University of Virginia Library

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]


Page 120

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


Page 121
8s: A8-Pp8. This was subsequently modified to incorporate the four-leaf cancel-
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


Page 122
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
  • 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-


Page 123
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.


See F. R. Johnson, A Critical Bibliography of the Works of Edmund Spenser Printed Before
pp. 11–18; and Hiroshi Yamashita et al., A Textual Companion to The Faerie Queene 1590
(Tokyo: Kenyusha, 1993), especially the appendices.


The best general treatment of Wolfe's printing career is still Harry R. Hoppe, 'John
Wolfe, Printer and Publisher, 1579–1601'. See also Ian Gadd, 'John Wolfe, bookseller and
printer', Oxford Dictionary of National Biography (Oxford: Oxford Univ. Press, 2004); I am grate-
ful to Ian Gadd for private comments on Wolfe's career. For an interpretation of the intellec-
tual property concerns developing out of Wolfe's battle with the Stationers' Company in the
1580s—concerns that might well have suggested Wolfe to Ponsonby's or Spenser's notice for
this job, and certainly advertised his reputation and capacity—see Joseph Loewenstein, 'For a
History of Literary Property: John Wolfe's Reformation', English Literary Renaissance, 18 (1988),


The tradition that The Faerie Queene was printed quickly has until recently been just
that—no more than an unsubstantiated tradition. The terminus a quo has usually been taken
to be the prefatory letter to Walter Ralegh eventually affixed to the poem, dated 23 Janu-
ary 1589 (presumably new style, which Spenser tended to favor in other contexts, and par-
ticularly in his secretarial work), and that ad quem the end of the calendar year 1590, which
given the imprint date, as well as the awarding of a royal annuity in February 1591, seems
very likely. On the other hand, a recent manuscript discovery suggests that the first three
books of the poem were in a final form (if not the final form of the 1590 edition) in 1588, and
Abraham Fraunce quotes from the text in his The Arcadian Rhetorike of that year; see Joseph
Black, '"Pan is Hee": Commending The Faerie Queene', Spenser Studies, 15 (2001), 121–124,
and D. Allen Carroll, 'The 1588 MS Commendation to The Faerie Queene', Spenser Studies,
16 (2002), 105–123. But as I will demonstrate below, there is good reason to think that the
evidence of the 1590 printed book itself points to a single, continuous, and expeditious print-
ing of the volume; and this is a point to which, at the conclusion of the present article, I will


D. F. McKenzie's learned and important argument in 'Printers of the Mind' notwith-
standing, it seems probable that F. R. Johnson's hypothesis—of a group of compositors setting
the entire text of the Spenser's poem with four skeleton-formes on two presses—is broadly cor-
rect. See 'Printers of the Mind: Some Notes on Bibliographical Theories and Printing-House
Practices', Studies in Bibliography, 22 (1969), 1–75; repr. in Making Meaning: "Printers of the Mind"
and Other Essays,
ed. Peter D. McDonald and Michael F. Suarez, S.J. (Amherst: Univ. of Mas-
sachusetts Press, 2002), pp. 13–85. As Johnson demonstrated and Yamashita et al. confirmed
(a position I will further corroborate and amplify at various points below), the printing of the
volume was exceedingly regular, with the four nearly invariable skeleton-formes employed in
a consistent system and the compositors assigned to consistent stints from the earliest stages
of the project straight through to the end. Given the special nature of the text as a stanzaic
poem—easily cast off and straightforwardly set, but requiring custom indentation—it seems
that the volume might have been securely set at speed, and that there were clear efficiency gains
to be had from doing so. That the printing was a commission from the publisher and bookseller
William Ponsonby, rather than Wolfe's own venture, may have constrained his scheduling. It
is important to recognize that a general acceptance of Johnson's hypothesis does not rule out
concurrent printing on either press, especially of ephemera, or suggest that both formes of the
inner sheet always went to one press and those of the outer to another (though this seems likely
to have been the practice), or that the same two (of Wolfe's four) presses were used throughout
the job.


In fact, there was one case of skeleton-forme inversion, after the printing of gathering
Bb; see below, pp. 132–133.


See F. R. Johnson, A Critical Bibliography of the Works of Edmund Spenser Printed Before
pp. 11–13.


This view has been substantially revised by Yamashita et al., who have determined,
through an analysis of word-spacing and spelling preferences, a consistent pattern of division,
within each regularly-printed gathering, of work between three distinguishable compositors
(X, Y, and Z). See A Textual Companion to The Faerie Queene 1590, pp. 406–408. In any
case, the pagination was probably added to the formes of both sheets when—that is to say,
just before—the chases were locked on the imposing stone, and the fact that the outer formes
of both sheets are off by the same amount supports (though it does not confirm, and certainly
does not confirm for the duration of the entire job) the idea that the inner and outer sheets
were machined concurrently on two presses. For pagination on the stone, see Fredson Bowers,
'Notes on Running-Titles as Bibliographical Evidence', The Library, 4th ser., 19 (1938), 313–338
(p. 317).


For a full description of the idiosyncrasies in the running-titles used to identify the
skeleton-formes, see F. R. Johnson, A Critical Bibliography of the Works of Edmund Spenser Printed
Before 1700,
pp. 16–17. Facsimiles of the running-titles are reproduced by Yamashita et al. as
part of a useful summary of the pattern of skeleton-formes used in the printing of the entire
volume (see A Textual Companion to The Faerie Queene 1590, pp. 423–425), an exercise in
reproduction that I have in part reproduced below (see figure 1).


As Bowers has argued, relying on evidence from Moxon, the skeleton-forme was prob-
ably dismantled after running off a given sheet, and re-assembled piece by piece around the
pages currently being imposed: 'in a sense, therefore, the same skeleton has been transferred
to the fresh letterpress, instead of the letterpress being added to the undisturbed skeleton.' See
Fredson Bowers, 'Notes on Running-Titles as Bibliographical Evidence', p. 319. The running-
titles, effectively part of the furniture, were thus transferred from one forme (Bowers' 'fresh let-
terpress') to another. In occasional instances where a running-title was not required in its usual
position in the forme, then, it must have been simply, but carefully, left aside during printing,
and then recovered for the imposition of the following sheet. That some running-titles, left aside
in this way at one point during the printing of gathering Bb of The Faerie Queene in 1590, were
altered suggests that these titles were not as carefully handled as other running-titles, similarly
suspended, at other points in the printing. Given the overall precision of the process, this de-
parture from usual practice invites investigation.


See Yamashita et al., A Textual Companion to The Faerie Queene 1590, pp. 405–411.


The following account summarizes the results of Yamashita et al. in their Textual
Companion to
The Faerie Queene 1590, pp. 407–408. D. F. McKenzie has rightly questioned
the reliability of evidence from punctuation- and word-spacing as a basis for identifying com-
positors and demarcating their stints, in 'Stretching a Point: Or, the Case of the Spaced-Out
Comps', Studies in Bibliography, 37 (1984), 106–121 (repr. in Making Meaning, pp. 91–106). It
should be noted that Yamashita et al. have used evidence of spacing in conjunction with other
types of evidence, such as spelling preferences, and have interpreted and tested their findings for
compositorial stints by the broader, and regular, patterns of press-work apparent throughout
the production of the volume. Their hypotheses on compositorial stints thus appear credible,
while always constrained by the usual limitations on inductive inference.