University of Virginia Library

FORMAT

"The Concept of Format" was published in Studies in Bibliography, 53
(2000), 67–115. In it I offered a new definition of format based on the
number of page-units selected to fill each side of a piece of paper or
parchment of a given size; this formulation makes the concept applicable
to manuscripts as well as printed matter of all kinds and periods. A few
years after my essay appeared, B. J. McMullin provided what he called
"an extended, if somewhat discursive, footnote" to it: "Some Notes on
Paper and Format," Bibliographical Society of Australia and New Zealand
Bulletin, 28.4 (2004), 92–104. He summarizes my arguments and gives
several illustrative examples, and I appreciate his careful consideration
of the points I made. I especially like his observation that, in cases where
the format is uncertain and requires explication, "we should not agonise
over ascribing a particular notation to the volume." His discussion does
not require any modification in my essay, but readers will benefit from
following his line of thinking about specific books.

I do wish, however, to comment on his final paragraph. Since there
are cases where the designation of format varies according to the bibliog-
rapher's way of approaching it (as when double-size paper is cut before
printing), he asks, "If format is an intrinsic characteristic should it ever
be subject to variant designations, dependent on how it is defined?" He
raises here a basic philosophical point, which relates to all description. If
we grant that objects do have intrinsic qualities (as opposed to what we
project onto them), those qualities can only be apprehended and reported
as filtered through our individual perceptions and judgments. Every part
of a bibliographical description necessarily involves analysis and interpre-
tation, and the format statement is no exception. Nor is there reason to
believe that, by devising a comprehensive concept of format, we might be
"trying to force discrepant items into a uniform mould." The fact that we
may not be able (or at least so far have not been able) to determine with
certainty the format of many nineteenth- and twentieth-century books
does not mean that the concept has failed. Indeed, a single concept is
necessary to make clear how situations differ.

The difficulty of determining the format of books printed on unwater-
marked wove paper (which became increasingly common during the sec-
ond half of the eighteenth century) has long been recognized, and recent
research on the transitional period before and after 1800 provides the
major analytical techniques to be added to those mentioned in my es-
say. McMullin, in another article ("Watermarks and the Determination
of Format in British Paper, 1794 –circa 1830," Studies in Bibliography, 56
[2003–04], 295–315), shows that format can be established for hand-

made wove paper manufactured after early 1794 because it does contain
a watermark as a result of a legal requirement. A refund of part of the
paper tax for exported books was available if a date-watermark were eas-
ily visible; thus moulds were constructed with such marks along edges,
usually one or both of the longer edges, sometimes in opposite corners
and sometimes in all four corners, and often with the maker's name as
well. McMullin provides diagrams showing the positions of these marks
in various formats, and by reference to these diagrams (which form a use-
ful supplement to Gaskell's diagrams in A New Introduction to Bibliography
[1972, 1974]) one can determine the format of books printed on such
paper. (A further diagram, showing edgemarks along the shorter sides of
the mould, is given by Carlo Dumontet in "An Unrecorded Position of
Watermarks in Early Nineteenth-Century English Paper," Script & Print,
35 [2011], 111–113.) That the dates in these watermarks must be used with
caution was made clear in an earlier article by Hilton Kelliher ("Early
Dated Watermarks in English Paper: A Cautionary Note," in Essays in Pa-
per Analysis, ed. Stephen Spector [1987], pp. 61–68): some papermakers
did not change the "1794" watermark in later years, and it could happen
that a mark would be postdated.

McMullin also discusses another development of the same period: the
growing use of machine-made paper. Although he indicates some ways
of distinguishing machine-made from handmade paper, along with the
possibility in some instances of identifying format by examining the pat-
tern of original edges in uncut copies, he notes that the cutting of paper
in edition-binding eliminates that possibility for most books beginning in
the 1820s. One clue that remains, however, is the seam marks (that is, the
imprint left in the paper by the seams joining the ends of the wire-mesh
conveyor belt that replaced the hand-held mould); and Catherine M. Ro-
driguez has explained how the pattern of occurrences of these seam marks
can sometimes enable one to learn the format (see "The Use of Web Seam
Evidence to Determine Format," Bibliographical Society of Australia and New
Zealand Bulletin, 28.3 [2004], 122–124).

A few years later, McMullin pursued this point in an impressive article
that should be read by every bibliographer dealing with pre-1850 books
printed on machine-made paper ("Machine-Made Paper, Seam Marks,
and Bibliographical Analysis," The Library, 7th ser., 9 [2008], 62–88). He
gives a clear description of the process of producing paper by machine
and the resulting characteristics of the paper (not only seam marks and
evidence of wire-belt repairs but also a thinning along the original edges
as a result of slippage of the pulp under the deckle straps). Seam marks
may be either vertical or horizontal in the leaves of books, or approxi-
mately so—but sometimes at angles that cause them not to run through

contiguous leaves in an unfolded sheet (as McMullin shows in three dia-
grams). The pattern of occurrences of seam marks in a book, when stud-
ied with the help of these diagrams and the standard imposition diagrams,
can occasionally enable the bibliographer to determine the format—but,
as McMullin cautions, additional evidence is usually needed to make a
conclusive determination (except when a seam mark spans two or three
gatherings). One must remember, in dealing with seam marks, that they
may be present in some, but not all, copies of a gathering and thus that
their pattern of occurrence will vary among the copies of an edition.
(McMullin also discusses a few other analytical uses of seam marks: see
below, under "Paper.")

Another difficulty related to format in books from the later eigh-
teenth and the early nineteenth century is posed by the increasing use
of eighteenmo (see above, under "Collation"), which raises the question
of how to tell whether a book gathered in sixes is duodecimo or eigh-
teenmo. Pamela E. Pryde has answered it by summarizing three imposi-
tion schemes for duodecimo in sixes and two for eighteenmo in sixes,
established by drawing information from eight printers' manuals (plus
Gaskell); and she adds two more for eighteenmo in sixes, with diagrams
("Determining the Format of British Books of the
Second-Half of the Eighteenth Century Gathered in Sixes," Bibliographical Society of Australia
and New Zealand Bulletin, 23 [1999], 67–77 [cited for a different purpose
in note 42 of my essay]).

All the imposition diagrams mentioned above supplement those re-
ferred to in note 82 of my essay. The relevant lists in my Introduction
to Bibliography (2002 revision) are sections 9D6, 9F3–4, 9G6, and 9H5
(pp. 267–268, 291–293, 309–310, and 315, respectively). My Bibliographi-
cal Analysis (2009) briefly summarizes the primary means for determin-
ing format on pp. 38–39, 47, 53, 57–58, and 100–101 (note 12). Since
format is a basic fact about every book, bibliographers should employ
all available techniques for trying to discover it; thus the progress that is
being made in learning how to use the evidence in machine-made paper
is particularly welcome.