University of Virginia Library

I

The practice of using printed figures to keep a record of the work done by particular pressmen[13] was employed extensively in the 1790's by several of the most prolific printers in Philadelphia. Such men as David Hall, William Sellers, and Thomas Dobson were making use of these figures at the beginning of the decade and at its end. One should assume, therefore, since they issued Bibles, statutes of Pennsylvania, and encyclopedias, that a large body of material exists, from which a thorough survey of American press figures can some day be drawn. Only a few works will be examined here, to serve as illustrations of the practice of several different printers at various times throughout the decade. First, three books printed by Hall & Sellers (in 1790, 1791, and 1798) will represent the work of one firm at different times and in different formats; then three Bibles, each by a different printer and in a different format, will provide some comparisons; three books printed by Thomas Dobson will act as an introduction to his great work, the Encyclopaedia; and, finally, a quick look at the complications of Mathew Carey's Bibles will serve as a transition to the next century.

Hall and Sellers, whose names first appeared in joint imprints in 1766, issued in 1790 a duodecimo Book of Common Prayer and Whole Book of Psalms (listed separately in Evans but printed together with continuous signatures). This book, gathered in sixes, contains the figure 1 twenty-four times, in every case on $6v:

Little can be deduced from this meager example. One may merely observe that the figures, when they occur, are placed invariably on the last page of a gathering but that nearly half the quires (twenty-two) are unfigured. The figures alone do not here provide such unequivocal evidence of half-sheet imposition as they do in some other volumes, although the fact that 1 occurs on the same page of five consecutive gatherings (a-c, A-B) —and later in three (M-O) —strongly suggests half-sheet imposition (for in the case of two half-sheets worked together the same figure would not be expected to occur in the same forme of two consecutive quires). However, the pattern for most of the volume (particularly from O through Ss) is an alternation between figured and unfigured gatherings; assuming that a second pressman worked without a figure, this arrangement of figures is consistent with the results obtained by working two half-sheets together. Surely the lack of a figure, in this instance, does represent the work of a second press; but whether that press printed every other sheet by half-sheet imposition, or the inner formes of two consecutive half-sheets, is a matter impossible to determine on the basis of the figures alone. In other words, evidence from the figures suggests that at least gatherings a, b, c, A, F, G, M, N, Tt, and Uu are the product of half-sheet imposition but that the other thirty-six gatherings could conceivably be the result of either method of producing half-sheets.[14]

A second Hall & Sellers example, from the following year, illustrates a regular division of labor in folio:

   

Fig.	1v	2r	1r	2v	Totals
1	6	19	3	8	25(i)	11(o)	36

In this second Hall & Sellers volume, as in the first, the only figure that appears is 1, and again the unfigured formes clearly represent work by another pressman, since the usual pattern is for each sheet to consist of one figured and one unfigured forme (and whenever both formes of a sheet are figured, both formes of another sheet are unfigured, so that the balance is retained). The long series of sheets (from F through Q) in which the inner formes alone are figured suggests offhand that composition was keeping pace with presswork, while those sheets in which both formes (or neither forme) are figured suggest that at those points composition was lagging behind or that one press was required briefly for another job; the alternating pattern (as in Q-S, Y-Aa, Cc-Dd) could also indicate an imbalance between composition and presswork, if the inner or outer formes were consistently the first formes through the press; but, if this condition did not obtain, then obviously other explanations could fit. In the absence of precise information about the figures in other books printed at the same time in Hall & Sellers' shop, it is difficult to make any reliable deductions about the relations between composition and presswork or the size of the edition.

Of course, certain rough calculations can be made. If one accepts the estimate that a colonial compositor could set approximately 600 ems (1200 ens) per hour,[15] then a forme of this book (roughly 8000 ens) would have been nearly seven hours' work; if the output of the press was about a token an hour (250 sheets, not perfected),[16] then nearly 1750 copies of a forme could have been machined in that same time. If then the appearance of figure 1 in many consecutive inner formes is taken to mean that the minimum time for presswork on any one inner forme was the time required for composing the next inner forme, the result would be to estimate the edition at 1750 copies (or perhaps between 1500 and 1750); however, if only one compositor were attempting to supply both presses, the time available for presswork would be twice as long. On the other hand, the appearance of figure 1 in three consecutive formes, two inner and one outer (as in Bb-Cc), might suggest, if two compositors were working, an edition of only 750 to 875 copies, or, with one compositor, an edition of 1500 to 1750 copies

(the same as that postulated for two compositors above). It is not necessary to assume that the pressmen were machining each forme during the whole of the time available, for Benjamin Franklin commented on the fact that, given the small size of editions in America, pressmen did not have constant work, as compositors did, and "must often stand still";[17] or pressmen might work on other jobs, such as blank forms or broadsides. Press figures, in other words, will become of greater usefulness in these calculations as more complete data are assembled about the work in a shop at any particular time and as more precise time limits can be established through external evidence.

One interesting feature of the volume is the appearance of figures (three times) on $1r—that is, on the same page as the signature—contrary to customary practice in both England and America. The usual locations in Hall & Sellers folios at this time (at least as illustrated by this book) are the recto and verso of the second leaf of each gathering; it is impossible, at the present stage, to assign reasons for the deviations, since they do not occur simply at those places where $2 presents an unusual situation (such as short text) and since they may be the result merely of fortuitous circumstances that made it more convenient to insert the figure in one place rather than another.

One more book printed in this shop, another volume of session laws characteristically issued as a folio in twos, shows that seven years later the practice in placing figures had shifted somewhat:

As in the earlier volume, the only figure which appears is 1, but the regular position for figuring the inner formes has become $1v instead of $2r (although figures appear on $2r a number of times near the beginning of the volume). As before, figures do appear on the first page of a gathering, along with the signature (four times: F, 2E, 2H, and the single leaf 2I at the end). But, contrary to earlier practice, the division of labor (assuming the unfigured formes to be the work of a second press) is not generally

to assign the two formes of each sheet to separate presses; normally here (129 times) the figure appears in both formes, with occasional interruptions when both formes are unfigured. A pattern sometimes seems to emerge (as in 3Z-4L, where alternate gatherings are figured in both formes and the intervening ones entirely unfigured), but it does not last for long, and certain runs of doubly figured sheets are quite extensive (as 7L-7Y, 8A-8I, 9G-9S, 2A-H). Perhaps the most interesting feature of the figuring is the positions of the figures relative to the footnotes, although the differences in these positions probably do not distinguish different pressmen but are only indications of the shifting ease or difficulty of inserting figures at particular points. There are a great many long footnotes in the volume, and the figures are 108 times placed above the footnotes (that is, between the text and the note) and 25 times below. In ten instances also the figures occur near the right margin, the normal position being near the left.

Three Bibles, printed at various times through the decade, may serve to represent the figuring of three more printers. The first of these Bibles, the work of William Young, a prolific Philadelphia printer who began issuing books in 1785, shows the unusual situation in which the signature page is the preferred one for figures:[18]

Although only four sheets (Dd, Ff, Gg, Ll) are figured in both formes, the fact that the same figure appears in both formes in those instances suggests that, in the majority of sheets, the unfigured inner forme was machined by a different man from the figured outer one. Perhaps this pattern supports the suggestion that figures were not used (and indeed were superfluous) when presswork was proceeding according to some previously arranged schedule and were needed only when there were deviations from it.[19]

Young's Bible of the following year, though gathered differently (with alternating quires of six and twelve leaves), still has its figures (when they appear) on $1r.

Another instance of figuring on $1r and $2r furnishes, at the same time, an example of the large quarto Bible. The 1792 Bible printed in New York by Robert Hodge and Samuel Campbell contains fifty-five figures:

Since the major part of the book is not figured, and since figures do appear at scattered points throughout (as in 4M, 4N, and 2A, after the regular figuring stops with 3N), one may take the volume as a further indication that figures were sometimes used to record occasional assistance by a press not originally assigned to the job.

One further Bible—representing a third format, the large folio—may be chosen from the work of the Philadelphia printer Jacob R. Berriman:

Although $1r had been a standard position for figures in the two previous volumes, it is clear that Berriman avoided the signature page, for the outer forme is almost exclusively figured on $2v, while the inner is figured extensively on both $1v and $2r. No figures appear in this book until sheet O, but after O there are only 38 unfigured formes; if figures are used only for unscheduled piecework, then the bulk of the volume falls into this category—which may conceivably be the case, since the presswork was started without figures (the first thirteen sheets). Another large folio Bible of this time which exhibits a similar pattern of figures is the famous "hot press" edition printed by John Thompson and Abraham Small of Philadelphia in 1798. Not only are figures lacking in the first 36 sheets, but there are large gaps later in the volume; where figures do appear, they are normally on $2ra and $2va, as in the Berriman work, and again only the

figures 1 and 2 are used (but with more variation in position on the page, since they often come at the right side, or in the center, of the lower margin).

The final series of examples from this period represents the work of perhaps the most enterprising of the Philadelphia printers before Mathew Carey—Thomas Dobson. Although his name appeared in imprints in 1786 and 1787, it was not until 1788 that he began printing some of the books he promoted, and another year or two before he printed any considerable number; but by 1790 he was ready to embark on the great Encyclopaedia, which he brought to conclusion in 1798 and which furnishes an extensive illustration of the sustained use of press figures. Before turning to that work, one may glance at a few other figured books from his shop during the same years. Thomas Percival's Moral and Literary Dissertations illustrates Dobson's practice, in half-sheet duodecimo, of figuring the last page of a quire:

The only deviation from the pattern, the 1 on F1v, is explained by the fact that F6v is blank. An earlier volume, in half-sheet octavo, also shows the consistent choice of the last page of a gathering for figures, though the figures occur more sporadically:

The deviation on E4r is due to the fact that the text is short on E4v, although somewhat short texts on Ii4v and 3C4v did not prevent the placing of figures on those pages. Dobson occasionally issued separately some of the longer articles in the Encyclopaedia, and the same year (1792) saw the publication of such an extract in A Compendious System of Anatomy (Evans 24206); gathered in eights, it too is normally figured (with a 3)

on the last page of gatherings. And a crown quarto, published the year after the Encyclopaedia, shows resemblances to the method of figuring in the large work:

As in the Encyclopaedia, the figures generally appear on $4r and $4v, with more outer formes than inner figured; and, as in the last volume of the Encyclopaedia, the figures stop altogether halfway through.[20] Another Dobson quarto, also figured with 1, 2, and 4, and figured on $4r and $4v, is the 1796 printing of The Four Gospels, annotated by George Campbell.[21]

Dobson's crowning work is of course the eighteen-volume Encyclopaedia, the largest job that any American printer had attempted up to that time. It is a reprinting of the third edition of the Encyclopaedia Britannica, with revisions and additions (the article on "America," for example, was completely rewritten by Jedidiah Morse). Dobson took each part of the Britannica as it was issued, set his men to work on it, and managed to keep pace with the Edinburgh printers, for his final volume appeared only a few months after the Edinburgh one. His original plan was to issue the work in weekly parts of five gatherings each. Printing began late in 1789, and the first weekly issue appeared at the beginning of January 1790. The tenth issue, released in early March, finished the first half of Volume I, and with

that issue the practice of weekly publication ceased; from then on, a half-volume was issued approximately every ten weeks. Although separate title pages, all bearing the date 1798, were supplied for each volume upon completion of the entire work in 1798, it is nevertheless possible to date with some accuracy the appearance of each volume and to deduce the amount of time available for the printing of each. To have completed a work of such magnitude so expeditiously, and to have fallen so little behind schedule, Dobson must certainly have been an efficient and industrious manager. The whole production is generally considered the greatest achievement in American printing before the nineteenth century; in the words of Lawrence Wroth, the Encyclopaedia "marks the end of printing in America as a household craft and the beginning of its factory stage of development."[22]

It is appropriate that such a work should be one of the first American instances of the consistent and extensive use of a professional device for keeping efficient records—the press figure. And, as an example, it is nearly an ideal one, for it is large enough to provide plentiful data and prolonged enough to represent one printer's work during a period of more than eight years; the volumes are comparable as printing jobs, since they are parts of the same work, and the knowledge of the time taken to produce each one is important external information of a kind not always available. Dobson did not use press figures at the beginning of his task, probably believing that the division into weekly parts itself provided a convenient enough way of keeping records. On $4v of every fifth gathering of Volume I through 3D (the work is quarto) appears an indication of the weekly issues—that is, "No 1" falls on E4v, "No 2" on K4v, "No 3" on P4v, and so on. After sheet 3D ("No 10"), when the weekly issue ceased, these numbers still occur through the rest of the volume (except that numbers 11, 15, 19, and 20 are missing), but press figures also make a somewhat tentative appearance in seven formes. In Volume II Dobson is still not sure, for "No 22" appears on K4v and "No 23" on P4v, but after that figures occur consistently through the volume and the following fifteen volumes. Then, in the final volume, figures appear only in the first half (through 3D); beginning with 3E, the signature page also carries the designation "Part II," and the practice of using figures seems to have been abandoned.

If the dating of these volumes established by Evans (in 33676-93) is correct, it is interesting to note the way in which the number of different figures in each volume correlates with the supposed speed of production. In Volume II, the first one regularly figured, only the figures 1 and 2 are used, and this is true also of Volume III; these volumes may have required

more than the predicted twenty weeks each to print, since only two volumes per year were issued in 1790 and 1791. Then, according to Evans, Volumes V, VI, and VII came out in 1792 and VIII, IX, and X in 1793; the press figures increase accordingly to 1, 2, and 3 in IV and V, and to 1, 2, 3, and 4 in VI through XI. When the pace again slows down to two volumes per year in 1794, 1795, and 1796 (Volumes XI through XVI), the number of figures decreases to three (1, 2, and 4). And 1, 2, and 4 are also the figures in XVII (mid-1797) and XVIII (early 1798). Even though the work is not divided equally among these pressmen, it is in accord with common sense to find that more pressmen were at work on those volumes which were produced more quickly. Such would not be the case if some of the pressmen machined only one or two formes for a given volume in odd moments between other jobs (as in XIV, where 3 appears only once), but in general each figure that appears represents a substantial amount of work.

Assuming again that the colonial compositor could set 1200 ens an hour, it would have taken him about 32 hours to set one forme, for the Encyclopaedia's closely-packed double column pages contained about 9620 ens each. If a pressman, machining roughly a token an hour, worked for 32 hours on one forme, the result would be an edition of 8000 copies, surely an unwarrantably large number. Furthermore, if an edition of this magnitude were postulated, the volumes could not possibly have been printed in the envisaged time; for a pressman, machining approximately eight tokens a day, six days a week, would require more than 41 weeks to complete 62 formes—and figure 4 appears in 62 formes of Volume VII, one of the volumes supposedly produced most rapidly. There is no reason to suppose, therefore, that the pressmen engaged in work on the Encyclopaedia spent their entire time on it—in fact, the Dobson Four Gospels of 1796 contains the same figures (1, 2, 4) as those in the Encyclopaedia volumes of that year. Nor is there reason to suppose that the copy could not have been cast off accurately (given the printed copy of the Edinburgh edition), so that more than one compositor might work at once.

Continuing with the example of Volume VII for a moment, the time taken to produce that volume had to be great enough to allow for the machining of 62 formes, since that is the largest number machined by any one man (even assuming all the unfigured formes to be printed by one man, there are only 44 of them). Because Volume VII falls in the middle of that series of volumes seemingly published with greatest speed, it may be safe, for purposes of argument, to assume that the originally planned rate—for the whole volume—of ten formes a week (or a volume in twenty weeks) was actually achieved. If, then, a man worked steadily for twenty weeks to complete 62 formes, the size of the edition would be in the neighborhood of 3870 copies; but of course he may not have devoted so much time to this job. Perhaps a more accurate basis for calculation is to begin with the notion of issuing five sheets (ten formes) per week. Since the

figures in later volumes (after Volume I) indicate generally a division of labor between the inner and outer formes, one may then assume that a pressman was expected, at most, to machine five formes in a week, a figure that suggests an edition of 2400 copies.

All these numbers are, at this stage, only rough guesses based on various feasible ways of manipulating the data; it is not possible, without more knowledge of other work in Dobson's shop at the same time and other external information, to be much more precise. But such calculations do suggest the help which press figures will some day be able to give. And further possibilities may suggest themselves to anyone who examines carefully the patterns in an extended series of figures. The press figures in Dobson's Encyclopaedia offer an opportunity, unique in American printing, for doing this, and summaries of them are presented below:

10. Encyclopaedia; or a Dictionary of Arts, Sciences, and Miscellaneous Literature. . . . Philadelphia: Printed by Thomas Dobson [1790-]1798.

• Volume I. [1790]. Evans 22486, 33676. ICU. 4: π2 a2 b4 A-5H4, 408 leaves, pp. [2] [i-iii] iv-xiv, [1] 2-799 [800].
• Summary of figures: Press figures appear only in five quires, as follows:                

  	(i)	(o)
  Sig.	4ra	4va
  ----	----	----
  4D	3	-
  4E	-	3
  4G	4	1
  4H	-	4
  4I	2b	4

• Volume II. [1790]. Evans 22486, 33677. ICU, IEN-M. 4: π1 A-4C4 dddd-nnnn 2 4D-5H4, 421 leaves, pp. [2] [1] 2-576 [40] 577-799 [800].
• Summary of figures: 51 of 100 sheets figured in both formes, 9 figured in one forme, 40 unfigured; 1 of 10 half-sheets figured (dddd with 2).

         

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	0	0	1	14/2	0	0	1	33/5	17(i)	39(o)	56
2	0	0	10/1	22/2	0	0	0	19/1	35(i)	20(o)	55
	-	-	--	--	-	-	-	--	---	---	---
Totals	0	0	11/1	36/4	0	0	1	52/6	52(i)	59(o)	111

• Volume III. [1791]. Evans 23351, 33678. ICU. 4: π1 A-5I4, 405 leaves, pp. [2] [1] 2-806 [807-808].
• Summary of figures: 78 of 101 sheets figured in both formes, 22 figured in one forme, 1 unfigured.

         

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	0	0	0	20/5	0	0	1/1	67	25(i)	69(o)	94
2	1	0	3	51/1	0	0	0	28	56(i)	28(o)	84
	-	-	-	--	-	-	--	--	---	---	---
Totals	1	0	3	71/6	0	0	1/1	95	81(i)	97(o)	178

• Volume IV. [1791]. Evans 23351, 33679. ICU. 4: π1 A-4H4 4I2 4I*2 4K2 4K*2 4L-5G4 χ1, 398 leaves, pp [2] [1] 2-793 [794].

---

138




• Summary of figures: 50 of 97 sheets figured in both formes, 41 figured in one forme, 6 unfigured; 3 of 4 half-sheets figured (two times with 2, once with 3).

           

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	1	0	2/1	17	0	0	1	33/1	21(i)	35(o)	56
2	2/1	0	2	26/1	0	1	2/1	28/1	32(i)	33(o)	65
3	1	0	1	8	0	0	0	10	10(i)	10(o)	20
	--	--	--	--	-	-	--	--	---	---	---
Totals	4/1	0	5/1	51/1	0	1	3/1	71/2	63(i)	78(o)	141

• Volume V. [1792]. Evans 24300, 33680. IEN-M. 4: π1 A-4X4 4X*4 (-4X*4) 4Y-5I4, 408 leaves, pp. [2] [1] 2-720 [6] 721-807 [808].
• Summary of figures: 45 of 102 sheets figured in both formes, 55 figured in one forme, 2 unfigured.

           

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	0	0	0	0	0	1	0	4	o(i)	5(o)	5
2	3	0	5	24	0	2	3	46	32(i)	51(o)	83
3	2	0	5	17	0	1	2	30	24(i)	33(o)	57
	--	--	--	--	-	-	-	--	---	---	---
Totals	5	0	10	41	0	4	5	80	56(i)	89(o)	145

• Volume VI. [1792]. Evans 24300, 33681. IEN-M. 4: π1 A-5G4 5H1, 398 leaves, pp. [2] [1] 2-794.
• Summary of figures: 64 of 99 sheets figured in both formes, 35 figured in one forme.

             

Fig.	1v	2r	3v	4r	1r	2v	3r	4v		Totals
1	6	0	2/1	16	0	0	2	24	25(i)	26(o)	51
2	0	0	3/1	12/3	0	1	0	23/2	19(i)	26(o)	45
3	2	0	3/1	12	0	0/1	2	32/4	18(i)	39(o)	57
4	1	0	1	4	0	0	0	4	6(i)	4(o)	10
	--	--	--	--	-	--	-	--	---	---	---
Totals	9	0	9/3	44/3	0	1/1	4	83/6	68(i)	95(o)	163

• Volume VII. [1792]. Evans 24300, 33682. IEN-M. 4: π1 A-5H4, 401 leaves, pp. [2] [1] 2-797 [798-800].
• Summary of figures: 57 of 100 sheets figured in both formes, 42 figured in one forme, 1 unfigured.

             

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	0	0	0	5	0	0	0/1	22	5(i)	23(o)	28
2	3/1	0	2	10/1	0	0	1	18/2	17(i)	21(o)	38
3	1	0	1	16	0	0	0	9/1	18(i)	10(o)	28
4	1	0	3	21/1	0	1	2	33	26(i)	36(o)	62
	---	-	-	--	-	-	--	--	---	---	---
Totals	5/1	0	6	52/2	0	1	3/1	82/3	66(i)	90(o)	156

• Volume VIII. [1793]. Evans 25450, 33683. IEN-M. 4: π1 A-5G4, 397 leaves, pp. [2] [1] 2-89 98-799 [800] [=792].
• Summary of figures: 38 of 99 sheets figured in both formes, 60 figured in one forme, 1 unfigured.

             

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	1	0	0	11/1	0	0	0	7/2	13(i)	9(o)	22
2	1	0	2	4	0	1	0	27	7(i)	28(o)	35
3	0	0	4	8	0	0	1	20	12(i)	21(o)	33
4	1	0	1	13/1	0	2	1	24/3	16(i)	30(o)	46
	--	--	--	--	-	-	-	--	---	---	---
Totals	3	0	7	36/2	0	3	2	78/5	48(i)	88(o)	136

---

139




• Volume IX. [1793]. Evans 25450, 33684. IEN-M. 4: π1 A-5H4, 401 leaves, pp. [2] [1] 2-800.
• Summary of figures: 43 of 100 sheets figured in both formes, 56 figured in one forme, 1 unfigured.

             

Fig.	1v	2r	3v	4r	1r	2v	3r	4v		Totals
1	1	0	1/1	17	0	0	3/1	22	20(i)	26(o)	46
2	0	0	2	10/2	0	0	1/1	20/1	14(i)	23(o)	37
3	0	0	1	4	0	0	0	12/3	5(i)	15(o)	20
4	2	0	2	9	0	1	0	22/3	13(i)	26(o)	39
	--	--	--	--	-	-	--	--	---	---	---
Totals	3	0	6/1	40/2	0	1	4/2	76/7	52(i)	90(o)	142

• Volume X. [1793]. Evans 25450, 33685. IEN-M. 4: π1 A-S4 T2 T*2 U2 U*2 X2 X*2 Y2 Y*2 Z2 Z*2 Aa2 Aa*2 Bb-4H4 *4I2 4I-5F4 5G2, 397 leaves, pp. [2] [1] 2-620 619-788 [=790] [791-792].
• Summary of figures: 22 of 92 sheets figured in both formes, 66 figured in one forme, 4 unfigured; 3 of 14 half-sheets figured (with 2, 3, and 4).

             

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	1	0	0	0	0	0	1	13	1(i)	14(o)	15
2	1	0	0/1	6/1	0	0	0	14	9(i)	14(o)	23
3	1	0	0	7	0	1	0	7	8(i)	8(o)	16
4	5	0	2	14	0	2	1	31/1	21(i)	35(o)	56
	--	--	--	--	-	-	-	--	---	---	---
Totals	8	0	2/1	27/1	0	3	2	65/1	39(i)	71(o)	110

• Volume XI. [1794]. Evans 26943, 33686. IEN-M. 4: π1 A-5K4, 409 leaves, pp. [2] [1] 2-814 [815-816].
• Summary of figures: 23 of 102 sheets figured in both formes, 72 figured in one forme, 7 unfigured.

             

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	0	0	0	2	0	0	0	2	2(i)	2(o)	4
2	2	0	1	7/4	0	2	1	26/2	14(i)	31(o)	45
3	0	0	0	3	0	0	0	2	3(i)	2(o)	5
4	7/1	0	3	24	0	3/2	0	21/3	35(i)	29(o)	64
	--	--	--	--	-	-	-	--	---	---	---
Totals	9/1	0	4	36/4	0	5/2	1	51/5	54(i)	64(o)	118

• Volume XII. [1794]. Evans 26943, 33687. IEN-M. 4: π1 A-3S4 3T2 3T*2 3U2 3U*2 3X2 3X*2 3Y2 3Y*2 3Z2 3Z*2 4A-5A4 5B2 *5B2 5C2 *5C2 5D2 *5D2 5E-5H4, 401 leaves, pp. [2] [1] 2-799 [800].
• Summary of figures: 29 of 92 sheets figured in both formes, 57 figured in one forme, 6 unfigured; 15 of 16 half-sheets figured (three times with 1, eight times with 2, four times with 4).

           

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	2	0	2	4	0	1	2	18/1	8(i)	22(o)	30
2	3	0	5	9/2	0	1	1	23	19(i)	25(o)	44
4	0	0	10/1	4	0	1	0	22/3	15(i)	26(o)	41
	--	--	--	--	-	-	-	--	---	---	---
Totals	5	0	17/1	17/2	0	3	3	63/4	42(i)	73(o)	115

• Volume XIII. [1795]. Evans 28628, 33688. IEN-M. 4: π1 A-5H4, 401 leaves, pp. [2] [1] 2-799 [800].
• Summary of figures: 16 of 100 sheets figured in both formes, 76 figured in one forme, 8 unfigured.

           

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	0	0	2	0/2	0	0	2	19	4(i)	21(o)	25
2	2	0	2	3	0	3	3	19	7(i)	25(o)	32
4	1	0	4	7	0	0	3	35/1	12(i)	39(o)	51
	--	--	--	--	-	-	-	--	---	---	---
Totals	3	0	8	10/2	0	3	8	73/1	23(i)	85(o)	108

• Volume XIV. [1795]. Evans 28628, 33689. IEN-M. 4: π1 A-5H4, 401 leaves, pp. [2] [1] 2-797 [798-800].
• Summary of figures: 8 of 100 sheets figured in both formes, 77 figured in one forme, 15 unfigured.

               

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
0	0	0	0	0	0	0	0	1	0(i)	1(o)	1
1	1	0	0	1	0	1	0	19	2(i)	20(o)	22
2	0	0	3	2	0	2/1	2	14/1	5(i)	20(o)	25
3	0	0	0	0	0	0	0	0/1	0(i)	1(o)	1
4	1	1	0/2	2/1	0/1	4	3	28/1	7(i)	37(o)	44
	--	--	--	--	--	--	-	--	---	---	--
Totals	2	1	3/2	5/1	0/1	7/1	5	62/3	14(i)	79(o)	93

• Volume XV. [1796]. Evans 30390, 33690. IEN-M. 4: π1 A-5H4, 401 leaves pp. [2] [1] 2-799 [800].
• Summary of figures: 21 of 100 sheets figured in both formes, 78 figured in one forme, 1 unfigured.

           

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	1/1	0	2	4	0	0	0	16	8(i)	16(o)	24
2	1	0	2	6	0	0	1	16	9(i)	17(o)	26
4	2	0	3	5/1	0	0/1	3	52/3	11(i)	59(o)	70
	--	--	--	--	-	-	-	--	---	---	---
Totals	4/1	0	7	15/1	0	0/1	4	84/3	28(i)	92(o)	120

• Volume XVI. [1796]. Evans 30390, 33691. IEN-M. 4: π1 A-5H4, 401 leaves, pp. [2] [1] 2-799 [800].
• Summary of figures: 25 of 100 sheets figured in both formes, 70 figured in one forme, 5 unfigured.

           

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	0	1	3	7	1	0	2	24/1	11(i)	28(o)	39
2	0	0	4/1	5	0	0	0	17	10(i)	17(o)	27
4	0	0	4	2	0	1	0	46/1	6(i)	48(o)	54
	--	--	--	--	-	-	-	---	---	---	---
Totals	0	1	11/1	14	1	1	2	87/2	27(i)	93(o)	120

• Volume XVII. [1797]. Evans 32088, 33692. IEN-M. 4: π1 A-5L4 5M2, 415 leaves, pp. [2] [1] 2-827 [828].
• Summary of figures: 17 of 103 sheets figured in both formes, 81 figured in one forme, 5 unfigured; the one half-sheet unfigured.

           

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	1	0	6	5	0	5	0	37/2	12(i)	44(o)	56
2	0	0	0	3	0	0	0	3	3(i)	3(o)	6
4	0	0	6/2	3	0	2	0	39/1	11(i)	42(o)	53
	--	--	--	--	-	-	-	---	---	---	---
Totals	1	0	12/2	11	0	7	0	79/3	26(i)	89(o)	115

• Volume XVIII. [1798]. Evans 33693. ICN, IEN-M. 4: π1 A-6C4 6D2, 475 leaves, pp. [2] [1] 2-945 [946-948].
• Summary of figures: 19 of 118 sheets figured in both formes, 30 figured in one forme, 69 unfigured.

           

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
1	3	0	2	3	0	0	1	22	8(i)	23(o)	31
2	1	0	2	5	0	0	0	18	8(i)	18(o)	26
4	0	0	2	1	0	0	0	8	3(i)	8(o)	11
	--	--	--	-	-	-	-	-	---	---	--
Totals	4	0	6	9	0	0	1	48	19(i)	49(o)	68

The fact which strikes one immediately about these figures is the consistency with which they are used over a period of eight years. Their usual positions remain $4ra and $4va; the deviations from these pages (or even from the left to the right sides of these pages) are comparatively few. Certain patterns appear at times (such as the split between 1 and 2 in II and III, or the long run of 4 in XVI, A-X, and they usually suggest that inner and outer formes, in general, were worked off by different presses. The matter of unfigured formes is inconclusive: the same figure rarely occurs in both formes of a sheet, suggesting that perhaps unfigured formes were the work of the same pressman indicated in the related figured formes; on the other hand, the same figure does occur in both formes frequently enough to raise doubts (as in XI, where a distinction is seemingly made between the single 4 in 3I-3O, etc., and the double 4 in 3T, 4T, and 5D). The figures seem to have been placed with some care, for they usually are found above the footnotes, just as the signatures are (though there are exceptions, as IX, Kk4v); in the two instances in which a figure appears on $1r (XIV, Kk1rb, and XVI, 4K1ra), it comes below the footnote, while the signature remains above. But it required some attention to detail to insert figures so high on the page as was necessary to precede the long footnotes on X, Hh4v and Ii4v. In some cases where figures appear on leaves other than $4, though by no means all, there is some peculiarity (such as short text or a long footnote) which would render the figure more unsightly than usual or more difficult to insert; in other instances a figure appears despite a short text (see X, T2r), and in one volume a figure even comes on the final page, just above the words "End of the Seventh Volume." Since the placing of figures may often have been a function of the physical arrangement of the shop (just which edge of the forme was easier to reach

at the proper moment),[23] the fact that three figures in the Encyclopaedia are upside down takes on added interest (figure 3 on VI, 3S4v; figure 4 on IX, Uu3v, and XIV, 5F2r). Systematic study of deviations from the usual positions, in conjunction with such accidents as turned figures, when undertaken for all the books in a shop at one time, may yield information about the mechanical set-up and physical organization of the shop.

An overall picture of the customs of figuring in Dobson's shop during these years, as represented in the Encyclopaedia, can be gained (from one viewpoint) by examining the habits of each of the pressmen who used figures, in terms of the number of times they figured each page:[24]

               

Fig.	1v	2r	3v	4r	1r	2v	3r	4v	Totals
0	0	0	0	0	0	0	0	1	0(i)	1(o)	1
1	19	1	26	136	1	8	19	395	182(i)	423(o)	605
2	22	0	52	223	0	14	17	369	297(i)	400(o)	697
3	7	0	16	76	0	3	5	132	99(i)	140(o)	239
4	22	1	46	114	1	20	13	387	183(i)	421(o)	604
	--	--	--	--	-	--	--	--	---	---	---
Totals	70	2	140	549	2	45	54	1284	761(i)	1385(o)	2146

The two least popular pages are $1r and $2r, but the reason is not the desire to avoid rectos, since the regular position for figuring inner formes is $4r and since the second most frequent position for outer formes is $3r. In other words, whenever $4r and $4v were not used, the chances favored $3r and $3v; the first two leaves were least often chosen. Of the four principal figures (the one o may be a mistake), number 3 is responsible for the smallest amount of work, while 1, 2, and 4 performed roughly the same amount. The resemblance between 1 and 4 is especially striking, both in the total number of formes worked (605-604) and the proportion of inner (182-183) and outer (423-421) —as well as in the fact that these are the only two figures to appear on $1r and $2r. Not only did the four men figure outer formes much more often than inner; they were more definite in their preference of the page on which to figure outer formes. Number 1 figured $4v in 93% of his figured outer formes; number 2, 92%; number 3, 94%; and number 4, 92%. But number 1 figured $4r in only 75% of his figured inner formes; number 2, also 75%; number 3, 77%; and number 4, 62%. The only departure from the general pattern is 4's preference of $2v over $3r as his second-choice position in the outer forme; but there are actually no striking deviations that serve to distinguish the four men.

To complement this somewhat static overview, one may examine various percentages of figured and unfigured formes and sheets as they change during the course of the eighteen volumes:

                                         

Vol.	% formes figured	% formes unfigured	% sheets figured in both formes	% sheets figured in one forme	% sheets unfigured	% inner formes figured	% outer formes figured	% inner figured formes figured on $4r	% outer figured formes figured on $4v
I	3	97	2	3	95	3	4	100	100
II	56	44	51	9	40	52	59	77	98
III	88	12	77	22	1	80	96	95	98
IV	73	27	52	42	6	65	80	83	94
V	72	28	44	54	2	55	87	73	90
VI	82	18	65	35	0	69	96	70	94
VII	78	22	57	42	1	66	90	82	94
VIII	69	31	38	61	1	48	89	80	94
IX	71	29	43	56	1	52	90	80	92
X	60	40	24	72	4	42	77	72	93
XI	58	42	23	71	6	53	63	74	88
XII	63	37	32	62	6	46	79	45	92
XIII	54	46	16	76	8	23	85	52	87
XIV	47	53	8	77	15	14	79	43	82
XV	60	40	21	78	1	28	92	57	95
XVI	60	40	25	70	5	27	93	52	96
XVII	56	44	17	79	4	25	86	42	92
XVIII	29	71	16	25	59	16	42	47	98
Averages
(based on III-XVII)	66	34	36	60	4	46	85	67	92

Volumes I, II, and XVIII, which are not figured all the way through, do not fall into the same pattern as the other fifteen volumes; volumes III through XVII do conform roughly to a pattern, with XIV the only notable exception (less than half its formes are figured). Since 85% of the outer formes are figured, as opposed to only 46% of the inner formes, one might argue, if one assumes unfigured formes to be the work of the same press as the preceding forme, that outer formes were normally the first laid on. However, if unfigured formes are not taken as the work of a separate press, then one would expect the percentage of sheets figured in both formes to rise at times when great speed was desired; yet, as the table shows, there is no steady increase in this figure for those volumes supposedly issued most rapidly (V-X), though it is true that the later figures are generally lower. Furthermore, when 66% of the formes are figured consistently over a period of years, it seems reasonable to assume that the figures represent a device for recording more than simply occasional piecework. The testimony of the figures provides a basis for reconstructing in detail the printing of the bulk of the Encyclopaedia, as soon as a few still doubtful points are cleared up. The answers to these questions (the significance of unfigured formes, the composition-presswork ratio, the length of time spent by pressmen in working off various formes) will emerge, it is hoped, from a study of Dobson's other work.

It remains to say a word about the other great Philadelphia printer of the time, Mathew Carey. Although he first issued a Bible (Douay) in 1790, the only Bibles of his in which I have discovered figures come after 1800. The story of Carey's numerous printings of the King James Bible from standing type after 1804 is a complex one which deserves detailed study in its own right, and it is not possible here, without the benefit of such a study, to do more than give the broadest outline of Carey's use of press figures. First of all, it may be worth calling attention to Joseph Charless, who printed Carey's first quarto King James Bible in 1801:

About the printing of this Bible a number of external facts are known. Composing and printing began in April, just after Charless' proposal of 15 April 1801: "Suppose 120 sheets of which I engage to execute one Sheet per day—Provided the Font is large enough to employ a Sufficient number of Compositors."[25] Since the job was finished in early September (Carey's account book shows a credit to Charless for the Bible on 26 September), or about 130 working days later, the original plan of a sheet per day was adhered to fairly closely. Although each forme contained about 38,000 ens, or about 32 hours' work in composition time, it would have been possible to compose a sheet in one day with six or seven compositors—not a surprising figure in the light of Charless' concern about keeping employed "a Sufficient number of Compositors." Further, the Carey account book and other documents fix the size of the edition at 2000 copies. Since the daily output of one press was about eight tokens, 2000 perfected sheets would be the daily product of two presses working together. The conclusion to which one is led, therefore, is that the unfigured formes in the 1801 Charless Bible were worked off by a second press. Figure 2 appears consistently in the inner formes; only if another press consistently machined the unfigured outer formes could a perfected edition-sheet be completed in one day.

As to the later Carey Bibles, only these observations can be made at this time: Carey continued to use press figures at least as late as 1818, for they appear in Bibles to that time with changes apparently reflecting the division

of labor in each printing (that is, they do not seem to be present simply because they were locked up in standing type at some earlier period and not removed when their usefulness ceased); in the earlier years (and at least as late as 1811), some of his quarto Bibles were gathered in eights, and in these the figures are found on $1v and $3v; in his regular quarto Bibles, the figures come exclusively on $1v; the figure 2 appears in the overwhelming majority of instances, but in the 1803 Bible figures 3, 4, 5, and 7 also occur (here the New Testament at least was printed for Carey by T. S. Manning); only rarely is $1v unfigured. Even a cursory examination of the Carey Bibles suggests that the full story of their printing will demand a study both of press figures and of impression figures (those three-digit numbers in parentheses). But with Carey we have moved into the nineteenth century.