University of Virginia Library

James Thomson's Sophonisba received its first performance at Drury Lane on February 28th, 1730; it had a great success, running for ten nights, out of which Thomson received three benefits. Millar paid 130 guineas for the publishing rights. The printing was the work of W. Bowyer, and from the evidence of his paper ledger (which is discussed in more detail below)[1] the first paper for the play arrived on the 27th, but printing apparently cannot have commenced before March 4th. On Monday March 9th the Daily Post announced:

On Wednesday next will be publish'd, Sophonisba. A tragedy. As it is now acting at the Theatre Royal in Drury-Lane. By Mr. Thomson. Printed for A. Millar . . . N.B. A small number is printed in 4to on large paper, for the curious. [Later advertisements add:] Price in 4to. 2s. 6d. in 8vo. 1s. 6d.

On Wednesday March 11th the date of publication had been postponed —the Daily Post reads 'To-morrow morning will be publish'd' and on the Thursday announces 'This day, at four o'clock in the afternoon, will be publish'd'. The Daily Journal gives the time as 'This afternoon at five'. The statutory nine copies were deposited at Stationer's Hall on March 11th, and those that survive contain no cancel; it therefore seems possible that the day's delay was caused by the need for cancellation discussed below.

Copies of both the quartos and octavos collate A-L4, with A2 a cancel in the first impressions; the octavos are printed by half-sheet imposition. It is clear that the quartos are a separate and earlier impression; Mr. Iolo Williams pointed out[2] that they have a list of five errata on L4 as opposed to three in the octavos, and this will serve for the moment. The type pages had to be reimposed for half sheet imposition and to adjust the margin before the octavos could be printed; the independence of the two impressions is also attested by the completely different press figures.

Though I have seen no uncut copies of the quartos it is clear from Bowyer's ledger that the ordinary quartos were printed on Crown paper which usually has small initials GF as watermark; the fine paper copies are


Page 205
on Writing Demy with a fleur-de-lys watermark. There was no need to make adjustments to the type between printing the two, and one cannot assign priority between them; any attempt runs against contradictory evidence:

1. On G1v, p. 42, a headband with a winged cherub as its centre appears in all Demy copies, but only some Crowns; in others it is replaced by a block with a rudimentary fleur-de-lys, which is also used in the octavos. The cherub block re-appears at the head of A3 in all copies, showing four nail heads; it apparently came loose from its mounting while the Crown paper was being printed and had to be replaced. This confirms that the octavos were printed after the quartos; shows that the preliminaries were as usual printed last; and that in this forme the Demys were printed first.

2. At the head of L2v there is a row of type ornaments. In all the Crown copies I know of, this is composed of 11 blocks; in the Demy copies one block has apparently dropped out; and in the octavos it has been replaced by its mirror-twin which does not properly complete the design. This evidence suggests that in this forme the Crowns were printed first.

This discrepancy of evidence need not disturb us. The use of the same press-figures throughout[3] shows that the printing of quartos was continuous irrespective of paper. It is perhaps rash speculation, but observing that one piece of evidence is in the inner and one in the outer forme, we could argue that the contradiction is explained by a logical printing sequence such as one in which the outer formes of Crowns were printed first, followed by the Demys; then the Demys preceded the Crowns in perfection with the inner forme.[4] For my part, neat as this explanation is, I would just as soon put the discrepancy down to the varying practice of pressmen.

Before discussing the different impressions in octavo it is as well to deal with the cancellation of A2 in the quartos and the first octavo impression. A2 carries Thomson's dedication to the Queen, and a comparison of the text as found uncancelled in a few quartos with that in the cancels reveals a number of textual changes.

The original dedication[5] read:

[A2r] MADAM, / THE notice, Your MAJESTY has condescended to take of the following Tragedy, emboldens me to lay it, in the humblest manner, at Your MAJESTY's Feet. And to whom can this illustrious Carthaginian so properly fly for protection, as to the QUEEN of a People, more powerful at sea than Carthage? more flourishing in commerce than those first Merchants? more invincible to conquest? [A2v] and, under a Monarchy, more free than a Common-wealth itself?


Page 206

But I dare not here attempt a Character, where both the great and the amiable Qualities shine forth in full perfection. All words are faint to speak what is universally felt, and acknowledged by a happy people. Permit me therefore only to subscribe myself, with the truest zeal, and veneration,

May it please Your MAJESTY, / Your Majesty's / Most devoted Subject, / and Servant, / JAMES THOMSON.

The revised setting[6] has the following variants:

as to a QUEEN, who commands the hearts of a People, . . . more secure against conquest? and, under a[A2v] Monarchy, . . . I dare not, nor indeed need I, here attempt . . .

MADAM, / Your MAJESTY's / Most humble, / Most dutiful, / And most devoted / Servant, / JAMES THOMSON.

As Dr. Todd points out to me, these changes are not only stylistic; the first change also gives a less ambiguous expression to Thomson's republican views.

In spite of the fact that Thomson is a classic example of a man with second thoughts,[7] it is unlikely that the cancel would have been made had it not been for an ill-advised attempt to make these changes by stop-press correction while the octavos were printing. In printing by half-sheets the recto and verso of the half-sheet are locked up side by side in one chase and printed together; when the sheets have been perfected from the same type the sheet is divided and two perfect half-sheets are the product. It appears that when the preliminaries of the octavo had been printed on one side, Thomson wished to revise the text of the dedication.

If the type is changed at this stage the inevitable result is that the revised recto and verso will perfect the unrevised verso and recto, and if one leaf is revised the half-sheets will each contain one revised and one unrevised page. This would be unsatisfactory but not disastrous; what caused chaos in the case of Sophonisba was that the corrections involved moving the three words 'and, under a' back from verso to recto; the final result was that half the copies had these three words repeated on the turn of the page, and the other half omitted them completely. One example of each case survives uncancelled,[8] and each in a copy which contains sheets of the second impression; these copies may have been put together after the main batch of the first impression had been sent for cancellation. There is no evidence that the corrections were made precisely at the end of the first run through the press; if corrections were made before the first run was complete, some copies would be printed with the full corrected text; if the corrections


Page 207
were made after perfection had begun, some copies would have the full unrevised text.[9] I have not located any copies of the first impression where both recto and verso are in the same state, but they may survive unnoticed.

However that may be, it is clear that the dedication [state a] had to be cancelled when recto and verso did not fit. In addition to the corrected text [state b1] already standing in type, the dedication was set up three more times [b2-4] so that four cancels could be printed at once (presumably still by half-sheet imposition). Thus there are four different settings of the cancel, of which one preserved the heading, factotum, and first seven lines of text of the original; and this setting [b1] was reimposed with the other preliminaries for use in later impressions.

It was not absolutely necessary to print cancels for the quartos which had already been correctly printed, though with the unrevised text; but presumably with the type at hand either Thomson or Bowyer decided to carry the improvements through by printing the cancels in quarto as well; all four settings are found in Crown copies, and they were probably also used for the Demys, though in these only two of the settings have so far been noticed. It should perhaps be noted here that from the evidence of Bowyer's paper ledger all the copies which required cancellation (both quarto and octavo) were handed over to an unidentified Harris on March 9th and a small new impression of 750 copies was delivered direct to Millar (the publisher) on the 12th, the day of publication. If Harris was responsible for the task of cancellation (and perhaps for stabbing the pamphlets), it is possible that the copies delivered direct to Millar were those first issued, though of a later printing.

With the octavos we come to more difficult ground; the general outline is clear, the details obscure. With the help of twenty copies lent by Messrs. Pickering and Chatto (and one in the British Museum) which could be collated side by side, it was not difficult to sort out variants; it soon became clear that some copies were mixtures of sheets from different printings. But out of twenty-one copies, sixteen fell into four groups of mutually consistent copies; each of the remaining five was a unique mixture. It seemed clear therefore that even though there were occasional sheets apparently common to two groups, each group had some sort of independent existence; and this view was encouraged by Nichols' mention of four 'editions'[10] (confirmed in general by Bowyer's paper ledger). I here refer to the four groups as four impressions. The reliability of my sample is statistically doubtful — for example it included four examples of one cancel and three of another, but none of the other two, though all four should be


Page 208
equally common. Details of the groups and mixtures appear in Table 2 below; Table 3 attempts to show the distinctions in each signature of each group.

Examples of each of the four impressions have been collated on the Hinman collating machine. In many sheets this reveals throughout a slight movement of the type such as would be compatible with tied up type-pages being reimposed with different furniture and often with new running titles; this is represented in Table 3 by a vertical rule between impressions of a sheet. In a number of cases the type is apparently undisturbed except for some small detail — a printer's flower damaged, a press-figure inserted or a slight change in spacing; these are indicated by a broken vertical rule and probably represent formes which have been kept locked up in their chases between impressions. In a few other cases there is no noticeable difference between states, and while there is no reason why locked up formes should not have been reimpressed as they stood, it is possible that overprinting of the sheet at an earlier time made reimpression unnecessary.

The behaviour of press-figures is far from helpful. Thus in sheet H the figure 7 on p. 56 appears in only some copies of the first and third impressions (and very likely in some copies of the second not seen by me); this is caused by frisket-bite, the figure being below the direction line — some copies show the figure half obscured. But in sheet E the same figure 1 appears clearly in all the first three states; there is a good deal of movement of the type between impressions 1 and 2 which makes reimpression certain despite the remaining press-figure; there is no movement between impressions 2 and 3, but in the circumstances the presence of the same press-figure is not a reliable indication of continuous impression. The same is true of impressions 3 and 4 of sheet K which are linked by a common press-figure.

There was considerable interchange of headpieces as the printing went on, though not sufficient to give a clear picture of the order in which sheets went to press. In the first impression ornaments used in B and C were used again in G H I and L, which suggests normal progressive printing. The same ornaments are found in the second printing of the sheets G to L but that from K was later used to fill the gap left in B during the first printing. The gaps in C were filled by printers' flowers and the headpiece from E — which was itself replaced from I. Thus in the second impression some at least of the later sheets were printed before the earlier; and of the earlier C was printed before E. The only rearrangement in the later impressions[11] was that the gaps left in I and K by the changes in the second printing were both filled by the same set of printers' flowers which were apparently used in I and then K in the third impression, while in the fourth they remained in K and were then moved back to I.

The only detailed evidence for the history of Sophonisba's publication


Page 209
is found in Bowyer's paper ledger now in the Bodleian, though this raises almost as many problems as it solves:

As mentioned above, the first 30 reams of Demy paper for the octavos were delivered before the first night, February 28th; the Writing Demy for the fine quartos on the other hand did not arrive until March 4th. Since the quartos were printed first, the printing must have started after this and have been completed by March 9th when the first copies were delivered to Harris and advertisements inserted in the press.

The first problem arises in considering the simplest entry, the Writing Demy quartos. 1 ream of paper (500 sheets) is not sufficient to produce the 50 copies delivered, for Sophonisba runs to eleven sheets in quarto. A hypothesis which fits certain other facts is that the original paper order was made on the basis that the book would only run to ten sheets; in that case one ream would have sufficed. But how the figures have been made to balance in the ledger must be pure guesswork.

The Crown quartos are much more of a problem. It appears from the ledger that 34½ reams were delivered; this should have produced 1570 copies rather than the 689 actual or 700 nominal in the ledger. If we could read the March 2nd entry as 6 half-reams the figures would be more reasonable. The half-reams would presumably be left-over oddments; the preliminaries are found with at least two watermarks, both different from those in the body of the book, which offers some confirmation. We could then assume that the original order for 28 reams was based on the estimate of ten sheets, and the six half-reams the additional paper for the additional sheet. We then have the more simple relationship of a supply of paper for 1400 copies and only 700 copies accounted for. Whether only half this number were printed or whether 700 copies were put on one side is still an unsolved problem. However, Thomson's Works in octavo, 1738, vol. 1. contained the 1730 sheets of Sophonisba with the title removed, a half-title being printed on T4 of the newly printed body of the volume. 1000 copies were printed as against 1500 of vol. 2,[12] and both octavos and Crown quartos


Page 210
(brutally cut down on all four edges and oversewn) were used to make up the volume. This suggests either that sales of the quarto were very bad (in which case a decision to halve the printing order would have shown foresight) or else that the missing 700 copies were held back with a reissue in a quarto volume of Works in mind. Sophonisba did indeed appear in vol. 2 of the quarto Works published in 1736, but this was a quarto of too large a size for the sheets to have been of use.

The thirty reams of Demy were probably first intended to produce 3000 copies; when it became clear that eleven half-sheets would be needed, an additional three reams were ordered. These did not arrive until March 9th, the date when the first copies were delivered, and the original printing appears to have been 2250 copies. This would have taken 25 reams, leaving five to add to the three delivered on March 9th. These eight would then suffice to produce the second impression (here 'second edition') of 750 copies delivered to Millar — the three days, March 9th - 12th, would have been a reasonable time to print these. The chief problem is the deliveries to Massey and Mrs. Dodd, totalling a thousand copies. These must correspond to the 11 reams received on March 11th, but the deliveries are not dated. It is not impossible that this thousand copies, the third impression, was completed between Wednesday morning and Thursday afternoon. If it were so, it is not surprising that sheets of the three impressions printed in a week sometimes got intermingled. The final order of paper arrived after publication day and the thousand copies resulting apparently lingered until November — indeed there was little demand for them thereafter, since they could still be reissued with a cancel title as 'The Second Edition' in 1735 (announced in the Grub Street Journal of April 3rd 1735 at 4d.) and again in vol. 1 of Thomson's Works, 1738. It is no doubt their poor sale which accounts for the fact (noted by the D.N.B.) that it was not until after 1738 that Thomson revised the offending line, 'Oh! Sophonisba! Sophonisba! oh!'