University of Virginia Library


Page 197


by James E. May *

This paper examines offset (sometimes called "set off") principally on the sixteen leaves that were cancelled or printed in partial gatherings for the first edition of Edward Young’s prose satire and homily The Centaur Not Fabulous (1755), printed by Samuel Richardson. Besides the frontispiece, the edition has the following partial gatherings and cancellantia: a confusing nine-leaf B gathering; two cancellans folds in the text, N1.N2 and R1.R2 (with parts of Letters III and V); half-sheet Cc4; and four singletons: title-leaf A1, final leaf Dd1, and cancellantia B3 (third leaf of the preface) and U1. In another article I have demonstrated where those leaves were imposed in relation to one another on three printed units.1 I placed leaves A1, B3, U1, and Ddi on a sheet shared by Cc4 (with A1 and U1 adjoining at their heads and with their fore-edges lying at the end of the sheet, Dd1 conjugate to A1, and B3 conjugate to U1). I identified cancellantia folds B2.B7 + 1 and B6.B7 as printed as a half-sheet either through work-and-turn or with four other leaves to form a whole sheet (with B2 and B6 adjoining at their heads), and cancellantia N1.N2 and R1.R2 as running laterally across a separate whole sheet as N1.N2.R2.R1. The work thus has the following collation 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, including ‘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)], 1 2-53 54-56 57-221 222-224 225-378. [Note. The reconstruction creates a Cc8 sheet containing leaves A1, B3, U1, and Dd1 positioned as follows: A1 =Cc3, B3 = Cc7,


Page 198
U1 =Cc2, and Dd1 =Cc6, with the leaf signed ‘Cc2’ occurring in the position of Cc4.]

After a disjunct frontispiece (printed separately) and a title leaf comes B1, conjugate to a section title that had been B8 prior to the cancellations and insertions in gathering B; then the cancel B2 conjugate to B7 + 1; then a disjunct B3 printed with A1, U1, and the final leaf Dd1; then the fold B4.B5 from the original B gathering; then a cancellans fold B6.B7; and then B7 + 1 and the section title B8. In addition, I showed that all leaves printed on the same sheet or half-sheet regularly ended up in the same copy. I supported this collation formula with evidence from press figures, watermarks, tranchefiles, and chain-lines in dozens of trimmed copies and confirmed it with additional evidence from the point holes and leaf dimensions of the untrimmed copy at St. John’s College, Cambridge.

In this article, while confirming with new evidence what I formerly argued about the cancellantia and partials of The Centaur, I will demonstrate the utility of offset evidence in reconstructing the pattern of imposition. For instance, the offset evidence will corroborate that all four singletons were printed at the same time within a sheet with Cc4. Offset from the cancellanda will also provide information about Young’s revisions, offering some variant textual readings. In addition, the analysis will offer insights into the printing, folding, and assembling of sheets of this edition in Richardson’s shop. It will, for example, indicate the extent to which these activities were handled consistently while also revealing interesting exceptions, as when the leaves for diverse gatherings were folded together before they were cut apart. So far as these insights involve only a single edition from one press the account is anecdotal, but the results are worth reporting because offset has not often been discussed in bibliographical literature. Finally, I make observations and raise questions about the nature, or behavior, of offset ink impressions. Only a handful of arguments have been offered from offset evidence.2 I have seen no study involving as many copies or cancels as this conducted for The Centaur, nor any study attending to the diverse folding of sheets, half-sheets, and quarter-sheets with partials and cancellantia. The Centaur is a good subject for studying offset both because it is a common book well printed by a good and orderly printing shop and because much is known from correspondence between Richardson and Young about the order in which sheets went through the press.

One is sometimes amazed, even irritated, by the backwards letter-press on the pages of eighteenth-century books, especially in photocopies. Some of that, particularly noticeable on thin paper, is "bleed-through," the seeping of ink from the other side of the leaf. In The Centaur, a well-printed book not prone to offset, one sees more bleed-through than offset, not only within and between regular sheets but in most pages of the partials and cancellantia. But some of the extrane-


Page 199
ous ink nonetheless is "offset," set off from the facing page, much the way a wet newspaper can leave a reverse text on a countertop. In particular, the backwards inking in a margin of a page will often be offset rather than bleed-through. The two situations can often be distinguished by color: offset, with this and other Richardson editions at least, has a yellowish-brown tint (e.g., B7 + 1r on [>] B2v of the ViWCF copy; see Appendix I for library codes). That tint often seems the stronger when the offset is from facing pages (e.g., exchanges between B3v/B4r and B5v/B6r in NcU and B5v/B6r in ViWCF).

Offset ink was transferred when sheets were folded, collated, stacked, compressed, and pounded at the printer’s and the binder’s shops.3 Of course, even after a book has been bound, offset often will be transferred to facing pages. This timing is especially evident when facing text appears on blank endpapers (e.g., offset from the final verso of The Centaur on the first back endpaper in CtY, LU, Mayi, May3, NNUi, PSt, ViWCF) or on inserted papers, especially when the leaves have been moistened. The capacity for offset transfer after binding is shown by the O-HF copy where, after the top outside corners of B5 and B6 were dog-eared, B5r received offset from B7r, and B6v received offset from itself. The cancellantia and partial gatherings in thirty-six of the sixty-nine copies of The Centaur that I have examined do not have this additional inking (other than from the frontispiece). But thirty-three copies, nearly half, do have a page or more of offset from pages that do not face them in the bound book.4

Four Singletons and Half-Sheet Cc 4

Most of the offset impressed onto the singletons A1, U1, B3, and Dd1 is from another singleton or a leaf of half-sheet Cc4. Furthermore, all the evidence offered by these copies is consistent with the singletons’ having been printed with Cc4 in the following arrangement of page settings:

Outer forme

A1r  Dd1v  Cc3r  Cc2v 
U1v  B3r  Cc4v  Cc1r 

Inner forme

Cc2r  Cc3v  Dd1r  A1v 
Cc1v  Cc4r  B3v  U1r 

The folds Cc1.Cc4 and Cc2.Cc3 occupy the positions normally occupied in an octavo sheet by 1.8 and 4.5, ensuring that, when detached, two folds left the


Page 200
half-sheet ready to be bound. The arrangement of the singletons’ half-sheet is not entirely arbitrary, for it ensured that A1r could be folded face up on the top with Dd1 on the bottom or that A1r could be protected within the fold under leaf U1, that all four rectos would be normally placed as rectos on the folded halfsheet, and, that, as in other gatherings, the first leaves of A and U often held the watermark and a tranchefile. As does other evidence, the offset inking patterns indicate that the singletons were printed in this arrangement and that they shared as described a full sheet with half-sheet Cc4. Those who wish to understand and to test the following observations and inferences are encouraged to take sheets of 8½ X 11" paper, label the leaves as indicated above, and fold them as separated half-sheets or whole sheets as indicated above and below.

To arrive at these conclusions, let us first establish that the singletons were on a half-sheet in the pattern described. To judge from the evident offset in copies examined, the half-sheet with singletons was sometimes separated and sometimes not separated from half-sheet Cc prior to folding, but, most often, as will be discussed below, the halves were separated first. This cut would save the labor of unfolding the sheet to separate the halves later.

A separated half-sheet’s first fold could be in four directions (either side could be folded forwards or backwards along its long or short axis). The most common fold for the half-sheet with singletons, usually done after the separation into half-sheets, was to bend the outer forme (that with A1r) inward along the short axis, reducing by half the longer dimension of the half-sheet. This fold brought the paired leaves A1r and U1v as well as B3r and Dd1v face to face, while leaving the pairs joined at their top edges. It accounts for combinations of offset found on two sets of facing pages in at least sixteen copies. (See Appendix 2 for a full list of offset examples identified.) Pages A1r and U1v exchange offset reciprocally in fourteen copies, while B3r and Dd1v do so in nine of the same copies, and ten overall. Several instances of offset occur in only one direction on these pages. Three times B3r shows up on Dd1v, and Dd1v appears once on B3r, A1r once on U1v, and U1v once on A1r.

When, as a separated half-sheet, this once-folded unit was folded a second time, A1v and Dd1r or B3v and U1r would have been brought face to face. The former fold led to an exchange of offset between A1v and Dd1r in CtY, FU, May3, and LdU-B (reciprocal in CtY, May3, and LdU-B), and allowed U1r, now on the outside, to exchange offset with Cc4v as occurred reciprocally in CtY, May3 and FU. The latter fold led to an exchange of offset between B3v and U1r in LU, MR2, NRU, and Ob. Presumably B3v and U1r were brought face to face in the NRU copy also, for its A1v received offset from Cc1r and its Dd1r received offset from Bb8v.


Page 201

If the unfolded half-sheet as described above was folded the opposite direction along the short axis, such that the inner forme was folded inward, then pairs A1v and U1r and B3v and Dd1r would exchange offset. This folding occurred in DLC, ICN1, OAU, and RPB copies. In DLC and OAU, both A1v/U1r and Dd1r/B3v reciprocally exchanged offset, and in ICN1 and RPB A1v has U1r’s offset and B3v and Dd1r reciprocally exchanged offset. If this folded half-sheet were folded a second time, it would result in offset exchanged either by A1r and Dd1v or by U1v and B3r. The latter fold occurred in DLC, ICN1, OAU, and RPB, where U1v and B3r exchanged offset. In DLC and RPB, A1r, left on the outside, exchanged offset with Cc4v (Dd1v lacks offset in both copies, and in both the frontispiece was stored between Bb8v and Cc1r). In OAU, A1v and Bb8v reciprocally exchanged offset, as did Dd1v and Cc1r.

Folding a half-sheet along the long axis certainly was not the common practice, but it happened surprisingly often. If the inner forme were folded inward along the long axis, then offset would be exchanged both between A1v and Dd1r and between B3v and U1r. This, in fact, happened to the half-sheet at CaOHM where A1v has offset from Dd1r, and B3v and U1r reciprocally exchanged offset. If a second fold occurred, it would place either A1r against U1v or Dd1v against B3r. In CaOHM the former is also seen: U1v has A1r’s offset. The three offset exchanges in the CaOHM half-sheet could not arise from any manner of folding the half-sheet first along the short axis.

If the outer forme of the singletons’ half-sheet were folded inward along the long axis, then offset would be exchanged between A1r and Dd1v and between U1v and B3r. The evidence for this fold involves both pairs of conjugate leaves in the NIC and NjP copies, where reciprocal offset exchanges occur between A1r and Dd1v and B3r and U1v. A second fold would place Dd1r against B3v or A1v against U1r. The former fold occurred to the NIC copy: Dd1r and B3v reciprocally exchanged offset, which left U1r on the outside of the folded unit, where it received offset from Cc4v. In NjP, there appears to be offset of a short catchword probably beginning "th" on A1v, suggesting that it was on the exterior, with B3v and Dd1r folded against each other. Just as for the CaOHM copy, so for the NIC and NjP copies no other fold but initially along the long axis can explain the offset pattern.

It seems unlikely that the initial fold of the half-sheets with the singletons was regularly along the long axis, for folding along the short axis to reduce the longer dimension at each step is the easier practice since it reduces arm work by reducing the size of the unit worked on. Furthermore, it is the conventional practice in folding a half-sheet like Cc4 into a unit ready to be bound. The preference for reducing the longest dimension by folding it at the short axis is apparent in those copies where the singletons and half-sheet Cc4 were folded together as an uncut sheet. Nevertheless, in three copies offset is recorded as exchanged between both leaf pairs facing across from one another along the singletons’ half-sheet’s long axis (NIC and NjP folding inner forme inward [A1v against Dd1r, U1r against B3v]; and CaOHM folding outer forme inward [A1r against Dd1v, U1v against B31]). Still, this long fold involves but relatively few copies with offset, for exchanges on multiple pairs of leaves and other evidence suggest at least twelve


Page 202
copies were cut into half sheets and then folded along the short axis (A1 against U1 and Dd1 against B3).5

On several occasions the singleton half-sheet was cut into quarter-sheets before being folded. This alone can explain the anomalous offset in the CSt and Owo copies. In Owo Dd1r’s offset is on A1v, and U1v’s exchanged reciprocally with B3r. Thus, the inner forme was folded inward in the first pair, but the outer forme was folded inward in the second. Since there is no evidence of resetting and reimposition, this pattern indicates that the half-sheet was cut into folds A1 .Dd1 and U1 .B3 before each was separately folded. These separate folds left A1r and B3v on the exteriors, and, stacked together, A1r and B3v received each other’s offset. In CSt the offset shows that, after cutting, A1 .D1, B3 .U1, and Cc4 were folded separately and stacked together. Offset exchanges indicate that both A1v.Dd1r and U1v.B3r were folded inward. Pages A1r and B3v, both on the outside of their folds, exchanged offset. And U1r, also an exterior page, received Cc4v’s offset. (This three-unit stack had Dd1v and Cc1r on the outside, neither receiving offset.)

Offset evidence also corroborates that the singletons and the leaves of Cc4, in positions defined above, were printed on a full sheet. The best evidence of offset exchanged within a full sheet occurs in the BrP, InU, KU, May2, May4, MB, and O-EF copies, either folded as a whole sheet prior to cutting in half or folded as if still uncut. In these seven copies the inner-forme pages of the singletons’ half-sheet bear offset from the inner forme of half-sheet Cc4. In May2 and MB all of the eight page-settings of the inner forme exchanged offset, resulting in the following pattern (bold indicates pages as actually printed; regular font indicates offset on those pages):

A 1 v   Dd1r  Cc3v  Cc2r 
Cc2 r   Cc3 v   Dd1 r   A1 v  
Cc1 v   Cc4 r   B3 v   U1 r  
U1r  B3v  Cc4r  Cc1v 

In InU and May4 seven of the eight pages in this forme received similar offset: the InU copy only lacks offset of A1v on Cc2r, and May 4 does not have offset of


Page 203
B3v on Cc4r. Six of the eight pages bear offset in the KU and BrP copies. In the KU copy, Cc2r and Cc3v do not carry offset, but they have transferred it to A1v and Dd1r respectively. BrP has all the exchanges caused by this fold except, back and forth, between B3v and Cc4r. O-EF has offset from the outermost leaves: A1v has Cc2r’s offset and U1r has Cc1v’s.

In six copies (BrP, InU, KU, May2, MB, and O-EF) the half-sheet with the singletons carries an offset exchange between the outer forme of the half-sheet (A1r with U1v and Dd1v with B3r; reciprocal in BrP, InU KU and May2; only reciprocal on the former pair in May4 and MB; and only on the latter in O-EF). This offset was exchanged as the result of a second fold: the sheet was folded such that the outer forme of the singleton half was bought down upon itself. The offset on the outer forme thus has the following pattern (bold indicates regularly printed pages; regular font indicates offset on those pages):

U1v  B3r 
A 1 r   Dd1 v   Cc3 r   Cc2 v  
U1 v   B3 r   Cc4 v   Cc1 r  
A 1 r   Dd1v 

This second fold would then have been followed by a third one. No offset indicates the direction of this fold, but it is likely to have been conventional in bringing Cc2v against Cc3r instead of Cc1r against Cc4v.

Other patterns of offset likewise provide evidence that the four singletons and the Cc4 half-sheet were printed together. The offset patterns in the seven copies just noted could have occurred were the whole sheet not cut before folding. Whatever advantages might accrue to that procedure, a disadvantage was that time would have to be spent unfolding the unit later. It is therefore not surprising that the offset patterns in most copies indicate that the printed sheet was first cut in half and then the half-sheets were folded separately.

One such pattern is evident in copies CaOHM, NIC, and NjP: here offset shows that the group of four singletons was folded along the long axis of the unit and then, in CaOHM and NIC, across the short axis. (The first fold results in these pairings of offset: A1v/Dd1r, B3v/U1r [CaOHM]; A1r/Dd1v, B3r/U1v [NIC, NjP]. In CaOHM the second fold yields A1r/U1v; in NIC, B3v/Dd1r.) The second fold and resulting offset could not have occurred without the full sheet first having been cut in half.

Another pattern of offset exchanged between pages of the two half-sheets likewise shows that the transfer could have occurred only if the half-sheets had first been separated. Except for the seven copies that seem to have been folded as full sheets, offset between the two major sections of the sheet occurs on only a single page of each half-sheet at a time. The particular pairings of pages from each half-sheet can vary, but the extant combinations preclude their having been folded while the singletons were still conjugate with half-sheet Cc4. For instance, CtY has offset exchanged between U1r (inner forme) and Cc4v (outer forme). If the half-sheets were separated, set face to face with U1r and Cc4v facing each other, and then folded together with the outer forme of the singletons inward so as to transfer offset according to the pattern A1r/U1v and Dd1v/B3r (as found in


Page 204
CtY), the resulting quarter-sheet-shaped unit could not be folded again to allow the exchange between A1v and Dd1r present in the CtY copy (because Cc2.Cc3 would intervene). CtY’s offset exchange between U1r and Cc4v had to occur after the two half-sheets were twice folded separately; subsequently U1r and Cc4v, now on the exterior of the folded half-sheets, met when they were stacked.6

In several copies with this "stacking" evidence the fact that the half-sheets were folded after being separated from each other is further attested by offset transfer within each of the half-sheets. In CtY and May3, the half-sheets were first folded in opposite directions: for the singletons half-sheet the outer forme was folded inward (A1r/U1v), while for Cc4 the inner forme was first folded inward (Cc1v/Cc2r and Cc3v/Cc4r mutually exchanged much offset). In the second folds, A1v and Dd1r were brought face to face, as were Cc2v and C3r; this positioned U1r (from the inner forme) on the outside, where it received offset from Cc4v (a page of the outer forme).


For an account of the handling of the printed sheets from the press through their being bound, see Philip Gaskell, A New Introduction to Bibliography (New York: Oxford Univ. Press, 1972), pp. 143-148. Gaskell’s account, as we will soon see, does not entirely fit the history of this edition, for sheets were separately folded to octavo size and collated into stacks for individual copies while still at the printer’s shop, an activity Gaskell places at the bindery.


Of sixty-nine copies examined, I noted no offset present in and from cancels and partials except between facing pages or the frontispiece in thirty-six copies: BirmU, C, CaOTU (only frtr<B7 + 1v), CLU, CLU-C (only frtr<B7 + 1v), CoU, Csj, CSmH, E, ICN2, L (only frtr<Dd1v), May1, May5, MH1, MH2 (frtr<B7 + 1v), MiU, MoU, MR1, MR3, NcU, NeU, NNU2, NStBU, O (frtr<B7 + 1v), Oa, OAshU, Occ, OCiU, OCU, O-HF, Ose, OTH, PSt, TnU (only frtr<B7 + 1v), TxU, WU. In addition, five copies, starred in the list of copy locations in Appendix i as having offset, aside from exchanges with the frt, have only one exchange between a page of the cancels and partials with another leaf: MdBJ (U1v<B3r); N (B5v<B7r); NBiSU (Pir><Dd1v [also: frtr<08v]); Ok (A1r<Cc4v); and WNs (P1r>Dd1v).


Some evidence for the long folding of both the singleton and Cc4 half-sheets is offered by CaAEU: U1v has B3r’s offset on the outer forme and Dd1r has Cc3v’s offset on the inner forme. If the whole sheet were not cut, this pattern could occur in two ways: the inner forme folded in half along the short axis, placing Cc3v against Dd1r, and then the half-sized sheet folded along the long axis inward on the outer forme of the singletons’ half, placing B3r against U1v and leaving the outer forme of Cc4 on the outside; or with three folds, accordion-like: with pages A1r and U1v on the far left folded inward atop Dd1v and B3r, and Cc2v and Cc1r on the far right folded inward atop Cc3r and Cc4v, and then the half-sheet-sized unit folded backward to place Cc3v against Dd1r and Cc4r against B3v. However, after a third fold this sheet would not be very serviceably folded for later separation and use. The alternative is to suppose the sheet was cut in half, the singletons’ half folded inward along the long axis, placing B3r against U1v, and then folding U1r against A1v, leaving Dd1r on the outside to receive Cc3v’s offset when the singletons’ half was slipped inside folded half-sheet Cc4 or when Dd1r was stacked against an incorrectly folded Cc4 with Cc3v on the exterior of the folded unit. In either case this pattern cannot be explained by a common efficient folding of the sheet or its half-sheets.

Cancellantia In Gathering B

Although the cancellans folds of the dedication and of the text never exchanged offset, cancellantia B2.B7 + 1 and B6.B7 exchanged offset among each other as did N1.N2 and R1.R2. Offset evidence confirms what was elsewhere shown ("Interrelating the Cancellantia," 255–258): that leaves B2.B7 + 1 and B6.B7 were printed on the same half-sheet in the following arrangement:

B6r  B7v 
B2v  B7 + 1r 

B6r and B2v were at the end of the original sheet, for tranchefiles run along their outer edges. If the two conjugate pairs so arranged were folded along the short axis of the half-sheet such that B6r was brought down on B2v and B7v down on B7 +1r, the result would be the offset pattern seen in seventeen copies, where B2v and B6r exchanged offset (reciprocal in all but May2 and Ob), as did B7v and B7 + 1r. Of course, only the B2v and B6r exchange can be offered as evidence of imposition and folding patterns, for B7v and B7 +1r would have exchanged offset after binding, from their adjacent placement in the bound book. The sensible second fold would be to bring B6v and B7r face to face and leave B2r upward, signature visible for identification. Because B6v and B7r would often exchange offset after being bound, other evidence must demonstrate this second fold occurred. In


Page 205
DLC, NRU, OAU, and RPB that evidence involves exchanges by B6v with B4r and by B7r with B5v, showing fold B4.B5 was temporarily encompassed by fold B6v.B7r. Other evidence for the second fold comes from B2r.B7 + 1v, left on the exterior to receive offset from outside the normal placement (in ICN1 and KU, B7 + 1v exchanged offset with B1r; in KU, B2r carries the offset of G8v; in RPB B7 + 1v received offset from K1r; and in NNU1, B7 + 1v has N1r's offset). No copy's offset shows that the second fold was made in the other direction, a circumstance that would have required refolding later.

One might suppose that all copies would have the cancellans B half-sheet folded in the sensible fashion just described, but many do not. Apparently, about a third of the time the half-sheets lay upon the table facing the folder with what might be called the outer forme upward:

B2r  B7 + 1v 
B6v  B7r 

Folding this half-sheet forward along the short axis, with B2r against B6v and B7 + 1v against B7r, would produce the offset pattern seen in six copies (CSt, CtY, IU, LdU-B, NIC, and ViWCF); all but one (ViWCF) have offset exchanged by both pairs of facing leaves. The exchanges are reciprocal between B2r and B6v within CSt, CtY, IU, LdU-B, and NIC, and reciprocal for the other pair in CSt, IU, LdU-B, and NIC. In NIC the second fold led to reciprocal exchanges between B6r and B7v, which left B2v and B7 + 1r on the outside (B2v subsequently transferred offset to B1r, and B7 + 1r transferred offset to L8v). In most copies with this same initial fold, the second fold turned B2v against B7 + 1r: B2v bears the offset of B7 + 1r in CtY and ViWCF, and the exchange is reciprocal in CSt, IU, and LdU-B). This double-fold left B6r.B7v on the exterior, able to exchange offset with the leaf of another printed unit, as B6r did with B5v in ViWCF, for example, and B7v exchanged offset with B1r in CtY and LdU-B. Copies with offset on B6r or B7v from outside the unit may provide evidence for the inefficient folding.

The inefficiency of leaving B6r.B7v on the outside of the folded unit is curious. It might be that the workman wasn't paying attention and continued folding as if he were still facing the inner forme, or perhaps he was aware of the inappropriate orientation of the half-sheet, having noted the signature "B2," but was too lazy to flip the half-sheets over or raise the center and fold. If the worker did not care that he was making a fold that would have to be unfolded, it is unlikely that he was the same fellow who would eventually bind the stacked sheets into a book, when the difficulty would have to be dealt with. The diversity of folding patterns might reflect multiple workers, or it might simply indicate a single individual rebelling against the redundancy of his task.


Similarly, in DLC and RPB copies, where offset exchanges occurred within the inner forme of the singletons' half-sheet, A1r on the outer forme has offset of Cc4v. In DLC, offset records two foldings of the singletons' half-sheet, with reciprocal exchanges in the first fold between A1v.Dd1r on the one hand against U1r.B3v on the other, and then U1v against B3r. The exchange between A1r and Cc4v in that copy therefore could only have occurred after the two half-sheets were cut apart and separately folded. In the RPB copy the singleton's half-sheet was certainly folded twice separately (five of the eight pages bear offset from other pages of the unit); here, the exchange between A1r and Cc4v can only be explained by supposing the two half-sheets were folded twice separately, A1r ending up on the outside and stacked against Cc4v. The NIC copy has offset linking U1r and Cc4v, but its singletons' half-sheet was folded twice separately, first along the long axis.

Cancellantia in Gatherings N and R

Offset evidence (as well as watermark evidence reported in "Interrelating the Cancellantia," 251, 255–257) indicates that cancellans folds N1.N2 and R1.R2 were not printed on one-half of a sheet, with top edges adjoining, as the B cancels were. Yet the two two-leaf folds often exchanged offset, implying that


Page 206
they somehow were printed on the same sheet and folded and/or stacked with one another. Pages N1r and R2v reciprocally exchanged offset in twenty copies; both leaves in the pairs N1r.N2v and R1v.R2r (with N1r facing R1v) exchanged offset in five others, and R1r and N2v exchanged offset in two more. Thus, pages of quarter-sheets N1.N2 and of R1.R2 exchanged offset in twenty-seven of the twenty-eight copies seen to have offset on more than one set of nonadjacent pages. Since in many copies leaves N1 and R1 have tranchefiles (presumably all did before trimming), we know they were each at the end of the sheet, parallel to the short axis. Since three quarters of copies examined have a watermark on N1 or R1 but no copy has a watermark divided between both leaves ("Interrelating the Cancellantia," 255–256), we can surmise they were not on the same half of the sheet, with N1 and R1 joined at their top edges. Rather, they were probably printed on the same sheet along the long axis N1.N2.R2.R1. Prior to folding, these two conjugate folds had already been separated from whatever originally occupied the other half of the full sheet and may have been cut in half into quarter-sheets before the folding.7 No offset from the extraneous work or works occupying the other long half of the sheet appears on N1.N2 and R1.R2 of any copy examined.

The offset evidence argues that they were arranged on the long axis specifically as N1v.N2r.R2v.R1r (or, flipping this unit end to end, R1v.R2r.N2v.N1r). The arrangement allowed for the quarter-sheets to be easily folded so that N1r and R1r faced upward. Also, placing leaves N1 and R1 on the outside edge increased the likelihood of their receiving a watermark. The arrangement is indicated by the reciprocal exchange of offset between both R1v/N1r and R2r/N2v in five copies (CSt, IU, NNU1, OAU, and ViWCF). This exchange may have occurred were the two pairs arranged on the uncut sheet with R1v.R2r.N2v.N1r face up and that sheet folded in half prior to cutting or, if cut, overlaid as if never cut, that is, with the inner side of fold R1v.R2r facing the outer of fold N2v.N1r. Then a second folding would reduce the unit to leaf size, engulfing one quarter-sheet within the other. Oddly in at least CSt, IU, OAU, and ViWCF copies, fold R1.R2 was on the exterior of N1.N2, though this would require a lifting fold instead of a turning fold. This left R1r on the outside, to exchange offset reciprocally with Z8v in CSt, IU, and OAU, and R2v also on the outside, to receive offset from Aa1r in CSt, IU, and OAU, and from Bb1r in ViWCF. In NNU1 both N1v.N2r and R2v.R1r bear no offset whatsoever, as if they had dried in quarter-sheet size before any second folding occurred. However, NNU1's quarter-sheets were


Page 207
probably folded a second time to reduce the pair of leaves to leaf size, for that seems implied by offset in CSt, IU, OAU, and ViWCF, and by the general pattern of stacking gatherings and units folded to leaf size. Curiously, we don't find any evidence that the opposite sides of these quarter-sheets were laid against each other: in no copy is offset exchanged by R1r/N1v or R2v/N2r. Such offset could have occurred if the quarter-sheets were so laid upon one another, and folded together with N1r facing up on top. Given the variety seen in folding the singletons' sheet and half-sheet with B cancellantia, the failure to find any instances of N1v.N2r folded against R1r.R2v (and only five copies with full exchanges between the reverse sides) suggests the four leaves were usually not cut into quarter-sheets prior to folding.

The most common offset exchange between N1.N2 and R1.R2, noted as reciprocal in twenty copies, is for N1r and R2v to have exchanged offset. Pages N1r and R2v were printed on different sides of the sheet, and so could not be folded face to face in the manner that N1r and R1v were in CSt, IU, NNU1, OAU, and ViWCF. But, if the two quarter-sheets were first cut apart, N1r and R2v could have been folded face to face in either of two ways.

First, the separated quarter-sheets could have been placed open against each other, with N1r facing R2v and N2v facing R1r. Then the stacked quarter-sheets could have both been folded with N1v turned against N2r (leaving R1v.R2r on the exterior) or with R1v turned against R2r (leaving N1v.N2r on the exterior). This method is indicated if offset appears as often between one pair (R1r and N2v) as another (N1r and R2v), and if offset from leaves other than these four tends to occur on N1v.N2r or on R1v.R2r. Yet there is almost no evidence for this folding. None of the twenty copies with reciprocal exchange N1r/R2v also has offset exchanged between their adjacent pages, N2v and R1r. Also, no copy seen has offset from outside pages transferred to both exterior leaves of this hypothesized fold, either R1v.R2r or N1v.N2r. Further, if the quarter-sheets were cut apart, one might expect that at times the pair would have been stacked with R1v.R2r against N1r.N2v, allowing R1.R2 to be folded as it ultimately would be with R1r outside, yet no offset shows this happened. So, there is no evidence for what might be thought a likely way to fold N1.N2 and R1.R2 economically (with a single cut and fold).

Pages N1r and R2v could also exchange in a second way after cutting. The quarter-sheets could first be folded separately, with the first versos turned against the second rectos, and then these leaf-sized folds could be stacked with the exterior N1r directly under the exterior R2v or with exterior R1r directly under the exterior N2v. But only Owo copy's limited offset suggests N1.N2 and R1.R2 were first cut, folded, and stacked this way: here N1v was folded against N2r, and R1v against R2r; then they were stacked against each other, with R1r facing N2v and exchanging offset. This arrangement left N1r on the exterior and able to exchange offset with B8v. (Owo's singletons half-sheet likewise was also cut into quarter-sheets before folding.) In contrast to this rare pattern, a similar arrangement is far more common. Here the offset occurs as if the folded quarter-sheet R1.R2 rather than N1.N2 was on top: the offset shows N1r placed against R2v, and R1r and N2v left on the outside to receive the offset. Of the twenty copies with offset reciprocally exchanged between N1r and R2v, sixteen have an exchange between R1r


Page 208
and Z8v and fourteen have an exchange between N2v and Aa1r, making nineteen of the twenty with at least one or the other. Furthermore, an exchange between R1r and Z8v also occurred in CSt, InU, IU, and OAU, and R2v exchanged offset with Aa1r in CSt, IU, and OAU. Thus, in total, offset of either R1r or N2v appears on either Z8v or Aa1r in twenty-three copies, that is, in well over half the copies where offset has been noticed (perhaps in as many as ninety percent of the copies with offset, for I initially failed to scrutinize Z8v and Aa1r for offset). Why, we might ask, if the quarter-sheets had been cut free, folded, and stacked, was the R1.R2 fold almost never (never except for Owo) placed underneath N1.N2 such that N2v and R1r exchanged offset? That this placement occurred rarely is suggested by there being no copy seen with offset exchanged between N1r and Z8v. Indeed, it is odd that, within quarter-sheets N1.N2 and R1.R2, only N2v and R1r have contact with Z8v and Aa1r. I believe there is a better explanation of how N1r and R2v were normally folded to exchange offset.

The offset exchanges N1r/R2v, R1r/Z8v, and N2v/Aa1r can be accounted for without cutting oblong N1.N2.R2.R1 into quarter-sheets if the unit was instead folded three times. With the side N1v.N2r.R2v.R1r face up, the workman moved from left to right. First he folded N1v down on N2r; then he folded N1r, now facing up, right and down on R2v, and then he either turned R1r downward and under or turned the three-leaf left-hand unit under R1, bringing R1v against R2r. If this was done, the inner formes of both folds R1v.R2r and of N1v.N2' would be facing each other, folded for use. Also, the leaf-sized unit would have N2v and R1r facing outward, both ready to receive offset, respectively, from adjacent leaves Aa1r and Z8v. (Perhaps R1r, not N2v, was face downward toward the front of the alphabet so that the joined and rounded edges of two folds were more identifiable as appearing among separated leaf edges.) The advantage of this folding would be that the two quarter-sheet folds would be linked, and the leaves the right size for stacking. Furthermore, the two folds were ready to guide the blade and were both on the same side of the folded unit. Surely it would take more time both to fold the cancellantia down to leaf size and to separate them into two folds than it would to just fold them three times. And, if one workman folded but did not cut, he could work on, even if a companion borrowed or was using the cutting tool. In any case, even if a blade or shears were at hand and the folder was tasked with separating these quarter-sheets, folding may routinely have preceded cutting as it created the line to cut along or the fold to slice outwards. That the N and R cancellans were folded three times without cutting is suggested by their being stacked against each other with R2v atop N1r in most copies. This method explains the offset exchanged reciprocally by N1r and R2v in the aforementioned twenty copies (and possibly in CaOHM), the offset exchanged by R1r and Z8v in at least sixteen of these copies, and the exchange between N2v and Aa1r in at least fourteen. All three consequent pieces of offset evidence occur in eleven copies (CaAEU, DLC, ICN1, LdU-B, May2, May3, MB, NIC, NRU, Ob, and RPB), and at least two occur in eight others (BrP, CtY, FU, KU, LU, May4, MR2, and NjP). Further, no contrary evidence occurs in any copy; in all of them, N1v faced N2r and R1v faced R2r, as need be the case. This hypothesis involving three folds without cutting also explains why R1r and N2v never were overlaid, exchanging offset.


Page 209

If this predominant offset pattern for N1.N2 and R1.R2 was produced without cutting, probably the other common patterns for the N and R cancellantia also did not involve cutting prior to folding. The exchange of offset between both R1v/N1r and R2r/N2v at CSt, IU, NNU1, OAU, and ViWCF could have occurred prior to cutting when the quarter-sheet folds were still joined. If this unit was laid out uncut as N1v.N2r.R2v.R1r, and N1v were folded right against N2r, and R2.R1 backwards with R2r against N2v and then R1v to the right downward against N1r, the two sets of exchanges would occur. And this would leave R1r. R2v on the exterior to exchange offset with other leaves as happened in CSt, IU, OAU, and ViWCF. In addition, the offset exchanges in Owo and possibly CaOHM could have been produced with three folds. Imagine that the oblong four-page unit had the reverse side up, R1v.R2r.N2v.N1r. Folding again begins from the left with R1v turned down on R2r and then these two leaves turned with R1r face down on N2v, producing the reciprocal exchanges in CaOHM and Owo. Then to produce patterns in Owo the right-most leaf with N1r facing up was folded down and under, turning N1v against N2r and leaving N1r on the exterior able to transfer offset to B8v. Or, to produce CaOHM's further exchange where N1r's offset was placed on R2v, N1r was not turned under but wrongly folded left over the three-leaf stack with R2v upward. This final fold would leave N1v. N2 r on the exterior—a blunder but possibly done. If the folding began with N1v.N2r.R2v.R1r facing up, three folds could also produce CaOHM's offset pattern but still with wrong pages (R1v.R2r) on the exterior. Neither option prepares CaOHM's R2v to exchange offset with Bb8v.


It seems likely that a folder separated a sheet into halves by tearing it or slicing it, not by cutting it with shears. It is difficult to identify the method of separation, for usually the top fold in an octavo was left sealed until trimmed by the binder or cut free by the owner of an untrimmed copy. The long exterior leaf-edge as that joining A1 with Dd1 or U1 with B3 is always hidden at the spine. Scissors would seem to leave a series of straight cuts not perfectly in line and require more time than a long-bladed knife or paper slicer (I've seen little evidence of scissors used by those assembling and binding sheets). In the only untrimmed copy of this book examined (the Csj copy), the top fold and outside edge on B3 and the half-sheet with B cancellantia were not scissored but cut free with a knife, leaving uneven edges with surpluses and deficits on adjacent leaves (Dd1's top complements B3's; B2's, B6's; and so on). The N and R cancellantia have had their top edges cut evenly on a straight line as by a slicer.

Gathering the Sheets

Setting aside the method by which N1.N2 and R1.R2 were folded, we turn to two interesting facts: both folds were placed together (no copy has evidence of their being stacked apart), and they were routinely inserted into the folded stack between sheets Z and Aa (twenty-three copies have offset exchanged by either R1r or N2v with either Z8v or Aa1r). Only in a few copies do folds N1.N2 and R1.R2 have offset from any other leaves than each other or Z8 and Aa1.8 The offset recorded for the N and R cancellantia therefore also raises the question of whether the placement in the stacked sheets reflects when the cancels were printed relative to the last gatherings. It is unlikely that they were printed before sheet Aa, for Young and Samuel Richardson's correspondence indicates that the press had Young's corrected sheets Aa and Bb before Young decided what to do with the preliminaries and the N and R cancellantia.9 But it seems possible


Page 210
that these partials were routinely inserted between the same folded sheets of a book to aid the binder in locating them and to double check that all units were included. If the folded sheets that were stacked to be bound as one copy had the first printed sheets on the bottom, then it would have been easiest to place the late-printed partials and cancels into divisions toward the end of the signature sequence (between sheets at the top of the stack). Placing them after Z would put them in a convenient position: at the end of an alphabetical sequence, and also possibly at the end of a conventional number of sheets gathered in the warehouse as a unit. (Gaskell reports that the sheets of long books initially "were divided in quires of 12–24 sheets" [p. 144n.]). However, the position of other late-printed partials within the stack does not assist the hypotheses that the N and R cancellantia were printed after sheet Z and before Aa or that partials were routinely placed after sheet Z.

The half-sheet with B cancellantia and the four original leaves for B were commonly inserted into the stack of folded sheets together. In at least fifteen copies offset was exchanged by original B leaves and the leaves of the B cancellans half-sheet. Though gathered together, the eight leaves were not yet rearranged into the order in which they were bound (for instance, B1v and B4r exchanged offset in CaOHM, CSt, and FU; B5v and B7r in DLC, N, NRU, OAU, Owo, and RPB; and B5v and B8r in CSt, CtY, FU, KU, May2, May3, May4, MB, NNU1, and ViWCF). Also, cancellans B3 was almost never inserted in the stack with the original or cancellans B sheets (in no copy is pre-binding offset found exchanged between B3 and other B leaves). Unlike the N and R cancellantia, the half-sheets with the B cancels and singletons were stacked in no single position. As offset in many of these copies shows, the other eight leaves of the B gathering were inserted in various positions between sheets G and N: after G in KU (G8v>B2r), after H in MB (I1r<B7 + 1v), after I in RPB (I8v<B1r and K1r>B8v), after L in DLC (L8v>B1r) and NIC (L8v<B7 + 1r), and after M in NNU1 (M8v>B1r). Note that the position of the B cancels closer to the front of the stack sheets might suggest they were printed before the N and R cancellantia.

Offset in some copies shows that the B cancels were gathered and stacked beside the singletons; for instance, in copy O, in addition to B6r's receiving Dd1v's offset, B7 + 1r's offset was transferred to the recto of the frontispiece—it appears as if Dd1 was slipped in between B2v and B6r, and the frontispiece was slipped in between B7v and B7 + 1r. But the leaves of the singletons and Cc4 do not bear offset showing they were inserted while damp into the stacks of folded sheets.

The frontispiece was stacked sometimes with the dedication leaves (it received offset from B7 + 1v in at least seven copies (BirmU, CaOTU, CLU-C, MH2, O, PSt, and TnU), sometimes with the singletons half-sheet (as in CtY, E, L, and


Page 211
MiU), and sometimes with Cc4 (DLC and RPB). The frontispiece and whatever it was gathered with ended up inserted in diverse places (and not regularly after sheet Z). Offset indicates that the frontispiece was inserted after X8r in MR2, after Aa8v and before Bb1r in LdU-B, and after Bb8v and before Cc1r in DLC and RPB. The singletons (B3 stayed with them) and half-sheet Cc4, often with the frontispiece, were placed sometimes between sheets O and P (in NBiSU and WNs Dd1v exchanged offset with P1r, and in NBiSU the recto of the frontispiece has O8v’s offset, too).

Finally, I would stress that the offset is being exchanged between pages in full sheets and half-sheets that have been folded completely into leaf-size units. Contrary to what Philip Gaskell describes as the routine process,10 in these copies offset is never exchanged between sheets folded in half and tied for the bindery for further folding by a binder. Rather, sheets were apparently folded, thrice for whole sheets, until they were reduced to octavo size. That is evident from much offset evidence regarding the stacking, such as from the offset of cancellanda on M8v, Q8v, and T8v, where offset was transferred because the sheets had been folded and collated to form books prior to the cancellations of N1, R1, and U1. Usually the partials printed together were inserted into this stack prior to their separation. However, sometimes offset indicates half-sheets of cancellantia were further cut into quarter-sheets before being inserted into the stacked sheets.


In Owo, N1r transferred offset to B8v; in ViWCF, R2v exchanged offset with Bb1r; and in CaOHM, R2v exchanged offset with Bb8v (but in CaOHM Aa1r has Y8v's offset, indicating that sheet Z was not stacked its proper place).


We know from Richardson and Young's correspondence that sheet Z was printed before 21 January 1755 and that N1.N2, R1.R2, and sheet Aa were not yet printed (The Correspondence of Edward Young 1683 —1765, ed. Henry Pettit [Oxford: Clarendon, 1971], pp. 416 419). In listing corrections needed in future editions, Richardson's letter of 21 January 1755 refers to the text as far as p. 325 (Z3r), making no suggestions for sheets Aa and Bb (pp. 416–417). In his undated response to Richardson, Young asks that "this last sheet [be seen] again before working off" (Pettit, p. 418). That sheet was probably Aa, but could have been Bb, to judge from remarks in a subsequent undated letter. In that letter, which Pettit dates four days later on 26 January 1755, Young remarks, "Tomorrow I expect the last Sheet" and promises, when he returns it, to speak of what to do about the dedication and the N1-N2 materials that Richardson would delete and alter (pp. 418–419). Thus, Young wished to complete full text sheets before dealing with partials and cancels. Finally, that the N cancels have the highest frequency of offset exchanges suggest they were the dampest unit when folding began, and thus the last printed.

Changes in the Text

One last revelation about the cancellation in The Centaur Not Fabulous is offered by offset evidence: it provides text for the cancellanda replaced. One instance concerns the changes to N1.N2. Richardson on 21 January 1755 advised Young: "In another impression methinks it were to be wished that all from, If this is a man of pleasure, p. 161 [N1r], to, from a higher hand, p. 163, were omitted, as it interrupts, by ludicrous images, emotions that were nobly excited."11 While Young kept the initial clause, the phrase ending the proposed deletion does not appear on or near p. 163. Richardson applauded as affective what preceded the clause "If this is a man of pleasure," offering no encouragement to alter the preceding two paragraphs on N1r. Yet p. 161 has relatively less text than is usual, with nineteen lines and three paragraph spacings (twenty lines and three such spacings is typical in the book), and the first spacing atop the page is redundant (there is one at the bottom of the previous page, M8v). One might therefore wonder whether something was removed on N1r. However, the offset of the cancellandum


Page 212
received by M8v in the CtY, DLC, IU, NIC, and ViWCF copies indicates the text on N1r was not substantially changed but that a line from cancellandum N1r was moved to the next page to fill in for material cut within N1v-N2v. These copies show that the cancellandum’s final line has the same reading as the first line of cancellans N1v: "total, is the transit of these Phaeton-" and the catchword ("tiades;") was the rest of that word, found now on the second line ("tiades:" with different punctuation). Also, the cancellandum’s third and second last lines (on N1r) read the same as the cancellans’ penultimate and final lines. In CtY, we can see on M8v that the final three lines of the preceding paragraph also remained unchanged ("gay ... expired."), and scattered lines on the page above are in the same position relative to one another as they are in the cancellans, suggesting that no changes other than to accidentals or substantives of few letters probably occurred. One can make out a few words in the top paragraph, as the same first line (on page and in paragraph). The CtY, DLC, and ViWCF copies’ offsets show that the cancellans has one fewer line because the spacing before and after the first paragraph was increased, not as a result of anything added to the cancellans’ text: in the cancellandum there was 4 mm of space between "161" and the final word in the first line, "poison’d" (ViWCF), whereas in the cancellans this distance is 10–11 mm. In ViWCF, offset indicates cancellandum N1r began as it currently does: "My principles have poison’d | my friend." The first four lines to "wife!—" are identical in cancellandum and cancellans, and the words "is there another" occur in the fifth line of both.

In addition, the cancellandum N2v’s text, to judge from offset in DLC on N3r, held the identical text on its last five lines. However, offset from the cancellandum on N3r of OAU indicates the page began with different text ("I mean, of our ... [word in italics not clear] because real men").

Offset of the cancellandum of R1r appears on Q8v in over a dozen copies. Clear offset in CtY, DLC, FU, ICN1, IU, and LdU-B indicate that the cancellandum of Rir carried seven lines, three fewer than does the cancellans, and that these lines are identical in the two versions, in accidentals as well as substantives. The cancellandum’s catchword was "world’s" (which begins line 8 of the cancellans). Young and Richardson’s correspondence does not indicate why leaves R1 and R2 were cancelled, but at least the offset from the cancellandum of R1r indicates that the two leaves were cancelled to add, not to cut, text. However, the addition was not that substantial since the cancellandum of R2v had one more line than the cancellans: the first word of the cancellandum of R2v was "turn" (indicated by offset in OAU on R3r), as is the first word on line 2 of the cancellans.

As for leaf U1, offset on T8v in DLC and FU shows that lines 1–3 of U1r are the same in both cancellans and cancellandum, but in line 4 the cancellans spreads out the words "lessen its formidable gloom" and then begins a new sentence with "I | have seen a Death-bed." Lines 4–6 of the cancellandum read: "lessen its formidable gloom. To Cen- | taurs, especially ... [least?] the among | them, who may be |" (beneath "who may be" in the line below is a word beginning "impe" [perhaps "impeach’d"] and directly below that in the next line is "to each"—none of these words following "gloom" appear in the cancellans


Page 213
paragraph). Offset (especially in DLC) indicates that the second paragraph of the cancellandum began with a sentence now cut: "But to return to our last bed, the | grave" (fifth and fourth last lines). Thereafter the fourth last through final lines have the same text as the cancellans. The verso of cancellans U1 appears not to carry significant revision. The cancellandum’s offset on U2r in DLC suggests that the last nine lines on U1v have not been changed and shows that the last two lines of U1v were identical except that "death;" in the last line ended with a comma.


Gaskell notes that a "gatherer ... walked along the line of sheets, taking off one copy of each in turn, until he had gathered a complete copy of the book in sheets. The book was knocked smooth at the edges and laid down, together with any inserts such as plates and cancellantia, and the process was repeated ... until all the sheets were taken from one of the heaps. The books were then collated to ensure that each was made up correctly, and they were finally folded in half (2°, 4°, and 8° across the longer side, 12° across the shorter side), pressed, and baled up for delivery or storage" (pp. 143–144). As noted above, Gaskell says that in long books this folding in half occurred "in quires of 12–24 sheets" (p. 144n.).


Correspondence, pp. 416–417. Richardson called for another cut affecting pp. 163–164 (N2r-N2v), and neither the quoted words beginning that section nor ending it remain.

Conclusions and Speculations

The discovery that sheets of The Centaur were folded and collated while the printing went forward leads to the cautious supposition that at least a few copies of books with cancels are likely to contain offset from the cancellanda (and that such instances should be sought in books known to contain cancels). Conversely, the presence of offset might indicate that cancellation has occurred. It stands to reason that the wetter the sheet (the more recently printed it was) the more prone it was to receive and transfer its image to the adjacent page. It would seem likely that offset is, as in this edition, more common in cancellantia and partials involving preliminary and final leaves than the regular sheets of the book, for these were usually printed late in the run and folded closer to the time of printing, rushed by the desire to publish, and thus folding occurred while sheets tended to be damper and more vulnerable to offsetting. Also, the cancellantia and partials were damper when placed into the sorted stacks of sheets for particular copies. Much offset can be identified as transferred while in a pre-bound stack of the folded sheets and not just during the folding of the sheets. For instance, in the May1 copy, T8v’s and U2r’s offsets are cast in the top margin on U1r and U1v respectively, and cancellans U1’s are cast in the bottom margins on T8v and U2r; the alignments are 10 mm above and below present position. Similarly, in O-HF R2v’s offset on R3r is too low to have been caused after the sheets were bound—the offset from R2v’s footnote is 3 mm lower than it is currently positioned in this contemporary-bound copy. Offset transference during the stacking period is sometimes apparent from offset’s being laid down obliquely. In O, B6v’s offset runs obliquely across B7r as not folded properly when stacked (the leaves are bound with lines parallel in the contemporary-bound copy, and the offset of B7v on B8r is straight and parallel to the bound book). Also, offset exchanged between sheets not adjacent, as between Y8v and Aa1r in CaOHM, must reflect exchanges during the stacking.

It seems likely that offset exchange is caused more by moist paper than by wet ink (ink was expected to set, not dry), either because the wetness of the transferring paper had somewhat dissolved the ink or, more likely, because the moist receiving paper was more absorbent (that paper was more absorbent when wet is the reason for wetting the sheets prior to printing). Nonetheless, the role of ink should not be entirely discounted. The amount of offset in sheets of The Centaur varies between copies as if partly the product of the amount of ink applied to the type-forme. The degree of inking on the same forme varied not only between the impression of one sheet and the next but across the forme during a single


Page 214
impression—thus, on a page one might see much bleed-through at the top of a leaf and none at the bottom. Sometimes a page-setting was inked enough that the printed page threw offset on other pages when it was folded, when it was stacked, and after it was bound; thus, in CaOHM we find R2v cast offset on N1r, also exchanged reciprocally with Bb8v, and R3r (similarly, CSt's R2v discharged offset to both N2v and Aa1r). Also, the sizing must affect the extent and depth to which a sheet absorbed ink.12 The thickness of a paper stock must affect the degree that moisture is retained (parenthetically, the thicker the paper, the less interference of bleed-through from the opposite side of the sheet). In The Centaur, all partials and cancellantia have the same paper stock as that found in most sheets of the book, one with an "NH" watermark.13 Conclusions about the impact of paper are difficult in part because even within the dominant paper stock the thickness of leaves varies considerably: sometimes one can look through the eighth leaf of a gathering to read the next signature and sometimes one cannot. Another consideration relative to moistness of the paper involves whether the two sides of a sheet were equally dry—if sheets were hung folded over a line as in many depictions, then one side might remain damper and more prone to the exchange of offset, something I have not investigated.

This study can offer a tentative answer to the question whether a damp leaf was more prone to discharge (express) or to receive (absorb) offset. Certainly in exchanges between cancellantia pages N1r, N2v, R2v, U1r, and U1v with M8v and other pages of full sheets that these faced, the cancellantia much more often discharged than received offset (as shown in the table below). The only unreciprocated exchanges between M8v and cancellans N1r involve M8v’s receiving N1r’s offset (from CaAEU and twelve other copies); the only unreciprocated offset between cancellans R2v and R3r involve R3r’s receiving R2v’s offset (from CaOHM and six other copies). Also, N2v received offset from Aa1r only in eight copies with reciprocal exchanges (CaAEU, ICN1, LdU-B, May3, May4, MB, NRU, and Ob), but it transferred its offset to Aa1r in six others (BrP, DLC, KU, May2, NIC, and RPB); in no copies did I note offset moving only from Aa1r to N2v (and, since N2v was a cancel, I would have consistently scrutinized it for offset). Similarly, I noted Aa1r’s receiving offset from R2v (CSt, IU, and OAU) but never did I find R2v’s receiving offset from Aa1r. (The disproportion in these exchanges suggests that the N and R cancellantia were printed well after sheet Aa.) Leaves B1.B8 and B4.B5 from the original printing of the B sheet were often stacked facing the B cancellantia; offset from such reveals the later printed cancellantia more


Page 215
often discharged offset to the early printed leaves than received it. For instance, although B1r has reciprocal exchanges with B7v (LdU-B) and B7 + 1v (KU), the only non-reciprocal exchanges involve its receiving offset from B2v (NIC), B7v (CtY, IU, ViWCF), and B7 + 1v (ICN1). B5v unreciprocally transferred offset to B8r in eight copies, yet it never did so to a page of cancellantia; it reciprocally exchanged with B6r or B7r in five copies and received offset from these pages in five others.

Exchanges between the frontispiece and other leaves also suggest damp leaves were more prone to transfer offset than to receive it. It seems likely that the frontispiece was produced outside of Richardson’s shop and, even more likely given the delays caused by Young’s revisions, well before publication (Young had asked Richardson about 11 August 1754 to have it cut [Correspondence, p. 411]). Other than a reciprocal exchange in MH2, no text leaf has been seen to have absorbed the image of the frontispiece, yet the frontispiece received much offset, not only from the title-page A1r (L, NBiSU, NjP) but from B3v in CtY, X8v in MR2, Aa8v in LdU-B, and Bb8v in DLC and RPB. Not surprisingly, the blank recto of the frontispiece received much offset: from B7 + 1v in CaOTU, CLU-C, MH2, O, PSt, and TnU; O8v in NBiSU; Bb1r in LdU-B; Ccir in DLC and RPB; and Dd1v in L, E, and MiU. It is noteworthy that with one exception (NBiSU) only leaves from the final gatherings printed transferred offset to the frontispiece. In NBiSU the frontispiece verso received offset from A1r, and P1r received offset from Dd1 v; thus, partials and cancels stacked close to O8 and the frontispiece probably provided moistness for the irregular exchange.

But the vagaries about inking and paper cannot alone explain which copies have offset. It must be significant that, if there is offset in one of the three printed units with partial gatherings and cancellantia (singletons-Cc sheet, the B cancellantia half-sheet, and the N1.N2 and R1.R2 folds), there is offset usually in the other units: that is true of twenty-four of twenty-eight copies with more than one offset exchange between a page of the partials and cancellantia and pages not adjacent to them, and the remaining four have offset exchanges within two of the units. (The exceptions to this pattern are the five copies noted in footnote 4 that have a single exchange, four of which involve the sheet including the title-page, Dd1, and the cancellantia B3 and U1 — very possibly the final sheet printed.) The frequency of offset (implying dampness at folding) on the three different sheets suggests they were printed concurrently, something possible at a shop with as many presses as Richardson’s and supported by the press figures: the sheet with Cc4 and the partials has press figures 1 on Cc3r and 4 on Cc3v; the half-sheet with B cancellantia has figure 7 on B7 + 1r; and cancellans R2r has figure 2. Also, the propensity either for all the partials in a copy to discharge and receive offset or for none of them to do so confirms what the paper evidence offered in my "Interrelating the Cancellantia" indicated: that all the replacement leaves in a specific copy were from the same original sheets (damp sheets, as distinguished from ones allowed to dry longer). The proportion of copies with that propensity suggests that half the copies were bound up soon after printing. It is easy to imagine the order was to "bind half," but perhaps the interruption for a weekend might have created sufficient time for sheets to dry thoroughly prior to folding. Since


Page 216
the second edition was felt likely enough to begin a revised setting prior to the publication of the first edition, no great time duration is likely to exist between those copies first bound up and those bound later.14


Leonard N. Rosenband, in discussing how paper was coated with sizing, notes that it is an important factor in the paper’s absorption (Papermaking in Eighteenth-Century France: Management, Labor, and Revolution at the Montgolfier Mill, 1761–1805 [Baltimore: Johns Hopkins Univ. Press, 2000], pp. 12–13).


All sheets have the same paper with "NH" watermark except for four sheets that have one of three paper stocks with fleur-de-lys watermarks centered on the chain-lines. All these three have chain-lines 25–25.5 mm apart but differ in the shape and the dimensions of the fleur-de-lys’s head, arms, and tail: 17–19, 39–42, 23–25 mm in sheet B; another with flattened head 15.5–17, 32–36, 23–26 mm in E and N; and a third with lanceolate fleur-head, 17, 40–44, 20–22 mm in Q and in many frontispiece leaves.


Copies Examined

Asterisks indicate offset exchanges with non-adjacent pages (other than the frontispiece) by at least one page of the partials and cancels. Leaf dimensions are in millimeters. The frontispiece (frt) is present unless "-frt" is noted. Irregularities in copies are indicated. Nearly all copies without binding information are in contemporary calf.

  • BirmU (Birmingham U.) rPR3782.C4 1755; 200 × 122; section title B8 bound before B1; rebound
  • *BrP (Brighton Central) Class 176; 195 × 123; cont[emporary] calf, rebacked C (Cambridge U.) N.17.25; 199 × 128; rebound; perhaps some offset overlooked
  • *CaAEU (U. of Alberta) PR3782.C39 1755; 194 × 125
  • *CaOHM (McMaster U.) B-623; 194 × 117–118; cont. calf, rebacked; lacking section title to Letter I
  • CaOTU (U. of Toronto) B-11 5583; 199 × 124; section title to Letter I bound before B1 CLU (U. of California, Los Angeles) PR 3782.C33 1755; 203 × 123; rebound in buckram
  • CLU-C (Clark Library) PR3782.C31; 202 × 124; section title to Letter I bound before B1
  • CoU (U. of Colorado, Boulder) Pettit 72; 205 × 128
  • Csj (St. John's College, Cambridge) P4.9.119; uncut, 208–216 [frt, 219] × 129–134; cont. calf spine with marble paperboards
  • CSmH (Huntington Library) 351165; 204 × 127
  • *CSt (Stanford U.) PR3782.C4; 194 × 119; with leaves N1-N2 bound after N8
  • *CtY (Yale U.) formerly Ik Y85 755, recatalogued as 1997/1114; 195 × 124–125; cont. calf, rebacked; some leaves on conservation stubs
  • *DLC (Library of Congress) PR3782.C4 1755; 195 × 122–123; cont. calf, rebacked E (National Library of Scotland) Newbattle 2053; 199 × 123
  • *FU (U. of Florida) 824.5 Y71c; 200 × 122
  • *ICN1 (Newberry Library) C.69.99; 194 × 123
  • ICN2 Y135.M6387; 195 × 124; cont. caf with spine label "YOUNG'S | WORKS"); bound before Comus: A Masque, 1750
  • *InU (Indiana U.) PR3782.C3 1755; 197 × 121; -section title to Letter I; cont. calf, rebacked
  • *IU (U. of Illinois) Nickell 318; 197 × 122
  • *KU (Uxs. of Kansas) C907; 194–195 × 124; -frt; cont. calf, rebacked L (British Library) 525.i.3; 200 × 124; rebound
  • *LdU-B (Brotherton Collection, U. of Leeds) Lt q Young; 198 × 124
  • *LU (U. of London) Sterling I [Young, E. 1755]; 194 × 123; cont. calf, rebacked May1 (James E. May) 195–197 × 124–25; -frt
  • *May2 196 × 124; cont. sheep, rebacked; acquired from David Strauss, 2006
  • *May3 197 × 122; cont. calf; from Church Green Books, 2008
  • *May4 193 × 124–125; cont. calf; from Cornell Books, 2009


Page 217
  • May5 193 × 117; cont. polished calf; from Francis Edwards, 2011
  • *MB (Boston Public) Defoe 30.755.YoBC; 195 × 124; section title to Letter I bound before B1; rebound
  • *MdBJ (Johns Hopkins U.) PO3782.C39 1755 [sic]; 199 × 118
  • MH1 (Harvard U.) Typ 705.55.892; 195 × 123–124; rebacked
  • MH2 *EC7 Y8460.755c; 201 × 125; section title to Letter I bound before B1; cont. calf, with arms of England stamped on front and back boards
  • MiU (U. of Michigan) PR3782.C4; 196.5 × 123
  • MoU (U. of Missouri) 828.Y8ce; 190 × 116; rebound
  • MR1 (John Rylands Library, U. of Manchester) 1110; 200 × 127
  • *MR2 R144667; 194 × 120
  • MR3 MAW CW 74; 200 × 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"; and "Charles Wesley junr | 1776"; title-page signed "CWesley 1756"
  • *N (New York State Library) N248.4 Y71; 196 × 125; -B6, replaced with cont. MS copy of the leaf; rebound
  • *NBiSU (SUNY at Binghamton) PR3782.C4 1755; 198 × 125
  • NcU (U. of North Carolina) PR3782.C4 1755; 200 × 122
  • NeU (U. of Newcastle)
  • *NIC (Cornell U.) PR3782.C3; 196 × 127; rebound
  • *NjP (Princeton U.) Ex3999.7.323.12; 197 × 121–122; cont. calf, restitched
  • *NNU1 (New York U.) PR3782.C4 1755; 198 × 121; cont calf; made up with A1 and perhaps also Dd1 from a sheet other than that on which the Cc4 present was printed, for both leaves A1 and Cc2 contain watermarks—Cc4 is regularly conjugate. However, because A1 and Dd1 lack any offset, all the offset noted for the singletons and Cc4 reflects the same printed unit: U1v and B3r exchanged offset, and U1r, left on the outside, exchanged offset with Cc4v.
  • NNU2 196 × 120; A1 signed ‘John LeGrand 1822"; rebound
  • *NRU (U. of Rochester) xPR 3782.C39m; 194 × 126
  • NStBU (SUNY at Stony Brook; PR3782.C4 1755; 198.5 × 123; rebound
  • O (Oxford U.) 141 j.106; 203 × 125–126; section title to Letter I bound before B1
  • Oa (All-Soul’s College, Oxford) qq.13.27; 196 × 121–22
  • *OAU (Ohio U.) PR3782.C4 1755x; 195 × 122
  • OAshU (Ashland U., Ohio) EL 159; 200 × 122; section title to Letter I bound before B1
  • *Ob (Baliol College, Oxford) 860.c.13; 200 × 124
  • Occ (Corpus Christi College, Oxford) 1704.Yo; 200 × 121 – 122
  • OCiU (U. of Cincinnati) PR3782.C4; 191 × 122
  • OCU (Ohio State U.) PR3782.C4 1755; 195 × 122; rebound
  • *O-EF (English Faculty Library, Oxford) XL.85.69 [Cen]; 197 × 125
  • O-HF (Harold Forster Collection, Bodleian Library, Oxford) Forster 51; 198 × 125
  • *Ok (Keble College, Oxford) 19116; 197 × 122; cont calf, rebacked
  • Ose (St. Edmund Hall, Oxford) 204–05 × 124–25; section title to Letter I bound before B1
  • OTH (Heidelberg College, Ohio) PR3782.C4 1755; 199 × 126; cont. calf, rebacked
  • *Owo (Worcester College, Oxford) B.B.2.6; 196–97 × 124; cont. calf; "MISCELLANY" on red spine label; bound with John Leland’s Reflections on the Late Lord Bolingbroke’s Letters on the Study and Use of History, 1753; James Hervey’s Remarks on Lord Bolingbroke’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 × 123–124; section title of Letter I bound before B1
  • *RPB (Brown U.) YQE Y8 Lamont; 196 × 121; cont. calf, rebacked
  • TnU (U. of Tennessee) PR3782.C4 1755; 201 × 127; section title of Letter I bound before B1

  • 218

    Page 218
  • TxU (U. of Texas, Austin) AK Y858 755c; 200 × 124–25; cont. calf, rebacked; section title to Letter I bound before B1
  • *ViWCF (Colonial Williamsburg Foundation) PR3782.C4 1755; 197 × 123
  • *WNs (Winchester College) 199 × 125; cont. manuscript addition correctly identifying footnote "*Sir S J" on p. 237 as to "Jeremiah Sambrooke of Gubbins in Hertfordshire"
  • WU (U. of Wisconsin, Madison) PR3782.C4; 197 × 122

The second edition, sharing some type set for the first edition’s B sheet (and most of its title-page setting), appeared forty-four days after the first edition, on 17 April 1755 (London Evening Post); it was being printed as early as 13 February (Correspondence, p. 420).


Offset Evidence in Copies Examined

This list is organized by the position of leaves within the book. It records offset transferred to other pages (>), reciprocal exchanges of offset with other pages (><), and offset received from other pages (<). To reduce the inherent redundancy in "transferred to" and "received by" entries, I have not duplicated information that would be repeated within a few lines. Exchanges between pages facing one another in a book have been excluded except for transfers involving cancellantia, cancellanda (when so labeled), and partials. As a caveat, it need be said that I was more zealous in searching for offset on the frontispiece, cancellantia, and partials than on the leaves of regularly printed gatherings, especially in copies examined early in my research.


B7 + 1v  BirmU, CaOTU, CLU-C, MH2, O, PSt, TnU 
O8v  NBiSU 
Bb1r  LdU-B 
Cc1r  DLC, RPB 
Dd1v  E, L, MiU 


><  A1r  MH2 (post-binding, presumably, an exception to the rule that the frt never transfers offset) 
A1r  FU, L, NBiSU, NcU, NjP, PSt (usually post-binding) 
B3v  CtY 
X8v  MR2 
Aa8v  LdU-B 
Bb8v  DLC, RPB 

A1r/ title

U1v  CaOHM 
><  B3v  CSt, Owo 
><  U1v  BrP, CtY, FU, InU, KU, LdU-B, LU, May2, May3, May4, MB, MR2, NRU, Ob 
><  Bb8v  OAU 
><  Cc4v  DLC, RPB 
><  Dd1v  NIC, NjP 
U1v  O-EF 
Cc4v  Ok 


><  B8v  CLU-C, PSt, TnU (due to binding B8 before B1) 
><  U1r  DLC, OAU 
><  Cc2r  BrP, May2, May4, MB 
><  Dd1r  CSt, CtY, LdU-B, May3 
U1r  ICN1, RPB 
Cc1r  NRU 
Cc2r  InU (reciprocity not checked), KU, O-EF 
Dd1r  CaOHM, FU, Ok, Owo 


Page 219


18v  RPB 
M8v  NNU1 
><  B7v  LdU-B 
><  B7 + 1v  KU 
><  L8v  DLC 
B7v  CtY, IU, ViWCF 
B7 + 1v  ICN1 


><  B2r  DLC, MH2, NBiSU, OAU 
><  B4r  CaOHM 
B2 r  CLU-C, May4, OCiU, Owo 
B4r  CSt, FU 


B8v  Ob, MB (B8, section title, is bound before B1) 
><  B6v  CSt, CtY, IU, KU, LdU-B, NIC, ViWCF 
G8v  KU 


><  B6r  BrP, CaOHM, DLC, FU, ICN1, InU, KU, May3, May4, MB, NjP, NNU1, NRU, OAU, RPB 
B6r  May2, Ob 
><  B7 + 1r  CSt, IU, LdU-B 
B7 + 1r  CtY, O-EF, ViWCF 


U1v  CaAEU 
Dd1v  LU, May3, MR2 
><  U1v  CSt, DLC, ICN1, MdBJ, NIC, NjP, NNU1, OAU, Owo, RPB 
><  Dd1v  BrP, CtY, FU, InU, KU, LdU-B, May2, NRU, Ob, O-EF 
Dd1v  MB 


frtv  CtY 
U1r  LU, MR2 
><  A1r  CSt, Owo 
><  U1r  CaOHM, NRU, Ob 
><  Cc4r  InU, KU, May2, May4, MB 
><  Dd1r  DLC, ICN1, NIC, OAU, RPB 


B1v  CaOHM, CSt, FU 
><  B6v  DLC, NRU, OAU, RPB 


B7r  O-HF (ll. 3–6, postbinding: caused by dog-eared fold of B5-B6 downward over B6v, causing B5 to face B7) 


B8r  CtY, KU, May2, May3, May4, MB (B8 is bound before B1), NNU1, ViWCF 
><  B6r  NBiSU 
><  B7r  DLC, NRU, OAU, RPB 
><  B8r  CSt 
B6r  CaOHM, MH2, ViWCF 
B7r  N, Owo 
B8r  FU 


><  B2v  BrP, CaOHM, DLC, FU, ICN1, InU, KU, May3, May4, MB, NjP, NNU1, NRU, OAU, RPB 
><  B7v  NIC 
B2v  May2, Ob 


Page 220


B7r  CaOHM, MH2, O (laid down obliquely in O) 
><  B2r  CSt, CtY, IU, KU, LdU-B, NIC, ViWCF 
><  B4r  DLC, NRU, OAU, RPB 
><  B7r  FU, L, MB 
B6v  O-HF (see B5r


B5v  N, Owo 
><  B5v  DLC, NRU, OAU, RPB 
><  B7 + 1v  CSt, IU, LdU-B, NIC, 
B7 + 1v  CtY 


B1r  CtY, IU, ViWCF 
><  B1r  LdU-B 
><  B6r  NIC 
><  B7 + 1r  BrP, CaOHM, DLC, ICN1, KU, L, May2, May3, May4, MB, N, NBiSU, NjP, OAU 

B7 + 1v/xv

B2v  CtY, ViWCF 
L8v  NIC 
><  B2v  CSt, IU, LdU-B 

B7 + 1v/xvi

frtr  BirmU, CaOTU, CLU-C, MH2, O, PSt, TnU 
B1r  ICN1 
B7r  CtY 
I1r  MB 
K1r  RPB 
><  B1r  KU 
><  B7r  CSt, IU, LdU-B, 
N1r  NNU1 


><  B5v  CSt 
B5v  CtY, KU, Maya, May3, MB (B8 bound before B1), NNU1, ViWCF 

B8v/xviii blank

B2r  Ob, MB 
L1r  CtY 
N1r  Owo 


B2r  KU 


B7 + 1v  MB 


B1r  RPB 


B7 + 1v  RPB 


B8v  CtY 


><  B1r  DLC 
B7 + 1r  NIC 


B1r  NNU1 
N1r  CaAEU, CtY, cancel-DLC, IU, NIC, landum ViWCF 
N1r  CaAEU, CaOHM, cancel-FU, L, May2, lans May3, MH2, NBiSU, NcU, NIC, NRU, PSt, TnU 




Page 221
B7 + 1v  NNU1 
B8v  Owo 
><  R1v  CSt, IU, NNU1, OAU, ViWCF 
><  R2v  BrP, CaAEU, CtY, DLC, FU, ICN1, KU, LdU-B, LU, May2, May3, May4, MB, MR2, NIC, NjP, NRU, Ob, O-EF, RPB 
R2v  CaOHM 

N2v cancellandum/164

N3r  DLC, OAU 

N2v cancellans/164

N3r  BirmU, FU, May2, May3, MH2, N, NBiSU, O, PSt (much), TnU 
Aa1r  BrP, DLC, KU, May2, NIC, RPB 
><  N3r  NcU, NRU 
><  R1r  CaOHM, Owo 
><  R2r  CSt, IU, NNU1, OAU, ViWCF 
><  Aa1r  CaAEU, ICN1, LdU-B, May3, May4, MB, NRU, Ob 


frtr  NBiSU 


Dd1v  WNs 
><  Dd1v  NBiSU 

R1r cancellandum

Q8v  BrP, CaOHM, CtY, DLC, FU, ICN1, IU, LdU-B, May2, May3, May4, MB, NIC, NjP, NRU, OAU, Ob (cancellandum text fully legible in CtY, DLC, LdU-B) 

R1r cancellans/225

Q8v  BirmU, BrP, CaAEU, CaOHM, Csj, CtY, E, IU, L, May2, May3, May4, MB, MdBJ, MH2, N, NBiSU, NcU, NIC, NRU, O, Oa, Ob, OCiU, O-EF, O-HF, OAU, PSt, TnU, ViWCF 
Z8v  MB, NRU, RPB 
><  N2v  CaOHM, Owo 
><  Z8v  CaAEU, CSt, CtY, DLC, ICN1, IU, LdU-B, May2, May3, NIC, NjP, OAU, Ob 
Z8v  FU, InU, LU, MR2 (reciprocity not checked in InU, LU, MR2) 


><  N1r  CSt, IU, NNU1, OAU, ViWCF 


><  N2v  CSt, IU, NNU1, OAU, ViWCF 

R2v cancellandum/228

R3r  OAU 

R2v cancellandum/228

N1r  CaOHM 
R3r  CaOHM, CtY, FU, MH2, N, PSt, TnU 
Aa1r  CSt, IU, OAU 
><  N1r  BrP, CaAEU, CtY, DLC, FU, ICN1, KU, LdU-B, LU, May2, May3, May4, MB, MR2, NIC, NjP, NRU, Ob, O-EF, RPB 
><  R3r  CLU-C, CSt, L, May1, MoU, NcU, O 
><  Bb1r  ViWCF 
><  Bb8v  CaOHM 


Page 222

U1r cancellandum/273

T8v  CtY, DLC (much), FU 

U1r cancellans/273

A1v  ICN1, RPB 
T8v  May1 (prebinding: offset is io mm below of present position of U1r; similarly U2r>U1v at point 10 mm below U2's current position), May2, (either cancellandum or cancellans: CaOHM, May3, May4, MB, MH2, NRU, OAU) 
><  A1v  DLC, OAU 
><  B3v  Ca  OHM, NRU, Ob 
><  T8v  L, MoU, NBiSU, NcU, PSt 
><  Cc1v  BrP, InU, KU, May2, May4, MB 
><  Cc4v  CSt, CtY, FU, May3, NIC, NNU1 
B3v  LU, MR2 
Cc1v  O-EF 

U1v cancellandum/274

U2r  DLC 

U1v cancellandum/274

A1r  O-EF 
U2r  CSt, FU, May3, May4 
><  A1r  BrP, CtY, FU, InU, KU, LdU-B, LU, May2, May3, May4, MB, MR2, NRU, Ob 
><  B3r  CSt, DLC, ICN1, MdBJ, NIC, NjP, NNU1, OAU, Owo, RPB 
><  U2r  CLU-C, CoU, CtY, L, MdBJ, NcU, PSt, TnU (all post binding) 
A1r  CaOHM 
B3r  CaAEU 


frtv  MR2 


Aa1r  CaOHM (sheet Z was not in its proper position in the stacked sheets) 


R1r  FU, InU, LU, MR2 
><  R1r  CaAEU, CSt, CtY, DLC, ICN1, IU, LdU-B, May2, May3, NIC, NjP, OAU, Ob 
><  Aa1r  NBiSU, NcU (suggesting nothing was inserted between them during the stacking) 
R1r  MB, NRU, RPB 
Aa1r  PSt 


><  N2v  CaAEU, ICN1, LdU-B, May3, May4, MB, NRU, Ob 
N2v  BrP, DLC, KU, May2, NIC, RPB 
R2v  CSt IU, OAU 
Y8v  CaOHM (sheet Z not inserted in its position within the stack) 


frtv  LdU-B 
><  Bb1r  NBiSU, PSt (also Aa8v<Bb1r in MH2, suggesting nothing was inserted between them during the stacking) 


Page 223


frtr  LdU-B 
><  R2v  ViWCF 


frtv  DLC, RPB 
Cc1r  BrP, ICN1, KU, May2, May3, MH2 
><  A1r  OAU 
><  R2v  CaOHM 
><  Cc1r 
><  Dd1r  NRU 


frtr  DLC, RPB 
A1v  NRU 
Dd1v  OAU 


U1r  O-EF 
><  U1r  BrP, InU, KU, May2, May4, MB 


A1v  InU, KU, O-EF 
><  A1v  BrP, May2, May4, MB 


Dd1r  CaAEU, KU 
><  Dd1r  BrP, InU, May2, May4, MB 


><  B3v  InU, KU, May2, May4, MB 


A1r  Ok 
><  A1r  DLC, RPB 
><  U1r  CSt, CtY, FU, May3, NIC, NNU1 
><  Dd1r  CLU-C, L, May2, MH2, PSt (post binding) 
Dd1r  May1 


A1v  CaOHM, FU, Ok, Owo 
Bb8v  NRU 
><  A1v  CSt, CtY, LdU-B, May3 
><  B3v  DLC, ICN1, NIC, OAU, RPB 
><  Cc3v  BrP, InU, May2, May4, MB 
Cc3v  CaAEU, KU 


frtr  E, L, MiU 
B3r  MB 
><  A1r  NIC, NjP 
><  B3r  BrP, CtY, FU, InU, KU, LdU-B, May2, NRU, Ob, O-EF 
><  P1r  NBiSU 
B3r  LU, May3, MR2 
P1r  WNs 
Cc1r  OAU 


Page 224

Many librarians answered questions about copies, some kindly providing reproductions, among whom I wish to particularly thank Jeffrey Barr, Margaret Foley, Melissa Mead, John Mustain, Richard Noble, and Bruce Swann. I thank my eldest son, Erich May, for help inspecting copies, Shef Rogers for suggesting that I use gloss the word "offset" with "set off" and employ the term "bleed-through," and David Vander Meulen for advice on recording offset and for much rephrasing, correction, and regularization.


"Interrelating the Cancellantia and Partial Gatherings in the First Edition of Edward Young’s The Centaur Not Fabulous," Studies in Bibliography, 53 (2000), 241-263; hereafter cited as "Interrelating the Cancellantia."


Allen Hazen, "The Cancels in Johnson’s Journey, 1775," Review of English Studies, 17 (1941), 201-203; William Todd, "The Printing of Johnson’s Journey (1775)," Studies in Bibliography, 6 (1954), 247-254; Peter W. M. Blayney, "Introduction to the Second Edition," in Charlton Hinman, ed., The First Folio of Shakespeare, 2nd ed. (New York: W. W. Norton, 1996); and my own study "Cancellanda in the First Edition of Steele’s ‘Poetical Miscellanies,’" Papers of the Bibliographical Society of America, 82.1 (1988), 71-82.