:: :: University of Virginia Library

Recently Mr. Fredson Bowers called attention to the fact that there are six different states of the imprint of Dryden's Troilus and Cressida (ist ed., 1679).[1] It is a tantalizing puzzle, and Mr. Bowers warned that his quite plausible reconstruction of the history of the printing must be only tentative because no solution can be proved beyond a doubt. It may be of some interest to look into the possibilities of a somewhat different theory as to what may have happened.

Each of the six imprints is introduced by the phrase "London, printed for", names Jacob Tonson and Abel Swall as publishers, gives their addresses in the same wording, and closes with the date, 1679. The two main forms of the imprint differ in that one names Tonson first and the other names Swall first. Mr. Bowers classifies the Tonson-first imprints as T1, T2a, and T2b; and the Swall-firsts as S1a, S1b, and S2. The distinguishing features of each are shown in the following table:

	T1	T2a	T2b	S1a	S1b	S2
Rule	A	B	A	A	B	A
London	A	A	A	B	B	B
Tonson	A	A	A	B	B	B
Swall	A	B	B	A	A	B

186

There are two different settings of type for each of the following elements in the imprint: the phrase "London, printed for" (London); Tonson's name and address (Tonson); and Swall's name and address (Swall). Swall A is followed by a period; Swall B, Tonson A, and Tonson B are each followed by a comma. Swall B has the same line ending in T2a and T2b, but it is better spaced in T2b. Also the rule above the imprint may consist of a 97 mm. piece (Rule B) which begins at the left margin but does not reach the right margin, or it may be extended to the right margin by an additional 12 mm. piece (producing Rule A).

Mr. Bowers' reconstruction of the history of the printing may be outlined as follows:

(1) T1 is printed first. Swall A period is placed before the date correctly and by choice.
(2) Required number of Tonson copies is printed.
(3) S1a is formed from T1 by removing London A and Tonson A, setting London B, moving up Swall A to follow it, and then setting Tonson B to follow Swall A. Swall A period is overlooked and remains now by error because it is followed by Tonson B instead of by date.
(4) S1a begins printing Swall copies.
(5) S1b results when the short piece of rule drops out leaving only Rule B.
(6) Need for more Tonson copies is discovered while S1b is printing. T2a is formed by using idle London A and Tonson A and setting Swall B.
(7) S1b finishes printing the required number of Swall copies. S1b is replaced by T2a. Printing of the second batch of Tonson copies begins.
(8) T2b results when the small piece of rule is again inserted producing Rule A, and Swall B is better spaced.
(9) Need for additional Swall copies is discovered while T2ab is printing or when it is about to be removed.
(10) T2b finishes printing the required number of copies in the second Tonson batch.
(11) S2 is formed from T2b by replacing London A with London B, removing Tonson A, moving up Swall B to follow London B, and adding Tonson B at the end.
(12) S2 prints required number of second batch of Swall copies.

There appear to be several difficulties with this hypothesis.

It is, for instance, not impossible that the Swall A period came by mistake from the comma box; the comma box lay in the case just above the period box and it would be easy enough to drop a period into the comma box

187

while distributing. Although a period before an imprint date is not uncommon, T1 is the only Troilus and Cressida imprint so punctuated. Possibly the Swall A period does not represent the compositor's intention in T1 any more than in S1a and S1b where it does, indeed, seem to be an error. If so, the period would have no value as an indication that Swall A was first used in T1 because the compositor who would not bother to remove the period from S1a and S1b where it was in error would scarcely bother to remove it from T1 where it would probably not offend.

Again, in step 3, it would seem that a simpler procedure and less time consuming would be merely to rearrange the type already in the forme for T1. Instead, the compositor pulls out London A and Tonson A and sets new London B and Tonson B to take the place which they might have occupied. This would be necessary if London A and Tonson A had pied, but they reappear unchanged later in T2a and T2b. Nor did he do it because he wanted to avoid shifting sections of the imprint; London A which was removed need not have been shifted at all, and Swall A which was left had to be shifted to a position following London B.

A similar objection seems to apply to step 11. Here again there is already in the imprint all the type needed, and only a rearrangement would be required. The compositor need not even do that. At step 7 he had replaced S1b entirely with T2a; in the same way he could now replace T2b with S1b. Instead, he removes only part of T2b, and shifts what is left so that he can insert part of S1b. It is, of course, unlikely that the printer of Moxon's day was completely efficient in everything he did; at the same time it is not impossible that he was as eager as any other man to get out of as much work as he could and that whenever he came across something which saved time he might tend to repeat it.

Mr. Bowers, it may be noted, was aware of these objections but he felt that even so his hypothesis provided the most reasonable solution of the problem. He was convinced that the Swall A period was used deliberately and correctly in T1. He rejected the idea of a complete new imprint in step 3 because this would mean to follow T1 with S2, and he thought it impossible to work out a convincing order for the remaining imprints if this were done. He suggested that the compositor may have intended to use a new imprint at this point, but delayed setting it until it was too late and then made the alteration in what may somehow have seemed to him the simplest manner even though it does not seem so today. Finally, he recognized that in step 11 the printer failed to use the lesson of step 6, but he felt that it is not necessary to force the compositor always to follow the same procedure in meeting an unusual situation.

188

The following hypotheses are offered as illustrations of what may be done with a somewhat different approach.

This approach involves two theories: (1) Both compositor and pressman would have saved time if while one imprint was being used at the press another imprint were set up complete and ready to replace it in the forme as soon as the required number of copies had been run off. The type of T1 is entirely different from that of S2, and T2ab is entirely different from S1ab. In either pair one imprint could replace the other with minimum delay. (2) The switch in parts of the two imprints first used might have resulted if, after the first batch of copies for Tonson and for Swall had been printed, one or more portions of a then idle imprint had been removed for use in some other book naming only one of the publishers.

Whatever the sequence, however, it seems necessary to make two assumptions: (1) T2b followed T2a because Swall B is better spaced in T2b. (2) In any two consecutive but typographically entirely different imprints the rule probably (but not necessarily) remained the same; otherwise the short rule was lacking two or more times instead of just once during the printing.

A schedule such as (A) or (B) below might be satisfactory. (It may be worth noting that Hypothesis B agrees with Mr. Bowers' feeling that the Swall A period had to appear first in T1.)

Hypothesis A

(1) S1a is printed first. Swall A period is taken from comma box and never corrected.
(2) S1b results when the short piece of rule drops out leaving Rule B. Before S1b is finished printing T2a is set.
(3) Required number of Swall copies is printed. T2a replaces S1b and begins printing Tonson copies. S1b is tied up or left in a galley.
(4) T2b results when short piece is put back in rule producing Rule A, and Swall B is better spaced.
(5) Required number of Tonson copies is printed. T2b is left in the forme or tied up with undistributed title-page type.
(6) Swall B is removed from T2b in title-page type to be used in another book.
(7) Need for additional Tonson copies is discovered.
(8) T1 is made up from what remains of T2b (London A and Tonson A) plus Swall A from idle S1b. T1 begins printing.
(9) Need for additional Swall copies is discovered.
(10) S2 is made up from what remains of S1b (London B and Tonson B) plus Swall B (now finished with other job).

189

(11)T1 finishes printing second batch of Tonson copies. S2 replaces T1.
(12) S2 prints second batch of Swall copies.

Hypothesis B

(1) T1 is printed first. Swall A period is placed before date by accident because it is taken from comma box, or possibly by choice. Before T1 is finished printing S2 is set.
(2) Required number of Tonson copies is printed. S2 replaces T1 and begins printing Swall copies. T1 is tied up or left in galley.
(3) Required number of Swall copies is printed. S2 is left in the forme or tied up with undistributed title-page type.
(4) Swall B is removed from S2 for use in another book.
(5) Need for additional Swall copies is discovered.
(6) S1a is made up from what remains of S2 (London B and Tonson B) plus Swall A from T1. S1a begins printing. Swall A period is overlooked and remains now by error because it is followed by Tonson B instead of by date.
(7) S1b results when short piece drops out during printing leaving Rule B.
(8) Need for additional Tonson copies is discovered.
(9) T2a is made up from what remains of T1 (London A and Tonson A) plus Swall B (now finished with other job).
(10) S1b finishes printing second batch of Swall copies. T2a replaces S1b and begins printing second batch of Tonson copies.
(11) T2b results when short piece is returned making Rule A, and Swall B is better spaced.
(12) T2b finishes printing second batch of Tonson copies.