It is a commonplace of Spenser studies that The Shepheardes Calender in its original printed form is an extremely visual text the typography and layout of which are not only complex but used deliberately to create meaning. These specific features are associated with particular elements (italic for arguments, roman for glosses, blackletter for poetry), type sizes are also used to structure the page and to make it easier to tell apart the different elements of the text. Furthermore, it is almost impossible for readers of the 1579 edition to confuse headings, arguments, eclogues, emblems and glosses because there are clear typographical markers for each. For example, the large italic type used towards the beginning for the poem ‘To His Booke’ is also used for emblems, the titles of eclogues, and for the running titles. As a result, this type assumes an important structural function, because it frames each of the woodcuts and thus helps to signal the beginning of a new eclogue. Occasionally, this function of the large italic type blends into its use in the running titles, whose purpose is of course to facilitate navigation through the text: for January, April, July, August, September, and December, the running title is printed directly above the woodcut and thus effectively doubles as the title:
Figure 1-2: running title doubling as title in 1579 (Ar and I3r)
Finally, its use in the emblems visually signals the conclusion of eclogues (and, in the case of ‘Merce non Mercede’, the conclusion of the entire book). Similarly, the small roman type used in the glosses first appears in the Epistle and the Generall Argument, meaning it is associated with explanation and contextualisation from the start.
Non-typographical elements also have a clear structural function: as with many early modern books, printers’ flowers are occasionally used simply as place fillers to avoid large empty spaces – for example on the title page, or after the March and June glosses. (A decorative device featuring a face framed by two cornucopias is used for the same purpose after the June and August glosses):
Figure 3-5: place fillers in 1579 (title page, C3r and Gr)
The main function of the printers’ flowers in The Shepheardes Calender, however, is to create a clear division between emblems and glosses, which are printed relatively closely together:
Figure 6: printers’ flowers separating emblem and gloss in 1579 (F4v)
Additionally, the Calender predominantly uses the same style of printers’ flowers for this purpose, which is identical to the one that also appears on the title page and in the border above ‘To His Booke’ (April and October are the only exceptions, as they use different styles of printers’ flowers). The use of decorated initials in The Shepheardes Calender is less clear and consistent: in the 1579 edition, they are taken from five different sets:
Figure 7-11: initials from different sets used in 1579 (C4v, ¶iiijr, Ar, I3r, M4r)
There is also no fixed position for them – the first two introduce the Epistle and Generall Argument, while in the eclogues, some occur at the beginning and others later in the text. This does not mean that their use is arbitrary, however: the way in which the decorated initials are being used loosely associates them with the presence of Colin, and specifically Colin’s songs. Out of the four initials at the beginnings of eclogues, three are the A of ‘A Shepheards boye (no better doe him call)’ (January), the L of ‘Lo Collin here the place, whole pleasaunt syte’ (June) – which reuses the initial of the Generall Argument – and the T of ‘The gentle shepheard satte beside a springe’ (December) which frame the story of Colin’s unhappy love in three out of the four eclogues in which Colin himself appears. By contrast, the remaining three initials that occur further into the eclogues (one in April, two in August) all involve invocations of Colin in his absence. The Y in April introduces his lay of fayre Elisa, recited by Hobbinol, and the same initial reappears in August, when Cuddie quotes another song of Colin’s, the ‘dolefull verse of Rosalend (who knowes not Rosalend?)’, after having settled the poetic dispute between Perigot and Willye by concluding that both are fine singers, though suggesting that neither is a match for Colin. Earlier, Perigot has been implying just that, by introducing the roundelay with the words ‘Sike a song neuer heardest thou, but Colin sing’. Correspondingly, the first line of the roundelay, ‘It fell upon a holly eue’ (spoken by Perigot) begins with an initial from the same set as the Y used when Hobbinol and Cuddie are quoting Colin’s songs, creating a visual echo of Perigot’s claim that the roundelay is comparable to one of Colin’s compositions. This is not to say that use of the decorated initials in the 1579 Calender is necessarily always deliberate or that its readers would have perceived it as such. Nevertheless, there are patterns to the way in which the initials are used, which effectively assign them the function of evoking Colin and Colin’s poetry and make them one of ways in which readers are reminded of Colin’s presence in the Calender even when he is not one of the speakers of the eclogue.
As well as facilitating the distinction between the different elements on the page, the typography and page layout of The Shepheardes Calender often signals the relative importance of the elements: for example, the page featuring the poem ‘To His Booke’ draws the attention of readers through the border of printers’ flowers at the top, but mainly through the fact that the type is so large that the 18-line poem fills the page (additionally, it is the only page in the book that names the ‘New Poete’). Arguments are printed in type smaller than that used for the eclogues, but larger than the type used for the glosses, implying a hierarchy of importance that seems borne out by the fact that the glosses are the only place in the book in which text is occasionally grouped into several columns in order to save space, as in this extract from the February glosses, which uses up to four columns:
Figure 12: multiple columns in 1579 glosses (B3v)
What can be seen from these examples is that the 1579 Calender uses typographical and non-textual features to convey meaning so deliberately that any change or omission amounts to a judgement regarding the significance of the changed feature. This poses a predictable dilemma for modern editions, which, as in all editions of early modern books will need to deviate from the original in ways that affect more than just the layout and typography: printing the eclogues in roman type and reducing, if not removing, the size differences (which publishers are likely to insist on, unless they are preparing a facsimile edition) means losing one of the most striking ways in which the 1579 edition visually distinguishes between different elements of the text. Similarly, choosing a two-column layout rather than a single column layout (a practical choice that allows for a more efficient use of space) will result in a ‘fuller’ page and a layout that makes the woodcuts appear disproportionately larger. Deciding to omit initials and printers’ flowers, on the other hand (routinely done in order to save space and to make the text easier to set), effectively relegates them to the role of mere decorative place-fillers, although they may have a structural function in the text as well. None of these problems are unique to The Shepheardes Calender, but its specific use of layout and typography means that typesetting the Calender will always amount to an act of reading and interpretation.
In this context, it is worth remembering that a text as visually complex as this presents a challenge for any reprint – not just for modern editions. As a matter of fact, the features described above only apply to the best-known edition: the first edition of 1579. As Spenser’s most successful work, however, The Shepheardes Calender was also his most frequently republished, and saw four quarto reprints during his own lifetime (in 1581, 1586, 1591 and 1597), as well as three published later as part of collected Works (in 1611, 1617 and 1679). There was also a bilingual Latin/English edition in 1653 that contained only the text of the eclogues and omitted everything else (including the woodcuts). It is the quarto reprints that are the most interesting in terms of layout. While all of them were prepared for the same publisher – John Harrison the younger – each was produced by a different printer, and every single one of them was reset, which means that between 1581 and 1597, several different people had the challenging task of setting the Calender. As a result, the different quarto versions of the book make for a particularly interesting case-study of early modern reprinting practices.
When resetting a book for reprinting, early modern printers and compositors faced a task that was not dissimilar to the one faced by their modern counterparts: they normally referred to a printed copy of the most recent edition, which meant that from the outset, they had some general idea of how the layout would work out and where changes might be made if necessary. Unlike their modern counterparts, of course, they did not have the luxury of planning the entire layout in advance but had to plan it gathering by gathering – eight pages at a time in the case of a quarto. Whenever they were behind, or ahead of, the page breaks of the previous edition, the compositors had to make adjustments as they went along, in order to catch up with the pagination. Catching up with the pagination mattered because the Calender is a quarto made up of 14 gatherings of eight pages each, which means that compositors had exactly 112 pages to work with. Going over 112 would have required an extra sheet of paper and increased production costs, while attempting to save eight pages to make do with one less sheet would have been impossible because of the woodblocks with the illustrations, which were a fixed size and restricted compositors’ options for moving text about.
While compositors were of course professionals, setting The Shepheardes Calender would have been a far from routine task. In addition to the woodcuts – which, unlike decorated initials or printers’ flowers, could not simply be dropped or substituted if necessary – there were the different type sizes for the different elements of the text, which made it more difficult to estimate how many pages would be needed for each month. To add to the difficulty, the eclogues and the glosses vary greatly in length: the shortest eclogue (January) takes up just two and a half pages including the argument, while the longest (May) requires slightly over nine; similarly, the shortest glosses (August) add up to less than a page, and the longest (April) to three and a half. Additionally, the proportion of eclogues to glosses does not remain the same throughout, ranging from slightly over one page of eclogue for each page of gloss (October) to the equivalent of eight pages of eclogue per page of gloss (August). The most straightforward solution to the challenges the Calender posed for compositors would have been simply to follow the layout and typography of the previous edition as closely as possible, i.e. to keep types and sizes as they were, to use initials and decorative elements of comparable size in the same places and to correct minor layout differences as soon as possible. For the most part, however, this was not what actually happened. Instead of playing it safe, the compositors engaged in the reprinting of the Calender made changes, some of which had significant knock-on effects on the layout.
What I would like to suggest in this essay is that at least some of those changes made to the layout and typography in early modern quarto reprints of the Calender are neither accidental nor motivated purely through practical concerns. I would argue that they indicate that printers and compositors involved in the reprinting actually ‘read’ and interpreted the text – not in terms of its contents, but in terms of its appearance. So paradoxically, those reprints are effectively reader’s responses from people who probably never read The Shepheardes Calender.
The first reprint was prepared in 1581 by Thomas East. The timing is interesting because it directly follows not just the first edition of 1579 but also the so-called Spenser/Harvey letters, in which the two ‘university men’ ‘G.H.’ and ‘Immerito’ discuss Immerito’s recently published book, perhaps leading to a surge in interest – or at least leading John Harrison to hope for such a surge. The remaining three quarto editions were all published at moments that could be regarded as strategically timed: the third edition in 1586, following the death of Sir Philip Sidney (perhaps an attempt to capitalise on the fact that some early readers believed the book to be by him), the fourth in 1591, the year in which the Complaints were published and subsequently called in, and the fifth in 1597, after the publication of the (commercially unsuccessful) second edition of The Faerie Queene.
The first striking difference to the first edition is the layout of the first page, containing the envoy ‘To his Booke’:
Figure 13-14: ‘To his Booke’ in 1579 vs 1581 (1579 ¶v and 1581 ⁂v)
As mentioned above, the poem is printed in an unusually large type in the first edition and consequently fills the entire page. The 1581 edition, by contrast, uses much smaller letters (the italic type also used in the glosses) and centres the poem on the page. There is little practical purpose to this: selecting a smaller type allows the compositor to save space, but it is space he cannot use for anything. So the decision to shrink and centre the poem seems to have been made by Thomas East’s compositor for aesthetic reasons, perhaps to create a visual resemblance to a dedicatory sonnet, which would be more common in this position in the book, and to remove the apparently excessive significance given to the poem by the oversized letters. Other visual changes in the edition concern the Epistle and Generall Argument, which take up the same number of pages as in 1579, but are spaced out more evenly, and begin with two decorated initials from the same set (unlike in -1579):
Figure 15-18: the first two initials in 1581 vs 1579 (1581⁂iir and ⁂iiijr and 1579 ¶ijr and ¶iiijr)
Overall, those changes result in a balanced, slightly more uniform-looking layout. In the main text, it quickly becomes clear that Thomas East’s compositor privileges certain elements over others. Arguments, emblems and glosses are treated in a way that suggests he considers them manoeuvrable bits of text that can be spaced out, compressed or shrunk at will in order to arrive at his desired page layout – but he is scrupulously exact in reproducing the layout of the eclogues proper. While he wraps the glosses around the eclogues without following the original page breaks, and uses them to avoid blank space or decorative elements (neither of which he appears to be particularly fond of), he keeps the woodcuts in almost exactly the same position and preserves all page breaks for the eclogues.
Figure 19-20: page layout in 1579 (H2v-H3r) vs. 1581 (Hiiv-Hiiir)
The main effect of the changes in the 1581 edition, then, is to visually highlight the eclogues and woodcuts as the most important elements of the text. Glosses and emblems are treated as annotations rather than as part of the text. In the image above, the argument is set in noticeably smaller type, but the compositor has not used this space to squeeze in an additional line of verse at the bottom of the page. There is also a slight erosion of the distinction between the different types used in the 1579 edition: some elements that were originally in roman type are now in italics, and vice versa (such as the headwords of the glosses and the proper names in the eclogues), and the few words of Greek (in the Generall Argument, the glosses of February, April, and July, and the May emblems) have been transliterated, probably because East’s print shop did not own any Greek type. Printers’ flowers – most notably the ones separating the glosses from the emblems – have been removed almost entirely; perhaps they were not in keeping with East’s ‘house style’. The surviving titles that were printed by East predominantly use printers’ flowers to form frames on title pages and to introduce or conclude paratextual materials, such as dedicatory epistles, prefaces, prologues, envoys or tables of contents, and occasionally as a border below or above the impressum. Printers’ flowers used to mark subdivisions within the main body of the text occur much more rarely in East’s books (and if they are used, they are never used consistently after or before every section). The 1581 edition of the Calender corresponds to this use of printers’ flowers: it has a frame on the title page and a row used as a border above the Generall Argument – but they are absent from the main body of the text, except for two rows used to fill some blank space after the end of the March glosses. On the two other occasions when there is a larger blank space to fill (after the June and August glosses), the compositor instead opts for a decorative pattern also featured on the title page.
In terms of type size, the most significant change is that the emblems are now set in substantially smaller type, and the italic type of equivalent size is reserved for running titles and titles of eclogues. Overall, however, the 1581 edition is a careful reprint that introduces changes – including the relegation of arguments and emblems to supporting roles – but otherwise manages to stay visually close to the original. As a result, pages containing no glosses, emblems or arguments look almost identical to the 1579 edition, though minor variations of spelling and punctuation nevertheless indicate the text has been reset:
Figure 21-22: page layout in 1579 (A2r) vs 1581 (Aiir)
The third edition of 1586 was prepared by another printer, John Wolfe, perhaps because Thomas East was too busy reprinting John Lyly’s Euphues, one of the hottest bestsellers of the decade, to accept the kind of rush job that the third quarto of The Shepheardes Calender appears to have been.
The erosion of types used for the different elements of the text is a gradual process across the re-printing history of the Calender, but the 1586 edition makes more of those changes than the others, indicating that Wolfe’s compositor did not assume there was any special significance to the types used – except for the blackletter of the eclogues, which he appears to have thought an integral feature of the text that needed to be preserved. He also seems to have considered the epistle less important than other elements, because he sets it in a much smaller type than the two previous editions. Consequently, he fits the whole epistle onto three pages instead of four and still has five lines’ worth of blank space at the bottom of the third page. This ‘saved’ space is then used on the Generall Argument, whose letters are nearly twice as big, so it takes up more space than the Epistle, although it is in fact a shorter text. This typographically privileges the Generall Argument over the Epistle and may be an indication that someone read the headings and decided something called ‘The generall argument of the whole booke’ was of greater significance – and consequently merited larger type – than a mere dedication. By the end of the ‘general Argument’, however, we can see that Wolfe’s compositor has misjudged the amount of available space:
Figure 23: trying to save space in 1586 ([viii]-Ar)
In this image, the text of the Generall Argument continues onto the first page of the January eclogue, and the compositor has clearly attempted to minimise the damage by reducing the type size in the final paragraph and putting the argument of the January eclogue in the smallest possible type. The overall visual effect is unconvincing, and despite his best efforts, the compositor is still four lines behind the two previous editions by the end of the page.
He does not attempt to compress the eclogue, however, and restricts himself to squashing together the glosses, the effects of which become visible by the end of January. Having used very small letters for the glosses, Wolfe’s compositor is now ahead of the two previous editions – as can be seen by the position of the February woodcut, which is a couple of centimetres higher up on the page. To use up this space again, he increases the February argument to nearly three times the size of the tiny January argument shown in the previous image.
Figure 24: typography of 1586 (A2v-A3r)
After this, the page breaks are largely identical to those of the previous editions until the start of the glosses, where the pattern of gaining more space than needed and squandering it again on the next page repeats itself. This continues until the end of the volume. Even on the final two pages Wolfe’s compositor continues to compress the glosses, even though there is no longer any need for it. So the first two lines of the concluding poem ‘Loe I have made a calender for every yeare’ are tagged onto the December glosses, even though there would have been more than enough space to put the entire poem on the final page. (Note also the mistake in the impressum, which wrongly gives the name of the printer for the third edition as ‘Thomas East’).
Figure 25: ‘Loe I have made a Calender’ in 1586 (N3v-N4r)
Overall, the page layout of the 1586 edition shows similar attempts to make sense of the different components of the text and assess their relative significance as the previous edition, although the strategies employed by Wolfe’s compositor suggest he was slightly out of his depth at times. While in the 1581 edition, the changes made to the layout and typography are carefully balanced so the overall visual effect of the text on the page is still close to the original, the layout of the 1586 edition looks rather rushed.
Nevertheless, the 1586 edition does introduce a feature that had not been present in the previous two editions: a set position for the woodcuts. In the editions of 1579 and 1581, only half of the woodcuts had actually been right at the top of the page. While the compositor for the 1586 edition misjudges the page layout for the first page of January, from April onwards he successfully juggles the layout to ensure that all woodcuts are placed at the top of the page, to create a more uniform layout for the eclogues. As an added bonus, this allows him to use up some of the extra space gained from putting the glosses into smaller type, since oversized arguments would not be sufficient for that purpose.
John Wolfe’s compositor does not learn much from his layout mistakes over the course of the book, and it is clear that he struggles to use the available space efficiently, which is why he is forced to leave several pages half-empty and fill them with a decorative device. The first time this happens is at the end of March (which has the longest eclogue but comparatively short glosses), as well as June (a short eclogue with comparatively long glosses), July (a long eclogue with long glosses), August (a long eclogue with very short glosses), and October (a short eclogue with long glosses). He also uses decorative devices or borders of printers’ flowers to fill less than half a page’s worth of empty space after April, May, and November. By comparison, Singleton’s compositor had only needed to use decorative place-fillers after four months (March, June, July, and August), of which only two (March and June) came close to leaving half a page unused. The second edition, which formed the basis for the one printed by Wolfe, and whose compositor, had clearly been wary of blank spaces on the page, had managed with just three, only one of which (August) had left half a page unused. The compositors’ contrasting attitude towards blank space becomes particularly apparent when comparing the double page H.iiv-H.iiir in the different versions:
Figure 26-28: empty space on H2v-H3r in 1579 vs. 1581 and 1586
The edition that succeeded Wolfe’s, printed by John Windet in 1591, the same year that Spenser’s Complaints were published, stands out among the quarto editions, because it is the only one whose compositor opted for playing it safe. Windet and Wolfe often cooperated, so it is possible that the compositor of the 1591 edition had been warned of the pitfalls of setting The Shepheardes Calender and decided to reproduce the layout of the previous edition as precisely as possible. Unlike East’s compositor, who had followed the layout of the previous edition for some elements, but not for others, this compositor follows the previous edition down to the page breaks for glosses, as can be seen from this double page in the 1586 and 1591 edition:
Figure 29-30: identical page breaks in 1586 and 1591 (D2v-D2r)
Apart from using a few different spellings (for example changing ‘perteyneth’ to ‘perteineth’, ‘florished’ to ‘flourished’, and ‘deuise’ to ‘diuise’), and using a smaller type for the catchword on 14v, Windet’s compositor has produced what could pass for an exact replica of the same double page in the previous edition. This is maintained throughout the book, so that the only pages from the two editions that can be told apart at a glance are the ones that contain decorative elements, since Windet’s print shop uses a different selection of printers’ flowers and decorative place-fillers. In a sense, the precision with which Windet’s compositor manages to replicate the layout of the 1586 edition is quite impressive. Nevertheless, the price for his facsimile-like approach is that he also ends up reproducing the previous edition’s mistakes, including a pagination error (labelling fol. 11 and 12 as ‘12’, which the 1591 edition, perhaps in a misguided attempt to correct the original error, increases to three by also labelling fol. 10 as ‘12’), an incorrect catchword on 36v (instead of ‘Other’, both editions have ‘That’, which had been the catchword on the corresponding pages in the 1579 and 1581 editions) and the tagging on of Colin’s emblem to the November eclogue without a space, so the words ‘Colins Embleme’ end up resembling a speech prefix. Windet’s compositor even goes so far as to replicate the awkward layout of the first page of January:
Figure 31: imbalanced layout of 1586 Ar reproduced in 1591
What is remarkable about the 1591 edition, then, is that it is the only one whose compositor follows his predecessor’s decisions down to every page break and attempts to preserve the layout and typography of the edition he is working from as accurately as possible, without introducing significant changes – even if it means reproducing mistakes.
The last of the quarto editions was printed in 1597 by Thomas Creede, who had printed Colin Clouts Come Home Againe and Sidney’s Defence of Poesie two years earlier. The 1597 edition is the first not to print the paratexts and the glosses in significantly smaller type. As a result, the Epistle takes up a full five pages, and the type of the glosses is nearly the same size as that of the eclogues.
Figure 32-33: eclogue, emblems and glosses typographically distinguished in 1586 (B2v-B3r) and 1597 (C3v-C4r)
This approach has an interesting effect on the text, because printing the glosses in such a large type visually signals that they are part of the ‘main’ text and no less important than the eclogues themselves. The 1597 edition’s treatment of the glosses is unlike that of the three previous editions, whose compositors had used the glosses to balance the layout (1581) or to save space (1586 and 1591), which suggests they considered them to be the least important element of the text. Through the greater similarity of type sizes, the visual distinction between arguments, emblems and glosses also becomes more blurred in the 1597 edition – an effect which is heightened by the fact that the text is almost entirely free of decorative elements. In practical terms, it means that Thomas Creede’s compositor soon encounters serious layout problems, because printing the glosses in such big letters has a knock-on effect. (This also explains the absence of printers’ flowers – Creede’s compositor cannot afford to use them.) For the rest of the volume he is playing catch-up and trying hard to stay within 112 pages. Unlike his predecessors, however, this compositor tries to save space on the eclogues rather than the glosses, so he squeezes as many additional lines of verse as possible onto each page. Creede’s compositor even makes a minor aesthetic sacrifice by dropping the last of the decorated initials that had survived all previous reprints – the A at the beginning of January, which allows him to fit seven more lines of verse onto the first page than the previous two editions had, and three more lines than the 1579 and 1581 editions.
Initially, the compositor’s strategy appears to work, although slowly but surely he falls behind. By the time he reaches April, the layout is still lagging by a full page, as can be seen from the fact that the April woodcut is placed on a right-hand page instead of a left-hand page. This is not a problem as such, since readers would not normally notice the lagging layout unless they happened to have an earlier edition to compare it to, but unfortunately for Creede’s compositor, the April eclogue has particularly long glosses, so he is eventually forced to do something that all of his predecessors had managed to avoid: place a woodcut right at the bottom of the page.
Figure 34: woodcut placed at the bottom of a left-hand page in 1597 (E4v-Fr)
If printed in this position, the woodcuts – whose function is after all to introduce the eclogues – end up visually resembling decorative place-fillers, which is why all previous compositors had tried to ensured there was at least room for a few lines of the argument below each woodcut. After a lot of scrambling to squeeze in as many extra lines as possible, however – as well as dropping a stanza from June, which allows him to gain a further six lines – the compositor of the 1597 edition finally manages to catch up with the layout of the previous edition just in time to place the August woodcut at the top of the page again.
At this point, the compositor must have thought the worst of his layout problems were over, because at the end of the August glosses he dares to leave some space unused for the first time, and fills it with his first (and only) row of printers’ flowers instead of placing the September woodcut and argument on the bottom half of the page.
Figure 35: ‘wasting’ space in 1597 (K2v-K3r)
After Creede’s compositor has successfully placed the August, September and October woodcuts at the top of a page, however, things start to unravel for him again, because the October glosses happen to be the second longest in the entire book. In the 1579 and 1581 editions, those glosses had taken up three and a half pages, the 1586 and 1591 editions had even managed to squeeze them into two and a half. The 1597 edition needs a little over four pages, so the inevitable happens – again. This time, placing the woodcut at the bottom of the page is arguably even worse, because it is on a right-hand page, which gives the woodcut the appearance of marking the conclusion of a section rather than the beginning of a new one. To make things worse, the positioning of the woodcut is also particularly conspicuous, because the bottom right corner of a right-hand page is the last section of the page to be viewed before readers turn the page, making it a particularly noticeable spot on the page (which is why full-page advertisements in print newspapers and magazines traditionally place the company logo in this spot).
Figure 36: woodcut placed at the bottom of a right-hand page in 1597 (M3v-M4r)
Luckily, November and December have relatively few glosses, so Thomas Creede’s hapless compositor just about manages to stay within 112 pages, though only by leaving out the impressum and compressing the poem ‘Loe I have made a Calender’ so that it looks almost indistinguishable from the glosses rather than like a counterpart to ‘To his Booke’.
Figure 37: the final two pages in 1597 (O3v-O4r)
Interestingly, despite all his layout woes, the one thing Creede’s compositor never tries to do is reduce the size of the glosses or switch to using small italics, which might have been used to save space in the glosses even if Creede’s print shop was short on small roman type. He would sooner commit one of the deadly sins of typesetting, it seems, by placing a woodcut at the bottom of a right-hand page.
After the 1597 edition, The Shepheardes Calender underwent one more major layout change, which effectively neutralised the dynamics of layout and typography in the quarto reprints that had caused the compositors so much grief. As Stephen Galbraith has persuasively argued, Matthew Lownes’ folio edition of Spenser’s works was a cost-efficient, commercial project. Consequently, the layout changes to The Shepheardes Calender had two main aims – to save paper (and money), and to standardise the layout to make it resemble that of the other works in Lownes’ Works, most notably The Faerie Queene.
Figure 38-39: standardised layout in 1611 (FQ 16-17 vs SC 44-45)
As can be seen from these images, the Lownes’ edition uses a two-column layout for the verse that superficially resembles the layout used for The Faerie Queene, the text that directly precedes The Shepheardes Calender in all surviving copies of Lownes’ edition. Glosses and arguments, on the other hand, are printed as a single column and in a slightly larger type. One of the typographical casualties of Lownes’ streamlining efforts is the characteristic blackletter type for the eclogues. All of these are objectively sensible choices for a ‘Complete Works’ edition, but the result is a relatively straightforward page layout that appears to highlight the arguments and the running title, and in which the woodcuts are so dwarfed by the text that they need additional borders for reinforcement. The result is a page layout that decidedly differs from that used in the quarto editions: the woodcut which in the quarto edition had accounted for nearly a third of the page now corresponds to less than a quarter, and the visual link that had existed between the image and the text in the 1579 quarto, because the width of the argument matches that of the woodcut, has been removed. Additionally, the woodcut has been separated off by two horizontal lines, which have a ‘stretching’ effect and serve to connect the woodcut to the added borders on its sides, changing the image’s proportions. Through the two-column layout for the eclogue, the 1611 page layout is also much more text-heavy: it features an additional four and a half stanzas of verse plus three speech prefixes (which take up two thirds of a page in the 1579 edition).
Figure 40-41: page layout of the first page of Aprill in 1579 (Ciiiv) vs 1611 (13). Figure 41 has been resized to facilitate comparison
It is perhaps unsurprising that after the folio edition, the layout-juggling seen in the quarto reprints ceased – all subsequent editions reproduced the standardised layout of the folio edition, and the next edition after Lownes’ (Jonathan Edwin’s 1679 Works of that Famous English Poet, Mr. Edmond Spenser) lacked the woodcuts, making it even more text-heavy than its predecessor. Through the changes introduced in the Lownes edition, the text became much easier to set (and arguably easier to read) but also less visually interesting – and thus it ceased to challenge its non-reader readership of early modern compositors.
The reprinting history of The Shepheardes Calender is not representative of early modern reprints, because it involves a succession of different compositors who tackled a complex text and had to consider the significance of its components. Not all early modern reprints were such a complicated business, of course; often it was a routine task. The numerous reprints of the two parts of Euphues that Thomas East, the printer of the 1581 Calender, churned out between 1579 and 1588, for example, do not significantly differ from each other visually, because Euphues had no illustrations and only required one size of type for the main text and another for the headings, and was thus a straightforward text to set. What the uncommon example of the Shepheardes Calender quarto reprints does, however, is to make the strategies and thoughts processes of early modern compositors visible, because each of them had to find his own solutions to the typographical and layout challenges that this book posed: the first made a judgement about the significance of components by preserving the layout of the eclogues but changing the rest, the second tried to standardise the positioning of the woodcuts at the expense of the glosses, the third attempted to preserve as much of the text’s appearance as possible, and the fourth made the bold move of placing the glosses on an almost equal visual footing to the rest of the text – possibly followed by the even bolder move to drop a stanza of verse whose absence might not be immediately noticed in order to remain within 112 pages.
Images from the 1579 edition (figures 1-13; 17-19; 21; 26; 40) are from the Folger Shakespeare Library and licenced under a Creative Commons Licence. Images from the 1581 edition (figures 14-16; 20; 22; 27; 32) are from the Harry Ransom Center, The University of Texas at Austin. The remaining images – from the 1586, 1591 and 1597 editions, and from the Lownes Works of 1611 – are all from archive.org and based on books located at the University of Illinois Urbana-Champaign.
I am very grateful to Joe Loewenstein for kindly letting me use images of the 1581 edition originally digitised for the Spenser Archive, and to the Harry Ransom Center for granting me permission to reuse those images without additional charges.
 The only exception to this are the May emblems, which are set in Greek type that is substantially smaller, presumably because there was only one size of Greek type available in Hugh Singleton’s print shop.
 The fourth eclogue featuring Colin (November) does not begin with a decorated initial, perhaps because the eclogue begins relatively low on the page and there would not have been enough space for it. The only initial to introduce a Colin-less eclogue is the D of ‘Diggon Dauie, I bidde her god day’ (September), possibly chosen because the initial features an image of King David.
 A particularly extreme example of changes made to the typography and layout of the book in a modern edition would be W. L. Renwick’s 1930 edition, which almost entirely dispensed with size distinctions, omitted the woodcuts, and introduced each new eclogue with a new title page, with the argument printed on its reverse. The Smith and de Selincourt edition of The Poetical Works of Edmund Spenser (first published in 1912 and reproduced multiple times), uses a two-column layout throughout, which is applied even to the arguments and glosses of The Shepheardes Calender, which had been printed in a single column in the folio edition. The Variorum edition, on the other hand, at first sight has the appearance of a facsimile and even reproduces the title page and original initials. On closer inspection, however, it omits all printers’ flowers or decorative elements, and the proportion of type sizes is different: for example, the type of the Epistle, which the editors have renamed ‘DEDICATORY EPISTLE’ in the running titles, is the same size as that of the eclogues (in 1579 it is the size of the glosses, so the first page of the Variorum Epistle contains nearly 200 more words; this is particularly noticeable when measuring the height of the initial V in relation to the type: in 1579, it corresponds to seven lines, in the Variorum edition, the height of the initial only corresponds to five lines). The type used for the eclogues, on the other hand is relatively smaller, so the pages of the Variorum edition on average contain two more lines of verse than the 1579 edition. Like most modern editions, it also attempts to standardise the layout and make it more calendar-like by placing all woodcuts at the top of a page – much like John Wolfe’s compositor did in the 1586 edition discussed below. The eclogues reproduced in the Norton selection of Spenser’s Poetry omit printers’ flowers and initials and uses the same type size for the Epistle and Generall Argument, arguments, eclogues and emblems. It also prints the glosses as footnotes and adds vocabulary glosses on the right-hand margin. The layout in Richard McCabe’s edition of the Shorter Poems preserves the glosses in their original position but reduces the distinction between type sizes and positions woodcuts at the top of a page, similarly to the Variorum edition. It omits all non-typographical elements except for the woodcuts. All of these choices – except the omission of the woodcuts, perhaps – can easily be justified through the practicalities of modern printing or the needs of modern readers, and most of them are likely to have been made by publishers rather than editors (who may well have wished to preserve more of the original appearance of the book).
 There are several indications that although the 1596 Faerie Queene was well received, it was not a commercial success. William Ponsonby, who outlived Spenser by several years, never reprinted The Faerie Queene between 1596 and his death in ca. 1604. Had the book been commercially successful, Spenser’s death in 1599 would have presented a convenient opportunity for Ponsonby to issue a new edition, but he appears to have made no attempt to do so. Ponsonby’s son-in-law Simon Waterson, to whom the printing rights for Spenser’s works were transferred, never published his own edition but instead sold the rights on to Matthew Lownes, whose first edition was printed in 1609. The collected works subsequently published by Lownes were a cost-efficient ‘Build-It-Yourself edition’ (as demonstrated by Galbraith), suggesting the 1609 edition was not a resounding success either. Several decades later, Thomas Fuller in his brief life of Spenser hinted that The Faerie Queene (which Fuller only calls ‘his book’) had been a commercial flop, speculating that it ‘had been more salable if more conformed to our modern language’ (The History of the Worthies of England (1662), 219).
 For a detailed overview of the typographical changes made in the different quarto editions, including measurements, see Jessica Beckman’s essay ‘Time, Reading, and the Material Text: Revising Spenser’s Shepheardes Calender’ (Spenser Studies 33 (2018), forthcoming), which examines how these changes affected the way in which readers were able to navigate the text.
 Thomas East printed titles from a wide range of genres, including reprints of Malor, Thomas Elyot’s Boke named the Governor and both parts of John Lyly’s Euphues, as well as medical titles such as Thomas Gale’s Certaine works of chirurgerie, sermons, several titles translated from French and Spanish and even sheet music (among other things, he published the works of William Byrd). None of this required the use of Greek type, however.
 His edition of The Shepheardes Calender falls into what appears to have been East’s most productive decade (1576-1586): although East continued to work into the 1600s, more than half of his surviving titles are from this decade. Of the 54 titles from this period listed on EEBO, 37 use printer’s flowers on the title page, 30 use printer’s flowers to mark paratextual materials (including some with multiple paratextual materials), twelve use them as borders around the impressum, and only eight use them for subdivisions within the text. Beyond the title frames, only six of the books contain more than five rows of printer’s flowers.
 For the significance of blackletter type and its connotations, see Galbraith’s essay ‘‘English’ Black-Letter Type and Spenser’s Shepheardes Calender‘, Spenser Studies 23 (2008), 13-40.
 Interestingly, the compositor seems to have been determined to keep the decorated initial despite his space constraints, perhaps out of aesthetic conventions. Early modern books rarely begin without a decorated initial of some sort, presumably because browsing customers were most likely to view the first few pages, so it was a strategically useful move for printers to place handsome illustrations or initials towards the beginning of a book, even if the rest was largely plain and undecorated.
 The title page correctly names John Wolfe as the printer.
 For a counterexample, see Creede’s reprint of Ralph Robinson’s translation of Thomas More’s Utopia (also 1597), where the marginal glosses are consistently of a smaller size and different type to the glossed text: small italic type to gloss the Epistle (printed in roman type) and small roman type to gloss the main text (printed in blackletter). The earlier editions of the book, of 1551, 1556, and 1557, had used blackletter throughout. It is possible that Creede’s decision to distinguish the different elements of the text through different types in his edition was partly influenced by The Shepheardes Calender.
 The lowest-placed woodcut of the previous editions had been the March woodcut in the 1579 edition, with only three lines of argument printed underneath (this was replicated in the 1581 edition). The 1586 edition, for all its shortcomings in terms of layout manages never to place a woodcut lower than the middle of the page.
 This omission may or may not have happened unwittingly. At the very least, the missing stanza is a fortunate accident for Creede’s compositor, because it saves him from having to place the July woodcut at the bottom of the page. The stanza, the last on 24r in the 1591 edition, which begins with the line ‘Nowe dead he is, and lyeth wrapt in lead’, is essentially an elaboration on the preceding stanza (‘The God of Shepheards Tityrus is dead’) and only referred to in one short gloss (‘Oh why’), which may explain why its absence went unnoticed in subsequent editions. If Creede’s compositor had decided to drop a stanza on purpose, this particular stanza would therefore have been a shrewd choice.
 One early modern publication that exploits this effect in a particularly interesting manner is the Epitaph vpon the death of the worshipfull Maister Benedict Spinola, printed by Thomas East in 1580 (STC (2nd ed.), 1057). The Epitaph consists of a single sheet, which features a poem printed in two columns which are separated and framed by four vertical rows of printers’ flowers. The bottom third of the page contains a woodcut of Spinola’s corpse wrapped in a shroud inside an open grave, surrounded by other graves and scattered skulls and bones. The bottom right-hand corner of the woodcut depicts a smaller skull facing towards Spinola’s grave, framed by a cross and a horizontal bone arranged so they form a little square with the border of the image.
 ‘Spenser’s First Folio: The Build-It-Yourself Edition’, in Spenser Studies 11 (2006), 21-49.
 The quarto reprints match the width of the woodcuts less precisely than the 1579 edition, because they are using type that is either slightly smaller (1581, 1586 and 1591) or slightly larger (1597). Nevertheless, it is clear that the compositors are using the woodcuts as a point of reference for the line width of the arguments, while the 1611 edition uses the added borders for this purpose.