University of Virginia Library

PRINTING THE FAERIE QUEENE IN 1590

by
Andrew Zurcher [*]

When Edmund Spenser's The Faerie Queene went into publication for the
first time in 1590, it was the largest work of English poetry ever seen
through the press by a living author. Spenser apparently had some experience of
printing-house proofing, garnered during what appears to have been a carefully-
organized printing, by Hugh Singleton, of the first edition of The Shepheardes
Calender in 1579;[1] but the pressures of producing The Faerie Queene—so far sur-
mounting Spenser's earlier verse as epic, in generic terms, towers over pastoral—
must have been immense. Although the layout of The Shepheardes Calender had
been somewhat complex, including a range of prefatory materials, a woodcut and
'argument' at the head of every eclogue, and a 'glosse or scholion' attached to the
back of each eclogue, the sheer personal stakes of the proper presentation of the
first instalment of Spenser's epic made this the riskier venture. In 1579, Spenser
had been a young man, personal secretary to John Young, Bishop of Rochester,
recently graduated with an MA from Cambridge, and with a promising career
ahead of him; by 1590, he had been living ten years in Ireland, during which time
he had published no further poetry, had retired from the comparative bustle of
Dublin to the anonymity of Kilcolman, near Cork, and had apparently lost most
of his once-promising patronage connections: Philip Sidney (d. 1586), Arthur
Lord Grey of Wilton (back in England from 1582), and the Earl of Leicester
(d. 1588). Spenser's return to public notice in 1590, with the advertisement of a
new patronage connection to Walter Ralegh, and—above all—his favorable re-
ception by Queen Elizabeth upon presentation of a manuscript copy of The Faerie
Queene (and himself?) at court, gave him a new opportunity to secure his status as
'England's arch-Poët'. It was an opportunity that he could not afford to lose.

With this in mind, the 1590 edition of The Faerie Queene stands out, from
one perspective, as a landmark example of a poet experimenting with print pa-
tronage conventions in an innovatory and ambitious way: the volume originally
contained a dedication to Queen Elizabeth seven commendatory poems, ten
dedicatory sonnets to aristocratic patrons, and an epistle from Spenser to Ralegh
giving important context for 'the Authors … whole intention in the course of
this worke'; the number of dedicatory sonnets to court and government lumi-
naries was later augmented to seventeen. The barrage of dedications and com-
mendations, alongside an open acknowledgment of the support of the queen
and the intimacy of her principal favorite, would seem a bravado display of a
poet's achievement. From another perspective, though, the 1590 edition of The
Faerie Queene almost looks like a critique, even a burlesque, of the contemporary
culture of literary patronage. The proliferation of dedicatory sonnets, especially
in the revised state, makes any one seem superfluous (in the spirit of the motto
from 'September' of The Shepheardes Calender: inopem me copia fecit), and simply
by its mass seems to outweigh what must have been the primary dedication
to the eponymous Gloriana, Elizabeth. In addition, all of the prefatory mate-
rial appears not at the front of the volume, but at the back—except for the
dedication to Elizabeth which was thrust, perhaps hastily and probably disre-
spectfully, into the only available space in the first gathering, the verso of the
title page.

Modern scholars have often debated the poetics, politics, and printer's exi-
gencies that might have led Spenser, or his publisher William Ponsonby, or his
printer John Wolfe, to lay out the prefatory matter in this final (mostly terminal)
form.[2] Some readers have supposed that Spenser had always intended his prefa-
tory material for placement at the back of the volume, answering the title page
(and the royal dedication) at the front of the volume in the same way a plural
body politic mirrors its single head and ruler. Others have supposed that Spens-
er's hand was forced by the unanticipated interest and patronage of the queen,
and that he had to devise a suitable format to accommodate existing debts of
patronage without insulting his royal patron or jeopardizing the newly-awarded
annuity (confirmed by patent in January 1591 at £50 p.a.) that she had promised
him.[3] Still others have suggested that the final arrangement was dictated not by
Spenser, but by Wolfe who, as an experienced printer and a very efficient and
perhaps impatient businessman, received the collection of dedicatory materials
too late to incorporate them into the opening gathering. These interpretations

have been variously cumbered and supported by the inter-dependence of the
decisions taken by Spenser or Wolfe regarding the first and last gatherings. The
decision to place the title page on the recto of the first leaf of the opening gather-
ing, A1, and to begin the text of the proem to Book 1 on the recto of A2, must
have coincided with the decision to relegate any prefatory material to the back of
the volume (in a full, two-sheet gathering eventually signed Pp), as Wolfe would
presumably never have planned to insert a mis-paginated eight-leaf gathering
of prefatory materials between A1 and A2. The need to place a dedication to
the queen at the front of the volume would seem the obvious, and perhaps the
only, reason to have displaced the prefatory material, and yet copies of the 1590
edition survived without the dedication to Elizabeth:[4] the royal dedication was
inserted on A1v as a stop-press correction to the inner forme of the outer sheet
of A. Finally, Spenser prevailed upon Wolfe to run off a single-sheet gathering,
Qq, containing seven additional dedicatory sonnets, to replace Pp6-7, perhaps
immediately upon 'completion' of the printing, or sometime reasonably shortly
thereafter. For a publication for which Spenser might have been supposed to
have taken especial care, the bibliographical record seems to tell a story of ac-
cident, confusion, and revisions—a story that has amply admitted of a range of
scholarly interpretations.

But if modern observers have been interested in the way Spenser handled,
or mis-handled, his dedications in 1590, at least one contemporary observer
seems to have found the whole business interesting, and of substantial explana-
tory power as a critique of the contemporary culture of literary patronage. As I
have argued elsewhere, Thomas Nashe furnished in his 1592 satire, Pierce Peni-
lesse His Supplication to the Diuell, a stinging parody of the printing of The Faerie
Queene in 1590 as a publishing event.[5] Nashe has many targets in his satire, but
the ad hominem invective is aimed above all at the cause of precise Protestants
(non-conforming ministers, and their Puritan followers), their now-dead cham-
pion Robert Dudley, earl of Leicester, and the man who had seemed to be tak-
ing up Leicester's old precisian clients, Walter Ralegh. Ralegh's star had fallen
disastrously in the late spring of 1592, when his clandestine marriage to Bess
Throckmorton was advertised to the Queen and made court gossip; by the time
Nashe produced Pierce Penilesse in September 1592, Ralegh had been committed
to the Tower. Pierce's patronage project—a scholar's desperate, Faustian turn to
the succour of the devil—combines Nashe's vitriolic contempt for Ralegh with a
series of parodic emulations of Spenser's layout of the 1590 edition of The Faerie
Queene, above all in repeated references to a collection of dedicatory sonnets (all
of them addressed to die dead) that Pierce, or perhaps Nashe, intends to deliver
to the devil, or the printer, to be inserted 'at the latter end' of his supplication.
In a lengthening string of insinuations, backhanded compliments, and explicit
criticisms, Nashe and his printers Richard Jones (first edition) and Abel Jeffes

(later editions) re-imagine the 1590 Faerie Queene as a physical witness to the same
fissures in print patronage politics that eventually caused both Spenser's disgrace,
upon the suppression of Complaints in 1591, and Ralegh's fall from power.[6] Nashe
concludes Pierce Penilesse with an explicit discussion of Spenser's own dedicatory
sonnets, in which he accuses Spenser of some act of 'forgetfulnes' while himself
celebrating the projected rise of a new literary and political patron, Ferdinando
Stanley, Lord Strange.

Nashe's sustained and often precise parody of Spenser seems to locate the
source of Spenser's eventual disgrace and patronage fiasco of 1591 in the printing
event of the 1590 edition of The Faerie Queene. The repeated emphasis in die early
prefaces and the final episde of Pierce Penilesse on the sonnets Nashe intended to
append to its 'latter end'—sonnets that Nashe does not send because, he claims,
he is kept prisoner by the plague in the country—suggests that the communi-
cations between Spenser and Wolfe broke down in 1590, and that this might
somehow have been connected to Spenser's unusual disposal of his prefatory
material, and in turn to his later patronage problems. Nashe's parodic account
gives us some sense of the reception of Spenser's volume, and the scuttlebutt that
surrounded its publication in the years after it appeared; but it is much harder to
know whether the layout of the volume reflected Spenser's intentions in the first
place, and what those intentions might have been. The bibliographical evidence
has, to date, resisted explanation: when and why Spenser or Wolfe decided to
move the dedicatory sonnets to the back of the volume, when and why they
decided to print the title page to the work in gathering A, when and why they
added the dedication to Elizabeth on the verso of the title page, when and why
the number of dedicatory sonnets was expanded—these related questions, the
answers to which would give us some purchase on Spenser's intentions behind
his layout, have seemed unanswerable. Scholarly work attempting to interpret
the layout of the 1590 edition of The Faerie Queene has proceeded in the absence
of the very bibliographical evidence the interpretation of which ought to provide
its basis. This paper returns, in light of Nashe's insinuations about Spenser's or
Wolfe's misjudgments in 1590, to the bibliographical evidence, attempting to
provide a sounder footing on which to judge Spenser's intentions for the layout of
his most important publication. As Nashe's jibes suggest, the story of the printing

of The Faerie Queene in 1590 is likely more complicated, and more revealing, than
bibliographers have, to date, led us to believe.

A View of The Faerie Queene in 1590: F. R. Johnson
and Yamashita et al.

The meticulous and enduring work of F. R. Johnson on the printing of the
1590 edition of The Faerie Queene has never been significantly challenged, and the
picture he first described of the printing process in Wolfe's shop has remained
the dominant view; recent expansions of, and some adjustments to, his analysis
have been supplied by Hiroshi Yamashita et al., particularly touching composito-
rial practice in the edition.[7] Ponsonby may have chosen Wolfe, with whom he had
never worked before, because Wolfe was an efficient printer with a considerable
capacity and Spenser's was a job with a tight timetable;[8] Wolfe had in 1586 also
printed, for John Harrison the younger, the most recent edition of Spenser's
The Shepheardes Calender. Ponsonby was developing a good relationship with John
Windet at about this date, but the latter may have been tied up printing Sidney's
Arcadia, and his shop did not have Wolfe's resources. Situated from 1588 in Sta-
tioners' Hall, with four presses, a team of experienced and apparently very able
compositors, and an ample stock of type and paper, Wolfe could offer fast and
reliable publication—his bilingual edition of Castiglione's The Courtier, like the
line of works by Machiavelli that he produced with false imprints, was published
to an exacting, and probably very expeditious, standard. As Johnson's discus-
sion of Wolfe's printing of The Faerie Queene demonstrates, and as the analysis of
compositional practice undertaken by Yamashita et al. has confirmed, the work
was parcelled up very precisely, and printed by an organized team with remark-
able consistency at what appears to have been a clip—so fast, it may be, that
Spenser did not have time to gather his thoughts about the dedicatory, commen-
datory, and epistolary material he had perhaps intended, though printed last, to
be bound at the front of the volume.[9]

Johnson's basic outline of the organization of the work is almost certainly
correct. Wolfe had four presses at this date, but he appears to have used only two
of them to produce The Faerie Queene. It is a given that a printer as experienced
and shrewd as Wolfe would have maintained a steady flow of production on his
presses; to keep the Spenser job going, Wolfe's compositors and press-men used a
system of four skeleton-formes, where two skeleton-formes were being printed at
any given time while the other two skeleton-formes were being set. The running-
titles in these four skeleton-formes were exactly retained (with a few very impor-
tant exceptions) throughout the 600-page job.[10] Because the edition was printed
as a quarto in 8s, each (two-) press stint produced half a gathering, running off
the full complement of one side of one sheet each; when the second set of skel-
eton-formes was completed and locked in the press, the sheets were perfected,
producing the full complement of a two-sheet gathering: 8 leaves (16 pages). As
Johnson noticed, the type from a completed forme was then redistributed at the
same time that the chase was re-dressed, so that the compositor could use the
single available signature on each skeleton-forme as a guide for the layout of
the pages on the next forme: the absolute regularity of this method, Johnson ar-
gued, meant that the skeleton-formes were not reversed, even once, throughout
the entire printing.[11] The collation of the edition up to the original final gathering
(Pp), which was altered after the printing had been concluded, is regular: 4o in

8s: A8-Pp8. This was subsequently modified to incorporate the four-leaf cancel-
lans, producing the final collation: A8-Oo8 Pp8 (-Pp6,7; Pp5 + Qq4). There are a
few comparatively minor anomalies in the running-titles, in the pagination, and
in the sigla, which Johnson accurately noted.[12] The stanzaic distribution across
pages is uniform throughout the entire volume: each opening contains exactly
seven stanzas of text (where the text is continuous). Ornamental boxes contain-
ing the inset arguments at the heads of cantos were designed to take up exactly
the same space as a stanza, no doubt in order to regularize layout and simplify
casting-off, and were never printed in the bottom position on a page; proems to
each of the three books were set on the recto and verso of a leaf; and a single
(verso) page was left between Books 1 and 2 (accommodating the woodcut of
St George and the dragon) and 2 and 3; otherwise, with a few important excep-
tions, the text runs continuously in each book.

Johnson supposed that the poem was printed continuously from start to fin-
ish, beginning with gathering A and running straight through to gathering Pp;
this assumption is retained by Yamashita et al. Noticing that, in gathering F, the
pagination is off by two in the outer formes of both sheets (affecting 8 pages in
total), Johnson argued that the inner formes were prepared by one compositor
or pair of compositors, and the outer formes by another compositor or pair of
compositors; the compositor setting the outer formes of F, Johnson reasoned,
had temporarily forgotten that the text of the poem began on A2r and not A1r.
Johnson imagined that one press was therefore used to print the inner formes of
both sheets, and the other the outer formes.[13] Johnson assigned the four skeleton-
formes used in the printing the numbers 1, 2, 3, and 4, and adopted a convenient
notation for inner and outer sheets, inner and outer forms: α i (outer sheet, inner
forme), α o (outer sheet, outer forme), β o (inner sheet, inner forme), and β o (inner
sheet, outer forme), which for conformity with his analysis I have retained. The
distribution of running-titles throughout the entire volume is remarkably consis-
tent, save for six anomalies.[14] The pattern adopted in gatherings C–S should be

considered standard for the print run, and runs (using the notation of Johnson
and, in parentheses, of Yamashita et al.):

         

skeleton-forme	forme
1 (I)	β o
2 (II)	α o
3 (III)	α i
4 (IV)	β i

The six exceptions are as follows:

• 1. Gatherings A and B are anomalous. Skeleton-forme I was used for both sides
  of the inner sheet (β o and β i) of A, and skeleton-forme II for both sides of the
  outer sheet (α o and α i) of A. Skeleton-forme III was used for both sides of the
  inner sheet (βo and β i) of B, and skeleton-forme IV for both sides of the outer
  sheet (α o and α i) of B.
• 2. Skeleton-forme HI was used to set both sides of the outer sheet (α o and α i) of
  gathering T. Skeleton-forme II was not used.
• 3. Skeleton-forme I was used to set both sides of the inner sheet (β o and β i) of
  gathering Cc. Skeleton-forme IV was not used.
• 4. The running-title on position 5r of skeleton-forme I was changed before print-
  ing gathering V: a period was added after 'The Faerie Queene'.
• 5. Some of the running-titles in skeleton-forme IV were altered after the print-
  ing of gathering Bb (they had been removed during the printing of Bb because
  Bb3v and Bb4r—the end of Book 2 and the beginning of Book 3—have no
  headlines).
• 6. Beginning with gathering Dd, skeleton-formes I and IV were switched, I there-
  after being used for β i, and β o.

Johnson offered no explanation for exceptions 1, 2, 3, 4, or 6; his explanation
for exception 5 is given above, though it should be noted that this explanation is
perhaps not as satisfactory as he suggested, given (as we shall see) that running-
titles were at other points removed from other skeleton-formes throughout the
printing process, and always afterwards carefully replaced.[15]

In their Textual Companion to The Faerie Queene 1590, Yamashita et al. have
documented and illustrated in detail the account of pagination, running-titles,
skeleton-formes, and other bibliographical data supplied by Johnson, adding to
this work their copious and compelling analysis of compositorial stints.[16] By dis-

criminating systematically between word- and punctuation-spacing, and between
regular orthographical patterns resulting from underlying spelling preferences,
Yamashita et al. distinguished three different compositors, who divided with
near-mechanical regularity the sixteen pages of each gathering into consistent
and equal shares.[17] Compositor X is thought to have set pages 1r—3r in gather-
ings C to Oo, with very few exceptions (H1r, H2v, N1r, Z1r, Bb1v, Ii1r, Kk1r, and
Kk1v); Compositor Y pages 3v—5v in gatherings C to Oo (though he was helped,
probably by X, in gatherings Hh and Ii, and on pages Ff4v, Ll5r, Nn4v, and Oo5r);
and Compositor Z pages 6v—8v in gatherings C to Oo, except for Y6v, which may
have been set by Compositor X. Page 6r (i.e. the eleventh page) in each gathering
appears to have been a 'swing' page, on which all three compositors may have
collaborated, variously in different gatherings. As for the (anomalous) first two
gatherings, Compositors X and Z seem to have divided A roughly evenly, while
all three compositors seem to have worked on B, largely according to the divisions
observed in the rest of the run. The compositorial analysis of Yamashita et al.
supports Johnson's picture of the extraordinary regularity of the printing of The
Faerie Queene in 1590: with only a few (and as we will see, explicable) exceptions,
the compositors setting the manuscript worked to a rigorously methodical divi-
sion of labor, very rarely varying their individual responsibilities.

Dedication, Vision, and Revision:
The Poet in the Printing House

While the computer-aided statistical work upon which the analysis of com-
position in A Textual Companion is based has undoubtedly added significantly to
our understanding of the printing of The Faerie Queene in 1590, the rest of the
analysis of Yamashita et al. perpetuates some of Johnson's unfortunate errors,
and adds new mistakes both of omission and commission, to our picture of the
use of the four skeleton-formes. These mistakes substantially cloud the evidence
of the order of imposition of the gatherings, which, it turns out, were not printed
sequentially. Yamashita et al. similarly stepped back from offering an interpreta-
tion synthesizing the various types of bibliographical evidence they had assimi-
lated from Johnson, and newly supplied themselves. As the following discussion
will make clear, there is a great deal of significant bibliographical evidence to
suggest that Spenser was intimately and presently engaged in the printing of the
text; that he substantially interfered with and delayed Wolfe's press run; that on
many substantial points of layout and composition, including the position of the

'prefatory' material and the dedication to the Queen, Spenser and the printer
remained unresolved until the last moment; and that many of the latest altera-
tions to the 1590 text (as would again be true in 1596) concerned contemporary
political allusions, and particularly the allegorical representations of Elizabeth.

One of the most ambitious and useful expansions of Johnson's work provided
in A Textual Companion to The Faerie Queene 1590 is in the form of a large table
that records in meticulous detail the disposal of the various running-titles, regu-
larly associated with one of the four skeleton-formes, throughout the volume. On
the basis of this table (a similar version of which Johnson must have generated in
his own research, though he did not publish it), the pattern of skeleton-forme use
can be reconstructed; correctly documenting the deployment of running-titles is
thus crucial to understanding the precise use of the skeleton-formes, and basic
to any interpretation of the order of imposition.[18] In all, Yamashita et al. identify
nineteen distinct running-titles, each of which is reproduced in facsimile in their
study. Distinguished by their typographical features, sixteen basic variants exist,
four to each skeleton-forme; for various reasons during the printing of the text,
according to these scholars, three new running-titles were composed and used in
the place of three originals. The two skeleton-formes regularly used from gather-
ing C to print the outer sheet of each gathering (II and III) remained remarkably
consistent throughout the press run. All of the changes to running-titles took
place in skeleton-formes I and IV, used to print the inner sheet of each gather-
ing, and were made in gathering V and after; the imposition of the inner sheet
became increasingly troubled after gathering Bb, and most of the changes to
running-titles took place between this point and the end of the job (Pp). These,
and further observations on the pattern of use of the running-titles, can be fol-
lowed in table 1, a corrected version of that compiled by Yamashita et al., which
also (for reasons that will become clear later) includes their findings on com-
positors' stints; for ease of reading the table, I have also reproduced analyses of
skeleton-forme construction (figure 1).

The first serious mistake, though it seems minor, in this presentation is that
these scholars have distinguished two running-titles—16 and 17—that are in fact
the same setting of type, and thus the same title; the printing of this title, used
regularly in gatherings C to Bb to print the recto of the sixth leaf (i.e. position 6r),
became intermittently corrupt during the course of the print run, such that by
the time of the later gatherings the terminal 'ne' of 'Queene' in 'The Faerie
Queene' seems regularly to have split and recombined, and appears deformed.
Yamashita et al. distinguish the damaged presentation of this title from the sound

FIGURE 1. The initial disposition of running-titles in the four skeleton-formes. The introduc-
tion of running-titles 11 (a modification of 10) and 19, along with the exchange of skeleton-
forme I for IV and the reversal of orientation of skeleton-forme IV, can be traced in table 1.
A similar figure appears in Yamashita et al. (1990).

FIGURE 2. Damage to running-title 16 in position 6r of gathering P (a) is similar to that
apparent in the same running-title in position 3r of gathering Ll (b). While the running-title
occasionally appears sound, as in position 6r of gathering K (c), or indeed in position 6r of
gathering T (not shown), it is clear from the regular fluctuations throughout die printing pro-
cess that a clean-state 16 is not to be distinguished from a foul-state 17. The copy illustrated
here is Cambridge University Library, Sel. 5. 101. Reproduced by permission of the Syndics
of Cambridge University Library.

presentation, and thus note repeated transitions from running-title 16 to 17 and
back again, particularly in the period of printing after gathering Dd, substantially
cluttering and confusing the course of events in this difficult area. While it is true
that the type in running-title 16 shows various and shifting states in this part of
the text, the range of distinction between its cleanest and its foulest presentations
here is no larger than, for example, its appearances in the inner forme of the
inner sheet in gatherings K to M. As is obvious from the illustrations presented
in figure 2,[19] the occasional deterioration and subsequent improvement in this

Table 1. Running-titles and compositorial stints in The Fearie Queene, 1590

                                           

Sig.	OUTER SHEET																INNER SHEET
	Outer forme								Inner forme								Outer forme								Inner forme
	1r		2v		8v		7r		2r		1v		7v		8r		3r		4v		6v		5r		4r		3v		5v		6r
A		X	1	X	4	X	12	Z		X		X	4	Z	12	Z		X	2	X	3	Z	10	Z	9	X	2	X	3	Z	10	Z
B	13	?	5	X	8	X	16	Z	13	X	5	?	8	Z	16	Z	14	X	6	Y	7	Z	15	Y	14	Y	6	Y	7	Y	15	?
C	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
D	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
E	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
F	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
G	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
H	18	?	1	?	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
I	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
K	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
L	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y
M	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
N	18	?	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
O	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
P	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
Q	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
R	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
S	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?
T	14	X	6	X	7	Z	15	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	10	Y	13	Y	5	Y	8	Y	16	?

                                   

V	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	11	Y	13	Y	5	Y	8	Y	16	?
X	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	11	Y	13	Y	5	Y	8	Y	16	?
Y	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	X	11	Y	13	Y	5	Y	8	Y	16	?
Z	18	?	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	11	Y	13	Y	5	Y	8	Y	16	?
Aa	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	11	Y	13	Y	5	Y	8	Y	16	?
Bb	18	X	1	X	4	Z	12	Z	14	X	6	?	7	Z	15	Z	9	X	2	Y	3	Z	11	Y					8	Y	16	?
Cc	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	9	X	2	Y	3	Z	11	Y	13	Y	2	Y	3	Y	11	?
Dd	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	16	X	5	Y	8	Z	13	Y	9	Y	2	Y	3	Y	11	?
Ee	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	19	X	5	Y	8	Z	13	Y	9	Y	2	Y	3	Y	11	?
Ff	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	16	X	5	Y+	8	Z	13	Y	9	Y	2	Y	3	Y	11	?
Gg	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	16	X	5	Y	8	Z	13	Y	9	Y	2	Y	3	Y	11	?
Hh	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	16	X	5	Y+	8	Z	13	Y+	9	Y	2	Y	3	Y+	11	?
Ii	18	?	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	16	X	5	Y+	8	Z	13	Y+	9	Y	2	Y	3	Y+	11	?
Kk	18	?	1	X	4	Z	12	Z	14	X	6	?	7	Z	15	Z	16	X	5	Y	8	Z	13	Y	9	Y	2	Y	3	Y	11	?
Ll	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	16	X	5	Y	8	Z	13	Y+	9	Y	2	Y	3	Y	11	?
Mm	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	16	X	5	Y	8	Z	13	Y	9	Y	2	Y	3	Y	11	?
Nn	18	X	1	X	4	Z	12	Z	14	X	6	X	7	Z	15	Z	16	X	5	Y+	8	Z	13	Y	9	Y	2	Y	3	Y	11	?
Oo	18	X	1	X			12	Z	14	X	6	X	7	Z	15	Z	19	X	5	Y	8	Z	13	Y+	9	Y	2	Y	3	Y	11	?

NOTE. Arabic numerals represent running-title, keyed to the scheme divised by Yamashita et al. (1990), with correction; blanks in the numer-
ical columns indicate pages where no running-titles were used. The letters X, Y, and Z record compositorial stints hypothesized by Yamashita
et al. (1990); Y+ represents Y working with limited support from another compositor, probably X; ? denotes apparently collaborative work, by
two or more compositors; blanks in these columns indicate pages where no text appears.

FIGURE 3. The substitution of a swash 'C' in 'Canto' in running-title 13 between position
5r of gathering Ll ((a), above) and 5r of gathering Mm (b) leaves a telltale marker providing
evidence of the sequence of imposition of gatherings. Because the swash majuscule appears
in position 5r of gatherings Dd (c), Ee, Mm, Nn, and Oo (not shown), but not in gatherings
Ff to Ll, it is obvious that Dd and Ee were printed out of sequence; the coincidence of the
introduction of running-title 19 (for 16) in gathering Ee, combined with its (only) reappear-
ance in gathering Oo, suggests that gatherings Dd and Ee were printed between Nn and Oo.
This pattern is confirmed by the substitution of a colon (for a full stop) at the end of running-
title 13 in position 5r of gathering Nn; this colon also appears in the same running-title in
gatherings Dd ((c), above), Ee, and Oo. The copy illustrated here is Cambridge University
Library, Sel. 5. 101. Reproduced by permission of the Syndics of Cambridge University
Library.

particular running-title, both between extant copies of the same gathering, and
between gatherings, is a usual feature of the work well before the period after
Dd, and we should therefore not distinguish between running-titles 16 and 17;
in short, 17 does not exist.[20]

Yamashita et al. make a related mistake in the same sequence of gatherings,
with respect to their documentation of running-title 13; this time it is an error of
omission repeated from Johnson's analysis. In gatherings Dd, Ee, Mm, Nn, and
Oo, running-title 13 appears in a slightly modified state on the recto of the fifth
leaf (position 5r) of each gathering; as will be immediately obvious from figure 3,
the italic majuscule 'C' of 'Canto' has been exchanged, in these gatherings, for
a swash majuscule. Given the near-absolute regularity with which running-titles

are generally retained throughout the printing of the text, the way this swash 'C'
seems to appear, disappear, and then reappear, should be treated with extreme
suspicion: there is no other place in the sequence where a change to a running-
title is subsequently changed again, and certainly no place where this correction
is then changed anew, in the same way. The fact that the introduction of this
swash 'C' to running-title 13 is coincident with another change to skeleton-forme
IV—the substitution of running-title 19 for 16 in gathering Ee—makes it all too
obvious what took place. This part of the text was not printed sequentially. Fol-
lowing the printing of gathering Cc, the compositors moved on to gathering Ff,
then proceeded through to Nn before returning to Dd and Ee; after completing
Ee, they finished the printing of the main part of the text of the poem with gather-
ing Oo. This sequence is confirmed by another minor change made to running-
title 13 in the printing of gathering Nn, where a colon was substituted for the full
stop after 'Queene'; this colon also, and only, appears in the same running-title
in gatherings Dd, Ee, and Oo. It is crucial to note that this pattern of imposition
was not immediately apparent in the analysis of Yamashita et al. because of the
confusion created by the false distinction of a supposed running-title 17.

There are three further pieces of evidence that confirm this non-sequential
pattern of imposition in the final gatherings of the volume, again information
overlooked or not collected by Johnson or Yamashita et al. The compositors
obviously had to adjust the running-titles every time they were used, updating
pagination and, less frequently, canto headings for all running-titles, and, on
two occasions, altering the book headings. Generally speaking the compositors
were conservative about changing the running-titles between formes, making as
few substitutions as possible. Thus, in an instance where 'Cant. II.' was to be
changed to 'Cant. III', the compositor might simply replace the full stop with
an 'I', to avoid unnecessary adjustment of the type spaces between the central
and peripheral elements of the headline. A result of this conservatism is that the
spaces between the pagination and the central element of the running-title, and
between the central element and the canto heading, tend to remain fairly con-
stant, but not perfectly constant, in width, from gathering to gathering. Measur-
ing these distances in each headline, from gathering to gathering, thus produces
a suggestive indication of the order of imposition of gatherings. Table 2 reports
the distances between the pagination (i.e. the outer element) and the central ele-
ment in the headline for running-titles 2, 3, 5, and 8 in gatherings Aa to Oo in
the Cambridge University Library copy (Sel. 5. 101). The evidence of running-
title 3 is probably too erratic to construe, and that from running-title 5 too
regular; but the patterns in the spacing of running-titles 2 and 8 are suggestive.
Running-title 2, in particular, is extremely conservative in the spacing of this part
of the headline, changing by only a single space between the printing of gather-
ings Ff and Oo. That this space was added when the running-title was used to
print gathering Ee, and retained for gathering Oo (following immediately after-
wards) seems the likeliest explanation for this pattern. In running-title 8, too, the
spacing in this part of the headline shows a clear pattern of consistency between
gatherings Aa/Bb and Ff/Gg, on the one hand, and Dd/Ee and Ll/Mm/Nn,
on the other. The pattern of spacing in these running-titles across gatherings,
then, tends to support the idea that gatherings Dd and Ee were printed between

Table 2. Distance between pagination and central element of running-title

                           

Running-Title 2		Running-Title 3		Running-Title 5		Running-Title 8
Cc3v	2.0 cm	Cc5v	1.90 cm	Aa3v	1.05 cm	Aa5v	1.60 cm
Cc4v	2.0 cm	Cc6v	1.80 cm	Dd4v	0.85 cm	Bb5v	1.90 cm
Dd3v	1.3 cm	Dd5v	2.0 cm	Ee4v	0.85 cm	Dd6v	1.75 cm
Ee3v	1.4 cm	Ee5v	1.65 cm	Ff4v	0.85 cm	Ee6v	1.75 cm
Ff3v	1.3 cm	Ff5v	1.55 cm	Gg4v	0.85 cm	Ff6v	1.90 cm
Gg3v	1.3 cm	Gg5v	1.60 cm	Hh4v	0.85 cm	Gg6v	1.90 cm
Hh3v	1.3 cm	Hh5v	1.65 cm	Ii4v	0.85 cm	Hh6v	1.55 cm
Ii3v	1.3 cm	Ii5v	1.55 cm	Kk4v	0.85 cm	Ii6v	1.55 cm
Kk3v	1.3 cm	Kk5v	1.65 cm	Ll4v	0.85 cm	Kk6v	1.55 cm
Ll3v	1.3 cm	Ll5v	1.60 cm	Mm4v	0.85 cm	Ll6v	1.75 cm
Mm3v	1.3 cm	Mm5v	1.60 cm	Nn4v	0.85 cm	Mm6v	1.75 cm
Nn3v	1.3 cm	Nn5v	1.60 cm	Oo4v	0.85 cm	Nn6v	1.75 cm
Oo3v	1.4 cm	Oo5v	1.60 cm			Oo6v	1.60 cm

Nn and Oo, a pattern consistent with, if not supported by, the patterns of spacing
in running-titles 3 and 5 across the same gatherings.

Two other pieces of evidence are more conclusive. Both relate to mistakes
in the running-titles not collated by Johnson or Yamashita et al. Running-title 8,
used in position 6v as part of skeleton-forme IV in gatherings Dd to Oo, includes
(like running-title 5) as its central element the phrase 'The third Booke of'. In
gathering Aa, which includes the latter part of canto xii of Book II of the poem,
this running-title naturally reads 'The second Booke of'. The compositor respon-
sible for updating this running-title for its first use in printing Book III of the
poem—in position 5v in gathering Bb—accidentally removed the space between
'The' and 'third', such that the running-title presents a central element reading
'Thethird Booke of'. By the time this running-title appears in gatherings Dd
and Ee, it has been corrected; but, crucially, it appears in its uncorrected state
in gatherings Ff and Gg (see figure 4). Given the patterns already noted in the
use of running-titles 3, 5, 13, 16, and 19, it is obvious that the correction to run-
ning-title 8 was made between the imposition of gatherings Gg and Hh, and that
it appears corrected in Dd and Ee because these gatherings were printed there-
after. Similarly, Johnson and Yamashita et al. failed to notice a substantial and
coincident error in running-title 6, used in position 1v between gatherings C and
Oo. The compositor charged with updating this running-title between Books II
and III neglected to change 'second' to 'third' in the transition between books;
the central element in running-title 6 thus appears incorrectly in gatherings Cc,
Ff, and Gg as 'The second Booke of', but appears correctly in gatherings Dd
and Ee (as in Hh through to Oo) as 'The third Booke of'. Again, given what we
know about errors in and changes to the other running-titles in this part of the
printing, it is obvious that this mistake was noticed and corrected between the
imposition of gatherings Gg and Hh, appearing in its corrected state in Dd and
Ee because these gatherings were not printed until after Nn.

The coincidence of so many mistakes and irregularities in this part of the
print job may have been due to fatigue, but it certainly had something to do with
the need to skip gatherings Dd and Ee, an obstruction to the regular printing that

FIGURE 4. The central element of running-title 8 in position 5v of gathering Bb reads
'Thethird Booke of' (a); this mistake in spacing also appears in running-title 8 in position
6v of gatherings Ff (not shown) and Gg (b), but appears corrected in all other gatherings,
including Dd (not shown) and Ee (c). This evidence suggests that the spacing in running-
title 8 was corrected between the imposition of gatherings Gg and Hh, and corroborates the
hypothesis that Dd and Ee were printed at some point after Ff and Gg. The copy illustrated
here is Cambridge University Library, Sel. 5. 101. Reproduced by permission of the Syndics
of Cambridge University Library.

no doubt threw the compositors off. It would not be particularly safe, at least at
this point, to speculate too nicely on the reasons why the press-men went directly
from gathering Cc to Ff, but a number of elements attending this decision should
be noted. First, the most likely reason for skipping the material contained in
these gatherings was that the manuscript— or an authorized manuscript—was
not available at the time that the printer needed it. Regardless of the speed or
expedition with which we suppose the job to have advanced, the only likely
explanation for this, if we assume that Wolfe had not mislaid some of his sheets
(and we will have cause, shortiy, for believing that he did not), is that Spenser had
taken the relevant parts of the text away. It is further obvious that Wolfe's press-
men would not have skipped two gatherings, or those exact gatherings, unless
the material withheld or unavailable spanned them both; if the missing material
had been in Ee, Wolfe might have printed Dd as normal, and then gone to Ff
or, if it had been in Dd only, he might have skipped Dd and begun again with
Ee. Given that canto iii of Book III is the only canto entirely contained within
these two gatherings, it seems likely that it was this canto to which Wolfe's men
did not, mid-run, have access.

Much more importandy, from a bibliographical perspective, it should be im-
mediately obvious that this sequence of events following on from gathering Cc al-
most certainly has something to do with the considerable anomaly of the printing
of that gathering. As noted by Johnson, and remembered above, skeleton-forme I
was used to print both the outer and the inner formes of the inner sheet of gather-
ing Cc. In the context of the regular and efficient four-skeleton-forme/two-press
system used to print The Faerie Queene in 1590, this doubling of a single skeleton-

forme to print both formes of the same sheet strongly suggests that printing in
this gathering slowed, and proceeded with only one press. After the printing of
Cc, as Johnson also noted, the pattern of skeleton-formes used for the inner sheet
was reversed, with skeleton-forme IV used to print the outer forme, and skeleton-
forme I the inner forme. The compositors working on The Faerie Queene would as
a matter of course have been imposing at least one forme while another was in
press, and would have taken up whichever skeleton-forme was to hand, as soon
as another had been laid on the press. Given that skeleton-forme I was used twice
in this gathering, it must therefore have been alternated, on a single press (i.e. not
necessarily on one actual press, but at single press power), with skeleton-formes II
and III. Because the type from this skeleton-forme—regularly used at this point
to print the outer forme of the inner sheet—had been distributed while the inner
formes of Bb were still being printed on two presses (outer formes, as Yamashita
et al. determined, were regularly printed before inner formes throughout the
volume[21] ), it was available, as usual, for the outer forme of the inner sheet of Cc.
At this point, too, skeleton-forme II, as usual, was available to set the type for
the outer forme of the outer sheet. When it became obvious that printing would
have to slow, the press-men moved to a one-press system, and skeleton-forme II
was printed first, followed by skeleton-forme I, now in sequence rather than, as
was usual, in tandem; when I was laid on the press, the type from II was redis-
tributed, but there had been plenty of time, already, to get on with setting the
inner forme of the outer sheet, using skeleton-forme III. After that had gone to
the press, the type from skeleton-forme I was redistributed. At this point, because
the compositors were working with a one-press system, there had been no need to
rush the setting of the inner forme of the inner sheet, and so skeleton-forme IV
had not, as was usual, yet been set out; because skeleton-forme I was now avail-
able, and contained the correct canto markings in its running-titles, it was re-
used. This process of imposing gathering Cc and the adjacent gatherings can be
summarized as follows:

         

	Press 1	Press 2
C–Bb	II→III	I→IV
Cc	II→I
	III→I
Ff–Nn	II→III	IV→I

The lack of urgency apparent in this slow process—the compositors would
have been comparatively idle, with respect to the Spenser job at least,
during the printing of the outer forme of the inner sheet—suggests that the decision had not
yet been taken to skip gatherings Dd and Ee, and return to a two-press system
with gathering Ff. The psychology of the slow-down in the printing of Cc, when
coupled to the subsequent events in Dd-Ee and Ff-Nn, seems difficult to ignore:
Wolfe and his press-men probably slowed down in Cc because they knew, at the
end of the gathering, they were going to hit a problem. That problem was a lack
of text. While it is of course possible that the reduction in power in gathering
Cc was unrelated to the decision to skip gatherings Dd and Ee—the demands

of some concurrent project may have required an extra press at the same time
that Wolfe's compositors hit a snag with copy in the following gatherings—on
balance, the coincidence of these unusual events suggests that they were related.
In the light of these telltale marks of non-sequential printing, it does not seem
surprising that the only reversal in the orientation of a skeleton-forme during
the whole press run—the reversal of IV between its use for the inner forme of
the inner sheet of Bb and its use for the outer forme of the inner sheet of Ff (see
table 1)—coincided with the slowing of the printing in Cc and the decision to
skip Dd and Ee. Of course, no signature appeared in the usual place (position 4r)
in skeleton-forme IV in gathering Bb, as this page between Books II and III of
the poem was left blank; but where a compositor might normally have judged
the orientation by eye without the signature, the gap occasioned by the double
use of I in Cc meant that the type from IV was probably distributed some time
before, rather than when, it was reimposed with new pages.

Oddly enough, this was not the first time in the course of printing The Faerie
Queene in 1590 that these beleaguered press-men had had to slow down, and per-
haps even come to a stop, nor would it be the last. As noted above, the dedication
to Elizabeth appearing on the verso of the title page in gathering A was almost
certainly added as a stop-press correction, as several copies survived without the
dedication. The evidence on compositorial stints gathered by Yamashita et al.
supports this hypothesis. As can be verified by comparing the data for gather-
ings A and B in table 1, Compositor X, who would have been expected to set at
least 1r–3r for gatherings A and B, did not have a clear hand in either 1r or 1v of
gathering B, but was helped in both cases. In every other gathering in the volume
(36 in all), Compositor X was solely responsible for setting these first pages. To
understand why X did not set these pages, and to appreciate its relevance to the
stop-press correction of the royal dedication, we must return briefly to another of
the anomalies in the printing of the volume that Johnson did not explain: unlike
the main sequence of regular imposition using the four-chase/two-press system,
gatherings A and B were printed with one skeleton-forme per sheet—i.e. using
the same skeleton-forme for both formes of the inner and outer sheets, for each
of the two sheets in each gathering: II for the outer sheet of A, I for the inner
sheet of A, IV for the outer sheet of B, and III for the inner sheet of B. As noted
above, in relation to the similar problem in Cc, the process of setting the type,
printing off the formes, and distributing the type made it convenient to alternate
skeleton-formes: the same skeleton-forme could not be used efficiently in back to
back press-stints. It is thus very likely that the first two gatherings were printed
in tandem:

         

Press 1	Press 2
II (± o A)	I (β o A)
IV (α o B)	III (β o B)
II (α i A)	I (β i A)
IV (α i B)	III (β i B)

This pattern would have left skeleton formes II and I available, during the impo-
sition of the inner formes of both sheets of B, to be re-set for gathering C, thereby

setting in motion the regular pattern of one-skeleton-per-forme imposition used
thereafter until gathering T. The absence of Compositor X from the early stages
of setting the type for both of his usual stints in gathering B, then, coincides with
the printing of the inner, and later the outer, formes of gathering A. In both of
these gatherings, substantial stop-press corrections had to be inserted: in the
outer forme, the narrow-spaced date of the title-page on 1r had to be re-spaced
to a centered, wide-state layout;[22] and in the inner forme, the royal dedication
had to be added on 1v. The second correction, in particular, would have taken
a reasonable amount of time; the fact that Compositor X was back on hand to
set the type for 2r of B suggests that (as the low incidence of the blank-state vari-
ant in extant copies suggests) this correction was made relatively soon into the
perfection of A.

The more obvious question to be asked about this scheme for printing the first
two gatherings in tandem is, of course, why it was followed at all. The sensible
pattern for using four skeleton-formes on two presses to set a quarto in 8s is, as we
have seen, the one followed for the majority of the press run—the advantage of
using each skeleton-forme once in each gathering is that it allows the compositors
to work on contiguous pages, rather than skipping from gathering to gathering,
and thus reduces the likelihood of formatting errors (as we have seen, the she-
nanigans of Dd and Ee would later coincide with many such errors). Thinking
narratively about the decisions taken at this stage of the printing, it seems likely
that, to begin with, the compositors intended to follow the same system for the
first two gatherings that they would take up, as soon as possible, in gathering C;
but that some obstacle prevented them from doing so when, having laid skeleton-
formes II and I on their respective presses, they turned to skeleton-formes IV and
III. The likely reason the compositors might have elected not to set the type for
the inner formes of both sheets of gathering A was, of course, the now-familiar
one: a final decision on the text for these formes had not yet been taken. Given
that the royal dedication was later inserted as a stop-press correction to 1v in
gathering A, it seems likely that the obstacle holding up the perfection of gather-
ing A a day or two earlier was this same problem. The dedication to the Queen,
completely in majuscules, would surely have been an element too conspicuous for
Wolfe merely to have overlooked, and subsequently corrected, especially when
he had not, yet, printed any other dedicatory or prefatory material for the vol-
ume; rather, the coincidence of an anomalous use of skeleton-formes, on the one
hand, and on the other a stop-press correction, suggests that Spenser, or perhaps
Wolfe, had not yet decided how, or whether, to incorporate the royal dedication.
Furthermore, the fact that the printing commenced at all, and commenced with
A, suggests that Wolfe was confident of the layout, or had no reason to be uncon-
fident, when the printing began; the indecision seems to have arisen during the
imposition of the outer formes of the outer and inner sheets of A, when skeleton-
formes IV and III were deployed to set gathering B rather than the inner formes

of both sheets of A. The psychology implicit in this pattern of imposition, again,
suggests not only that manuscript materials, or authorized manuscript materials,
were not available, but that the status of these materials changed suddenly at a
particular, and definable, point of the printing.

We are now in a position to make some limited inferences about the two
interruptions to the sequence of printing so far discussed. In both cases, the
compositors seem to have slowed down—in the case of the first two gatherings,
Compositors X and Z worked without Y on the text of A, while in Cc the printing
seems to have temporarily stalled and to have proceeded at single press. In the
first case, the compositors seem to have encountered an unexpected snag after
they had committed themselves to printing the gathering (beginning with the
outer formes of A), while in the second case, they seem to have slowed produc-
tion of Cc in anticipation of an upcoming difficulty in the following gathering.
With this in mind, it is almost certainly significant that both of these instances of
interruption in the printing correspond to elements in the text particularly closely
associated with Elizabeth herself: the obstacle in gathering A seems certainly to
have been the royal dedication, while the cantos at the center of the ultimately
deferred gatherings, Dd and Ee—Book III, cantos ii and iii—deal with Brit-
omart's vision of Artegall in the magic mirror, and her subsequent visit to Merlin
who, forecasting Britomart's famous progeny, concludes his prophecy with the
'royall virgin', Elizabeth. The pattern and implicit psychology of the slowing, the
ultimate ordering, and the selective skipping of the printing in both of these cases
suggest that Spenser was intimately involved in at least this part of the produc-
tion process (that is to say, intimately involved insofar as he was meddling, and
holding the whole process up!), and that he was particularly anxious about parts
of his text that spoke directly to, or explicitly represented, Elizabeth herself. If the
reason for skipping gatherings Dd and Ee was that Spenser had withheld canto iii
of Book III—and what else could have caused Wolfe to complicate his sequence
in such a way, a way that would lead to the introduction of several errors?—then
it seems likely that Spenser modified the text of this particular canto during the
press run itself. He could not have meddled much, as the canto still needed to fit
precisely into the available 10 leaves—though it is worth noting that, where the
canto ends on Ee8r, the compositors left an unusually large gap at the end of the
page, which they filled with a type ornament rather than commencing (as they
had done for Book III, canto iii itself) with an ornamental box and the first four
lines of the opening stanza of canto iv. Wolfe's hands were already tied at this
point by the extant, fixed state of Ff1r—canto iv had to begin on Ee8v—but the
unusually generous layout of Ee8r suggests that the original casting-off may have
provided for one more stanza in canto iii than Spenser ultimately supplied.

Spurning at a Dead Dog: Patronage, Politics, and Allegory

Of Johnson's original six anomalies in the otherwise consistent layout of the
skeleton-formes and running-titles in the printing of The Faerie Queene in 1590,
then, several have already begun to disclose information about the sequence of
printing the gatherings of the poem. The doubled pattern in the deployment of
skeleton-formes in gathering Cc, interpreted alongside fresh evidence of related

inconsistencies in the appearance of running-titles in contiguous gatherings, sug-
gests that this kind of anomaly provides evidence of a slowing, or perhaps even
a halt, to the printing at such points in the text. A similar pattern of doubling
in the skeleton-formes in gatherings A and B seems likely to be associated with
the stop-press corrections made to the recto and verso of A1, witness to another
point in the printing where the compositors were probably suddenly faced with a
problem in their immediate copy. But there is one further point in the sequence
of gatherings, in gathering T, where the pattern of usage of the skeleton-formes
shows exactly the same kind of doubling, a pattern that could only have occurred
if the compositors were working very slowly, and at single press. The evidence
of interruption in the printing of gathering T takes us back, again, to Nashe's
satirical comments on the printing of The Faerie Queene in 1590, and once again
implicate the tangled threads of Spenser's allegory, his dedications, and the pa-
tronage politics of his carefully-tendered epic.

Toward the close of the 'priuate Epistle of the Author to the Printer'
that Nashe added to the second edition of Pierce Penilesse, he vomits up a few
characteristically dark sayings about his satirical intent in the ensuing work, com-
ments that serve above all to link Pierce Penilesse to The Faerie Queene, and both
texts to the recent Martinist controversy, to the legacy of the Earl of Leicester,
and to the now-fallen Ralegh:

As Charles Nicholl has demonstrated in his careful reading of the 'Tale of a
Battle-Dore' in Pierce Penilesse, much of Nashe's satire is directed, in an ad hominem
attack, at the memory of the Earl of Leicester (d. 1588).[24] Leicester's legacy as a
patron of 'Swiss' bishops and precisian preachers was politically charged in 1592
for the same reasons that it had been charged in 1591, when Spenser published
Complaints, including Virgil's Gnat, 'long since dedicated to the most noble and ex-
cellent Lord, the Earle of Leicester, late deceased'. Leicester's public support for
precisian causes, such as his patronage of the radical 'puritan' preacher Thomas
Cartwright, or his support for bishops and archbishops like Edmund Grindal,
was compromised by the salacious, and likely true, stories of his rapacious sexual
appetite and the occasionally murderous means he compassed to satisfy it. The
cupiditas that supposedly characterized his sexual and 'personal' life was taken by
his detractors as evidence of the self-interested ambitions and designs that Leices-
ter had for Elizabeth herself, for the crown, and for the realm; Leicester's sexual
licentiousness was read by his contemporaries as a sign for his own ambitions for
power, and for his final disregard for common weal in the face of his own self-
satisfaction.[25] It was thus essential for allies and survivors of Leicester's factional
support of precisian causes—Walsingham, the Sidney family, Ralegh, Robert
Beale, Sir Francis Knollys, and others, probably including Spenser himselfS—to
defend Leicester's reputation, and to vitiate the supposedly automatic link be-
tween cupidity and self-interested political ambition.

This project became the more urgent for the Sidney faction in the early 1590s
as a result of the death of Sir Francis Walsingham (d. 1590), and the rise of Sir
Walter Ralegh. Ralegh's character was open to the same ad hominem attacks as
Leicester's had been: the same stories of sexual depravity, of enforced maids, of
illegitimate children, and of an almost psycho-sexual enchantment of Elizabeth
herself circulated in anti-Protestant and anti-Ralegh circles. Ralegh's association
with the Throckmorton family, through his clandestine marriage to Bess Throck-
morton, upon its public exposure in the summer of 1592 suddenly realized his
threat as a new Leicestrian patron of puritans: Bess's uncle Job Throckmorton
was probably the author of most, if not all, of the Martin Marprelate tracts of the
late 1580s, and the family had strong ties to the Cartwright-Penry-Udall circle
responsible for a dramatic upsurge in precisian preaching and nonconformity
in the early 1590s.[26] To attack Leicester publicly between 1590 and 1592 was to
attack Ralegh covertly; to defend Leicester publicly, as Spenser did, was in the
same period to defend Ralegh.

When Nashe promises to be an 'euil Angel' to anyone who disquiets him
without cause, in his preface to the second edition of Pierce Penilesse in the au-
tumn of 1592, he is responding directly to a very specific and charged bit of the
historical allegory of The Faerie Queene, published two years previously—one as-
sociated directly with the legacy of the dead Leicester, and the new incarnation
(at that time) of his political and religious agenda in Walter Ralegh. The begin-
ning of Book II, canto viii of The Faerie Queene finds Guyon, the patron knight
of temperance, lying in a swoon outside the delve of Mammon: oppressed with
the 'exceeding might' of the temptations offered him 'below the earth', Guyon's

'life did flit away out of her nest', leaving him apparently dead and defenceless
to the predations and infamies of his enemies. Pyrochles and Cymochles, the two
sons of Acrates whom Guyon had defeated in cantos v and vi, discover his body,
and set about despoiling it, determined to take their revenge upon his honor, if
not upon his life. Guyon's only protection comes from his guide the Palmer, to
whom his body is committed in trust, at the opening of the canto, by the 'faire
young man', an angel sent by God to preserve Guyon's life and person. As the
angel tells the Palmer:
Where Spenser promises Guyon an angel to defend and succour him 'euermore',
Nashe promises his own enemies an 'euil Angel, to haunt them world without
end'. Guyon's, and the angel's, 'foe' is any man who will dishonor Guyon's dead
(in fact merely sleeping) body, an abstract quickly realized, upon the angel's
departure, when Archimago, Pyrochles, and Cymochles appear full of 'coles of
contention and whot vengeance' (FQ II.viii.11).

It is with the arrival of this trio of despoilers that Nashe's other two references
to this passage—to 'mettal' and to the 'dead dogge'—become clear. Pyrochles
and Cymochles quickly set about 'disarraying' Guyon's body of his shield and
helm, deaf to the pleas of the effectless Palmer, but are immediately interrupted
by the arrival of Prince Arthur, whom Archimago recognizes as their natural
enemy. Archimago has by this point purloined Arthur's sword, intending (as he
indicates in stanza 18 of canto iii) to bestow it on the braggart Braggadocchio;
Pyrochles, addressing himself swordlessly to confront Arthur, demands Mord-
dure of Archimago, but is refused by the enchanter because the sword's power is
'contrarie to the worke, which ye intend':

Pyrochles, defying Archimago, seizes the sword and tries to fight Arthur with it;
sure enough, at the two critical points in the ensuing battle when Pyrochles deliv-
ers the death-blow to the Prince, the 'faithfull steele such treason no'uld endure',
and Arthur survives to triumph over both brothers. Nashe's joke about being 'of
another mettal' locates his comments on satire, alongside the 'euil Angel' allu-
sion, as a direct reference to this passage from The Faerie Queene: when his metal
bites, Nashe indicates, it will not swerve.

But it is the third part of the allusion—'who spurneth not at a dead dogge?'—
that takes us furthest and most fruitfully into Nashe's reading of this episode, and
discloses the meaning of both Spenser's historical allegory and Nashe's criticism
of it. On their first arrival on the scene, the two sons of Acrates engage the
Palmer in the same argument—over Guyon's honor—in which they later con-
test Prince Arthur. Pyrochles, disdaining to be impeded by a weak and sanctimo-
nious palmer, will not stay like his brother to argue, but pushes straight in:

It is clear, given the intertextual conversation in which we can now situate this
episode from The Faerie Queene, that Guyon's swoon, during which his detractors
seem to despoil and dishonor his apparently dead corpse, functions as a historical
allegory for Leicester's patronage of the precisian cause. Leicester's intemper-
ance, like Guyon's weakness in the delve of Mammon and later in the Bowre of
Blis, remained the Achilles heel of the Sidney-Walsingham faction at court, and
their clients in Parliament. In the period after his death, Leicester's old causes
were exposed to the mercy of his detractors, defended only by the ineffectual
intervention of less powerful men, Palmer figures like Walsingham, Sir Francis
Knollys, and Walsingham's brother-in-law Robert Beale. Nashe's brief comments
on satire in his preface to the second edition of Pierce Penilesse rely on an intimate
knowledge of canto viii of Book II of The Faerie Queene: the triple allusion to the
guardian/evil angel, the 'metall' of Nashe's wit, and the 'dead dog'—for Nashe,
as for Spenser, a reference to Leicester—position Pierce Penilesse as a satirical
response to this episode from Guyon's quest.

But why should Nashe take such careful pains to focus the reader's attention
so precisely on a few stanzas from Spenser? Why should both authors associate
Leicester with a 'dead dog'? And what has this to do with the interruption of the
printing of The Faerie Queene in gathering T? The opening volley of Nashe's at-
tack on Leicester in 1592, like Spenser's defence of Leicester in 1590, relies on a
history of further intertextual associations, both Biblical and historical, that join
Leicester's sexual intemperance to his political ambition. Anti-Leicestrian at-
tacks proliferated in manuscript and occasionally print works throughout the late
1570s and 1580s; the most notorious example is the well-known 1584 The copy of
a letter written by a Master of Art of Cambridge (or Leicester's Commonwealth, as it came

to be known), which the Privy Council sought to suppress, and which spurred
Sir Philip Sidney to the composition of a curious written defence of his maternal
uncle. Leicester stood accused in Leicester's Commonwealth of having murdered
his first wife, to make himself available to Elizabeth; of having poisoned the first
Earl of Essex, so that he might marry Lettice Knollys, his wife; and of cajoling or
enforcing many of Elizabeth's younger maids and ladies. The charge that he had
poisoned Walter Devereux, Earl of Essex, in Dublin in 1577, made for particu-
larly cunning libels from the intelligentsia, as Leicester could be associated in ap-
parent innocence by a range of intertextual allusions with two other figures, one
Biblical and one historical, who had cleared their way to a woman by making off
with an inconvenient husband. King David, as recorded in 2 Samuel chapter 13,
directed his general, Joab, to put Uriah the Hittite in the vanguard of an as-
sault on Rabbah, so that, Uriah having been dispatched, David might take to
him Uriah's wife Bathsheba, already pregnant with the king's child. The Roman
emperor Heliogabalus, similarly, alienated his Praetorian guard by indulging in
extravagant spectacles of underage sexual depravity with men, women, animals,
and the wife of a man who he had specially murdered for the occasion. As Cas-
sius Dio, Herodian, and Aelius Lampridius report, when the Praetorian Guard
finally assassinated Heliogabalus, at age eighteen, they desecrated his body by
dragging it about the streets of Rome before chopping it up and dumping it in a
privy, after which it was flushed through a sewer to the Tiber.[27]

These two stories matter to Spenser because they became part of the stan-
dard attack of Catholic libels on Leicester. The author of Leicester's Commonwealth,
for example, has one of the interlocutors of his dialogue, the scholar, comment
on Leicester's supposed escapades:

The author of this influential manuscript and print tract against Leicester pro-
voked Sir Philip Sidney to write a defence of his uncle, which, crucially, picked
up on the charge of Leicester's beastliness, and focused on the description of
him as a 'dog' and a 'bearwhelp'.[29] The association of Leicester with bears is a
direct reference to the Dudley family badge—a bear and ragged staff—but it
is also conveniendy linked to a well-known passage from 2 Samuel, chapter 17,
where the counsellor Hushai dissimulatingly advises David's son Absolom not to
attack his father in his flight, 'for … thou knowest thy father and his men, that
they be mighty men, and they be chafed in their minds, as a bear robbed of her
whelps in the field: and thy father is a man of war, and will not lodge with the
people.' 2 Samuel is also the source, in two locations, for Spenser's two uses in
Book II of The Faerie Queene of the derisive insult, 'dead dog'; the first Spenserian
use comes in Trompart's submission to Braggadocchio in canto iii of Book II,
which echoes the submission of Mephibosheth, the son of Jonathan, to David
in 2 Samuel chapter 9, verse 8: 'What is thy seruant, that thou shouldest looke
vpon such a dead dogge as I am?'[30] The second Spenserian use comes, as we (like
Nashe) have seen, in Pyrochles' impatient dismissal of the Palmer in canto viii
of Book II, when the two sons of Acrates begin to strip Guyon of his arms. The
emphasis here on the honor due to the fallen, in association with the use of 'dead
dog', echoes the other place in 2 Samuel where the insult occurs, chapter 16,
verse 9. Abishai offers to punish Shimei (a son of the house of Saul) for abusing
David and throwing stones at him in his ignominious flight from Jerusalem, say-
ing, 'Why doeth this dead dog curse my lorde the King? let me goe, I pray thee,
and take away his head.'[31]

Where Pyrochles' dismissal of the Palmer invokes the example of David dis-
honored by Shimei, Cymochles gets his turn at the same argument when he
upbraids Arthur for intervening, this time invoking another source:

In repeating the argument, and making the association between the dishonor-
ing of Guyon's body, on the one hand, and 'outcast dong' on the other, Cy-
mochles moves the association from David to Heliogabalus—whose body, as
we have seen, was hacked into pieces, thrown into a privy, and finally flushed
through the sewers of the Tiber. When Arthur finally receives from the Palmer
the sword of the fallen Guyon, and addresses himself to the fight, Spenser ties off
the set of intertextual associations in an epic simile, recalling at once the 'bear
whelp' of Leicester's Commonwealth and the whelpless bears and lions of 2 Samuel
chapter 17:
As God delivered David from the revolt of Absalom, despite his sins and
'bloody deeds', so God, Arthur reasons, must be the one to judge Guyon's appar-
endy posthumous honor. Unlike Heliogabalus, whose body had been disgraced
and dismembered, and then made as 'the outcast dong' in a Tiberian latrine,
Guyon's honor must be preserved. The intertextual associations make it obvious
that, for at least the brief period of this episode, the fallen hero whom Arthur
is defending is (at the level of the historical allegory) none other than Robert
Dudley, Earl of Leicester, whose honor was still a sensitive political issue for the
precisian preachers and divines (to say nothing of the poets) who had depended
on his patronage. Crucially, too, Guyon will shordy wake from his apparent
death, to carry on renewed with his quest; similarly, Ralegh (as it was thought in
1590) would take up the precisian cause. Where Pyrochles offers the example of
David, and Cymochles the precedent of Heliogabalus, Arthur rebuffs them both,
and preserves Guyon's honor and agenda—to fight another day.

That Nashe should have fixed on this episode of The Faerie Queene, with its
strong links to Leicester's Commonwealth, to Sidney's Defence, and to 2 Samuel and
the various Roman histories of the reign of Heliogabalus, suggests that he as-
sumed his readers would be able to reconstruct the intertextual associations on
which his satire depended. This is not the far-fetched assumption it might at first
seem: Leicester's Commonwealth circulated vigorously in the 1590s,[32] when it seems
to have been as popular as in its early heyday of the mid 1580s; it was still con-
sidered relevant enough in the early years of the seventeenth century to warrant
Thomas Rogers's verse redaction in Leicester's Ghost (ca. 1601–1604). Similarly,
the 'dead dog' passages from 2 Samuel continued to have important resonances,
for Spenser at least, throughout the 1590s—and though it is a complicated story,
it is worth pausing for a moment, with the patience of an archaeologist's brush,
to excavate it. Frank B. Evans demonstrated as early as 1965 (in this journal)
that Spenser introduced a substantial revision to Book V during the 1596 print-
ing of The Faerie Queene, rearranging the last two cantos to accommodate new
material allegorizing in the history of Sir Burbon the defection of Henri IV (i.e.
of Navarre) from the Protestant faith.[33] As Evans showed, Artegall's meeting

with Sir Sergis in canto xi of Book V—affecting stanzas 37—43—was added
during printing as a transitional sequence that would enable Spenser to move
the matter of Sir Burbon from canto xii back to canto xi. After this change was
made (in some haste, it seems: note the clumsy transitional stanza 36 in canto xi
of Book V), Spenser obviously forgot to update the argument to canto xii, which
continues to suggest that the final canto includes both the Burbon material and
the defeat of Grantorto.[34] As A. C. Hamilton has observed in his annotations to
this passage, it is obvious that Spenser was interested in consolidating the mate-
rial of canto xi—including the rescue of Belgae, or the Netherlands, by Arthur's
hand—in order to provide a comprehensive account of Elizabeth's defense of
Protestant causes on the continent. What Evans did not observe and Hamilton
does not note, but which continues to make sense of the printing of the first part
of the poem six years earlier, was that this material was almost certainly also
moved back from canto xii so that Spenser might make space to accommodate
an addition to the final canto—a set of stanzas not mentioned in the (unchanged
original) argument to the canto, and one that appears after another obviously ter-
minal, unaltered stanza. What is now stanza 27 of canto xii was almost certainly
originally intended as the final stanza of Book V: Artegall 'through occasion' is
called away from the reformation of Irena's government before completing his
task, and the final alexandrine sounds the perfect, tight-lipped thud that Spenser
would have wanted for Artegall's (failed) conclusion: 'So hauing freed Irena from
distresse, / He tooke his leaue of her, there left in heauinesse.' Suddenly, though,
a new section was available—the episode describing Artegall's return from the
salvage island and defaming by the hags Envy and Detraction—beginning with
another clumsy segue: 'Tho as he backe returned from that land …'. Fun-
nily enough, this passage recalls again, six years later, the same episode from
2 Samuel, chapter 16 that had guided revision in 1590: Artegall's restraint of
Talus in the final stanza 43 echoes David's restraint of Abishai, when he sought
permission to punish Shimei of the house of Saul for throwing stones at David
and reviling him:
This time, no Prince Arthur is on hand to protect the honor of the reviled mag-
istrate, as he passes back to the court of his sovereign, recalled, disgraced, and
beaten. Much had changed since 1590 in the Protestant patronage circles in
which Spenser had been wont to move: Walsingham had died, Grey had died,
Ralegh had fallen from favor, Essex had been obliged publicly to disavow his
precisian clients, Robert Beale (after his disastrous attack on the High Com-
mission in the 1593 Parliament) had been exiled from court, and Sir Francis
Knollys, the last surviving champion of Puritan causes on the Privy Council, had
been silenced and banned from the Queen's presence, and had, in 1596, finally
died.[35] Those who had faithfully served Elizabeth's Protestant foreign and do-
mestic policy had been ruined and impoverished, or were dead, and were being
defamed; where Arthur had reassured the Palmer in 1590 that he could rely on
him for his 'knights last patronage' (FQ II.viii.26), no such patron was available
in 1596. Spenser was determined to insert a complaint, and that complaint would
refer the reader back, through its intertextual links with 2 Samuel, to the other
magistrate so nearly left succourless in Book II, Guyon as the Earl of Leicester.

Nashe's interest in canto viii of Book II in 1592 was surely motivated by his
overall plan to set up Pierce Penilesse as an attack on Leicester and Ralegh based in
a satirical account of the patronage fiasco of the 1590 edition of The Faerie Queene.
Nashe may or may not have known that the printing of gathering T, which con-
tains canto viii of Book II, was marred by some anomaly that probably indicates
a slowing or even breakdown in the printing process; but he almost certainly
recognized that Spenser had meticulously revised this part of the poem so as to
include a pointed historical allegory defending the reputation of Leicester and
allying that defence not only with Prince Arthur, but with one of the only direct
Christian interventions, and certainly the most extended, in the entire gothic
world of The Faerie Queene. The pattern of skeleton-forme usage in gathering T,
as in A and B, and later in Cc, shows a doubling (here, of skeleton-forme III on
both formes of the outer sheet; II was not used) that indicates that printing must
have slowed, or perhaps even stopped. The psychology of this slowing is however
to be distinguished from the likely unfolding of constraints and solutions in the
earlier and later cases: where the doubling in gatherings A and B involved sheets
from two gatherings that could be alternated on the two presses, keeping up the
pace, and where the later doubling in gathering Cc was followed by a reverse of
skeleton-formes I and IV in the imposition of the formes of the inner sheet from
Dd (or, as we have seen, Ff) onward, in the case of gathering T it was skeleton-
forme III, and not II, that was used to impose both formes of the outer sheet, with
no corresponding shift in the use of skeleton-formes I or IV. As III would have
been in use to impose the inner forme of the outer sheet of S during the setting
of type for the first formes of T, it seems very likely that the delay evidenced in
the doubling of skeleton-formes in gathering T was specifically associated with
the outer sheet of the gathering—otherwise the compositors would have been
able to use skeleton-forme II to impose the outer forme of the outer sheet during
the printing of S, rather than waiting to impose the outer sheet until the printing
of the outer forme of the inner sheet of T (using skeleton-forme I). The process

of imposing the formes in gathering T, and in the adjacent gatherings, can be
summarized as follows:

             

	Press 1	Press 2
C–S	II→III	I→IV
T	I (β o)
	III (α i)
	IV (β i)
	III (α o)
V–Bb	II→III	I→IV

A hitherto unnoticed mistake in the pagination in gathering V indicates that
the inner forme of the outer sheet of gathering T was printed before the outer
forme of the outer sheet. Position 8r in gathering V reads '217' for '317', a mis-
take that would have been impossible if the conservative compositors had reset
running-title 15 from an immediately preceding use at position 8r in gathering T
(where the pagination reads '301'); the '2' of '217' is rather clearly retained from
an immediately preceding use at position 7r (page 299) of gathering T, where
running-title 15 was used in the printing of the outer forme of the outer sheet.
This particular '2', it should be noted, is different from the '2' that appears in
the pagination for running-title 15 in position 8r of gathering S; had the outer
forme of the outer sheet of T been printed before the inner forme, we would have
expected to see this '2' retained, but instead the '3' of'301' (at position 8r of T)
intervened. That the compositors did not then merely switch skeleton-formes II
and III after printing the inner forme of the outer sheet of T, and continue print-
ing at pace from that point, suggests that there was still residual lag, or some
further delay, with the outer forme of the outer sheet, which was, again, imposed
during the printing of the inner forme of the inner sheet (using IV), on skeleton-
forme III; it is also possible that, having reduced production of the poem to a
single press during this period, Wolfe had set up another short job on the idle
press, which forced the compositors to work in straitened circumstances on the
Spenser job until gathering V. It should be noted that, if the doubling effect in T
had been merely the result of Wolfe's decision to co-opt one of the presses tem-
porarily for another (concurrent) job, we should expect to find skeleton forme II
in use for the outer forme of the outer sheet of T, and not III.

When we map the bibliographical evidence of the sequence of printing in
gathering T against what we have seen of the topical intertextual allusions of
Book II, canto viii, the picture seems reassuringly clear: nearly all of the intertex-
tual allusions to 2 Samuel, to the history of Heliogabalus, and to the anti-Leicester
polemics like Leicester's Commonwealth occur squarely within the bounds of the inner
sheet of T; the outer sheet contains, in its share of the latter half of the gathering,
the fight between Arthur and the sons of Acrates, up to and including the epic
simile of the 'robbed whelps' at stanza 40. Given that skeleton-forme II re-appears
as usual in gathering V, and that its use throughout the rest of the printing is en-
tirely consistent, it seems likely that the decision to use skeleton-forme III for both
formes of the outer sheet of T was possible, and was ultimately taken, because the
copy from which to set the gathering was not available when II became free, in
the normal way, during the imposition of gathering S. The pattern of skeleton-
forme use in gathering T, in short, indicates that the printing actually stopped be-

fore T was printed. When it restarted, it proceeded for the space of one gathering
on a single press; and the compositors for some reason completed setting the inner
sheet first, followed (irregularly!) by the inner forme of the outer sheet—perhaps
because the Argument to canto viii, which appears in position 2v, in the outer
forme of the outer sheet, was not yet complete. In any case, the confusion attend-
ing the printing of T, and the salience of 'the canto viii problem', seem to have led
the compositors to make one of the more prominent mistakes in the running-titles
observed by Johnson: positions 1r, 1v, and 2r of gathering T (running-titles 6 and
14, twice) all mistakenly read 'Canto VIII' for stanzas actually still part of canto
vii. This is a mistake consistent with the fact that the outer forme of the inner sheet
(fully within canto viii) was unusually the only thing printed, using skeleton-forme I,
before the inner forme of the outer sheet was imposed; compositors used to mak-
ing only partial changes to the canto headings across transitional formes could
be forgiven, in the confusion of imposing gathering T, for forgetting that, having
imposed a forme fully within canto viii, they needed to return to a forme only
partly contained in that canto. It should be noted that it was during the printing
of this gathering that skeleton-forme I was corrected by the addition of a period
to 'The Faerie Queene' in running-title 10 (11), which may have something to do
with the delays in this part of the printing, and with the recognition that these
confusions had already led to careless mistakes.

If anxiety about matters of 'patronage' led to the eleventh-hour revisions of
Book II, canto viii, that later so exercised Thomas Nashe, it would not be the
last time in the course of printing the poem that such anxieties would interrupt
Wolfe's better business sense. Although the patterns of skeleton-forme usage seem
to have no further secrets to divulge, a very odd error in pagination—duly ob-
served by both Johnson and Yamashita et al., though not interpreted—suggests
at least one further, perhaps final hiccup in the printing after the headache of
gathering Cc. At position 8v of gathering Ii, running-title 4 inexplicably reads
'600' where it ought to read '510'. This is not the kind of mistake that a composi-
tor would make if Ii were being printed sequentially after Hh (where the page
number at position 8v, the last time running-title 4 had supposedly been used,
reads '494')—we might expect an error of '4' for '5' in the hundreds place, per-
haps, but certainly not '600'. The mistake makes sense, though, when one places
it next to position 5v in gathering Pp, the 'final' gathering of 'prefatory' materials
that Wolfe supposedly printed last; here the page number reads, in exacdy the
same setting of type, '600'. As it happens, position 5v in Pp corresponds exactly
to position 8v in Ii, if we assume that skeleton-forme II was used, between the
imposition of gatherings Hh and Ii, to print Pp; in fact, no other explanation can
plausibly account for this irregular mistake in pagination.

Although it is unclear why Wolfe decided to interrupt the 'regular' sequence
of printing after Hh in order to print Pp, it seems likely that the impetus came not
from a need to produce the commendatory and dedicatory materials, but rather
from a need to stall before the printing of Ii. Wolfe had already by this point left
Dd and Ee aside, and would not return to them until after Nn, and he may thus
have felt it unwise to skip Ii, as well. This gathering contains nearly all of Book III,
canto vii—fewer than three stanzas at the end of the canto spill over onto the
first page of gathering Kk. Wolfe must have had at least part of canto vii in hand,

because the canto begins on 8v of Hh, but it may be that the end of the canto was
missing, or that Spenser had asked to make revisions. The narrative of Florimell's
escape from the witch and her son, and of Satyrane's rescue of the Squire of
Dames from the giantess Argante, would not seem at first glance particularly
sensitive material that needed a timely reworking. Another possibility, of course,
is that the decision to delay Ii was somehow associated with the earlier omission
of Dd and Ee; Wolfe may have seized the opportunity to stall the printing after
Hh in the hope that Spenser would complete and deliver the revisions to Dd and
Ee at that point. Something, in any case, interrupted progress.

What is far more problematic, and interesting, about this interruption is
that this error in pagination supplies certain proof that all of the commendatory,
dedicatory, and epistolary material that would go on to appear at the back of
the poem was printed well before the end of the print run. The printing of Pp
midway through Book III indicates that the disposal of the 'prefatory' material
in a final gathering was certainly not an afterthought, but rather a deliberate
choice, taken at a time when Spenser was obviously regularly in contact with
Wolfe, or even physically present in the printing house. Given that the printing
of the cancellans, gathering Qq, re-used the type from the relevant settings in Pp,[36]
it seems tempting to assume that the printing of Qq followed directly after the
printing of Pp: surely Wolfe's compositors would have needed to redistribute the
type from the dedicatory sonnets before moving ahead with the printing of Ii.
But in a printing house as well-stocked as Wolfe's, it may not have been neces-
sary to do so, especially if the compositors came to anticipate a later revision to
the dedicatory sonnets (perhaps because an irate or frustrated poet, hearing of
the disordered imposition of his dedicatory materials, had insisted that further
sonnets had been intended, and would soon arrive);[37] and one piece of evidence
suggests that, in fact, the printing of Qq did not take place until after the comple-
tion of the rest of the volume. Had Spenser supplied Wolfe with the additions to
the dedicatory sonnets immediately upon the completion of the printing of Pp,
it seems likely that Wolfe's compositors would have reprinted the sonnets using
the same skeleton-formes; because Qq has no pagination—the eight pages of
the cancellans needed to replace a four-page cancellandum, and would therefore
inevitably have thrown off the existent pagination—the compositors would have
removed the '600' from its place in running-title 4 before printing the outer
forme of the outer sheet of Ii. Had this number been removed, it would certainly
not have been restored. The likeliest construction of events, then, seems to be
that Wolfe ordered the dedicatory material printed before the imposition of Ii,
perhaps because of a potential delay with that gathering, or because he was still
trying to buy time for the completion of the material needed for Dd and Ee; but,
finding that his peremptory production of the dedicatory materials was based on
incomplete copy (regardless of whether Spenser himself had finished the sonnets,

or not), that Wolfe determined at the end of the job to print a cancellans in order to
recover the omitted dedicatory materials. The retention of standing type from Pp
suggests that Wolfe was brought to recognize his mistake during, or immediately
upon finishing, the printing of Pp.[38]

Whatever the fine details of the timing of the cancellandum and the cancellans,
the pagination mistake evident in running-title 4 (at position 8v in gathering Ii)
seems to leave it indisputable that Spenser had a firm hand in the printing of
the dedicatory materials, and that he was involved in the decision to place them
at the back of the volume. He seems, too, to have planned this placement from
the beginning. What is more, the confusion surrounding the imposition of these
patronage-soliciting poems, along with 'A letter of the Authors … to Sir Wal-
ter Raleigh' and the commendatory verses, went hand in hand with—and was
probably caused by—similar kinds of confusion attending on other places in
the printing of the poem, where Spenser seems to have withheld or amended
his manuscript, during printing, in order to handle with particular care the deli-
cate topical allusions and historical allegories shadowing the Queen, Ralegh,
and Leicester. The overall picture that emerges from the physical witness of the
printed book, with the bibliographical record of a disordered sequence of gath-
erings, frequent and sometimes serious mistakes, and several periods of slowing
or even stopping during the printing process, is one of miscommunication, last-
minute revision, and a near desperation to position The Faerie Queene carefully
with respect to the royal and aristocratic patrons who could help a rising poet,
and lend support to the ideological positions and factions with which we now sup-
pose him to have been aligned. Nashe's derisive mocking of the 1590 printing of
The Faerie Queene in Pierce Penilesse, then, seems to have figured Spenser's travails
shrewdly and exacdy. Bold to secure a royal dedication, bold to present his new
epic to a gallery of powerful potential patrons, and bold to adjust his poem, even
at the eleventh hour, to accommodate detailed topical allusion participating in
contemporary religious and political factional debates, Spenser was perhaps too
bold. By 1592, when Nashe took a gleeful swipe at Spenser's, and Ralegh's, for-
tunes, the poet was in disgrace and all of his bids for preferment, like the glories
of his Complaints, lay in ruins.

Is it possible that I have created here merely another insubstantial 'house
of cards', neatly but inconsequentially constructed from a chain of supposals?
Any study that seeks to reconstruct historical events, much less the human inten-

TABLE 3. Summary of the order of printing The Faerie Queene, 1590

                       

Signature	Press 1	Press 2
A–B	α o A (II) α o B (IV) α i A (II) α i B (IV)	β o A (I) β o B (III) β i A (I) β i B (III)	The use of the skeleton-formes in these two gatherings establishes their characteristics. For the anomalous deployment of the skeleton- formes here, see pp. 133–135.
C–S	α o (II) α i (III)	β o (I) β i (IV)	The 'regular' pattern of skeleton-forme use. There is no evidence to suggest that sequential printing of the gatherings was interrupted in this part of the run.
T	β o (I) α i (III) β i (IV) α o (III)		The serial use of skeleton-forme III twice in this gathering indicates that printing must have gone forward on a single press; see pp. 144–145.
V–Bb	α o (II) α i (III)	β o (I) β i (IV)	A return to the 'regular' pattern.
Cc	α o (II) β o (I) α i (III) β i (I)		The serial use of skeleton-forme I twice in this gathering indicates that printing must have gone forward on a single press; see pp. 131–133.
Ff–Hh	α o (II) α i (III)	β o (IV) β i (I)	A return to the 'regular' pattern. The printing of Ff–Nn before Dd–Ee is suggested by a number of mutually reinforcing anomalies in running-titles 5, 6, 8, 13, 16, and 19; see pp. 128–130.
Pp	β i (II) α o (III)	α i (IV) β o (I)	The use of I, III, and IV here is conjectural; both the printing of Pp between Hh and Ii, and the use of II, are suggested by the pagination mistake at position 8v in gathering Ii; see p. 146.
Ii–Nn	α o (II) α i (III)	β o (IV) β i (I)	A return to the 'regular' pattern.
Dd–Ee	α o (II) α i (III)	β o (IV) β i (I)	The late printing of these gatherings, taken alongside an unusual gap in the stanza layout on Ee8r, suggests that Spenser revised the manuscript of the poem during printing; see pp. 130–133, 135.
Oo	α o (II) α i (III)	β o (IV) β i (I)	A return to the 'regular' pattern.
Qq	[no data]		Qq was not printed directly after Pp (see pp. 147– 148), but might have been printed (for example, after Nn) at another medial point in the run.

tions associated with them, from bibliographical remains must pause deliberately
over this possibility. Four skeleton-formes were certainly very consistendy used
through the printing of this volume, with only one substantial change intro-
duced to a running-title over thirty-seven gatherings, and only one reversal of
the orientation of a skeleton-forme; this alone does that prove that the edition
was produced quickly, continuously, and at double press, with a single trio of
compositors working in a dedicated fashion on this project alongside two teams
of press-men. On the other hand, the coincidence of a pagination mistake across
the outer formes of both sheets in gathering F suggests concurrent machining of
these two formes at least, and, anyway, a system of four skeleton-formes makes
most sense when working at double press. The decision to devote two presses to
the job, especially considered in light of the regular nature of the stanzaic text,
and the evidence of absolute consistency in setting and printing the volume, sug-
gests speed; moreover, Wolfe printed the title-page of the volume on the recto
of A1, giving the volume a publication date of 1590 right at the beginning of the
printing process—which may suggest that he did not anticipate the kind of ex-
tended, open-ended schedule characteristic of the concurrent printing practices
described by McKenzie in 'Printers of the Mind'. The anomalies in the order
of imposition, discussed extensively above and summarized in table 3, similarly
indicate an impatience to get on with the job: why skip Dd and Ee, to return to
them later, if the whole job might have been stalled until the copy was available?
Why skip to the printing of Pp before Ii, except to keep production going dur-
ing the period of confusion occasioned by the omission of Dd and Ee? As noted
earlier, the job might have slowed between the printing of gatherings S and T, at
least in part, due to the demands of some concurrent job in the printing house;
but given that the inner forme of the outer sheet of T was, irregularly, printed
first, that the allegorical material contained in gathering T was highly topical
and sensitive, and the subject of Nashe's later jibes at Spenser, and that the
dense literary and Biblical allusions of this part of the poem would again occa-
sion late-stage revisions to Book 5 in 1596, it appears probable that we are not
here witnessing a simple interruption of the Spenser job to take up some other,
more pressing, printing project. In short, it is possible—as it always is—that the
bibliographical evidence witnesses nothing more than a series of contingencies in
a busy printing house; but the convergence of textual witness, allegorical content,
biographical record, and contemporary response leads us, I think reliably, toward
a series of meaningful interruptions—rather than mere discontinuities—in the
printing of The Faerie Queene in 1590. A house of cards is also sometimes a house
of cards.