University of Virginia Library



Portions of this study derive from chapters 1 and 2 of my unpublished doctoral dissertation, entitled Procedures for Determining the Identity and Order of Certain Eighteenth-Century Editions (University of Chicago, 1949).


"Press Numbers as a Bibliographical Tool: A Study of Gay's The Beggar's Opera, 1728," Harvard Library Bulletin, III (1949), 198-212. For a further consideration of the problems discussed by Mr. Knotts, see my review in Philological Quarterly, XXIX (1950).


I have adopted the term "press figures" in deference to original contemporary usage in the printers' manuals. Perhaps a more accurate signification would be "pressman's mark," for, as I shall endeavor to show, the symbol identifies the man rather than the press, and exists in the form of letters and numbers as well as figures. Since this term is of my own coinage, however, and since it might be confused with "printer's mark" or "press mark," both of which have other denotations, I use the customary reference.


Cf. Ronald B. McKerrow, An Introduction to Bibliography (1927), pp. 81-82; R. W. Chapman, "Printing with Figures: A Note," The Library, 4th ser., III (1922-23), 175-76; and the latter's citation from several printers' manuals.


C[aleb] Stower, The Printer's Grammar (1808), pp. 386, 376.


Ellic Howe (ed.) The London Compositor (1947), pp. 70, 72; Stower, op. cit., p. 418.


Howe, op. cit., p. 31.


Bernard Mandeville, The Fable of the Bees, ed. F. B. Kaye (1924), II, 395.


Knotts, op. cit., p. 202.


This estimate is based on the statistics cited in Howe, op. cit., p. 133.


See Stower, op. cit., p. 433 for tables indicating the charge in hours as estimated from the number of sheets printed and perfected.


Tabulations of the figures for the Spectator, Fable of the Bees, and Temora are entered in the appendix. For the Fable of the Bees it would appear that half press is indicated by a figure on one forme of a sheet, full press by a figure on both formes. Note that when 8 is operating at full press (I-L or O-Q) it accomplishes almost twice as much as 7 and 5 at half capacity (M-N).


This explanation also suffices for Colley Cibber's The Careless Husband (Tonson, 1735. 120, A-D12 E6) and for Steele's Dramatic Works (Lintot, 1723. 120, A-L12 M6). Both have two sets of headlines and numerous figures.


Though of the shape and size of an octavo, this play has horizontal chainlines (so I am informed) and thus collates A 4 B-K4. My examination has been confined to a facsimile copy of MiU reproduced by the Augustan Reprint Society (Series V, No. 3, 1949).


Other instances of what seems to be an unnecessary reiteration of the same figure, but here throughout the impression and only for one forme of each sheet, appear in the first editions of Peacock's Nightmare Abbey (1818—sheet H unfigured, K figured twice) and Coleridge's Fears in Solitude (1798).


The figure cited is that for the year 1668. Howe, op. cit., p. 33. Though the Act limited the master printers to twenty, there is no evidence that they were reduced to that number at any time during the period that the law was on the books.


Cf. an anonymous report to Francis Place, cited in Howe, op. cit., p. 133.


Where the figures are apparently used for this latter purpose, as in the Fables, the 4th edition of the Fable of the Bees, and in the 6th and 7th volumes of the Spectator, the time represented may be a week. For these books it is plausible that the men completed, in a six day period, three or more sheets at full press, three or less at half press. The evidence is so ambiguous, however, and so variously recorded by the pressmen as to be almost worthless.


It has been suggested to me that the figures might identify the compositors as well as the pressmen, and thus serve as a means for insuring the return of the wrought-off formes to the proper persons for distribution. This would be a plausible assumption were it not for the fact that the figures are highly variable, not present in some formes of an impression, constant in others, shifting in still others, and changing in every forme of a reimpression. Thus for any one setting of type, presumably the work of a single compositor, there may be in the course of its use as many as eight different figures, all identifying the succession of men at press, but none the man who composed the type. An exceptional case is evident, however, in the initial gathering of Mandeville's Fable of the Bees. Part II (1730, 120—the second edition?), where a figure ‡ appears at the foot of every page from A7 to A11v. Professor James L. Clifford has offered what seems to be the only reasonable explanation for this phenomenon. An apprentice has composed these ten pages and is held responsible for whatever correction is necessary and for the eventual distribution of the type.


This comment should be qualified by the observation that pirates may infrequently attempt to duplicate the figures of their copytext as well as the ornaments and typography. Cf. Dr. Giles E. Dawson's article on "Three Shakespearian Piracies" in Papers of the Bibliographical Society of the University of Virginia, I (1948), 47-58.


For examples other than those noted in this article, see for (1) and (2) Greville Worthington, A Bibliography of the Waverley Novels (1931), p. 12 passim; for (3) Kaye, op. cit., II, 394-95. Occasionally the figures remain constant within an impression but move within the forme. Thus in Gay's , figure 7 appears in some copies at M3v (NN-Berg [2]), in others at M1v (MB, NN), though the headlines and letterpress are unaltered. Similarly, figure 3 in the second volume of Fielding's Amelia (120, 1752) may appear at position F6v (CtY) or F7 (CtY, ICU, NN-Berg[2]). Presumably these figures were extracted by the inkballs and replaced.


The only positive evidence would be that of the headlines, which would reappear according to pattern in both of the variants resulting from the circumstances described in (1), but only in the first of the variants envisaged under (2).


Kaye, op. cit., II, 394-95.


The disclosure of these editions requires supplementary information, which is in preparation for future publication. Of the several editions of the First Epistle, the first is unrecorded, and the second consists of a single impression, not two, as Griffith believes (items 458, 467). Either edition of the Second Epistle corresponds to the general description offered by Griffith under item 447.


Throughout this paper my terminology approximately conforms, I trust, to the several definitions set forth in Dr. Bowers' Principles of Bibliographical Description (1949). From the available evidence concerning planned overprints of certain eighteenth-century books, I am, however, unwilling to abandon McKerrow's classification of these as "editions" and to assume with Dr. Bowers (p. 48) that they represent states. For books printed before 1695 it is no doubt true that since legal restrictions operated to prevent an excessive issue of copies from a single setting, the necessary resetting for overprints "presumably occurred," as Dr. Bowers remarks, "during 'continuous printing' and concerns the text of only a small portion of the book." But the lapse of these restrictions brought about a modification of procedure which is evident in numerous books of a later period. In one, Gibbon's Decline and Fall (cited by Dr. Bowers [p. 48, n. 7] as an unusual case), the resetting admittedly involves only about one-third of the text; but this amounts to the impression of 500 sheets on fifty-four formes—or to some 13,500 perfected sheets, in all, which could hardly have been processed simultaneously with the work on the original issue. In a second the resetting, comprising 50% (160 pages) of the text, was certainly not undertaken until after the overprinting of the remainder, for the author was demanding publication of the first issue at the earliest possible date. In a third, resetting constitutes 60% (192 pages) of the text, and some of this was continuously impressed for the "third edition." And for others, as in the several editions of Pope's Epistles, and of The New Ministry cited in the text, printing apparently proceeded up to the last sheet or so before resetting or reimpression was begun. Quite obviously, then, the time-consuming process of recomposing and reprinting early sheets for most, if not all of these books must have been subsequent to the initial publication of the copies produced according to the original plan. Thus the evidence for this period is sufficiently complete, I believe, to deny the necessity and the practice of resetting during continuous printing, and to reject the designation of "states" for books in this category. Actually, if this designation must apply, it will be necessary to subsume under these "states" variant impressions and other "substates"; and this is quite impractical. I have, therefore, accorded such books the status of editions.


An instance of successive reimpressions in which the figures are dropped occurs in Gay's The Beggar's Opera. See Knotts, op. cit. More often, however, as Dr. Bowers reports for books printed in the sixteenth and seventeenth centuries, the entire type matter, excluding the headlines and quads, is tied together if intended for reuse. "Notes on Standing Type in Elizabethan Printing," Papers of the Bibliographical Society of America, XL (1946), 210. See further note 30 below.


Bowers, Principles, p. 110, n. 44.


See the respective author-bibliographies by J. E. Norton (p. 38 passim), S. L. Gulick (p. 97), W. M. Sale (pp. 68, 73-76), and Arthur E. Case, Bibliography of English Poetical Miscellanies 1521-1750 (1935), items 435(1)(a) and (1)(b).


Stower, op. cit., p. 384.


The composition of a work running to approximately six sheets was often completed before return of letter. Howe, op. cit., pp. 92, 361, n. 1. For all examples discussed in this section two assumptions are implicit: first, that sufficient chases were available in this period for retaining the pages of several formes, and secondly (as this may be demonstrated in the headlines of The Imposters and Alfonso), that the headlines are an integral part of the type-page, not of the forme, and thus are transferred from one impression to another.


Stower, op. cit., p. 386.


Ibid., p. 385.


Howe, op. cit., p. 381.


The editions of this pamphlet here identified as "A" and "B" represent two of some eight editions, only four of which are registered in Gulick's bibliography. "A" has not been previously recorded; "B" corresponds to Gulick No. 301.


The "second edition" is actually a reissue (reimpression) of the first. In both issues the figures appear on the same type pages, indicating, as mentioned before, that they were tied up with the letter-press between impressions. Where the figures are in the same position within the page that fact is noted by italics.


The circumstances of production for these and other unrecorded variants of Alfonso are so complex as to preclude analysis in this article. See my dissertation, pp. 123-24.


In the Princeton copy of the fourth issue the difference between the other gatherings and H-I is accentuated by a change in the paper from wove to laid.


The only exception I have seen, besides the two editions of Burnet's Reflections, is volume 7 of the Spectator. Cf. appendix.


The only copy I have seen (NN) collates 80, A4 B-T8 U8(—U8?) X-2R8 2S6 4A8 3B8. Dr. Case [op. cit., item 172 (6) (a)] correctly infers from the irregularity in the signatures and pagination that sheets N—P, 4A (described by Case as 3A) and 3B must have been printed before the rest of the book. Supporting evidence is represented by the figures for these sheets, and also by the paper, which is distinctly different from that used elsewhere. The figures and the paper also distinguish sheets B—D, but since these conform to the new system of pagination, they were doubtless printed after the others and simultaneously with those machined according to the later plan of imposition.


It should be noted, however, that although it is impossible to demonstrate the employment of two shops for this poem, at least two compositors were at work in different sections, one who spells the name "EVELINA" (in D8v-F1, G4v-M1v), and one who spells it "EVILINA" (in A4v, C2v-C4, F1v-G4v). Furthermore, this distinction roughly corresponds to that between the unfigured (A-G) and figured (H-N) portions of the text.


"Notes on Cancel Leaves," The Library, 4th ser., V (1924-25), 252.


If it is permissible to consider information not to be derived from the book—i.e., for the Appeal, the fact that the whole of sheet E was also cancelled [Thomas MacKnight, History of the Life and Times of Edmund Burke (1860), III, 407-9]—then the precise collation for this insignificant little pamphlet will read: 80, A 2[=K4.5] B8(± 'B5') C8(± C5) D8 (-D8+'D7') E8(±) F8 G8(-G8+'G7') H-I8 K8[-K4.5=A 2].


Dr. Gulick identifies the first edition (item 305) as existing in two states, one with and one without errata on H4v. Actually, as the figures show, there are two impressions, with the first in two states. The whole question of the bibliography and authorship of this and the other Hanover pamphlets attributed to Chesterfield is so complex as to require separate consideration elsewhere.


Regulation 3 for pressmen, Stower, op. cit., p. 386.


The figure cited for this work is the only one that appears, except for the reiterated figure mentioned in fn. 19.


The watermark, a twelve-pointed star measuring 3 cm. across, almost certainly identifies the paper as of Genevan origin, since the law in that locality specifically required the mark to be in the center of the sheet. Cf. C. M. Briquet, Les filigranes (Leipzig, 1923) II, 324, 349. Professor Allen T. Hazen, to whom I am indebted for this reference, also informs me that its appearance in this edition of the Voyage, printed in 1755, is most unusual, for it had been supposed that no Italian or Swiss paper was used by British printers after 1740. Corroborative evidence of the w/m position is to be found, first, in the other edition, which was printed just before this one [cf. Strahan's ledger entries, cited in J. Paul de Castro, "The Printing of Fielding's Works," The Library, 4th ser., I (1920-21), 257-70], on ordinary lily paper (except for A 4), and obviously according to imposition (a); secondly, in the preliminary gathering A 4, which is of the same setting of type in each edition and has, for both, half of the star w/m at the outer edge of A3, an impossible position for a mark normally placed; and thirdly, in the final gathering N6 of this edition, where half the mark appears at the outer edge of N5, again an impossible position for a normal mark.


If scholars would check the figures instead of haunting bookshops for "lost" editions, they might discover the copy they seek on their own shelves. Thus, to cite a trivial example, the unknown but presumed "second edition" of William Whitehead's Variety (1776, 4°) is disclosed as a corrected reimpression of the first.

Figures   Copies  
"Edition" A  7-7, 12-7, 15-6, 18-10  DCL IU MH PU 
"Edition" B (reimpressed)  7-9, 15-5, 18-10, 20-6  ICU NjP NN 
"Third Edition" (reimpressed)  7-6, 8-2, 15-5 18-10, 20-6  ICN 
Here again, the figures prove their usefulness, not only in identifying a variant, but in suggesting, prior to inspection, the kind and sequence of variants, and what would seem to be an overprinting of certain sheets.


The figures reveal, for instance, mixed sheets in MH copy of Akenside's Pleasures of Imagination NN (Berg) copy of Johnson's Taxation No Tyranny, and PU copy of Smollett's Humphry Clinker.