The 'Second Issue' of Shakespeare's Troilus
and
Cressida, 1609
by
Philip Williams, Jr.
IT HAS LONG BEEN KNOWN THAT COPIES OF THE 1609 quarto of
Shakespeare's Troilus and Cressida differ: some copies have
a title-page on which it is stated that the play '. . . was
acted by the Kings Maiesties | ſeruants at the Globe' whereas
in other copies this title-page has been can-celled[1] and replaced by a half-sheet signed
a. The new title-page (a1) changed the
original 'THE |
Hiſtorie of Troylus | and Creſſeida' to 'THE | Famous
Hiſtorie of | Troylus and Creſſeid', deleted the
statement about performance at the Globe, and added in its place
'Excellently expressing the beginning | of their loues, with
the conceited wooing | of Pandarus Prince of
Licia.'
The remaining portion of the title-page (from 'Written by
William Shakeſpeare' on) was printed from the standing type of
the original title-page.[2] The
second page of the cancellans (a2) contains an address
to the reader (concluding on a2v) in
which the play is said
to be 'a new | play, neuer stal'd with the
Stage, |
neuer clapper-clawd with the palmes | of the
vulger'.
Except for these differences all copies are identical; hence those
having the original title-page are usually considered 'first issue'
and those in which the cancel has been effected, 'second
issue.'
[3] Three copies of the quarto
having the original title-page but lacking
a
2 are
extant.
[4] One copy contains both the
original title-page and
a
2.
[5]
Eleven copies in which the original title-page has been cancelled
and replaced by the half-sheet
a
2 are
extant.
[6]
The question of exactly when and how the half-sheet signed
a
was printed has never been investigated,[7] nor has a satisfactory explanation
why
some copies of the book appear with and some without the second
title-page been advanced.[8] Critics,
indeed, have seemed content with the implicit assumption that some
copies do not contain the cancel since they were sold at a time
prior to the printing of the cancellans half-sheet.
It is the purpose of this article to show first that the
half-sheet signed a was very probably printed along with
the
final half-sheet M of the text as a pre-publication cancellans, and
secondly to offer a plausible explanation why in some copies the
cancel was not effected.
Fredson Bowers[9] has shown
that
bibliographical evidence can often be used to prove that different
parts of a book were or were not printed on a full sheet and
subsequently cut in half. In a book having a collation like 4°,
A2 B-M4 N2, with
A2 containing the title-page and
preliminary matter and the text ending on N1 or N2, the possibility
that A2 and N2 were printed together on
a full sheet has been
recognized for some time. Indeed, it has come to be accepted as the
normal method of printing in such circumstances. But a quarto
half-sheet gathering can also be printed by half-sheet imposition,
the four type-pages being imposed together and a full sheet printed
and perfected by this same forme, the two halves later being cut
apart to provide identical half-sheets. Dr. Bowers shows that it is
dangerous to assume that two half-sheet gatherings appearing in the
same book were printed together unless running-title evidence
confirms
this assumption, and he demonstrates that when the running-titles
from only one forme used to print a preceding full sheet of text
are used in both formes of the final half-sheet, this half-sheet
must have been printed by itself by half-sheet imposition. If the
running-titles from both formes used to print a preceding full
sheet of text appear in the two-leaf gathering, the half-sheet was
not printed by itself by half-sheet imposition but instead with
something else.
Twelve running-titles contained in three skeleton-formes were
used in printing Troilus and Cressida, from which two
skeleton-formes were drawn to print gathering L. In inner L, the
following running-titles appear: L1v (IV), L2 (IX),
L3v
(III), and L4 (VIII). In outer L, the following running-titles
appear:
L1 (X), L2
v (I), L3 (VII), and L4
v (II).
Two running-titles
appear in M
2:(X) on M1 and (IV) on
M1
v. It will be noted that
the running-title from L1
v (inner forme) appears on
M1
v
(inner forme), and the running-title from L1 (outer forme) appears
on M1 (outer forme). The final two-leaf M gathering of
Troilus
and Cressida cannot therefore have been printed separately by
half-sheet imposition. Since M was not printed in this manner, it
is necessary to assume that another half-sheet was printed along
with it.
[10] I suggest there is strong
evidence for believing that this half-sheet was
a
2, printed
in the same formes with half-sheet M
2.
First, it will be remembered that the lower portion of the
title-page contains standing type from the original title-page.
This suggests, although it does not prove, that the cancelling
title-page was printed soon after the original title-page (A1). At
any rate, the interval of time between the printing of the original
title-page and the cancelling title-page was so short that the type
from the original title-page had not been distributed.
Second, buttressing the typographical evidence for the
relatively continuous printing of the cancel with the body of the
book is the evidence of the paper. Sheets A-M consistently contain
a watermark of a gauntlet with the third finger surmounted by a
cross. Since in certain copies where a watermark appears in the
a half-sheet, the watermark is invariably this same
gauntlet,
the inference follows that the cancel was printed on the same lot
of paper Eld bought for the rest of the book, and as a consequence
that it cannot be separated by any very long period of time from
the printing of the text sheets. Indeed, this evidence, as well as
that of the standing type from the title, may be more narrowly
applied, for both would ideally obtain
if
a and M had been printed in the same formes. Under
such
conditions it is ordinarily found that one of the two separated
half-sheets in any single copy will contain a watermark and the
other will not. In the observed copies, this is what we find; for
example, in the Folger copy the watermark appears in
a
but not
in M, whereas in the Yale Elizabethan Club copy the watermark is in
M but not in
a.
[11] Thus the
watermarks demonstrate at least that the printing of
a
could not
have been long delayed after the completion of the book, and their
evidence is not inconsonant with the hypothesis that both
a
and
M were printed in the same full sheet.
[12]
Thirdly, the evidence supplied by the type-page measurements
supports the belief that a
2 and
M2 were printed together.
If a
2 were printed along with
M2, using the two skeletons
from L, the measurements of the type-pages in a should
conform
with the measurements of the type-pages of L. If the measurements
do not coincide, the inference would be that
a
2 and M2
were not printed together. The measurements do coincide, and
therefore the a type-pages seem to have been composed in
the
same printers' stick used for the text so that they would fit
without adjustment of the furniture into the skeleton-formes used
to print L and M.[13]
In view of the bibliographical evidence that has been presented,
it seems more than probable that a
2 was
printed
simultaneously with M2, the type-pages having been
imposed as is
indicated below:
If this theory of how a
2 was printed
is accepted, the time
at which it was decided to cancel the original title-page can be
determined with some precision. The decision must have been made
after outer A had been printed but before either forme of M had
been printed. 11 1/2 sheets intervene between outer A and M. If
only one press were used and if the edition ran to the maximum
1250-1500 copies, we may estimate that L was completed
approximately 15 working days after printing on A began.[14]
We must now attempt to explain why three of the extant copies
contain the uncancelled title-page (A1) whereas in eleven copies
the cancel of A1 has been effected and a
2
substituted. (The
Yale Elizabethan Club copy is, of course, aberrant and would align
itself with the eleven copies in which the cancel was made.)
Although it is dangerous when dealing with only fifteen copies to
make much of percentages, the three surviving copies in which the
cancel has not been substituted suggest that in the original
edition possibly a sizable number of copies existed in this
state.
Several hypotheses must be considered before coming to the
explanation I think to be the correct one. Troilus and
Cressida was printed by George Eld for two publishers, Richard
Bonion and Henry Walley.[15] It is
possible, though perhaps unlikely, that one of the two publishers
definitely preferred the original title-page with its reference
(rightly or wrongly) to performance at the Globe by the King's Men.
If so, he could have directed his binder to ignore the cancel in
the copies allotted to him. Or—much less likely—both
publishers, acting jointly, may have deliberately allowed some
copies of the book without the cancel to be issued, holding back
the cancelled copies with the expectation that the new title-page
and preface would stimulate sales at a later date. This particular
speculation should not be taken too seriously, for it is unlikely
that the publishers, having gone to the trouble and expense of the
cancel as a part of continuous
printing of the quarto, would deliberately issue copies without
that cancel. Moreover, there is no evidence that the altered form
of the title would have stimulated sale.
Unless we are willing to suppose that one or both of the
publishers deliberately issued some copies without the cancel, we
must look elsewhere for an explanation. Our knowledge of early
17th-century binding practices is, unfortunately, limited; but it
seems probable that the answer to our problem lies here.[16] Most of the copies of
Troilus and
Cressida were correctly
bound:
a
2 was added and A1 cancelled.
But it is probable that
in some copies, through accident or error,
[17] the cancel was not made.
The most recent attempt to define issue and
variant
state reaches the general conclusion that there are only two
major classes of reissue: (1) post-publication alterations in the
publishing or selling arrangements as indicated by a cancellans
title-leaf; (2) post-publication alterations or additions in the
book accompanied by or confined to a cancellans title-leaf to
assist in stimulating sales of old sheets. Additions or alterations
made to constitute what may be called ideal copy should be
considered as 'states,' whether made before or after publication.
Among such 'states' are specifically placed cancellans titles
"printed as part of an original sheet (whether of the preliminaries
or of the text) to perform the same function as a press-variant
title."[18]
If a
2 and M2 were
indeed printed simultaneously, and
if
the copies containing A1 but lacking cancellans
a
2 are the
result of binding error or accident to some of the half-sheets and
not demonstrable as a distinct publishing effort, the implications
are clear: the so-called 'first' and 'second issues' of
Troilus
and Cressida are not separate issues at all; and they should
therefore be treated as W. W. Greg treats the similar case of
The Dumbe Knight (1608),
[19]
that is, as variant states of only one issue.
But the classification, while important, is after all basically
only one of estimating accurately the true conditions of printing
and publication. The bibliographical evidence which can be brought
to bear to assess the cancel in the first quarto of Troilus and
Cressida indicates very strongly that the alteration in the
title and the consequential printing of the address to the reader
were decided on and manufactured before any copies of the book
could be issued. No collateral evidence exists which would lead to
a conclusion that intentional separate and simultaneous issues were
made of copies in the two forms: reissue is, of course, a practical
impossibility owing to the circumstances of printing.[20] If the bibliographical evidence is
accepted as a sufficient demonstration, we must alter our views
materially concerning the time at which this cancel was printed
and, to some extent, the circumstances which dictated it.
Notes