University of Virginia Library

Search this document 


  

expand section 
expand section 
collapse section 
  
  
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
Marginal Rules as Evidence by Ernest W. Sullivan, II
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
  

expand section 

171

Page 171

Marginal Rules as Evidence
by
Ernest W. Sullivan, II

Edwin Eliott Willoughby, in The Printing of the First Folio of Shakespeare (1932), first suggested the importance of marginal-rule evidence to group printing sequences (p. 17), to determine the order of sheets (p. 42), and to detect interruptions in the printing process (p. 44). In The Printing and Proof-reading of the First Folio of Shakespeare (1963), Charlton Hinman's detailed analysis of evidence provided by "centre rules" (I, 138-150) and "box-rules" (I,154-171)[1] markedly expanded the usefulness of such evidence beyond Willoughby's original suggestions and remains by far the most thorough treatment of marginal-rule evidence; however, because the Shakespeare Folio rules are a special case, his treatment of marginal-rule evidence should not be generalized without knowledge of its limitations. Furthermore, since his comprehensive description of the Shakespeare Folio printing often utilizes many kinds of significant evidence, it does not emphasize that marginal-rule evidence alone can solve a variety of bibliographical problems or that, in some fairly common circumstances, important problems cannot be solved without such evidence.[2] Using the quarto first edition of John Donne's Biathanatos (1647), I shall discuss some kinds of problems (initial point of composition, priority of different settings of the same sheet, partial-sheet imposition, and order of imposition


172

Page 172
of preliminaries) which may require analysis of marginal-rule evidence to obtain a valid solution and will develop some guidelines for the general application of such evidence.

Marginal-rule evidence occurs in three forms: (1) changes in the individual marginal rules; (2) changes in the marginal-rule skeleton, the pattern of marginal rules on a single page; and (3) changes in the rule skeleton-forme, the marginal-rule skeletons comprising an entire forme. The changes involve either a change of state or a change of arrangement, with changes of state, such as a break in a rule, the introduction of new rules into a skeleton, or of new skeletons into a skeleton-forme, generally providing the more reliable evidence.

Each skeleton on every page (except sigs. ¶1-¶2v) of the first edition of Biathanatos contains three vertical lines (one edging the text and running list of contents, another edging the other side of the text and running list of contents and separating them from the page number and marginalia, and a third outside the page number and marginalia) plus four horizontal lines (one connecting the tops of the three vertical lines; two 10-12 mm. below this top line: the first, 85-88 mm. long, under the running list of contents and between the two vertical lines edging the text; the second, 17-18 mm. long, underneath the page number and between the two vertical lines which mark the outer edges of the text and marginal material; and one, beneath the signature and connecting the vertical line bottom ends). The compositors had twelve such basic skeletons but nearly always used eight of them in two skeleton-formes. The collation for most copies of the first edition reads: ¶4 (*)2 πA4 [A2 mis-signed g2], A-Z4, Aa-Dd4 Ee2, 120 leaves.[3] The usual arrangement of the contents is: ¶1: [blank]. ¶2: Title [verso blank]. ¶3: Epistle Dedicatory. (*)1: List of Authors Cited. πA1: Table of Contents. C1: 'THE PREFACE.' D1: 'THE FIRST PART.' K2: 'Second Part.' V1: 'The Third Part.' Dd4: 'Conclusion.'

Marginal-rule evidence is ideal for determining where composition began. In the case of Biathanatos, different running-titles on sheets ¶, (*), πA, C, and D make running-title evidence irrelevant, and the different type fonts on sheets ¶, (*), πA, and C eliminate type-face evidence except for a decision between sheets C and D, where the evidence is inconclusive. Furthermore, the assumption that composition from a manuscript would begin with the text would lead to the mistaken inference that composition began at sheet C, where Donne's argument begins with "THE PREFACE Declaring the Reasons, the Purpose, the way, and the end of the AVTHOR."[4]


173

Page 173

A change of state in a marginal rule, however, shows that composition began at sheet D with 'THE FIRST PART.' Comparison of the various skeletons on a Hinman collator proves the sig. T1v skeleton to be identical with that on Compositor B's[5] sig. S1v except that the top horizontal rule on sig. T1v has a series of very sharp bends in the first 35 mm. of its left end. Logically, of course, the mere existence of different states does not establish their printing order, and particularly with works possibly composed in a non-alphabetical sequence, great caution should be exercised in drawing conclusions about priority from such evidence;[6] however, in this case, the unbent state of the rule would certainly appear to be the earlier, and it also seems unlikely that composition would begin at sig. T1v in a work set from manuscript. Thus, all pages (and, because the skeletons show up on each sheet, all sheets) with the sig. T1v skeleton were printed after Compositor B composed and printed sig. S1v.[7] Since the sig. T1v


174

Page 174
skeleton appears in sheets T-Dd and πA, A-C, sheet C must have been imposed after sheet D, which has the unbent state of the rule (sig. D3v). If sheets D-R were printed alphabetically, D is the earliest sheet with the unbent state of the rule and must have been the first sheet imposed. All sheets E to R also have the rule in its unbent state, so none of them could have stood in type simultaneously with sheet D; thus, D must have been imposed before any other sheet was composed, making D the first sheet typeset.

Evidence from marginal rules can solve another problem of compositional order such as the priority of variant settings of the same sheet: in Biathanatos, the order of the two settings of sheet S cannot be determined without such evidence.[8] Changes of state in skeletons and skeleton-formes establish that Compositor A reset Compositor B's setting of sheet S and that he did so after beginning, but before finishing, the composition of sheet T. Examining Compositor A's S setting in DLC copy ND 0332947, we find that he set both formes in a single skeleton-forme,[9] only two rules of which had definitely appeared in previous sheets. These two previously used rules, the top horizontal rule (characterized by a bump in the middle and a hook at the right end) on sigs. S1 and S2 and the longer middle horizontal rule (having a broken bump 17 mm. from its right end) on sigs. S1v and S2v, both occur on Compositor B's sig. S4v. Since this sig. S4v skeleton appears in sheets D-R, it must be the earlier skeleton, and those skeletons on Compositor A's sigs. S1-S2v must have been created from its rules.

Having determined the order of the settings, we can establish when the resetting took place. The fact that the rules on Compositor A's setting of S came from Compositor B's sig. S4v skeleton implies that Compositor A began his resetting after the sheets having Compositor B's setting had been printed. Sheet T, set by Compositor A in all copies, has the same skeleton-forme as Compositor B's sheet S;[10] hence the skeletons used on all previous sheets by Compositor A as well as on sheet S by Compositor B were available to Compositor A when he began to set sheet T—why did he not use these skeletons or skeletons scavenged from them (in addition to the two rules from Compositor B's sig. S4v skeleton), and why did he not continue using two skeleton-formes as usual? The answer to both questions is


175

Page 175
that before he reset sheet S, Compositor A had already begun composing sheet T (but had not finished it, else the rules from Compositor B's sig. S4v which appear on his sig. T4v would not have been available for use on his sigs. S1-S2v)[11] with the regularly used skeletons; thus, these skeletons (except, of course, that from sig. S4v) were full of type for sheet T, leaving Compositor A with fewer rules than usual, perhaps only enough to make part of a single skeleton-forme.

The marginal rules, then, on sheets R, S, and T prove that Compositor A's sheet S is the resetting and that he began the resetting after he started composing sheet T and finished the resetting and its printing before he completed composing sheet T. Indeed, since both settings derive from the printer's manuscript, and are therefore not directly related textually, neither the priority of Compositor B's setting nor the time at which Compositor A's resetting occurred could be established without the marginal-rule evidence.

Marginal-rule evidence, as well as running-title evidence,[12] can solve problems involving partial-sheet imposition, particularly whether half-sheets were printed together or separately. For example, collation in a normal copy of Biathantos runs ¶4 (*)2 πA4, A-Z4, Aa-Dd4 Ee2; leaving the bibliographer to determine whether sigs. Ee1-Ee2v and (*)1-(*)2v were printed together on a single sheet or separately as half-sheets.[13] Examination of the skeletons shows that a single quarto skeleton-forme, composed of the skeletons from sigs. Dd4, Dd4v, Dd1, and Dd1v, was used for both formes: the outer forme contained Ee(o) and (*)(i), and the inner, Ee(i) and


176

Page 176
(*)(o). The presence of both half-sheets in each skeleton-forme proves that they were printed together on a quarto sheet. After imposition, the sheet would have been cut in half, each half folded once, and then the folded half-sheets bound in the correct place in the volume. This sequence can be easily verified: if the binder failed to cut and fold a sheet containing the above arrangement of half sheets Ee and (*) but simply folded such a sheet in normal quarto fashion, sigs. (*)1-(*)2v would appear between sigs. Ee1v and Ee2, as they in fact do in the copies listed in footnote thirteen. Running-title evidence could not solve this problem: the title on sigs. Dd1 and D1v is 'Part. 3. Dist. 5. Sect. 7.'; sigs. Dd4, Dd4v, Ee1-Ee2v have 'Conclusion'; and sigs. (*)1-(*)2v have no running-title.

Marginal-rule evidence should prove particularly useful in determining the order of composition and printing of preliminary materials: as Hinman says, "Rules, moreover, unlike ornaments and other kinds of typographical materials [excluding type faces] that occasionally provide evidence of order, are ubiquitous" (I, 157);[14] whereas type-font and running-title variation in most preliminaries limit their use in determining order. We have seen that half sheet (*), containing the List of Authors, was printed before the other preliminary matter, so the preliminaries could not have been composed and printed in the order in which they were usually bound. The remaining preliminaries whose order remains to be ascertained include the title page and Epistle Dedicatory (sheet ¶), the Table of Contents (sheets πA, A, B) and 'THE PREFACE' (sheet C). Because these preliminaries do not occupy overlapping sheets, any order of composition would have been possible, and the final determination of order is therefore rather complex, requiring a combined analysis of rules, skeletons, and skeleton-formes.

The skeleton-forme of the outer forme of sheet πA also appears on both formes of sheets A and B; hence the break 38 mm. from the left end in the top horizontal rule in the skeleton on sigs. A1, B1, and B2 proves that sheets A and B were printed after sheet πA. This same skeleton appears on sig. C1 (though the rest of the C outer skeleton-forme differs from the outer skeleton-forme on sheets πA, A, and B) with the top horizontal line unbroken, proving that the outer (and presumably, the inner) forme of sheet C was printed before any of sheet A.[15] Since a πA, C, A, B sequence


177

Page 177
would require the compositor to stop after printing one-third (sheet πA) of the Table of Contents, mix the skeletons from both skeleton-formes of sheet πA to build the sheet C skeleton-formes, set up and print the Preface, rearrange four skeletons from the sheet C skeleton-formes into their previous positions in the sheet πA outer skeleton-forme, and then compose sheet A, we can assume that sheet C was printed after sheet Ee-(*) and before sheet πA. Sheet ¶, containing the title page (sig. ¶2) and Epistle Dedicatory (sigs. ¶3-¶4v), has sigs. ¶1, ¶1v, and ¶2v blank, sig. ¶2 in a double frame of rules from the skeletons on sigs. πA3v and πA4, sigs. ¶3 and ¶4 in the sig. g2 (πA2) skeleton, and sigs. ¶3v and ¶4v in the sig. πA1v skeleton; thus, sheet ¶ was set in a single skeleton-forme composed of rules and skeletons from the sheet πA inner skeleton-forme, with sigs. ¶3 and ¶4v imposed on one side of the quarto sheet, sigs. ¶2, ¶3v, and ¶4 on the other, and the sheet folded in normal quarto fashion.[16] Having established the C, πA, A, B sequence and that sheet ¶ has rules and skeletons which appear in the πA inner skeleton-forme (available while sheets A and B were being printed in the πA outer skeleton-forme), one can safely conclude that sheet ¶ was composed and printed either while sheets A and B were being composed and printed or shortly thereafter. Marginal-rule evidence, then, proves that the preliminary sheets were composed and printed in the following order: Ee-(*) (List of Authors); C ('THE PREFACE'); πA, A, B (Table of Contents); and ¶ (title page, Epistle Dedicatory).

Even when running-title and type-face evidence was made useless by the different running-titles and type fonts in each part of the preliminaries, marginal-rule evidence determined the printing sequence for the preliminaries in Biathanatos. Furthermore, the printing sequence strongly suggests that logic or intuition may generally be insufficient to solve such problems: one would not expect that the first sheet of the body of the text (Preface, sigs. C1-4v) would be printed after the last sheet of the body of the text and even after some preliminaries (List of Authors, sigs. (*)1-2v) or that the preliminaries listed first in the collation (title page and Epistle Dedicatory, sigs. ¶2-4v) would be printed simultaneously with or after the preliminaries listed last (Table of Contents, sigs. πA1-4v, A1-B4v).

The previous analyses primarily used evidence from changes of state in rules, skeletons, and skeleton-formes. If one exercises some caution in conjecturing


178

Page 178
priority from variant marginal-rule states, such evidence can be applied to problems of composition and imposition order in any volume. Evidence from a change in arrangement of marginal rules, skeletons, or skeleton-formes can solve the same problems exactly as evidence from changes in state did, but the less fixed the rules in the skeletons and the skeletons in the skeleton-formes, the less reliable the evidence from changes in arrangement. A rule is "fixed" in its skeleton if essentially undisturbed during the setting-up or distribution of the type; a skeleton, in its skeleton-forme, if basically unaltered during the composition, imposition, and distribution of the forme.[17]

Evidence from fixed rules and skeletons can, under certain conditions, identify pages or formes in the same printing sequences: sheets in alphabetical sequence having one rule positioned differently than on other sheets in another alphabetical sequence implies that all sheets having one arrangement were printed before or after all those having the other. If, however, either group of sheets is not in an alphabetically continuous sequence, then evidence from changed arrangement is not conclusive: a given arrangement of marginal rules, "X," for some forme, "A," could become a different arrangement, "Y," for forme "B," and still later randomly recur in the original arrangement, "X," in forme "C," causing the unwary bibliographer to place formes A and C in the same sequence. Hinman recognized this possibility, but dismissed it on the basis of his experience with the Shakespeare Folio: "Almost never, moreover, when rules took up new positions in a given forme did they resume exactly their former positions in some later forme. Hence a given arrangement of rules serves to define a group of formes belonging to the same printing sequence: we may ordinarily [he provides two counter-examples] be quite sure, that is, that all the formes showing exactly the same arrangement of their box rules were printed either before or after all the formes that show a different arrangement" (I, 157). "Almost never" is not "never," however, and even in a work with generally fixed rules like the Shakespeare Folio,[18] evidence from changes in arrangement of rules or skeletons needs corroboration conclusively to identify printing sequences. One should always remember that the more evidence of changed rule and skeleton arrangement found in a work believed to have fixed rules and skeletons, the less reliable that evidence is.


179

Page 179

Moreover, the rules and skeletons in Biathanatos show that Hinman's affirmation of the reliability of evidence from changes in arrangement should not be generalized to works having partially fixed rules and skeletons. Previous discussion has shown that Compositor B's sheet S, Compositor A's sheet S, and Compositor A's sheet T belong to three separate printing sequences. The fact that the skeleton from Compositor B's sig. S4v appears on Compositor A's sig. T4v, then, proves that identical rule arrangements on pages in works with partially fixed rules and skeletons need not imply membership in a single printing sequence. Indeed, the number of rules and steps involved in Compositor A's tearing down Compositor B's sig. S4v skeleton to construct skeletons for sigs. S1-S2v and then dismantling them to build his sig. T4v skeleton make the odds against identical arrangements in the two skeletons high enough that one could speculate that Compositor A deliberately made his sig. T4v skeleton look like Compositor B's sig. S4v skeleton.

Just as identical arrangements of marginal rules can exist on pages belonging to different printing sequences in a work with partially fixed rules, so can identical arrangements of skeletons in skeleton-formes exist on formes belonging to different printing sequences, as the skeletons on sheets E to S in Biathanatos will show. The skeleton arrangement in the E outer skeleton-forme also appears in the F inner skeleton-forme and that in the E inner skeleton-forme in the outer of F. This skeleton arrangement continues through sheet O, but the skeletons are switched on P, making its arrangement of skeletons and skeleton-formes the same as that in E. The sheet P skeleton rules, however, show extensive deterioration not found in the sheet E rules but traceable in the sheets between E and P; thus E and P cannot belong to the same printing sequence. This sheet P skeleton arrangement continues through R, but when Compositor B began to work on sheet S, he reversed the sheet R skeleton-formes; as a consequence, his setting of S has the same skeleton and skeleton-forme arrangement as sheet O, but, obviously, sheets F-O and Compositor B's sheet S cannot be one printing sequence, and sheets E, P, Q, and R, another. Identical arrangements of rules, skeletons, or skeleton-formes, then, do not conclusively prove that pages, formes, or sheets having the identical arrangements belong to a single printing sequence in works with partially fixed rules, skeletons, and skeleton-formes.

Even though it requires certain precautions, evidence from changes in state or arrangement of marginal rules, skeletons, and skeleton-formes can be extremely useful not only because it may prove helpful or even necessary in solving a variety of problems, but also because it is readily available. Frequently, such evidence is immediately obvious to the unaided eye, but even very slight variations can usually be detected with a Hinman collator. If the analysis involves only a single volume or if the bibliographer cannot gain access to physical copies of necessary volumes, xerox copyflow made


180

Page 180
from microfilm as described in George Guffey's article, "Standardization of Photographic Reproductions for Mechanical Collation," PBSA, 62 (1968), 237-240, and used on a Hinman collator is sufficiently accurate to perform any of the marginal rule analyses in this discussion.[19]

Notes

 
[1]

Hinman's terms, "centre rules" (which run vertically down the page center between two columns of text) and "box rules" (surrounding the running-titles and text), derive from James S. Streck's article, "Center Rules in Folio Printing: A New Kind of Bibliographical Evidence," SB, 1 (1948-49), 188-191. Streck proves that the "center lines," unlike the "box rules," are not an integral part of the skeleton-forme in the 1647 Beaumont and Fletcher Folio; Hinman proves the same for the Shakespeare Folio. The Streck and Hinman terminology, however, is inappropriate for a general theory of marginal-rule evidence: it describes only a single arrangement of rules in works having some rules occasionally distributed with the type and others usually fixed in the skeleton, and the implication that rules surrounding the text were used differently from those in the body of the text could be seriously misleading—even in the two Folios, any given center rule usually appears with the same box rules, and the box rules are frequently not part of the skeleton-forme. In an effort to develop a more general approach, I shall treat all rules simultaneously and so have named them for their usual function, indicating a margin of some sort.

[2]

For example, Guy A. Battle, in "The Case of the Altered 'C'—A Bibliographical Problem in the Beaumont and Fletcher First Folio," PBSA, 42 (1948), 66-70, shows that marginal rules may be more reliable than running-titles in tracing skeletons and in "A Bibliographical Note from the Beaumont and Fletcher First Folio," SB, 1 (1948-49), 187-188, that in at least one instance rule evidence could determine the printing order of formes on a sheet.

[3]

Two issues of the first edition exist differentiated by a cancel title page. Copies with their sheets bound differently are listed in footnote 13.

[4]

Ordinarily, one would expect material identified as "THE PREFACE" to be part of the preliminaries, and, indeed, the compositor began at sig. D1 (see below) with "THE FIRST PART. OF LAW and Nature," showing that he initially considered "THE PREFACE" part of the preliminaries. The Table of Contents, however, lists "THE PREFACE" as the first part of the text. Furthermore, the unusual signing of the sheets in the Table of Contents (πA, A, B) and the pagination beginning on the second sheet of the Table of Contents at sig. A1 strongly suggest that the compositor realized that sheet C, containing "THE PREFACE," was not part of the preliminaries before he began setting the Table of Contents: when the compositor signed his first sheet 'D' and numbered its first page '25', he planned to sign and number the pages of three sheets (a perfect calculation for the Table of Contents) as part of the regular sequence of signatures and page numbers, but when he realized that "THE PREFACE" (which happened to fill exactly one sheet, C) began the text, he had to move the Table of Contents back one sheet, designating its first sheet 'A' (signed in italic) and beginning pagination on its ninth page. This initial confusion of the status of "THE PREFACE" also answers the query of Geoffrey Keynes as to why "The pagination begins very oddly on the fifth leaf of the Contents list" (A Bibliography of Dr. John Donne, 4th ed. [1973], p. 119).

[5]

Two settings of sheet S exist: Compositor A set all of the sheets in the Library of Congress (DLC) copy ND 0332947 and every sheet except S in all other copies of Biathanatos with which I am familiar; Compositor B set only the S sheet found in all other copies. The available evidence strongly implies that Compositor A, using the printer's manuscript, totally reset Compositor B's setting of sheet S to make the quarto stylistically uniform. Compositor B, perhaps an apprentice, in an apparent attempt to reproduce the literal appearance of the printer's manuscript, failed to indicate its intentions with respect to quotations, punctuation, and capitalization in the manner conventional in printed books with the result that his pages differ greatly in appearance from those set by Compositor A. For a complete analysis of the two settings of sheet S, see pages 136-142 of my dissertation, "A Critical, Old-spelling Edition of John Donne's Biathanatos" (UCLA, 1973), University Microfilms order #73-16,704.

[6]

Willoughby sounds a further warning about evidence from bent rules: "owing to . . . the ease with which they could be both bent and straightened, rules cannot furnish evidence as conclusive as type" (p. 42). Hinman, on the other hand, is quite confident about inferring priority from the mere existence of two states: "Almost invariably, moreover, it is at once quite clear which is the earlier and which the later of any two states" (I, 155-156).

[7]

Pages containing the sig. T1v skeleton actually were imposed even later, after Compositor A reset sheet S; see the discussion of the order of composition of sheets S and T below.

[8]

Failure to consider marginal-rule evidence probably played a part in Charles Mark's incorrect ordering of these very settings in his dissertation, "John Donne: Biathanatos: A Critical Edition with Introduction and Commentary" (Princeton, 1969), pp. clxxxix-cxc.

[9]

The following pairs of skeletons are identical: S2v and S1v, S3 and S4, and S4v and S3v. S1 and S2 have the same rules but not in the same arrangement.

[10]

The one difference does not affect the validity of the analysis: the sig. S1 bottom horizontal line appears turned end for end and upside down on sig. T1, an arrangement maintained in the rest of the sheets.

[11]

If Compositor A had finished sheet T before resetting sheet S, these rules would have been available from sig. T4v however, had he already finished sheet T, Compositor A surely would have used its skeleton-formes rather than introduced a new skeleton-forme with several new rules in addition to the two obtained by partially dismantling the sig. T4v skeleton, tearing apart the new skeleton-forme after printing sheet S, and exactly rebuilding the sig. T4v skeleton for use on sig. V4v. Compositor A could not have returned to reset sheet S after finishing the rest of the sheets: marked deterioration in later sheets of the top horizontal rule used on sigs. S1 and S2 and longer middle horizontal rule on sigs. S1v and S2v eliminates any possibility that sheet S was reset after they had been printed.

[12]

See Fredson Bowers' article, "Running-Title Evidence for Determining Half-Sheet Imposition," SB, 1 (1948), 199-202, reprinted in Essays in Bibliography, Text, and Editing (1975), pp. 254-257.

[13]

Geoffrey Keynes's copy preserves stubs proving that sigs. (*)1-(*)2v and Ee1-Ee2v were imposed on the same sheet (A Bibliography of Dr John Donne, 3rd ed. [1958], p. 93), and the following copies retain the quarto sheet intact with signatures running Ee1, Ee1v, (*)1, (*)1v, (*)2, (*)2v, Ee2, Ee2v: British Museum (George Thomason's copy), Folger Shakespeare Library, Harvard University Library (Augustus Jessopp's copy), and William Andrews Clark Memorial Library. In many cases, however, such evidence may not exist, the available copies may have been rebound, the bibliographer may not know about or have access to the necessary copies, he may have to work from photographic reproductions, or he must forego unbinding the available volumes, leaving him no recourse but marginal-rule evidence. Even in our case, marginal-rule evidence is necessary to reconstruct completely the processing of this sheet.

[14]

Biathanatos demonstrates this ubiquity: even though its preliminaries do not require skeletons (pagination does not even begin until sig. A1, and the preliminaries lack marginalia), they appear on every printed page except the title page.

[15]

Since the break does not appear in the top horizontal line on sig. A2, the marginal-rule evidence does not prove that the inner forme of A could not have been printed before sheet C, but this possibility seems rather remote: the compositor would have had to stop in the middle of the Table of Contents, arrange the sheet A skeletons plus four other skeletons into the sheet C skeleton-formes, compose and print the Preface, dismantle the sheet C skeleton-formes, rebuild the skeleton-forme for sheet A, and return to composing the Table of Contents, all for no reason.

[16]

With ¶1, ¶1v, and ¶2v blank and ¶2 unsigned, one might think that the title page was printed on a separate sheet, but the rules prove that it and the Epistle Dedicatory were printed on a single sheet in the first issue of Biathanatos (which has the undated title page—the second issue carries a 1648 date). Significantly, the presence of the rules from sigs. πA3v and πA4 on the undated title page and the absence of any rules from other pages of Biathanatos on the 1648 title page establishes the priority of the undated title page and the order of the issues. As in the case of the Ee-(*) sheet, proof that the undated title page was printed as a single sheet with the Epistle Dedicatory is simple, given access to appropriate copies, but for a bibliographer working with rebound copies or photographic reproductions, the marginal-rule evidence is essential.

[17]

Marginal rules tend to become fixed for the same reasons running-titles do: "The convenience of using the same skeleton in different formes for speed and for consistency of register (instead of constructing a new skeleton for each forme of the book) made the running-title in effect a part of the furniture" (Fredson Bowers, "Notes on Running-Titles as Bibliographical Evidence," The Library, 4th ser., 19 [1938], 316).

[18]

Hinman concludes that the Shakespeare Folio box rules were part of the skeletons (I, 153) and that these skeletons were unaffected by composition and distribution, so that the box rules were fixed: "The components of this skeleton were not only then transferred directly from the old forme to the new but were placed in exactly the same relative positions in the new forme that they had occupied in the old" (I, 156).

[19]

To insure reliability, the dimensions of the copyflow reproduction should be checked by comparing the reproduction with a physical copy on a Hinman collator.