Bibliographical Evidence from the Printer's Measure
[*]
by
Fredson Bowers
TYPOGRAPHICAL MEASUREMENTS HAVE always been of
important service
in bibliography, as instance the basic uses for identification to
which incunabulists put the measurement of twenty lines of type.
The present study is not concerned, however, with measurements of
type for the purposes of identifying fonts held by printers, but
instead with certain inferences which the investigator of the
presswork of sixteenth and seventeenth-century books may draw on
occasion from identifying the length of the printer's stick, or
measure—sometimes, but not always, in conjunction with
alterations in the overall type-page opening in the skeleton-formes
of a particular book. This study is confined almost exclusively to
Restoration play quartos, but only because I have been working
closely with these for several years and have been able to keep
records of measurements in some hundreds of books. Except for the
final section devoted to the identification of compositors setting
in relay, there is
perhaps nothing very new in the evidence advanced; but since no
formal study has, I think, been made of this kind of evidence, it
is perhaps useful to bring together in one place a maximum of
information even though some part of it is familiar to most
analytical bibliographers.
In his stick, or measure, the compositor set the type from his
cases, and from this stick he transferred as convenient a series of
composed lines to his page-galley. In setting a mixture of verse
and prose, he often used two measures, one short and one long, the
longer being the full width of his type-page. Which measure was
used to set any given line may have considerable textual
significance, as Mr. George Williams has shown in this present
volume in his "A Note on King Lear, III.ii.1-3."[1]
In early times this compositor's stick was made in various fixed
lengths, but at some indeterminate period before Moxon's treatise
the adjustable stick came into use. It is generally believed that
with the adjustable stick, at any rate, and probably with standard
widths of the fixed stick, the compositor owned his own measure.
Whether this is an absolute fact is not essential for the present
argument, but it may be remarked that close examination of a number
of Restoration play quartos does not disclose the interchange of
measures between compositors during the course of setting books
where variant measures may be identified. Whether fixed or
adjustable, sticks were likely to vary among themselves by as much
as two millimeters even when intended to be used in setting the
same width of type-page. This is understandable when the
difficulties are taken into account either of two compositors
adjusting their sticks identically or of the artisan carving two
wooden fixed sticks to give
an absolutely precise opening for each.
This small variation in the long measure was of little
consequence in the printing. It was too small to be seen by the
eye,
and it did not prevent type-pages composed in two such variable
sticks from being imposed in the same skeleton-formes: the wedges
seem easily to have taken up the difference and provided equal
pressure within any portion of the forme. There would be a limit of
tolerance, of course. My observation has been that up to about two
millimeters difference may be taken as normal, although I have seen
measures varying up to three millimeters used in setting type for
the same formes: when more widely variant measures are found, one
will usually discover that different skeleton-formes contain the
type-pages of such unequal width.
Measuring to detect these variant sticks is not always easy. The
bibliographer must take account of the fact that different letters
were cast on different parts of the body of the type and that he
must choose roughly similar letters at the beginning and end of
successively measured lines if he is not to be thrown off by
non-significant variation of as much as a millimeter. This is
important, for often he must work with variance between two
compositor's sticks of as little as a single millimeter. Moreover,
the measurement of no one line on a page can be trusted to identify
accurately the stick used for that page owing to the fact that
compositors seem frequently to have justified a line by a final
thin space. Catchwords alone are the least trustworthy of all, and
should not be employed except in cases of necessity: my observation
has been that justifying by means of a thin space after the
catchword was a fairly common operation. Finally, owing to the
variable tightness
with which the quarters of the forme could be locked up by the
wedges, some normal differential, usually of about a millimeter, is
often encountered between type-pages set in the same stick. These
are severe difficulties, and for some pages are often serious
enough to make measurement untrustworthy when variation between
sticks is slight and the compositors did not set according to a
reasonably fixed pattern.
The most elementary and easily discerned cases which can
be determined by measurement occur when (a) a book is divided in
half between two compositors and each simultaneously sets his
portion; (b) printing of a book is so materially interrupted that
when work is resumed a different measure is inadvertently employed.
When running-titles are present, the basic fact of division is
ordinarily demonstrable without requiring the evidence of
measurement except as a corroboration. A book in which one complete
portion is printed with a certain set or sets of running-titles and
another portion with a completely different set or sets has
manifestly been printed in different skeleton-formes. However,
these books are useful for demonstrating the validity of the
evidence provided by the printer's measure since the two portions
are not always set with a different number of lines per page or
with a different font. Moreover, the measure alone can sometimes
decide whether (a) or (b) above obtained with a given book.
A typical example is John Crowne's Calisto (1675), in
which simultaneous two-section printing is demonstrated by the
faulty casting-off of copy which resulted in the second press
beginning with sheet H although subsequently the first press
concluded its section with sheet F. This simultaneous setting and
printing is also indicated by the running-titles, which are in
lower-case in sheets B-F but in full capitals in sheets H-L.
Although the font remains the same, the measurement of the
type-page in the first section is 36 lines, 169(182) x 113, 94R,
and of the second section 38 lines, 179(190) x 109 mm.
Running-titles are not always present, however, to indicate such
a division, and in these cases the type-page measurement may be the
only available evidence. Thus in Abraham Bailey's The Spightful
Sister (1667), which is without running-titles, one observes
that the text in sheets B-E is set with a printer's measure of 113
mm., but from sig. F1 to the end of the book on sig. 14v
the
measure jumps to 130 mm.
A question often arises whether a book has been simultaneously
set in two sections, or whether the break between
two portions, as indicated by the type measurement, is only a sign
of an interruption in the
seriatim printing, or else of
another compositor taking over not necessarily after a delay. In
some cases the same sets of running-titles, and thus the same
skeleton-formes, continue regularly throughout a book although at
one point there is a shift in the measure which indicates
composition by a different workman. A typical book is Peter
Bellon's
The Mock-Duellist (1675), which is printed with two
skeleton-formes per sheet, these same two skeletons being
maintained throughout; yet sheets B-F are set with a 120 mm.
measure and sheets G-I with a measure of 121 mm. In such a book the
inference is probably that with sheet G another compositor, who
intended to set his stick to the same measure, took over the work.
In general, one is likely to conjecture that any interruption of
the printing sufficient to cause a single compositor to adjust his
stick again after working
on some other book would most likely have been sufficient to cause
the skeleton-formes to be broken up—but in many books only the
conjecture is possible.
However, there is a kind of evidence which can be used
decisively in two-section books without running-titles or in books
where a change in running-titles and thus in skeleton-formes
indicates the possibility either of simultaneous two-section
printing or else of a marked interruption in the printing. In a
first edition, especially, the normal inference is usually that
separate preliminaries were printed as the final operation. For
certain first editions reasonable demonstration of this fact can be
made, as when an errata list is present in the preliminaries, or
when the text begins on A1 or else on B1 but with preliminaries
occupying more than one gathering. In some two-section books the
evidence is singularly neat. Thus in Calisto, mentioned
above, the three-sheet preliminaries signed A4
a-b4 were set
in the 109 mm. measure used to print sheets H-L but not B-F, and
thus one can safely infer what the signing would lead one to
expect, that these
sheets were machined after the last
sheet of the text had been wrought off, in this case by the second
press.
There is still an ambiguity in such books, however, for this
pattern could also result when there had been an interruption, or
when without interruption a second workman or press had been
substituted.[2] When, on the other
hand, in a two-section book one finds that the compositor setting
the first section also set the preliminaries, somewhat less
question can arise, for unless the preliminaries were set and
printed first, this allocation of composition could result only
when a book was simultaneously printed. In Thomas Southerne's
Oroonoko (1696), for instance, the text begins on B and the
preliminaries are confined to sheet A. On sig. E1 we find the
measure changing from the 111 mm. of sheets B-D to the 113 mm.
measure of sheets E-M. Here the case at first sight is not certain,
since the markedly unequal division of the book seems to militate
against simultaneous setting in two sections; and lacking other
evidence one might be led
to suspect that the appearance of the 111 mm. measure in sheet A
should be accounted for by the view that the preliminaries were
printed first, even though the book is a first edition. Yet other
evidence suggests simultaneous printing.[3]
On the other hand, when a book seems to have been broken rather
neatly in half between two compositors, and the compositor of the
first section set the separate preliminaries, the evidence is all
in favor of simultaneous printing. This is the case with The
Spightful Sister, where the text division is B-E and F-I, or
four sheets to each compositor, with half-sheet A
set by the first. Another example is Rochester's
Valentinian
(1685), divided B-G and H-M, the three sheets of the preliminaries
also being set by the compositor of the B-G section.
Some rather odd books offer the most positive evidence.
Occasionally in two-section simultaneous printing one press would
assist the other in cleaning up the job. A first-rate example is
John Crowne's The Married Beau (1694) in which the text
division is B-F, the type-page measuring 46 lines, 187(198) x 108
mm.; and G-K, the type-page being 47 lines, 190(201) x 115 mm.
Gathering F consists of three leaves, the fourth having been
excised. The preliminaries require the four-leaf sheet signed A
plus an unsigned disjunct fifth leaf. When we find sheet A set with
the 115 mm. measure used for G-K, but the disjunct preliminary leaf
set with the 108 mm. measure used in B-F, and when we also find
that the pagination of the book skips from 38 on F3v to 41
on
G1, the case is clear. The book was simultaneously printed in two
parts, with the second press printing sheet A but the first press
machining the odd preliminary leaf as F4, its text copy not being
sufficiently extensive to
fill the four leaves of final sheet F.
Another and more complicated example is found in Thomas
Southerne's The Disappointment (1684) in which the original
assignment to two presses had been text sheets B-E and F-I.
Gathering E is composed of three leaves, the first two conjugate
and set in the measure used for sheets F-I; but sig. E3, disjunct,
is set in the different measure used for B-D and also for
preliminary sheet A, this last having its fourth leaf excised. The
highly irregular gathering E has been mistaken for a cancellans,
but a rather complex chain of bibliographical evidence can be
constructed to show that the first press was delayed in its
printing between sheets C and D, and though gathering E had
originally been assigned to it (the second section clearly having
started printing with F), to finish the book expeditiously the
second press swung over after printing sheet I and the two presses
joined to print E. The second press machined
E1.2 by half-sheet imposition while the first press was printing
sig. E3 in the A4 position of the preliminaries.
[4]
Evidence as to the measure becomes more difficult when the
preliminaries consist only of a disjunct title-leaf or a half-sheet
with preliminary text set in a short measure. However, records I
have kept of several hundred books show that in most cases the
title-page was set in the same stick used for the text (or for the
rest of the preliminaries), and thus that its measure will
ordinarily be the same. When, in order to give room for large
display type, the title seems to have been set directly in page
galley and with an abnormally wide measure, often one will find, as
in Dryden's The Indian Emperour (1681), that the imprint has
been composed in the printer's text stick and therefore can be
compared with the measure in other parts of the book. In other
cases when the whole title and imprint seem to have been set in a
longer stick than that used elsewhere in the book, preliminaries
like dedications, forewords, dramatis personæ, and so on will
usually conform in measure
to one or other section of the text. If, on the contrary, as in
John Bancroft's Henry the Second (1693), the separate
preliminaries and title are set in a different measure from the
text, we may suppose—according to their nature—either that
they
were set last after some delay or, as with the second edition of
Dryden's The Spanish Fryar (1686), that setting of the book
began with the preliminaries but a larger measure was employed to
squeeze rather extensive material into one sheet. In reverse, we
find Dryden's The Rival Ladies (1693), the title,
preliminaries, and first two pages of text (B2-2v) set in a
117
mm. measure, but on B3 the measure shifting to the 126 mm. used
thereafter.
Sometimes rather interesting facts about the printing may be
deduced from the study of the printer's stick. The first edition of
Thomas Shadwell's The Squire of Alsatia (1688) shows a
printer beginning with the typographical plan of a page measuring
38 lines, 176(189) x 110, 93.6R, and setting sigs. B1-C1
v
according to this layout. Then, since the play is a long one, he
apparently felt the need to expand his page to save paper and
presswork, and set C2-4
v as 39 lines, 180(194) x 115 mm.
[5] Presumably he then found the page
too
crowded, since with sig. D1 he kept the longer measure but settled
on 38 lines and the original vertical type-page opening of 176(189)
mm.
The anonymous play The Triumphs of Virtue (1697)
is
unfortunately without the running-titles which might assist in
solving its printing, but the facts of its typography may supply
some bases for conjecture. The book is a quarto signed
A-H4 and
paged 1-4 57 68-40 33-55 56 [=64]. The pagination numerals
in the headlines are in smaller type in sheets E-H than in A-D. The
type-page in A-D measures 44 lines, 182(193) x 108, 82.6R; that in
sheet E, 44 lines, 178(188) x 113, 80.8R; that in F-H, 47 lines,
192(202) x 118, 80.1R. The pagination suggests that the book was
originally planned to be split between two presses in sections A-E
and F-H, and that the second press beginning with sig. F1 paged it
33 on the assumption that pagination would start with page 1 on
sig. B1, whereas in fact it begins on sig. A3 with page 5. Although
one might be tempted to conjecture that the smaller font was
adopted by the first compositor in order to compress into the
single sheet E
rather more copy than had been allowed for in the casting-off, the
change in the measure and also in the whole type-page opening (thus
presumably in the skeleton-formes) militates against this view.
Since the size of the pagination figures in the headlines
associates sheet E, instead, with imposition by the compositor of
the second section, one might apply the same theory to him, but
again the measurements do not encourage this attempt. One fact is
clear, at any rate: although sheet E had originally been assigned
to the first press, actually the second compositor imposed it, the
machining taking place
after the conclusion of the F-H section (as indicated by the
pagination). One may possibly speculate that the completely
different typography of sheet E, showing the construction of a new
skeleton-forme, may have resulted from the confusion of the
pagination between the two sections, so that when the first
compositor came to page 40 (D4
v) he believed he had
joined the
two sections of the book, since sheet F of press two began with
page 41. Only very much later, when the sheets were actually
collated—perhaps even for binding—was it discovered that a
sheet of text had, in truth, not been set, and thus sheet E may
have been composed and printed at a considerably later time to
complete the book, certainly at a time after the original
skeleton-formes had been broken up.
Important as it is for a study of the presswork to identify the
compositors of two or more contiguous sections of a book, one of
the more striking examples of the usefulness of the printer's
measure occurs when this evidence assists in identifying the
compositor and also the place of printing for cancels and other
separate material originally imposed elsewhere in one forme. Under
most circumstances the evidence of the measure alone is not
decisive, but certainly a study of cancels shows that the odds are
against any material added to a book at a later date than the
original printing being set in an identical type-page opening, but
more especially in precisely the same measure.[6] Thus the fact that the measure of
cancellans leaf G1 in Dryden's The Indian Emperour (1667)
is
that for the rest of the book assists in the belief that it was
printed as leaf K4, missing in the seven recorded American copies.
Just so, the measure makes it a certainty
(evidence of running-titles here assisting) that disjunct sig. E3
of Southerne's The Disappointment was printed by the first
press as leaf A4 and excised from the preliminaries to be
bound in its proper position. Similarly, although in this case the
fact can be proved by an aberrant copy, there would have been
strong reason to conjecture that the 1681 cancel title-leaf for
Crowne's
The Misery of Civil War (1680), which
transformed
it into the reissue
Henry the Sixth: The Second Part, was
printed as leaf K4 of
Henry the Sixth: The First Part (1681)
since the cancel title was set to the 114 mm. measure used in that
book. For these reasons Dr. Philip Williams in his "The 'Second
Issue' of
Troilus and Cressida, 1609," earlier in this
present volume, found the fact that the cancel fold in the quarto
was printed in the same measure as the text very comforting to
buttress the evidence of the running-titles that this same fold,
and not a part of some other book, was the material which was
undoubtedly printed in the same formes with half-sheet M.
We come, finally, to an unexplored and difficult use of the
measure as bibliographical evidence. As Dr. Charlton Hinman first
demonstrated,[7] his results later
being confirmed by Dr. Philip Williams,[8] spelling tests can be applied with
some
certainty to distinguish the work of different compositors setting
a book seriatim in relay. In this connection, the printer's
measure can usefully be employed on some fortunate occasions as
powerful corroborative evidence, and it may even become primary
evidence when on disputed pages the spelling tests are ambiguous or
when, as in the later seventeenth century, the growing uniformity
of spelling may make spelling tests of doubtful value. The prime
difficulty of the evidence of the measure lies in the fact that
when the two sticks were not in perfect adjustment, the variation
between them is sometimes no more than one millimeter, although
less difficulty is encountered
when the variation is two millimeters or the seeming maximum three
millimeters.[9] Moreover, uneven
shrinkage in
the paper may cause apparent variation to upset one's calculations,
as well as uneven pressure from the wedges.
As a test of the validity of this evidence, I chose of the three
quartos analyzed by Dr. Williams those two which are available in
photographic reproduction, Shakespeare's 'Pied Bull' Lear
(1608),[10] and Robert Armin's
The
History of the Two Maids of More-clack (1609),[11] both printed by Nicholas Okes.
Within
a slight non-significant variation without pattern and apparently
dependent on the tightness of the wedges, the short and long
measures in Lear of 80 and 93 mm. are constantly
maintained,
and this corresponds with Dr. Williams's spelling evidence
demonstrating beyond all question that only one of Okes's two
compositors set this play.
In The Two Maids of More-clack Dr. Williams found
that
the usual pattern was for compositor B to set
$1-2v of
each sheet, and compositor A $3-4v.
Measurements disclose
that compositor B used a measure of 88-89 mm. whereas
compositor A used a measure of 90-91 mm. From sheet C
on,
these measures coincide precisely with the identification of the
two compositors by spelling tests save in the two instances, sigs.
H3 and H4v, where Dr. Williams felt the trend of the
spelling
evidence enforced breaking the pattern and assigning H3 and
H4v
to compositor B, although they would normally have
constituted part of A's assignment. Since in both these
pages the 90-91 mm. measure indicates that A actually set
these pages and that the regular pattern was maintained in this
sheet, the evidence of the printer's stick proves a useful
counterweight as a check on spelling tests in cases of doubt.
Although the check of measure
against spelling as an identification is invariable in sheets C-I
except for these two pages, there is some difficulty in sheets A
and B;[12] nevertheless, the
consistency with which the evidence
of the measure operates in the other seven and a half sheets in the
book demonstrates that it can be highly effective.
Some results accrue when the evidence of the measure is applied
to Restoration play quartos where spelling tests would be doubtful.
Crowne's The Country Wit (1693) is a difficult book because
it is hard to decide whether certain variations of a millimeter in
one compositor's measure are non-significant, or whether they
represent the stick of a third compositor. If we take the more
difficult but probable view that three compositors were associated
with this book, we find that compositor I with a measure of
120-121 mm. set both formes of sheet A (which contains text as well
as prelims) and then dropped out for two sheets. Compositor
II, with a measure of 122 mm., then took over and set both
formes of sheet B. The third compositor III, with a measure
of 123 mm. seems to have set C1-2, and thereupon II and
III alternate, II apparently setting
C2v-3v,
III C4, and II C4v.
Compositor I
returns to
set both
formes of sheet D, followed by II setting both formes of
sheet E. Compositor III, enters with F1-2 and
F4-4v,
II interposing with F2v-3. Compositor
I set
G1-3v and perhaps the rest of G although II
seems to have
composed G4 and just possibly G4v. Gathering H, very
curiously
(since this is a second edition), is only a single leaf and is set
by III. Perhaps there was confusion in imposition even
though the pagination is continuous: it may be significant that
with the re-entrance of compositor I on G1 the two skeletons
which had printed each sheet of the book exchange their formes in
sheet G, this arrangement carrying over to sheet I. At any rate,
compositor
II set I1-2
v and
III
concluded the
book
with I3-4
v. I cannot guarantee the absolute accuracy of
every
page of this assignment since the tolerances are sometimes very
fine between compositors
II and
III; but this is
what
I make of the Harvard copy, and I am inclined to believe that a
pattern develops which is accurate in the main and which is not
inconsonant with a reasonably exact identification of compositors.
In the Harvard copy the paper of sheets D and G, set by compositor
I, seems to differ from that in the rest of the book.
From several other plays Thomas D'Urfey's A Fond
Husband
(1677) may be selected. Here a fairly regular pattern is
established of about four to five type-pages apiece between two
compositors in relay using measures of 112 mm. and 113 mm.
respectively. This play is especially interesting because, although
it is a first edition and thus set from manuscript, the evidence of
the measure seems to indicate that for the first two text sheets
(possibly to get formes as quickly as possible at the start to the
waiting press or presses) the compositors cast off copy and set by
formes. Thus the 112 mm. measure set the outer formes of sheets B
and C, and the 113 mm. measure the inner formes. Thereupon they
begin to alternate, the 113 mm. measure beginning by composing most
of sheet D, both formes.
The evidential value of the measure is not invariable for there
are numerous books almost certainly set by two compositors whose
sticks were so nearly equalized that measurement cannot distinguish
them. Negatively, therefore, the evidence must always be equated
with that of the presswork as shown by running-titles, or as Allan
Stevenson has demonstrated,[13] by
watermarks, before an invariant measure may be
taken as indicating the presence of only one compositor, spelling
tests not having been applied.
[14]
When, however, positive evidence is available that two measures
were used in the composition of a book, the analytical
bibliographer may find the information thus gained to be of
considerable value in any number of unsuspected ways to which his
ingenuity may lead him.
[15]
Notes