117-118; cont. calf; rebacked; lacking section title to Letter I

CaOTU (U. of Toronto) B-11 5583; 199 x 124; section title to Letter I before B1; cont. calf

CLU-C (Clark Library) PR3782.C31; 202 x 124; section title to Letter I before B1; cont. calf

CoU (U. of Colorado, Boulder) Pettit 72; 205 x 128; cont. calf

Csj (St. John's College) P4.9.119; uncut, 208-216 [frontispiece, 219] x 129-134; cont. calf spine with marbled paper boards

CSmH (Huntington Library) 351165; 204 x 127; nineteenth-century half calf with marbled paper boards

Cst (Stanford U.) PR3782.C4; 194 x 119; leaves N1-N2 bound after N8

CtY (Yale U.) formerly Ik Y85 755, recatalogued as 1997/1114; 195 x 124-125; cont. calf; rebacked; some leaves on conservation stubs

DLC (Library of Congress) PR3782.C4 1755; 195 x 122-123; cont. calf; rebacked

FU (U. of Florida) 824.5 Y71c; 200 x 122; cont. calf

The Hague (Koninklijke Bibliotheek) 234 m 22; 197 x 120; cont. calf remounted on new boards

ICN (Newberry Library) 2: 1) C.69.99; 194 x 122; cont. calf; 2) Y135.M6387; 195 x 124; cont. calf; with spine label

InU (Indiana U.) PR3782.C3 1755; 197 x 121; lacking section title to Letter I; cont. calf; rebacked

IU (U. of Illinois) Nickell 318; 197 x 122; cont. calf

KU (U. of Kansas) C907; 194-195 x 124; lacking frt; cont. calf; rebacked

L (British Library) 525.i.3; 200 x 124; rebound

LdU-B (Brotherton Collection, U. of Leeds) Lt q Young; 198 x 122; cont. calf

LU (Sterling Library, U. of London) Sterling I [Young, E. 1755]; 194 x 123; cont. calf; rebacked

May (Author's copy) 195-197 x 124-125; lacking frt.; cont. calf

MB (Boston Public) Deofe 30.755.YoBC; 195 x 124; section title to Letter I before B1; rebound

MdBJ (Johns Hopkins U.) PO3782.C39 1755 [sic]; 199 x 118; cont. calf

MH (Harvard U.) 2: 1) Typ 705.55.892; 195 x 123-124; cont. calf; rebacked; 2) *EC7 Y8460.755c; 201 x 125; section title to Letter I before B1; cont. calf; with arms of England stamped on front and back boards; rebacked

MiU (U. of Michigan) PR3782.C4; 196.5 x 123; cont. calf

MnU (U. of Minnesota) 824.y8 OCe; 187 x 122; rebound

MoU (U. of Missouri) 828.Y8ce; 190 x 116; rebound

MR (John Rylands U. Library, Manchester) 3: 1) 1110; 200 x 127; cont. calf; 2) R144667; 194 x 120; cont. calf; 3) MAW CW 74; 200 x 125; washed, resized, and resewn in 1986; preserved endpaper signed “Mel[issa?] Hotham June ye 20th 1755”; in another hand, signed “Wesley Jan. 29. 1756 The Gift of Miss Melly Hotham”; title-page signed “CWesley 1756”

N (New York State Library) N248.4 Y71; 196 x 125; lacking B6, replaced with cont. manuscript copy of the leaf; rebound

NBiU (SUNY at Binghamton) PR3782.C4 1755; 198 x 125; cont. calf

NcU (U. of North Carolina) PR3782.C4 1755; 200 x 121; cont. calf

NeU (U. of Newcastle); 821.55 You; 194 x 120; cont. calf

NIC (Cornell U.) PR3782.C3; 196 x 127; rebound

NjP (Princeton U.) Ex3999.7.323.12; 197 x 121-122; cont. calf; restitched

NNU (New York U.) PR3782.C4 1755; 198 x 120; rebound


Page 263

NRU (U. of Rochester) xPR3782.C39m; 194 x 126; cont. calf

NSbSU (SUNY at Stony Brook; PR3782.C4 1755; 198.5 x 123; rebound

O (Bodleian Library) 141 j.106; 203 x 125-126; section title to Letter I before B1; cont. calf

Oa (All Soul's College) qq.13.27; 196 x 121-122; cont. calf

OAU (Ashland U., Ohio) EL 159; section title to Letter I before B1; cont. calf

Occ (Corpus Christi College) 1704.Yo; 200 x 121-122; cont. calf

OCiU (U. of Cincinnati) PR3782.C4; 191 x 121; cont. calf

OCU (Ohio State U.) PR3782.C4 1755; 195 x 122; rebound

O-HF (Harold Forster Collection, Bodleian Library) Forster 51; 198 x 125; cont. calf

Ose (St. Edmund Hall) 204-205 x 124-125; section title to Letter I before B1; cont. calf

OTH (Heidelberg College, Tiftin, Ohio) PR3782.C4 1755; 199 x 126; cont. calf; rebacked

Owo (Worcester College) B.B.2.6; 196-197 x 124; cont. calf; “MISCELLANY” on red spine label; bound with James Hervey's Remarks on Lord Bolinbroke's Letters on the Study and Use of History, 1752, and Peter Whalley's A Vindication of the Evidences and Authenticity of the Gospels from the Objections of the late Lord Bolingbroke, 1753

PSt (Penn State U.) PR3782.C4 1755; 202 x 123-124; cont. calf; section title of Letter I before B1

RPB (Brown U.) YQE Y8 Lamont; 196 x 121; cont. calf; rebacked

TnU (U. of Tennessee) PR3782.C4 1755; 201 x 127; section title of Letter I before B1; cont. calf

TxU (U. of Texas at Austin) AK Y858 755c; 200 x 124-125; section title to Letter I before B1; cont. calf; rebacked

ViWCF (Colonial Williamsburg Foundation) PR3782.C4 1755; 197 x 122; cont. calf

Wns (Winchester College) 199 x 125; cont. calf

WU (U. of Wisconsin, Madison) PR3782.C4; 197 x 122; cont. calf


The first edition octavo, indexed in the ESTC as T113256, has the full title The Centaur Not Fabulous in Five Letters to a Friend, on the Life in Vogue and was printed for Andrew Millar and Robert and James Dodsley. The edition was first announced published in the London Evening Post of 1-4 March 1755 (bound, 5s.). In the same year, with “Five” altered to “Six” in the title, Millar and the Dodsleys published a revised second edition, also an octavo (15-17 April, LEP), and a revised third edition, a duodecimo (17 November, Public Advertiser; 3s. bound). Additions to the text lengthened the second edition sufficiently that, although often a page-for-page reprinting, it is several leaves longer, collating 8°: (frt. +) A 2 B-Cc8. Peter Wilson and John Exshaw brought out a duodecimo edition in Dublin in 1755, called the “Third Edition,” evidently published before the London third edition. Young revised The Centaur once again for its reprinting in Volume 4 of his The Works of the Author of The Night Thoughts (1757). Millar and James Dodsley brought out a “fourth edition,” in duodecimo, in 1765. Thereafter, the work was not frequently reprinted except within collections of Young's works. Unlike the authorized editions of 1757, 1762, 1767, and 1774 where The Centaur is in Vol. 4 after Night 9 of The Complaint, unauthorized works editions usually place The Centaur in a volume intended to be sold separately. The Dublin publishers G. and A. Ewing and partners placed it alone in Vol. 4 of their four-volume Works (1764), which lacks any works title and begins with the title-page for The Centaur. The Centaur shares the fourth volume with three prose items not included in the authorized collections (Conjectures on Original Composition, The Vindication of Providence, and the “Essay on Lyric Poetry” from Imperium Pelagi) in the unauthorized Works in Prose of the Reverend Edward Young (London: P. Brown, H. Hill, and S. Payne, 1765), a separate issue of Vol. 4 of a four-volume works with this same false imprint. In 1770 these four prose pieces were reprinted as The Centaur Not Fabulous, and Other Prose Works of the Reverend Dr. Edward Young (Edinburgh: Martin and Wotherspoon, 1770), a separate issue of Vol. 4 of their four-volume works, also issued with a false imprint (London: S. Crowder, C. Ware, and T. Payne, 1770). These four prose works were reprinted in 1774 with a variant of the 1765 false imprint (London: P. Brown, H. Hill, and S. Bayne [sic]). A six-volume Works printed in 1774, issued with various imprints (one for J. Donaldson in London, and four others placed in Edinburgh: for Alex. Donaldson; for J. Dickson; for C. Elliott; and “by and for” Colin Macfarquhar), includes The Centaur in Volume V with two poems excluded from the authorized works. The first American edition of The Centaur, a duodecimo, was printed in Philadelphia by T. Stephens and W. W. Woodward in 1795. The Centaur was translated into German in 1755 ( Der nicht fabelhafte Centaur in sechs Briefen [Leipzig: J. G. Dyck, with a separate Frankfurt issue]) and into Dutch in 1768 (De Centaurus geen Verdichtsel [Amsterdam: P. Meijer]).


Henry Pettit, ed., The Correspondence of Edward Young 1683-1765 (Oxford: Clarendon Press 1971), p. 404; hereafter cited as “Pettit.” For a judicious, comprehensive account of The Centaur's composition and publication, see Harold Forster's Edward Young: The Poet of the Night Thoughts 1683-1765 (Alburgh, Harleston, Norfolk: Erskine Press, 1986), pp. 268 and 280-287.


Samuel Richardson: A Biography (Oxford: Clarendon Press, 1971), 432n.


That within the dedication the second edition shares the same essential page settings as the B cancellantia is evident from type alignment and damaged type throughout. Note, for instance, that the “g” cut at top in ldquo;Pegasean” (B6v/xii.3) and the “d” cut in “untamed” (B6v/xii.4) are also in the second edition on B5v, and that the “a” cut at top in “scandalized” (B7r/xiii.1) and the “a” cut in “Travels” (B7 + 1v/xvi.2) are also in the second edition on B6r and B7v. The first- edition page settings end up as different pages of the second edition, for in the latter the dedication is begun on leaf A2 conjugate to the title-page.


The following pages of the first edition of The Centaur have press figures: B7 + 1r/ xv-7, C8v/16-2, D8r/31-5, F1v/50-2, G1v/66-4, H2v/84-3, 16r/107-3, K6v/124-2, L5r/137-3, M8r/159-5, N3v/166-2, O2v/180-3, P1v/194-3, Q2v/212-1, R2r/227-2, R6v/236-3, S8r/255-1, T3v/262-4, T7r/269-2, X6v/300-1, Y7v/318-1, Z6r/331-1, Aa4v/344-2, Aa7v/350-7, Bb1v/354-7, Cc3r/373-1, and Cc3v/374-4. In all copies seen, press figures appear on both the inner and outer formes of only gatherings T and Cc. Most but not all copies have two figures in gatherings R (although one is on cancellans R2) and Aa (this variation was missed by Eaves and Kimpel [432n.] and by William Todd [“Observations on the Incidence and Interpretation of Press Figures,” Studies in Bibliography, 3 (1950-51), 193]). Six copies lack the figure on R6v (CSmH, FU, ICN1, MR3, NRU, and Oa), and only the top of the figure is visible in CaOTU, NcU, and OCU, suggesting it was accidentally lost. A third of the copies examined (18 of 58) lack the press figure on Aa7v/350 and so have only one figure in Aa: C, CoU, Csj, CtY, LU, May, MH1, N, NeU, NIC, NjP, NRU, NSbSU, OAU, O-HF, PSt, MdBJ and Wns. Only gatherings E and U have no press figures in all copies, and U may have had a figure on the verso of the cancellandum of U1.


The preliminary, disjunct, cancellans, and concluding leaves of the uncut copy at St. John's College, Cambridge, have the following features (all measurements are in mm): Frontispiece (219 x 126): chain-lines 76-77/3 [i.e., 76-77 for 3 chain-widths]; tranchefile 11; no watermark.

A1 [t-p.] (213-215 x 137): chain-lines 81/3; tranchefile 13-14; 8-9 of the top of “NH” watermark (completing that on U1); glued to B1.

B1. section title-page (212-215 tall): chain-lines, 101/4; no tranchefile or watermark; closest chain-line to the fold on B1: 17 (matching B4).

B2 (212-213 x 135): conjugate to B7+1; chain-lines, 79-80/3; tranchefile 12 at top; 14 at bottom; no watermark; ragged top edge matching that on B6.

B3 (212-213 x 132): chain-lines, 79-80/3; no tranchefile or watermark; uneven top edge matching that on Dd1; glued to leaf B2.

B4.B5 (215 x 270): chain-lines, 101-103/4; no tranchefile or watermark; closest chain-lines to the fold: 19 on B4; 28-29 on B5; point hole at 165 down B5.

B6 (212-215 x 134): conjugate to what appears to be B7; chain-lines, 80/3; tranchefile, 13- 14; no watermark; ragged top edge matching that on B2.

B7 and B7+1 (215 x 133; 212 x 131): chain-lines, 80/3; no tranchefile or watermark; point hole on B7+1 114 down outer edge.

N1.N2 (208-210 x 128, 131): chain-lines, 79/3; tranchefile on N1, 12-13; no watermark; point hole on N2 shared with R2 at 165-170 down outer edge; sewn in, with stubs behind N2.

R1.R2 (210-213 x 124, 134-135): chain-lines, 78/3; tranchefile, 13; 9 mm of the bottom of the “NH” watermark (same half as U1); point hole on R2 shared with N2 at 165-170 down outer edge; sewn in, with stubs behind R2.

U1 (212-213 x 137): chain-lines, 105-106/4; tranchefile, 12-13; 4.5 of bottom of “NH” watermark visible (same portion as on R1 and completing A1's half); glued to stub behind it.

Cc4 (213-214 x 133): chain-lines, 105/4; tranchefiles on Cc1 and Cc2, 12-13; no watermarks; parallel chains on Cc1, Cc2; point holes on Cc3 110 down and on Cc4 153 down on outer edge.

Dd1 (215 x 132): chain-lines, 108/4; no tranchefile or watermark; uneven top edge matching that on B3; glued to leaf Cc4.


A New Introduction to Bibliography (New York: Oxford Univ. Press, 1972), p. 128.


By measuring the distance out for chain-lines on known conjugates in regular gatherings, one first determines how much space is usually swallowed up in the fold (often this adds 1 mm to the apparent width). For instance, if one finds that the distances to the first chain-lines out from the fold on leaves known to be conjugate add up to 22 mm when the usual chain-width is 26 mm, then one need allow more for the fold. If easy measurement from the fold is not possible, one can measure in from the outside edge to the closest chain-line to the fold and subtract that from the total width. One must remember that the trimmed leaves of a normal book will often vary progressively from beginning to middle to end by one or two millimeters, that rulers vary slightly in their measurements, and, that if the leaves are glued in several millimeters from the fold and the measurements begin just behind the fold, measurements could easily be off by 3-4 mm. For a good discussion of measuring chain-lines and other identifiable features of paper and a system for identifying paper moulds by these features, see David L. Vander Meulen's “The Identification of Paper without Watermarks: The Example of Pope's Dunciad,Studies in Bibliography, 37 (1984), 58-81. Vander Meulen's proposal that one reconstruct the chain-line model for the mould offers a surer but more time-consuming method than my reliance on the position of the first chain-line from the fold.


Watermarks are only seen in the leaf A1 (not U1 also) in 9 copies: CaAEU, CoU, KU, MR3, NcU, Oa, O-HF, PRB, and ViWCF; and only seen in U1 in 7 or 8: CLU-C, ICN2, InU, N, O, Wns, WU, and possibly The Hague (where I think I identified one, but Librarian Paul van Capelleveen did not); and in the 11 copies wherein leaves A1 and U1 share a watermark noted above in the text, for a total of 27 (possibly 28) of 58 copies seen. A watermark is shared by leaves B2 and B6 in 11 copies listed in the text and appears in either B2 or B6 in 15 others: C, CLU-C, L, LdU-B, MB, MiU, MnU, NeU, NIC, Oa, OCU, Ose, PSt, RPB, and WU, for a total of 26 of 58 copies seen. A watermark is shared by leaves Cc1 and Cc2 in 10 copies listed in the text and appears in either Cc1 or Cc2 in 14 copies: C, CaOHM, ICN1, IU, L, May, MdBJ, MoU, NIC, NjP, NNU, NRU, OCiU, and TnU, for a total of 24 copies seen.


For checking details in their libraries' copies, I thank the following: Paul van Capelleveen of the Koninklijke Bibliotheek, James Corsaro of the New York State Library, Geoffrey Day of Winchester College, Gail Greve of the Colonial Williamsburg Foundation's John D. Rockefeller, Jr., Library, Jeannine Green of the U. of Alberta, Jacob Hand of the U. of Cincinnati, Jonathan Harrison and Elizabeth Quarmby Lawrence of St. John's College, Cambridge, Tim Johnson of the U. of Minnesota, Mike Kelly of New York U., Margaret Kulis of the Newberry Library, James Lloyd of the U. of Tennessee, Kris McCusker of the U. of Colorado at Boulder, John Mustain of Stanford U., Richard Noble of Brown U., D. W. Riley of John Rylands U. Library of Manchester, Christopher Sheppard of the Brotherton Collection, U. of Leeds, Bruce W. Swann of U. of Illinois, and Stephen Tabor of both the Clark Library and the Huntington Library. I also acknowledge my son Erich May's help in examining copies and editor David Vander Meulen's many corrections to and suggestions for the essay.

Un-numbered 264

Page Un-numbered 264