University of Virginia Library

Shared Printing, Printer's Copy, and the Text(s) of Gascoigne's A Hundredth Sundrie Flowres
by
Adrian Weiss [*]

Shared printing introduced complications into the relationship between book production and textual transmission that produced bibliographical evidence of varying degrees of importance to bibliographers and textual editors. In general, two categories of shared books obtain. The first consists of the large majority of books in which the sharing operation proceeded smoothly and left no significant bibliographical evidence beyond that which demonstrates that a book was shared among two or more printers and permits the identification of one or more of the printers.[1] These two fundamental insights provide a sharper focus for bibliographical analysis in regard to the disposition of copy during production, the linkage of the book to the workmen and procedures of the respective shops and their potential effect upon the textual transmission process, and the temporal component of production. Taken alone, however, the detection of sharing does not in itself reveal anything about the temporal component of production. When supplemented by other forms of evidence accessible in most books, sharing may actually indicate an expansion of production time rather than the time-saving which it seems naturally to imply. The temporal issue usually can be settled with certainty by defining the printing context of the book in the primary printer's

shop. In general, this difficult and time-consuming proposition seems unnecessary in the absence of important textual or dating problems.

George Gascoigne's A Hundreth Sundrie Flowres provides a rather unique example of a second category of shared books, in which the printing process produced bibliographical evidence relevant to unresolved textual problems. In such books, it is essential to establish the facts about the shared printing operation so as to define the temporal component of production before proceeding; otherwise bibliographical analysis must rely upon circumstantial considerations that are external to the production of the book and exhibit no necessary physical connection to the evidence found in the book. Such an approach is bound to overlook, misconstrue, or adjust the bibliographical evidence to fit a hypothesis comprised of selected external circumstances and presumptions about early printing.[2] This is especially true of A Hundreth since the book contains an unfinished text and other textual indications of time-related problems on the part of the author.

A Hundreth is a quarto collating A4 B4(-B1,2) C-X4, 2A-Y, Aa-Hh4, Ii4(—Ii4). Pagination: 208 leaves, pp. [8] 1(=B3) . . . 36(=F4v) . . . 45(=G1) . . . [82-83 misprinted as 83-82] . . . 164(=X4v); 201(=2A) . . . 376(=2Y4v) . . . 377(=Aa1) . . . 443(=Ii2) . . . 445(=Ii3). The significant contents are: A2-3: 'The Printer to the Reader.'; A4: title, 'SVPPOSES'; B3-K1v: text; K2: title, 'IOCASTA'; K3-X4v: text; on X4v: imprint 'Printed by Henrie Bynneman | for Richarde Smith.'; 2A1: title, 'A discourse of the adventures | passed by Master F. I.'; 2A1-M3: text; 2M3v-S4v: 'The deuises of sundrie Gentlemen' [running title 2M3v-4: 'A translation of | sundrie Gentlemen.']; 2S4v-Ee2v: 'The deuises of master Gascoyne' [running title: 'The deuises of | sundrie Gentlemen.']; Ee2v-Hh4v: 'Dan Bartholmew of Bathe.'; Ii1: title [for misplaced narrative link 'The Reporter'], 'This should have bin placed in the | dolorous discourse, before the Supplication | to Care in Folio. 430.'; Ii2, 'The reporters conclusion unfinished.'; Ii3: imprint, 'IMPRINTED AT LOND-| don for Richard Smith.'

Two forms of typographical evidence establish the key facts about the printing of a shared book: (1) the composite of normal and variant letters that characterize a specific font at some point during its lifetime; and (2) the sequence of changes or transformations that occur in a font's composition during an extended period of usage. In this instance, analysis of the fonts

found in the book permits identification of Henry Bynneman and Henry Middleton as the printers and, furthermore, reveals that the book was shared in an asymmetrical pattern created by Bynneman's recall of the shared section from Middleton. Second, although typographical evidence provided by font transformations usually is useful only for locating a shared section in a shop's production schedule, A Hundreth is extraordinary in this respect because the sequence of transformed states of Bynneman-Y1 demonstrates an expansion of production time far beyond that required to print the book even if allowance is made for several intervening book-length interruptions. Given this insight, the bibliographical and textual analysis of the book can proceed according to the certain knowledge of delayed production instead of the standard assumption of rapid printing at maximum efficiency.

This understanding of the printing operation is crucial to resolving the textual problems present in A Hundreth. The extended production period during which the book was printed was adequate for major stages of textual evolution in regard to textual issues including the dates of composition and printing, modifications to the text such as revisions, deletions and omissions, changes in the author's concept of his overall collective text, and problems created by the delivery of printer's copy in segments. The recovery of Gascoigne's intention regarding the number and sequence of texts depends upon settling the issue of Gascoigne's involvement in the printing process. Whether Gascoigne intended to include the two plays in a particular location in the book has been the fundamental problem since C. T. Prouty eliminated the two plays from his edition, attributing their appearance to Bynneman's desire to enhance the book's profitability. The rearrangement of the texts in The Posies (1575), the second edition, indicates that Gascoigne was concerned with producing an overall impression by combining the separate compositions in a particular order and publishing the collective text as a single unit. An editor must determine whether Gascoigne attempted to achieve a similar effect in A Hundreth before resolving several relatively minor textual issues.

This paper is intended to contribute to an awareness of the potential relationships between shared printing and textual issues. My primary focus is upon a reconstruction of the printing context of A Hundreth as the basis for interpreting the bibliographical evidence of textual evolution. I have described in previous papers the procedures to follow in establishing the essential facts about a shared book: (1) the recognition of the preliminary clues to sharing; (2) the analysis of typographical and ornamental evidence that can clarify the divisions of labor and establish the identity of one or more of the printers; (3) a survey of the primary printer's books in order to identify his section(s) of a book; (4) a survey of books printed by likely sharing candidates in the hope of identifying the sharing printer(s) and their sections. Once these facts have been established, analysis can proceed to forms of evidence which reveal the temporal component of production in the sharing situation. The subsequent reconstruction of the printing context of a shared book usually establishes the date of the printing operation, and

in some instances such as A Hundreth, the printing dates of its sections. In the following discussion, textual issues will be addressed as the emerging picture of the printing operation warrants.

I

In general, the analysis of a shared book is complicated by the justifiable inference of time-related production pressure; otherwise sharing would have been unnecessary. Two interrelated circumstances seem to underlie the sharing strategy. First, A Hundreth represented a time-consuming job with its 52 sheets. Such a long book invited sharing, particularly in a healthy business climate when intervening jobs could very likely materialize during printing. A printer would rather not tie up his press(es) for a long time on a single project with jobs appearing regularly. Similarly, preliminary negotiations for future projects could lead a printer to schedule a book with sharing in mind. Second, sharing could be undertaken to satisfy a publisher's desire for rapid printing so as to capitalize upon current but transient public interest aroused by such events as dramatic performances, notorious murders, political events, and the latest developments in the siege of Ostend.[3] In short, the sharing strategy can imply either an expansion of production time through serial printing of sections because of interruptions or a compression through a preplanned concurrent shared printing. The analysis of shared printing proceeds in several stages beginning with the discovery that a book was shared and the definition of the divisions of labor in it, followed by the attempt to identify the printers. Several forms of evidence can be used to locate the book or its sections in the production schedule of the shop(s) of the primary and/or sharing printers. Usually such evidence at least provides a relative temporal reference for the reconstruction of a printing operation.

The detection of sharing is aided by the presence of preliminary "clues" at the boundary between shared sections including: (1) bibliographical anomalies or a change in setting style; (2) a shift in ornamental stock; and (3) a change in fonts. These clues provide an initial insight into the possibility that a book was shared in a particular pattern but require verification by the typographical evidence provided by font analysis. The typographical evidence can reveal further divisions within the sections indicated by the preliminary clues. Several preliminary clues suggest that A Hundreth was shared in at least two clearly defined sections. (1) The full imprint "Printed by Henrie Bynneman | for Richarde Smith." appears in X4v (page 164) at the conclusion of Iocasta and directly preceding "Master F.I.". (2) This imprint is followed by a repeated alphabet. (3) Although the pagination is sequential, it jumps from page 164 (X4v) to 201 (2A1). (4) A second partial imprint "IMPRINTED AT LON-/don for Richard Smith." appears at the end of the book on Ii3.

(5) In addition, ornaments and initials appear in both apparent sections, a fact which calls for a survey of books by the printer identified in the imprint in order to establish which of the pieces he owned and hence his responsibility for the section(s) in which they appear. The five identifiable initials and one tailpiece seen in A-X can be found without difficulty in Bynneman's other books 1572-74. This evidence supports Bynneman as printer of the section. Similarly, two of the three identifiable initials in the second section (up to 2M3v) can be found in proximate books by Henry Middleton and his partner Thomas East.[4] The identical "printer's flowers" at A3 and 2M3 are a possible source of confusion since these cast ornaments were quite common in the period and could be misinterpreted as evidence of the same printer in both sections. The ornamental evidence clearly points to a sharing boundary at the first imprint on X4v but typographical evidence is necessary for final confirmation of the sharing pattern. (6) A second sharing division at 2S-T is suggested by two setting features. First, the recto element of the running-titles in 2M4-S4 is invariably "sundrie Gentlemen." Beginning with 2T1, "sundry Gentlemen." alternates with "sundrie Gentlemen." in the pattern produced by single-skeleton imposition (i.e., "sundry" moves from 2T1-2, "sundrie" from 2T3-4). A preliminary analysis of running-titles working in both directions from the boundary reveals another fundamental difference: at least three skeletons were used in alternating fashion in 2M-S with a standard measure of 86mm (except in 2S1v,2,4v). The measure shifts to 92mm in 2T1 and thereafter. These changes strongly imply a sharing boundary although other explanations are possible.

Analysis of the fonts in A Hundreth confirms the implicit boundary at 2S-T and reveals that the book was shared in three asymmetrical sections with Bynneman printing A-X, 2T-Ii, and Middleton the middle section 2A-S. A rather common problem arises from the fact that both printers used same-face pica black letter text fonts, pica italic emphasis fonts, and Guyot doublepica roman and italic for sub-titles and running-titles. Nonetheless, the evidence provided by the remaining fonts used for emphasis, headings, and quotations is unequivocal in this instance because of their uniqueness, the alternating patterns of usage, and foul-case cross contamination:

The typographical evidence, in short, demonstrates a division of labor in which Bynneman printed two sections of nineteen sheets (B-X) and fourteen sheets (2T-Ii, A), and Middleton eighteen sheets (2A-S).

The sharing divisions revealed by the preliminary analysis may suggest a particular sharing strategy. In general, the division of labor seen in a shared book usually establishes the minimum time required for printing the book as defined by the largest number of sheets done by any one of the printers. Simple sharing patterns (e.g., four sections of two gatherings each, two sections of four gatherings each, etc.) such as are frequently encountered in play quartos prove nothing about production time but seem to suggest concurrent printing in the context of a deliberately planned time-saving sharing strategy.[8] However, asymmetrical sharing patterns may suggest serial rather than concurrent printing of sections in the absence of other plausible explanations for the resulting divisions of copy.[9] Markedly disproportionate sections raise doubts about the time-saving strategy and in fact may signify the primary printer's effort to offset delays caused by an interruption during which valuable production time was lost. This inference seems especially warranted if the primary printer prints the first and a later section in a book, as is the case in A Hundreth. Hence, the identification of the sharing printers as well as their respective sections is quite important in the temporal context.[10] For example, distinguishing the three sections of A Hundreth without the identification of Bynneman's type in the first and third sections would produce a

reasonably symmetrical pattern with a plausible textual rationale for the divisions of copy. The first boundary (X4v) occurs at the conclusion of a self-contained text (Iocasta); the second boundary (2S4v) yields a nearly equal number of sheets (nineteen vs. eighteen) and occurs at a point with no textual significance (i.e., within a text, but not at its end), implying that the boundary was arbitrarily chosen to produce the symmetry. This division produces a shorter third section but the added effort of dividing the book into equal sections would make no sense in terms of the overall time-saving resulting from concurrent printing by three printers and the fact that the need for accurate casting-off invited errors. On the other hand, the identification of Bynneman's type in the third section virtually eliminates concurrent printing at least in regard to his sections. The sharing pattern thus suggests either (1) concurrent printing of B-X and 2A-S followed serially by 2T-Ii, A; or (2) serial printing of the first two sections followed by Bynneman's recall and completion of the job. In short, the pattern produced by the identification of the sharing printers can provide circumstantial evidence of serial rather than concurrent printing and thus establish a minimum production time. However, more reliable forms of evidence are required to confirm the temporal implications of a sharing pattern.

The date of printing and the production time required to print a book are essential components of a printing reconstruction that aims at settling time-related textual issues and relating the printing operation to external circumstances such as events in the author's life, literary influence according to the priority of two texts, the possible role of censorship in modification to a text, the relationship among a sequence of editions, the relation between publication vs. performance, and other similar issues. The sharing situation increases the possibility of finding timing evidence since two (or more) shops are involved. The sharing printer's books may provide the necessary evidence that is lacking in books by the primary printer. Four forms of evidence are useful for establishing the date of printing and the production time by locating a book or section(s) of a shared book in the production schedule of the shop(s) involved in its production. The results of the analysis of A Hundreth for these forms of evidence are summarized in the following paragraphs.

(1) Documentary information such as entries in the Stationers' Register and dates found in prefatory materials can provide a temporal reference point. However, such dates must be verified by independent means.[11] A reconstruction often must remain relative because a concrete date cannot be established for the various points in a printing operation. At best, a sequence of dated entries in the Stationers' Register can provide a fairly certain realtime reference for inferences about the schedule of books printed during the

proximate period. Unfortunately, printers' and publishers' practices varied considerably: some consistently entered their books while others did not. The problem is universal for the early part of the Elizabethan period (roughly pre-1580) before the practice of dating entries was adopted: an indication of the year of entry seems to have sufficed. Books entered during the proximate period to Bynneman, Smith, Middleton, and East follow the early practice for books actually entered, and not all were. The three imprints found in A Hundreth lack the usual annual date, but two specific dates appear in the prefatory materials. "The letter of G. T. to his very friend H.W." is dated "this tenth of August, 1572." followed by "H.W. to the Reader" bearing the date "the xx. of January, 1572[/73]." These dates will be assessed later in the context of the production schedule.

(2) Progressive damage to ornamental stock is useful for locating a book in a shop's production schedule.[12] Beyond that, the recurrence of the same ornament in a sequence of gatherings or in two shared sections by the same printer usually indicates serial rather than concurrent printing because one ornament cannot be on two presses at the same time. In longer books containing many repeat appearances of ornamental stock, concurrent printing thus can be disproved for many of the sheets. Serial printing of all sheets can thus be reliably inferred given an interlocking network of recurrences. A single ornament repeats in A Hundreth but within Bynneman's first section and thus reveals nothing about the printing relation to his second section.

(3) Watermark evidence found in contextual books can frequently establish a book's position in the production schedule. In certain situations, the sequence of job-lots of papers used in books during the proximate period yields watermark evidence which can distinguish between concurrent and serial printing of sections in either shared or unshared books, indicate the point at which copy was divided in shared printing, and provide an indication of delays in production.[13] Unfortunately, the papers in A Hundreth

comprise a homogeneous job-lot consisting of very closely related varieties of a "cuffed-hand-pointing-at-star" from one continental source. Hence the watermark evidence is ambiguous and reveals nothing about the division of copy and whether printing was concurrent or serial. Moreover, the ambiguity extends beyond A Hundreth and obscures the location of the book in the production schedule in Bynneman's shop. This lot of related watermarks appears exclusively in at least thirteen of Bynneman's books from 1572-74 and then in The Posies. To make matters worse, several of Middleton's books also make exclusive use of the lot.[14] A high-resolution examination of one of the watermarks through the sequence of books, however, might reveal progressive states of deterioration of the mould which could be useful for dating.

(4) Three kinds of typographical evidence are useful for distinguishing serial vs. concurrent printing and for dating.

First, the value of an identified recurrent types survey in the temporal context is generally limited to: (1) those shared books containing two sections printed with the same font by one printer; or (2) books printed by one printer in which one of his fonts appears in sections separated by a section

printed in his second (and/or third) font.[15] The movement of a group of identified types through the sequence of gatherings comprising the two shared sections demonstrates serial printing of the sections. The typographical samples of Bynneman's three pica roman fonts in A Hundreth are too small to yield recurrent types.

Second, the method of imposition as revealed by the movement of running-titles can distinguish serial vs. concurrent printing. The single-skeleton method of imposition found throughout Bynneman's two sections unequivocally demonstrates serial printing. On the other hand, the fairly regular alternation of three skeletons imposed according to the single-skeleton method in Middleton's section clearly raises the possibility of concurrent printing on his three presses for most of his section. However, a few irregularities suggest that only two presses were involved.

Third, typographical evidence provided by a sequence of font transformations is quite valuable in establishing the relative position of a book in a shop's production schedule.[16] This form of evidence provides the key to the reconstruction of Bynneman's shop schedule and the dating of the sections of A Hundreth.

II

In certain long books (such as A Hundreth) with early and later sections printed in the same font, a sequence of font transformations can establish the relative positions of the sections in the shop schedule as defined by the respective states of the font. Bynneman-Y1 appears in a fairly large number of books in 1572-73 so that it is possible to plot the progressive contamination of the font over a period of several months. The necessary condition for transformation by fouling is met by Bynneman's simultaneous use of his three pica roman fonts during the proximate period in A Hundreth and other books. Moreover, the fonts frequently alternate within a book, a situation which invites fouling. Additionally, replenishment with new wrong-face letters produces a new state of the font. The fact that Bynneman-Y1 appears in A Hundreth as an emphasis font presents the major methodological problem since the resulting typographical sample is quite small.[17] As a general rule, the size of the typographical sample in a target book is more or less irrelevant as long as the evidence of the transformation state appears. A large sample of a font is needed only for establishing the proportions of the normal vs. foulcase letters and the expected ratio of appearances in textual units such as the page or gathering.[18] Once that probability is established, the necessary size of a sample diminishes to the level at which foul-case letters begin to appear,

regardless of how many are found. Thus the very small font sample in A Hundreth Sundrie Flowres is adequate since the contaminating letters appear. On the other hand, the absence of contaminating letters in a very small sample should not be taken as proof that the font is in the early state.

A font transformation in itself proves nothing about the temporal component of production unless it can be located in the shop's schedule. A stage of fouling can happen almost instantaneously in relative terms: the compositor requires only a few reaches to the wrong-font case to effect the transformation. Hence, no assumptions can be made about the amount of time required for a transformation. As will be seen below, radical transformations to a font can occur during the setting of a single page. The discovery of a transformation in progress in a given book establishes a definite temporal reference point in the shop's schedule but presents a problem of interpretation regarding the amount of time that elapsed before the appearance of the newly transformed font in the next book. In the absence of contrary evidence, it seems reasonable to assume that the printing of a short book in which a transformation occurred proceeded uninterrupted to its end. However, it is possible that a later, short book could have interrupted the printing of a long book with a large number of sheets following the transformation sheet.

Since transformations occur in stages, the process can produce clear evidence of the sequence of groups of books as defined by the state of the font that appears in each group. However, the transformation evidence cannot distinguish the sequence in which the books within a group were printed. Given a group of four books, then, the target book (or section of a shared book) may be separated from the transformation book by the total number of sheets in the two other books. It also could have been printed directly after the transformation book. In short, the evidence provides only a prior terminus for the printing of a book in the shop's schedule. In some cases, extraordinary circumstances may strongly suggest that the transformation book was completed without interruption.

In general, changes in a font's overall composition during its lifetime provide the evidence for distinguishing transient and permanent transformations and the sequence of states. The following analysis of the transformations of Bynneman-Y1 defines the printing context of A Hundreth in terms of the sequence of the groups of books printed during 1573. The evidence reveals an additional unsuspected division within Bynneman's second section at Cc-Dd.

Pre-1573. Bynneman-Y1. Bynneman-Y1 first(?) appeared in A dictionarie STC6832 in 1570 lacking both the 'W w' and 'fh' ligature which forced alternate settings of 'VV vv' and 'short-s h'. A few S-face 'k1' fouled the font in 1570-71 but were purged. A transient transformation in the font occurred in the setting of 'w' during 1572. Y-face 'vv' was invariably set during 1570-72 except in four books in which the compositor(s) alternated the S-face crimped 'w4' with Y-face 'vv' (T. shippe STC5952; A Speciall STC11759; Eclogues STC22991; and An Answer Q1 STC25427, B-I). Such a restriction of contamination to a single sort or two permits distinguishing this state (1572) of

Y1 from the later Y1c state (1573) where the 'w4' reappears in the context of S-face 'vv' and foul-case letters in at least eleven other sorts. The 'w4' was almost completely purged after I3 of An Answer Q1 STC25427 where the permanent transformation to Y1a occurred with the addition of S-face 'g1' while the 'w4' was still being set. A few 'w4' remained resident to produce a total of nine appearances in K-Y where the font is in the Y1a state. In addition, two appearances (D, I) of the F-hybrid short second-stem 'w' occur (from an oldface 86mm roman last used in 1570). Two remaining books bearing a 1572 imprint and containing adequate Y1a samples set exclusive Y-face 'vv' and 'g1' (A Paraphrase STC19137.5, 27 sheets; Poeta STC19139, 27 sheets).[19]

Group I. STC11635a (B-X). 21 sheets. Bynneman-Y1a appears in two books from 1573. Certaine Brief STC19060 (one sheet) seems to have been printed at this time. The setting of Y-face 'vv' is invariable in A Hundreth STC11635a (B3-X4v, twenty sheets containing Supposes and Iocasta). Only two S-face 'g1' (F3v, L3v) and one 'fi' ligature (F3v) appear in Y1a stage directions. Despite this limited sample, it is clear that STC11635a was printed after the introduction of the S-face 'fi' ligature in 2A of An Answer Q1 STC25427. Typographically speaking, STC11635a belongs with I-2A, * 2-3v of STC19137.5 and 2E, A of STC19139 which were probably printed contemporaneously with STC25427.

Group II. Bynneman-Y1b. 78 sheets. The transformed Y1b state is found in Group II consisting of five books totalling 78 sheets (Art of Reason STC 15541, Garden STC12464, De Furoribus STC13846, Historia STC20309 and approximately A-F of An Admonition Q3 STC25429) and is characterised by mixed S- and Y-face 'vv' and 'v' settings along with an expanded foul-case cluster. The transformation into Y1b actually occurred in 2A-P, A, bc of Q1 STC25427 which was completed either prior to or immediately after Christmas vacation. The S-face 'vv' was introduced at low-density levels late (2I3v, 2L3, 2M4v, 1A2) in STC25427; the 'c f k1' and 'ct f1 fl fi fl' ligatures also entered the font at this time. The non-appearance of the 'vv' and other letters in STC11635a, 19137.5, and 19139 is consistent with their low-density in STC25427. The 'f' and 'f1 fl ct' ligatures, for example, skip one or all of the

full-prose settings in 2M4v-2N1v while the 'g1' and 'fi' ligature appear in all (and in STC11635a). The expanded cluster of Y1b begins appearing late in STC15541 and STC12464, where the S-face 'd1' first appears. The limited use of Y1b in the Latin 20309 prevents S-face 'vv' settings until the Index (X-2C) and the 'f1' ligature appears by Q1; the S-face 'n3' is added in X2 and appears along with the resident 'd1 g1 vv' and 'ct f1 fl fi' ligatures. STC 20309 was probably the final book in Group II and overlapped Q3 STC 25429 A-F.

Group III. Bynneman-Y1c, STC11635b (2T-Cc). 127 sheets. Group III consists of seven books totalling 127 sheets (An Answer Q3 STC25429, G-2X; Certaine Godlie STC25010; A Christian Instruction STC24788; Margarita STC23003; A Hundreth STC11365b, 2T-Cc; An Exposition STC24171; The Supremacie STC3737, A-E2v) distinguished by the reappearance of S-face 'w4' in the context of S-face 'vv' and the expanded foul-case cluster of Y1b. STC25429 is the first book in 1573 in which Bynneman-S1 with the 'w4' alternates with Bynneman-Y1b (STC11635a required no 'w4' in the S1 settings). The extensive use of the two fonts within type-pages resulted in low-density 'w4' contamination. For example, an S1 quotation in K2:1-15 is followed by a Y1c quotation in K2:24-36 where the 'w4' appears at lines 32 and 36 and K2v:5. Moreover, the alternation increases the S-face 'vv' to new high-density levels (e.g., S-face/Y-face: 4/6, 10/7, 3/6, 9/29, 17/3, 15/12, 16/7, 19/13 etc.). In addition, a printing of STC25429 in proximity to STC25427 is suggested by the occurrence of the S-face 'k1' (STC25427: 2A3v, 2L2v; STC25429: F4, H2, M3) and F-hybrid short second-stem 'w' (STC25427, D4, I1v; STC25429, G4-4v, K2, 2E3v) only in these two books. The typographical evidence indicates that A Hundreth STC11635b (2T-Cc) was printed after STC25429. The Y1c sample in 2T-Cc amounts to about 48 lines ("Gascoigne" appears in about half of these). The ratio of S-face 'vv' to 'w4' to Y-face 'vv' settings (18/4/22) requires that the contamination have occurred in a previous book since Bynneman-S1 was not used after STC11635a B-X. Bynneman-S1 appears along with Y1c in two books other than STC25429 in Group III but neither presented an opportunity for contamination. Y1c is limited to T2:2-18 of Certaine Godly STC25010. In A Christian STC24778, S1 is used exclusively for emphasis to Q4v followed by a Y1c stint to 2A where the S-face 'vv' and 'w4' are already at high-density levels. In short, the setting of An Answer Q3 STC25429 produced the contamination of Bynneman-Y1c as it appears in the remaining books of Group III.

The introduction of S-face 'P' clearly is attributable to compositorial inconsistency in setting emphasis names in STC11635b; for example, Y1c serves this function in (repeat alphabet assumed) T2, T4, U1, U2, and U3 while the compositor turns to Bynneman-S1 in T2v, T3v, and T4v, and then sets the S1 "C[ae]sars" and "Greekes" separated by a Y1c "Menelaus" in U2v. A similar mixing occurs in Aa1-2v: the compositor set three Y-face and two S-face 'P' in Aa2 and six Y-face and one S-face 'P' in Aa2v; the latter appears in "Pencoyde" where the Y-face lower-case is used. Some of these S-face 'P' were mis-distributed into the Y1c case to reappear in STC23003 and STC

22241. An Exposition STC24171 clearly was printed simultaneously with T. Supremacie STC3737 A-E2v. The Y1c cluster appears in emphasis settings and the terminal imprint; the title imprint, however, exhibits two of the newly introduced Guyot short second-stem 'w'.

Group IV. Bynneman-Y1d, STC11635c,d (Dd-Hh; Ii, A). 173 sheets. The transformation into Bynneman-Y1d occurred in the massive T. Supremacie STC3737 (153 sheets) where Bynneman-Y1c,d, -S1, and the Guyot 76mm roman alternate as emphasis and quotation fonts. Bynneman-Y1c sets S-face 'w4' and the mixed Y1c and S1 'vv' to E2v. The newly acquired Guyot short second-stem 'w' begins replacing the 'vv' in E3 and is set almost exclusively thereafter including numerous quotation settings of up to 1½ pages in length. Likewise, the newly acquired Guyot 'fh' ligature replaces the 'short-s h' settings; low-density S-face 'x2' appears as well. The importance of discovering a font transformation in progress cannot be overstated. In addition to establishing a firm point of reference in the shop schedule, the discovery provides the correct context for interpreting this kind of typographical evidence. The fact that the new Guyot 'w' appears in A Hundreth STC11635d in two Y1d poem titles in Dd1:8-10, 19-21 (seven Guyot 'w', no 'vv') followed by a Guyot 76mm roman editorial link (Dd1:28,29) could suggest that the compositor fouled the Y1d poem titles from the Guyot case while setting the page. The observed transformation in STC3737 proves that the Guyot short second-stem 'w' was already in the font in state Y1d. Therefore A Hundreth 2Dd-Ii, A, A Briefe STC11985, and T. Closet STC7623 were printed after the transformation to Bynneman-Y1d in T. Supremacie STC3737. The final appearance as Bynneman-Y1e in T. First Part STC22241 exhibits the increasing density of the cluster as well as additional S-face letters as is typical of the final depletion of a font. It does not appear to have been used after this book.

III

In the absence of dates of entry in the Stationers' Register for Bynneman's books from 1573, the reconstruction of the printing context of A Hundreth and the dating of its sections must rely upon the temporal reference points established by the sequence of transformations in Bynneman-Y1 and any additional timing evidence which illuminates the production schedule. Several variables can affect the accuracy of such a reconstruction. In this instance, the survey included only those books listed in the outdated P. G. Morrison's Index of Printers and Publishers for Bynneman, Middleton, and East: the new index (Volume 3 of the revised STC) may reveal additional books which could shift the temporal reference points somewhat. Second, some of Bynneman's output may be hidden in undetected shared sections in books by printers other than Middleton and East. However, given the identified output from Bynneman's shop, such a shift would be negligible. Third, the method of defining the temporal reference points is of necessity based upon an average weekly production rate calculated from the annual production rate. This method cannot account for the possibility of radical fluctuations in production rates from week to week due to various mechanical factors such

as equipment failures and the difficulties presented by certain kinds of texts. For example, learned Latin texts probably required more time than vernacular ephemera; the outstanding quality of books such as De furoribus STC13846 and Historia STC20309 attest to this likelihood. Overall, the reconstructed schedule for 1573 seems realistic and, furthermore, the reference points for the sections of A Hundreth could vary considerably without affecting our understanding of the textual problems.

The identified total annual production of 466 sheets in 1573 is comprised of the 434 sheets in books containing Bynneman-Y1 and an additional 32 sheets in four other books. An average weekly production rate of 9.3 sheets per week is calculated on the basis of fifty working weeks per year (one week holidays at Christmas and Easter). This weekly average yields an average of 1.55 sheets per day (approximately 300 working days per year), a figure which reasonably approximates the output to be expected from a three-press shop over the long term.[20] The following assignment of specific reference points is necessary for the sake of argument and overlooks the reality of weekly variations in production and the possibility that part or all of Q2 STC25428 was printed in January 1573. The thirty-two sheets in books not containing Bynneman-Y1 have arbitrarily been placed at year's end: thus the reference points could be early by as much as 3.4 weeks.

Group I. Bynneman-Y1a, STC11635a. The printing of the twenty-one sheets of Group I, including A Hundreth STC11635a, required a minimum of two weeks. Two factors suggest that Group I extended into early February. Bibliographical evidence suggests that the printing of A Hundreth STC 11635a (B-X) was interrupted at least once, and probably twice, resulting in minor delays which are negligible in regard to overall production time and the issue of textual evolution. The fact that the pagination error between F4v (page 36) and G1 (page 45) dropped eight page numbers (37-44), or one gathering, suggests a calculation error of the kind that could be expected following the interruption by an intervening job. A compositorial misimposition of signatures and/or page numbers from the last-distributed forme presumes a serial distribution followed by imposition and is easily detected because the error usually affects only the sequence of signatures and/or page numbers.[21] However, the error at pages 36-45 cannot have occurred

through such a mis-imposition of numbers salvaged from either previous forme. In fact, this process was more likely to eliminate the error since a goodly supply of "3" and possibly a "9" (from "29" in the outer forme) would have been on the stone, requiring only a new "7 8" to complete the decade. A return to A Hundreth after an intervening job produces a different situation. Foremost is the fact that all type would have been distributed, thereby eliminating any erroneous transfer of page numbers. Second, the perfected sheet F would have been removed to the warehouse. Third, copy would have been marked to indicate that sheet G was the next to be set. However, the compositor's mark consisted only of the typical elongated bracket or slash followed by a marginal notation "G" or "G1".[22] The marking of page numbers was irrelevant in copy set seriatim for single-skeleton imposition; the error proves that no page number was marked. Given standard shop practice for a manuscript containing a title and preliminaries, the compositor would assume that the play-text began on A3r which would be assigned page number 1, resulting in a four-page negative offset for the calculation. F would then be the sixth sheet running to page 48 less the offset of four pages, or page 44. Sheet G thus began with page number 45. In short, the pagination error occurred because the compositor assumed that the text began in sheet A rather than B.

The evidence of an interruption between sheets E and F is less certain. First, the furniture of the skeleton surrounding the running-titles underwent some modification with the result that the relationship between the running-titles of the inner and outer formes shifted laterally about five letters from the previously accurate registration. Second, Bynneman-S1 was fouled extensively in the speech prefixes of B-E with Bynneman-Y1a 'E P' and Guyot 'C D'. The frequency of these foul-case letters drops radically in sheet F and after (for example, all fourteen 'D' in sheet G are correct-face). The Y1a and Guyot capitals seem to have been purged at a single time following sheet E rather than during the distribution of each previous fouled forme; if so, a slight interruption probably resulted, given the difficulty of the task of purging wrong-face roman capitals. The single-skeleton method of imposition in use in B-X points to a single compositor who could not simultaneously attend to composition and purging: an interruption was thus a necessary condition for the latter.[23]

Group II. Bynneman-Y1b (running total: 99 sheets). The seventy-eight sheets of Group II required about 8.4 weeks for printing prior to the transformation to Y1c in STC25429. The two learned Latin texts (Historia STC

20309, 28 sheets; De Furoribus STC13845-46, 14 sheets) printed in this group probably extended printing time beyond the fourth week of March and into April.

Group III. Bynneman-Y1c, STC11635b (running total: 226 sheets). A combination of factors suggests that STC11635b was printed in late May at the earliest. The regular alternation of three skeletons imposed according to the single skeleton method indicates that the printing of STC25429 G-2X proceeded without interruption into the second or third week of May, given 4.6 weeks for its forty-three sheets plus Easter week vacation. The typographical evidence noted above is complemented by the circumstances of publication. The royal furor caused by the surreptitious publication of An Admonition to Parliament Q1-2 STC10847-48 in June-August 1572 and A Second Admonition STC4713 in mid-November finally culminated in the Queen's Proclamation of 11 June 1573 after an unsuccessful year-long attempt to ferret out the secret Puritan press and suppress the books. An early reprinting of An Answer is understandable given the market situation created by the high-profile Admonitionist controversy.[24] Bynneman simply would not have interrupted the printing of Q3 STC25429 G-2X to continue work on A Hundreth and, in fact, gave STC11635 2A-S to Middleton in order to avoid an interruption of some book. The sharing of A Hundreth could also have been occasioned by two other longer books of Group III (STC25010, 21 sheets; STC24788, 33 sheets). Overall, the printing of Group III extended about 25.3 weeks in 1573, or through the second week of July.

Group IV. Bynneman-Y1d, STC11635c,d. The printing of the 173 sheets of Group IV following the transformation to Bynneman-Y1d in T. Supremacie STC3737 extended from mid-July through the first week of November. No evidence is present to indicate when the remaining seven sheets of A Hundreth (Dd-Ii, A) were printed during this period, but they represented

a minimal interruption in the printing of T. Supremacie STC3737 and probably were worked off when copy arrived, given the knowledge that early printers frequently worked on more than one book during a period in order to maintain an overall work flow. Long books such as STC3737 were of necessity susceptible to interruptions by short jobs of several sheets (such as Dd-Ii, A) which would be treated as single units of production. STC3737 very probably began printing before copy for STC11635c,d arrived; common sense suggests that Bynneman would have simply printed off the remainder of A Hundreth rather than beginning T. Supremacie if remaining copy for the former was in hand.

IV

The reconstruction of the production schedule demonstrates that the printing of A Hundreth extended through at least eight months with long delays between Bynneman's three sections (B-X, 2T-Cc, 2Dd-Ii, A) as distinguished by the transformations in Bynneman-Y1 and the times required for printing the intervening books. In general, delays raise an obvious question about the temporal relationship between the delivery of printer's copy and the printing of the sections of a book. Although presswork ultimately defined the rate of production, the delivery of copy was an overriding factor since printing could only proceed with copy in hand. Hence, no problem occurred if all copy was delivered at the commencement of printing or if parts of copy arrived before copy was exhausted. A delay in printing thus may indicate belated delivery of copy in segments. A second factor must also be recognised. Even if completed segments of printer's copy for A Hundreth were in hand from the start, the assortment of vain ephemera represented by the texts of A Hundreth could easily have been set aside, if occasion demanded, to make way for high-priority theological, political, historical, and learned texts such as were printed during the eight-month period.

Three temporal components related to printer's copy must be defined in the context of the sharing situation: (1) when the division of copy occurred in the primary printer's production schedule; (2) whether the shared sections were printed serially or concurrently; and (3), whether the resort to sharing points to a preplanned time-saving strategy. The transformation evidence absolutely eliminates the possibility that Bynneman's three sections were printed concurrently. However, since concurrent shared printing obviously requires that completed copy be in-hand for casting-off, division, and transportation, the printing location of Middleton's section 2A-S is the critical issue in the overall relationship between printer's copy and the evolution of Gascoigne's sequence of texts. If Bynneman actually divided copy after sheet B, producing nearly equal sections of nineteen and eighteen sheets and the distinct possibility of concurrent printing from completed printer's copy of the two plays, "F.J.", and "The Devises", these texts had to have been composed, revised, and fair-copied by early February, and nearly so by the fictitious date "xx January 1572[/73]." Since typographical and watermark evidence for dating Middleton's section is lacking, it is necessary to generate

a circumstantial case by interpreting relevant bibliographical evidence in the context of the sharing situation represented by the production of Groups I-III. In general, a section in a shared book must be analysed (1) as a unit of production in itself, and (2) as a sub-unit of the shared book as a whole in an attempt to identify the probable factors which led to the sharing strategy.

Several considerations eliminate any likelihood that Middleton's section 2A-S was printed during Group I along with STC11635a. First, the absence of internal bibliographical and textual anomalies reveals that Bynneman's first section, containing the texts of Supposes and Iocasta, was treated as a single production unit printed from manuscript fair copies. Both texts were in a final, stable state since no significant revisions or modifications (except for a very few changes in words) are found in the second edition. Furthermore, the extant manuscript copy of Iocasta agrees almost exactly with the text found in A Hundreth and pre-dates the printing of the book.[25] Gascoigne inserted the date "1566" and the identification "of Grayes Inn Esquire" into the titles of the texts in The Posies, details which correspond to known facts about this period of his life. Second, the interruption of printing at F-G (and possibly E-F) bears no relationship either to internal textual divisions or to a probable sharing situation. Moreover, the transition between the two texts at K1v-K2 is perfectly normal: printer's copy presented no difficulty to the compositor. Third, running-title analysis reveals that the single-skeleton method of imposition was employed throughout Bynneman's sections of the book. Given his three presses, Bynneman's choice of this slowest method of production is definite evidence that he saw no need for haste in printing the twenty-sheet unit either before or after the interruption at F-G. In short, the printing of B-X proceeded normally despite the interruption. Hence it is impossible to argue for the sharing of A Hundreth 2A-S during the printing of B-X.

In contrast, Bynneman's employment of the sharing strategy is very plausible in terms of the production circumstances of Group III. The sharing strategy was primarily motivated by the delivery of copy rather than the usual time-related factors. First, the division of copy into Middleton's section 2A-S and Bynneman's section 2T-Cc indicates serial rather than concurrent printing of the sections. The conclusion of the "F.J." narrative in 2M3 presented a perfectly logical textual boundary for the division of copy into nearly equal segments for concurrent shared printing: the maximum time-saving would result from sections of twelve and thirteen sheets (2A-M, N-Cc). Hence, the sharing was not based upon a preplanned, time-saving strategy since the actual division produced sections of eighteen and seven sheets. Second, the possibility that the sharing was intended to offset lost production time can also be rejected. At least two months passed between the printing of STC11635a and the transformation into Bynneman-Y1c. The sharing of 2A-S could in no way compensate for this amount of time. Nor is it reasonable to assume that Bynneman would tolerate a two-month delay before

deciding to share a book in order to avoid a further delay of about eighteen days if he actually had copy in hand all the while. Furthermore, it is unlikely that Middleton was unable to take on the shared section during this two-month period given the output from his shop in 1573. Similarly, Bynneman printed only seven sheets (2T-Cc) before once again interrupting the job for as much as two more months. Since only seven sheets remained to complete the text (2Dd-Ii, A), it stands to reason that, having already resorted to sharing, Bynneman would have passed the remaining copy to Middleton if he had it in hand. Third, the virtual certainty that the printing of An Answer Q3 STC25429 took precedence over A Hundreth corresponds to the time-related pressure of the classic sharing situation.

In short, the evidence indicates clearly that Middleton's section was printed during Group III along with STC11635b. The delays in printing resulted from the staggered delivery of printer's copy in segments in late January and mid-May. Taken alone, this fact very strongly implies that the composition of "F.J." and "The devises" occupied Gascoigne during most of the intervening period since the preparation of printer's copy required a week or two at the most. The probability that composition delayed the delivery of copy and printing is further suggested by evidence that these two texts were delivered in separate manuscript segments. In general, setting from such copy could cause compositorial confusion about the appropriate setting style and thus produce anomalies at the junction of contiguous sections of copy. The potential significance of graphic layout and typographical conventions as evidence of compositorial interpretation of copy has not received much attention although bibliographers are familiar with these matters. The setting anomalies at the junction of manuscripts containing "F.J." and "The devises" are identifiable in the context of the shop practice of employing (1) a hierarchical titling convention to graphically indicate the relationships among sub-texts in a book, and (2) a particular setting format to mark the end of such a textual unit.

Graphic conventions vary according to format.[26] In quarto settings, the main titles of independent texts such as Supposes, Iocasta, and A Hundreth are set in 7mm roman capitals and/or lower-case, with additional lines of title text set in smaller sizes of type in descending order, ending usually with the pica roman in the imprint. Preliminaries (e.g., "The Printer to the Reader"), components (e.g., "The Prologue or argument", "Epilogue") and sections of the text are headed by sub-titles in either roman or italic (compositorial preference) double-pica (5mm capitals). The running-titles are usually set in double-pica but sometimes in pica; however, the same size of type is used consistently in the running-titles throughout a book.

Two significant titling anomalies occur in A Hundreth 2A-Ii which indicate

junctions of segments of manuscript copy. According to "The Contents," A Hundreth contains four textual sections linked together by G. T.'s editorial frame, which presented a titling difficulty: once the text began with the double-pica sub-title "A discourse of the aduentures | passed by Master F. I." in 2A1, the continuous flow of G. T.'s editorial commentary between sections eliminated the opportunity to insert double-pica sub-titles for the remaining sections: "The devises of sundry Gentlemen," "certayne excellent devises of Master Gascoyne," and "Dan Bartholmew." Therefore the shift in textual sections could only be indicated by a shift to the appropriate new running-title. This occurred except for "devises of Master Gascoyne" and "The Reporter" (see later discussion). As a basis of comparison for the following analysis, the normal setting format for transitions between sections of the text in a continuous manuscript linked by G. T.'s editorial frame can be seen in Ee2v ("devises of Master Gascoyne" to "Dan Bartholmew").

G. T.'s link between "F.J." and "The devises" unequivocally implies continuous copy with a transition directly to the Ariosto translation: "I will begin with this translation as followeth." However, the compositor's end-of-text setting of this link is extremely improbable for continuous unbroken manuscript copy which included "The devises" as well as "F.J." The link is about evenly divided between 2M2v-3 (292-293) and if set at full measure, the carry-over text would occupy a third of the page. After the first ten lines on M3, the compositor narrowed the measure with each successive line and centered the lines so that the text converged in a "V" to a final line 20mm in width (see Prouty's setting); the "G.T." signature occurs at mid-page. He then vertically centered a rectangular setting of printer's flowers measuring approximately 26mm x 52mm between the "G.T." and the signature and catchword. In terms of standard practice for prose texts in Middleton's shop (and Bynneman's as well), this setting format is reserved for the end of an independent text or a sub-text with no textual relationship to the subsequent text.[27]

It is unlikely that the compositor was responsible for the insertion of the flowers at this point. Such layout details probably were marked in copy by the printer. In one rare manuscript, the printer Richard Field marked the copy of Sir John Harrington's Orlando Furioso to instruct the compositor to

insert an ornament at the transition between sub-texts but failed to mark a page-break. The compositor's treatment of the transition is illuminating in the present context. Harrington's instruction reads: "I would have [yow?] immedyately in the next page after the fynyshinge of this last booke, with some prety knotte. to set down the tytle, and a peece of the Allegory as followeth in this next page." In other words, Field should graphically indicate the end of "this last booke" by leaving the rest of the page blank (as he had been instructed and had complied at an earlier textual division), beginning the "Allegory" on the next page with an ornament over the sub-title, followed by text for the remainder of the new page. It seems clear that setting from the continuous copy influenced the compositor to duplicate with a continuous setting of "this laste booke," the "prety knotte.", the sub-title, and text on the same page.[28] Similarly, Middleton's compositor duplicated his physical copy by using the end-of-text format and shifting to a new page. His misinterpretation of the transition clearly points to setting from a new manuscript headed not by a sectional sub-title, but by the title of the first poem. Rather than adopting one of the three formats used thereafter in setting titles of poems, he set the first poem title as a section sub-title.[29] The first line of "A translation of | Ariosto allegorized" appears in double-pica italic on the running-title line of 2M3v (the second line in pica italic) along with page number "294" in double-pica roman. The choice of double-pica italic was dictated by the practice which distinguished a sub-title from a running-title: for example, the sub-title "Epilogue" is italicised in X4 but appears in roman as the running-title in X4v. Further, the running-title appropriately shifts to "sundrie Gentlemen." in facing recto 2M4 with the new first element "The deuises of" in verso 2M4v. (Interestingly enough, "A translation of | sundrie Gentlemen." would probably make sufficient sense to a compositor's or corrector's eye to let pass.) He then began the poem with a small 10mm 'W' used frequently by Middleton to head sub-sections of texts. It is essential to recognise that the compositor had both manuscripts in hand and set seriatim at this point. Having solved the transition problem, he then set an entirely unique double-pica italic signature "M.iii" and catchword "swash-A" from the case which supplied the sub-title at the top of 2M3v. This is the only possible explanation since the case had not been used since the beginning of "F.J." in 2A1.

In short, the transitional link's implication of continuous manuscript copy is simply unacceptable in view of the end-of-text setting anomaly and

the triply anomalous treatment of the beginning of "The deuises" (the double-pica sub-title itself, the use of the title of the poem as the sub-title, and the insertion of the ornamented initial). Overall, it seems likely that Middleton was given the two manuscripts without any explanation of their textual relationship. He and the compositor were left to their own ingenuity in resolving the puzzle presented at the junction of the manuscripts.

Bynneman's compositor responded similarly at the junction between the manuscripts containing "Dan Bartholmew" and the correction notice for insertion of the misplaced editorial link. Two setting features at the end of this text in Hh4v clearly imply an end of the "Dan" manuscript: a rectangular tail-piece (9mm x 43mm) fills the empty space (lines 26-35) and no catchword appears. The absence of a catchword for Ii1 indicates that Hh was printed before the final manuscript arrived: otherwise the catchword would have been included as a matter of course. The considerable amount of space remaining in Hh4v could easily have accommodated the correction notice set in pica roman (as it should have been) and several lines of the first stanza of the misplaced editorial link. That Bynneman expected more material is clear from the absence of an end-of-book imprint. Moreover, the manuscript sheet headed by the correction notice was interpreted by the compositor as a new text and not as a sub-section of "Dan Bartholmew". First, the running-title changes to "The Reporter.", corresponding to the entry "Last of all the reporter." (without page reference) in "The contents". Had the compositor (or corrector) checked against the printed sheet where the correction notice locates the misplaced editorial link, the correct running-title "Dan Bartholmew | of Bathe." would have been used. Second, the correction notice itself rather than "The Reporter." is set in Ii1r as a sub-title. "This should have bin placed in the" appears in double-pica italic on the running-title line (with page "441"), followed by a line in 96mm roman and then a line in pica roman. The actual sub-title of the misplaced narrative link ("The Reporter.") is set normally in the Guyot pica roman and probably was the source of the running title in Ii1v, Ii2v-3. "The reporters conclusion unfinished." likewise is set in double-pica italic on the running-title line of Ii2 and the correct catchword appears in Ii1v, indicating that copy was in-hand. (The absence of a page reference in "The contents" is perhaps attributable to the fact that no page number appears alongside the "conclusion unfinished.": whoever supplied page references probably searched for a doublepica sub-title corresponding to that in "The contents", so the page number 441 in Ii1 alongside the correction notice was irrelevant.) The setting of the two headings strongly suggests separate manuscript sheets rather than a single (probably) bifolium containing both texts. The final link is set in 96mm roman (a concluding flourish) in the "V" format, followed by the "FINIS" and the end-of-book imprint.

Given the fact that two compositors in different shops mis-interpreted the beginning of separate manuscript segments in the same manner, it seems likely that Gascoigne's style of heading a new manuscript or new sheet was

the common source of compositorial confusion. A variety of factors suggests possible manuscript divisions at four other points.

(1) The delay in printing evidenced by the shift from Bynneman-Y1c in Cc4v to Y1d in Dd1r implies segmented copy although no unquestionable setting anomalies occur. The text in Cc4v ends with the first line of the third stanza of poem No. 72; copy for the remainder of Nos. 72-73 that was carried to Dd1r was in hand so that the correct catchword ("If") could be set. However, the transformation evidence categorically excludes the possibility that the letterpress for Dd1r was set and left standing at the time Cc was printed. A minor anomaly occurs. The compositor set the final "poem" (a single line, not given a number by Prouty) following No. 73, "Quoniam etiam humiliatos; am[oe]na | delectant", in 96mm roman type. The turn-under was unnecessary and "delectant" is atypically off-centered to the left. The significance of this setting must remain a matter of opinion. The standard practice of setting foreign language quotations in emphasis type called for a shift from black letter, and given the pica roman links before and after the "poem," the only logical option was pica italic. If, however, the Latin quotation ended a manuscript, a compositor might be tempted to shift to the larger font size (as in the concluding editorial link). Further, it seems likely that the Latin verse marks the end of this compositor's stint. The titles for Nos. 72-73 are set in Bynneman-Y1d while the title of "Gascoignes voyage into Hollande" initiates the use of the Guyot 76mm pica in A Hundreth; it is used in the rest of the book for emphasis, poem titles, and editorial links except for a Y1d stint at Ff2v-Hh2v. Overall, the botched setting and the shift in poem title fonts suggest that the first compositor finished the copy carried over from Cc4v, but a second compositor was given the new manuscript and set to Ff2v.

(2) The absence of anomalies at the transition to "Dan Bartholmew" (Ee2v:8) suggests continuous copy extending from "Gascoynes voyage" (Dd1) to the end of "Dan" in Hh4v. The 76mm Guyot transitional narrative link in Ee2v is set normally with diminishing line lengths, and the running-title correctly shifts to "Dan Bartholmew | of Bathe.".

(3) Several factors raise suspicions of a manuscript junction at 2S-T but the evidence is ambiguous. First, the editorial source of the text and location for new running-titles is unclear. "The contents" indicates that a new subtext "certayne deuises of master Gascoyne" begins in S4v (repeat alphabet assumed) but the running-title is unchanged, the only such anomaly in the book. The transitional link begins "I will now deliuer vnto you so many more of Master Gascoignes Poems," an obvious source of a new recto element for T1. However, Middleton's compositor set the transitional link and the first twelve lines of "Gascoigns Anatomie" before Bynneman's compositor continued setting on T1. The oversight seems more probable if two separate manuscripts passed at different times between Bynneman and Middleton. Conversely, it seems more likely that Smith or Bynneman would have indicated the new recto element if copy was continuous, provided that the "so

many more of" in the link was not confusing. Second, the movement of running titles into sheet S reveals that the moderate irregularity obtaining in M-R degenerated into chaos. It is clear that outer-S was the last forme to be imposed, and somewhat later than would be expected. However, the need to adapt or construct a skeleton to accommodate the very long lines in the texts of S1v-2,4v is as plausible an explanation for the apparent slow-down in presswork as the late delivery of copy for "the deuises of Master Gascoyne." Finally, the fact that Bynneman recalled copy at this point may be a mere coincidence. All things considered, the evidence leans toward inferring that the slow-down in imposing sheet S after the quite regular machining of A-M and the general regularity of M-R indicates the compositor's recognition that he was approaching the end of in-hand copy which would have lacked material for about half of the final page S4v.[30]

(4) Finally, a major textual omission occurs in 2Y3 (373). The sub-title of "Gascoignes De profundis" (No. 65) appears at the bottom of 2Y2v (not anomalous, see 2O2v, P2v, R1v) along with the catchword for the editorial link in 2Y3 instead of the psalm: the sonnet which introduces the psalm follows the link. The link begins "The occasion of the vvrighting hereof . . ." in the editor's standard reflexive mode of commenting upon the previous composition. Compositorial omission can be ruled out: the setting sequence of prose editorial link, poem title, then another prose editorial link is an inherent contradiction to a compositor who has been interpreting copy for shifts from 67mm roman in the links to pica roman poem titles to black letter poem texts. Moreover, 2Y2v-3 would have been set seriatim regardless of the overall method of setting. The setting format for the links is normal as well. In short, the only reasonable conclusion is that the psalm was not in the manuscript copy and, moreover, the omission is not attributable to a manuscript junction. Prouty's collational data ("Critical Notes," pp. 277-290) reveal a dramatic rise at this point in errors later corrected in The Posies that clearly points to a deterioration in the quality of printer's copy. According to the chronological sequence of the poems detected by Prouty, a chronological division occurs here: Nos. 66-67 can be assigned to early 1572 prior to Gascoigne's first departure to Holland in May 1572. It seems likely that the omission and subsequent errors are attributable to Gascoigne's attempt to compose the links and to revise the recent poems during the faircopying process.

In summary, the bibliographical evidence reveals that copy was delivered to Bynneman in at least six separate manuscript segments: (1) Supposes (B3-K1v), (2) Iocasta (K2-X4v), (3) "The adventures of Master F.I." (2A-M3), (4) "The devises" (2M3v-Dd1r) (5) "Gascoignes voyage" and "Dan Bartholmew"

(2Dd1-Hh4v), (6) and the misplaced narrative link (2Ii1-1v); two additional sub-divisions are possible: (7) the "devises of Master Gascoyne" (2S4v-Cc4v), and (8) "The reporters conclusion unfinished" (2Ii2-3). The fact that the manuscript junctions, as indicated by the bibliographical evidence, correspond exactly to textual divisions cannot be attributed to mere coincidence. When placed in the context of the long delays between sections and the minimum production time of eight months, the segmentation of printer's copy clearly indicates that the composition of the editorial frame and the "F.J." narrative, and the organization of previously written materials into the editorial frame, occurred during the January-August period after Smith undertook the publication project. Thus the date affixed to G. T.'s letter, "this tenth of August, 1572", is a component of the fiction of surreptitious publication created in the preliminaries of "The adventures of Master F.I." and "The Printer to the Reader". It is intended to allow for a credible passage of time for H. W. to have decided upon publication, produced a faircopy for the printer, made arrangements with the fictitious publisher "A.B.", and gotten the project underway by "the xx. of January.1572[/73]". Six months is far more time than would have been required, so this date must be explained otherwise. The date is fictional in the sense that Gascoigne had not yet composed either "F.J." or H. W.'s letter. Hence, it seems clear that Gascoigne picked this specific date because, as the reconstruction demonstrates, the project probably began shortly afterwards. Similarly, the various references to printer's copy as "this written booke" (50:42), "this written regyster" (220), "the originall copie" (51:7), "the first coppie hereof" (47:10), and "this recorde" (48:7), as well as the illusion of a single continuous manuscript sustained in G.T.'s editorial frame, are purely fictional.[31]

V

It is necessary at this point to broaden the scope of this discussion to an examination of the combinations of bibliographical and textual evidence which demonstrate that Gascoigne neither was involved in the printing of the book nor saw any printed sheets. Thus Gascoigne's absence during the entire printing operation is the key to perceiving the relationships between the sharing situation and the bibliographical facts produced by the piecemeal evolution of the texts and the modifications in Gascoigne's overall plan for the book during the eight-month period. There can be no doubt that the texts evolved given the most remarkable insight provided by the printing reconstruction: that Gascoigne failed to complete one sub-text despite the minimum printing time of eight months. In the final editorial link, G. T. notes that he had "not recovered a full ende of this discourse" in reference to "The reporter's conclusion vnfinished.". This is clearly not a fiction. The completed version in The Posies added fifty-four stanzas to the original ten. Gascoigne, in short, had planned a considerably longer concluding link for "The Reporter" but abandoned it after just beginning its composition or the revision of an existing draft. Other internal textual evidence of the aborted plan for "Dan" is present. Beyond that, the failure to complete "Dan" probably is the reason for the modification of the original plan to include it as an independent sub-text. This late modification produced four major anomalies: (1) the cancellation of B1-2; (2) the anomalous imprint in page 164; (3) the pagination gap between pages 164-201; and (4) the separation of "The Printer to the Reader." from the text of "F.J." The evidence of Gascoigne's absence follows.

(1) The pre-textual materials of A Hundreth contain statements that are equivalent to components of a bibliographical description. Two fundamental discrepancies occur which categorically exclude the possibility that Gascoigne examined either the printed sheets or the completed book. Circumstantial evidence is provided by the only known fact regarding Gascoigne's whereabouts during the period: a record of his attachment to an English regiment in Holland by 25 May 1573.

First, the belated delivery of copy of "The Printer to the Reader," probably along with the "F.J." manuscript in mid-May or thereafter, doubtlessly produced the displacement of the letter from its intended position directly preceding the letters of H. W. and G. T. The "printer" begins in A2:7 (unpaginated): "Master H.W. in the beginning of this worke, hath in his letter (written to the readers) cunningly discharged himselfe of any such misliking . . ." (47:6-7). H. W.'s letter actually appears in 2A1-1v (page 201). The

"printer" obviously is a fictional character with a role in the surreptitious publication subterfuge. Equally obvious is the fact that Bynneman did not write the letter. As printer, he definitely was responsible for organizing and overseeing the printing of the sequence of texts and probably responsible for supplying page numbers for "The contents" by matching sub-titles with the pagination found in the printed sheets. The sub-title in "The contents" as printed reads: "Thirdly, a pleasant discourse of the adventures of master F.I. . . . 201." Furthermore, the "printer" claims that "hauing wel perused the worke, I find nothing therein amisse" (47:20-21). Bynneman did not write the letter, but simply printed it in its proper location as called for by its title: at the beginning of the book despite its undeniable connection to the text on page 201. Bynneman would have put it in sheet A regardless of where it appeared in copy.

The letter therefore reveals two components of authorial intention. Gascoigne intended that "The adventures of master F.J." appear first in the book. Moreover, his references in the letter to Supposes and Iocasta leave absolutely no doubt that the two plays were to be printed along with "F.J." but after it. In short, Gascoigne wrote the letter in the mistaken belief that his directions for the sequence of texts had been implemented by Smith and Bynneman. Although title pages are notoriously untrustworthy in regard to descriptive materials which appear beneath the main title, this title page confirms the statements of the letter. The main title is A Hundreth sun-|drie Flowres bounde | up in one small Poesie, corresponding to the title contrived in "H. VV. to the Reader." to refer to the collection of four sub-texts commencing with "F.J." "F.J." therefore must appear first in the book (I know of no exception to this rule). Moreover, specific reference is made to translations of Euripides and Ariosto: Euripides' only contribution is Iocasta; the reference to Ariosto is very probably to Supposes although the short translation of the allegory also appears at 2M3v. The title page references to translations by Ovid and Petrarch present an unresolvable crux but may indicate that at some point Gascoigne planned to add these texts to the book.

Second, the correction notice heading the misplaced narrative link in Ii1 was written by Gascoigne, but someone with access to the printed sheets provided the page reference "in Folio. 430." The notice is quite specific as to the proper textual location for the insertion of the link ("in the dolorous discourse" and "before the Supplication to Care." [my emphasis]) but would be ambiguous to either Smith or Bynneman. The misplaced link appears in The Posies at a location corresponding to page 420 in A Hundreth where it makes textual sense. Memorial confusion on Gascoigne's part probably led to the ambiguity of his confusing reference to the intended textual location.[32]

(2) In general, a repeated alphabet and/or pagination at a sharing boundary is an expedient solution to the problem of joining two shared sections without restrictions imposed by printing sequence and textual length. The two sections can be printed in any sequence and the first section can be extended indefinitely without causing confusion as to the order of texts, which is then established by the binding sequence. The flexibility is due to the absence of a sequential alphabet and/or pagination that locks the two sections into a defined order. Although the alphabet repeats in Middleton's section of A Hundreth, the pagination does not. Rather, Bynneman's first section B-X concludes at page 164 and Middleton's section begins at 2A1 with page number 201. Bynneman obviously intended to produce a book with sequential pagination for the reader's benefit as was his standard practice. Compositorial error can be ruled out: neither Middleton nor his compositors would supply a page number if none was marked in the received copy. Moreover, casting-off was a precise art essential to efficient printing, but the process categorically required copy to cast off.[33] The pagination gap, in short, is evidence that

Bynneman had copy of a "missing text" in hand that was cast-off to produce the page number 201, which was passed along with copy to Middleton.

The length of the "missing text" varies according to whether casting-off occurred before or after the gap at pages 36-46: if cast-off before, the length is 44 pages or 5½ sheets vs. 36 pages and 4½ sheets if after (note also that the correct page number for X4v is 156, not 164 as printed). The excess half-sheet normally would be used for the preliminaries if needed. The eventual omission of the "missing text" is clear evidence of a modification in the original plan for the book as presented to Bynneman. Gascoigne's withdrawal of a completed text seems quite improbable. Conversely, it is likely that the "missing text" was given to Bynneman in partially completed form with revisions and/or additions to be supplied later. The necessary assumption is that the text was completed in one of two senses: (1) the text required only minor revisions with little or no effect upon total length; or (2) a specific amount of text was to be added as defined either by a draft or an outline. An accurate casting-off would be greatly facilitated by a plan to add a precise number of stanzas of a fixed length to a partially completed verse text. A survey of Gascoigne's published work reveals only one candidate text that corresponds in length to the 5½ sheets yielded by an early casting-off. The fifty-four stanzas of uniform length that were added to "The reporter's conclusion unfinished" for the completed version of "Dan Bartholmew" as seen in The Posies brings the total length to forty-four pages or 5½ sheets. This assumes that the misplaced narrative link (Ii1-1v) was completed prior to casting-off; the assumption is clearly reasonable since it is the third sub-text in "Dan" and the second to be composed for the project (the enclosed poems are all earlier compositions).

The fact that "Dan" was omitted from the planned location in pages 165-200 can be attributed to two related factors. Without the additional fifty-four stanzas, the text fit neither length of gap (36 or 44 pages), but this would be of no concern to Gascoigne—it was the printer's problem. More importantly, when Gascoigne abandoned the composition of "Dan", it could no longer stand alone as an independent text. So he attached it to "F.J.," "The devises," and "devises of Master Gascoigne" through the simple expedient of G. T.'s quite short transitional editorial link (Ee2v:8-12). As delivered to Smith, "Dan Bartholmew" is a sub-text with its title embedded in G. T.'s editorial frame. The textual oddity produced by this solution points to Gascoigne's original conception of "Dan Bartholmew" as an independent sub-text (as it later appeared in The Posies). The editorial function of "The Reporter" exactly duplicates G. T.'s with the result that the global editor (G. T.), who is anthologizing poems by various authors, happens to incorporate a local editor's ("The Reporter") collection of poems by a single author. "Dan Bartholmew" simply does not belong in G. T.'s editorial realm. Furthermore, it can be argued on literary grounds that "Dan Bartholmew" was Gascoigne's initial experimentation with the narrative structure consisting of the incorporation of poems into an editorial framework. Perhaps

Gascoigne laid "Dan" aside to perfect the approach in "F.J." and simply ran out of time, leaving "Dan" unfinished.[34]

(3) The cancellation of B1-2 and the insertion of the anomalous imprint in page 164 clearly seem related to the modifications of Gascoigne's plan which produced the pagination gap at 164-201. The imprint in page 164 is significant evidence of Bynneman's preparation of copy for setting at the beginning of the project and reveals that, at that time, he viewed B-X as a self-contained production unit containing the texts of Supposes and Iocasta and preliminaries. As yet, Bynneman was unaware of any additional texts for the book. The terminal imprint at page 164 was marked in the manuscript by Bynneman at this time and indicates that the "original book" was limited to these two texts. The distinction between title-page and terminal imprints provides the key clue. The publisher provided and controlled the text for a title-page imprint and could clearly suppress the identification of the printer. Th title imprints of Bynneman's books as often follow the format "AT LONDON, | Imprinted for Richarde Smith." as the full format "H. Bynneman for H. Toy." and in some instances "Imprinted by F. Coldocke." the publisher.[35] (The omission of the date could reflect Smith's inexperience since A Hundreth was his first book.) On the other hand, the terminal imprint seems to have been a printer's tactic used usually to claim his work. Such imprints appear in two locations: (1) most frequently at the end of a book (as in Ii3r of A Hundreth), and (2) occasionally at the end of an initial or intermediate section in a shared book. Were it not for the pagination gap 164-201, the imprint "Printed by Henrie Bynneman | for Richard Smith." on X4v would be a normal intermediate imprint followed by the sharing printer's section. As such it would indicate that Bynneman knew that more text was to follow and that he would pass it to a sharing printer. However, the reconstruction shows that the text of "F.J." was not delivered until mid-May and was passed to Middleton because of production circumstances and not as part of a preplanned sharing strategy. The imprint further reveals that Bynneman was unaware of the fact that the "missing text" was to be included in the book; otherwise, he would not have thus marked the end of the Iocasta manuscript. In short, this terminal imprint remained in copy as an oversight which the

compositor faithfully set in X4v. The point at which Smith presented the "missing text" for casting-off is unclear, but there is no doubt that Bynneman still expected to print it in the assigned location when he passed copy for "F.J." with page number 201 to Middleton in mid-May.

The cancellation of B1-2 most probably is the end product of Gascoigne's evolving text and his failure to deliver the "missing text." Three facts are clear: (1) the use of Bynneman-Y1a shows that sheet B was definitely printed as part of B-X and probably first; (2) sheet A containing the independent-text title "Supposes," "The names of the Actors," and "Prologue" was printed last since it uses only the Guyot 76mm roman introduced in the final section (Dd-Ii) and contains "The contents." with page references; and (3), the whole sheet B was printed, followed later by the cancellation of B1-2.[36] Presumably the cancellation caused the resetting and movement of the preliminaries of Supposes to A4-4v. It seems reasonable to assume that these materials appeared in the cancelled B2-2v since no revisions or additions occur in The Posies; the formats in the preliminaries of both plays remains stable as well. Since the sheet is signed B, Bynneman clearly intended to prefix a signature A, probably as a half-sheet of preliminaries including the standard errata list and contents. The contents of the cancelled B1-1v must remain conjectural, but given Gascoigne's penchant for (usually lengthy) prefatory letters to various segments of his audience, common sense suggests that such a letter appeared in B1-1v. Gascoigne's prefatory letters always contain specific, detailed references to the subsequent text(s). Hence a prefatory letter (or letters) delivered along with the manuscripts of the two plays and printed at the outset of the project simply would not introduce the book as a whole. However, the simple expedient of shifting B-X to the end of the book would have solved the problem since the letter then would introduce the rest of the book and reproduce Gascoigne's intended sequence of texts as well.[37] This solution was unavailable in only one circumstance: part of the letter (or one of the letters) carried over to B1 from sheet A. (This was no problem since casting-off permitted printing a divided text in reverse order.) Such an incomplete text would categorically require the radical solution of cancellation and the financial loss sustained thereby. In the final analysis, this view must remain conjectural, but the logic is difficult to reject: something in B1-2 required cancellation because of the change in Gascoigne's plan.[38]

VI

The reconstruction of the printing of A Hundreth provides a firm basis for editorial decisions toward a critical edition. Bynneman and Smith essentially produced a bibliographical record of the author's evolving concept of his collection of texts. Gascoigne began the project with the two plays and concluded it with an explicit reference to them in "The Printer to the Reader." Despite the intermediate changes in his plan, the two plays are an integral component of the book. A key stage occurred when Gascoigne halted the separate publication of the two plays in February. Why Smith agreed to this is puzzling since a division of copy into two books cost about the same to print and he could have begun realizing a return in February. In any event, a critical edition must include the two plays, but in the order representing the latest stage of Gascoigne's plan—after the four sub-texts of A Hundreth Sundrie Flowres. The minor problem of the original "The contents" can be resolved by moving this text to the textual notes with appropriate comment. It should be preserved since the bibliographical information is quite important. The insertion of the "De profundis" under its orphaned sub-title needs no defense: if Gascoigne's comment in "Phylomene" can be believed, it was a quite old composition and existed at the time printer's copy was prepared. The placement of the misplaced editorial link (Ii1) is clearly a problem: its location in A Hundreth may signify the order of composition of "Dan Bartholmew," but Bynneman was responsible for placing it in Ii1 rather than after the concluding link, where such corrections normally were inserted. It seems reasonable to put it in the proper location with a textual note including the correction notice. Finally, the modern reader responds to the extratextual components of graphic layout and titling just as did the Elizabethan. Prouty's duplication of the end-of-text settings at the junctions of segments of manuscript copy (see his "V" format and flowers in p. 106, corresponding to 2M3, and the flowers in p. 218, corresponding to Cc4v) illustrates the incorrect impression conveyed by the transmission of a printer's errors into a modern edition. The editorial prose should be set continuously to signify the unbroken sequence of sub-texts. Similarly, the original running-titles ought to be amended and appear over their related sub-texts. It also seems desirable in a critical edition to indicate marginally the original page breaks.

Notes