| ||
As the late Allan Stevenson observed, the study of paper has until recently been historical rather than bibliographical, "and most of the attempts to draw useful inferences out of watermarks (and other aspects of paper) have been uninformed and unscientific."[1] There have, of course, been exceptions to this rule, among the most notable of which was Sir Walter Greg's classic work on the misdated Pavier quartos of Shakespeare.[2] And Stevenson's own inspired contributions to the study of paper, which culminated in his investigation of the Missale speciale,[3] have done much to make the bibliographical analysis of paper evidence more sophisticated and reliable.[4] The principal end of both Greg's work on the Shakespeare quartos and Stevenson's on the Missale was dating, and for this it was necessary for them to compare the paper in several books. Other scholars have used paper evidence to locate textual disruptions, by examining the patterns produced by folding single sheets of paper in specific formats. In the folio format, for example, a sheet of paper is folded once to produce one leaf with a watermark at its center and one unwatermarked leaf. In quartos, the double folding of the sheet splits the watermark so that roughly one half of the mark appears centered at the gutter on each of two conjugate leaves, while the other two leaves of the sheet are unwatermarked. Robert W. Chapman and others have realized that violations of these patterns indicate cancellation or some other disturbance of the gathering.[5] What has not been appreciated,
To illustrate this method of analysis, I shall refer to Cotton MS. Vespasian D. viii., the N-town Cycle of mystery plays, a small quarto dating from the second half of the fifteenth century. The plays in this text are patently composite, and there is good evidence that the cycle was in part compiled as the manuscript was transcribed;[6] it is therefore well suited to demonstrate some of the possible applications of the approach under discussion in this study. The manuscript is in a British Museum binding and the leaves have been cut and separately mounted on guards, save for a few bifolia which are mounted on single guards. Most of the original signatures and quire-marks have been cut away. Later, perhaps Cottonian, quire-marks do appear, but the quiring they indicate is very odd: according to these marks, six quires contain only two leaves each, while other quires contain three, six, eight, twelve, fourteen, fifteen, sixteen, eighteen, nineteen, and twenty leaves. By analyzing the watermark sequences, one can test the accuracy of the quire-marks, and in the process shed light on the compilation of the codex and the plays themselves.
As already noted, each folded sheet in the quarto format produces two watermarked leaves, each with roughly one half of the mark, and two un-watermarked leaves. The order in which these leaves actually appear in old manuscript books varies, probably owing to the different positions of the sheets when they were folded (various methods of folding the sheets may also have been a factor, but this seems less likely, and the results would have been the same in any case). The double folding of sheets in quarto produced four distinct patterns of watermark distribution, as I shall try to show. For the purpose of this demonstration, let us assume a fixed manner of folding: first folding the sheet in half and forward, toward the scribe, along a line parallel to the shorter side; then folding it forward again, along a line parallel to the other side. If a sheet sitting in the position illustrated by Figure 1 is folded in this way, the four leaves (let us call them "panels" in this context) will
Before looking for these watermark sequences in Vesp. D. viii., we must consider one further test of the method by which sheets were folded: the appearance of the mould-sides and the felt-sides of the paper. The watermark, chain-lines, and, to a lesser extent, wire-lines of the mould made indentations into the side of the paper which faced down upon them. This side of the sheet is called the mould-side or "right" side, while the other side is called the felt-side or the "wrong" side. Obviously, the
Now let us turn to the watermark and mould-side sequences in Vesp. D. viii., as represented by the following chart:
A | quire: | ||||||||||||||||||
r | v | r | v | v | r | r | v | v | r | v | r | r | v | v | r | r | v | r | v |
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 |
A1 | -- | A1 | -- | B1 | -- | B2 | -- | -- | A1 | B1 | -- | -- | A2 | -- | A1 | -- | B1 | -- | B1 |
B | quire: | ||||||||||||||||||
r | v | v | r | v | r | v | r | r | v | r | v | v | r | v | ? | v | r | r | v |
21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 |
-- | A2 | -- | -- | -- | A1 | A2 | A1 | -- | B1 | A1 | -- | B1 | B2 | B1 | -- | -- | -- | B2 | -- |
C | quire: | ||||||
v | r | v | r | v | r | v | r |
41 | 42 | 43 | 44 | 45 | 46 | 47 | 48 |
C1 | -- | -- | C2 | D2 | -- | -- | D1 |
D | quire: | E | quire: | F | quire: | ||||||||||||
v | v | r | v | v | v | v | r | r | r | r | v | v | v | v | r | r | r |
49 | 50 | 51 | 52 | 53 | 54 | 55 | 56 | 57 | 58 | 59 | 60 | 61 | 62 | 63 | 64 | 65 | 66 |
-- | C1 | E | F | C1 | -- | -- | C2 | C2 | -- | -- | -- | -- | D2 | D2 | -- | -- | D1 |
G | quire: | |||||||||||||||||
v | v | v | v | r | v | r | r | r | r | v | v | v | v | r | r | r | r | r |
67 | 68 | 69 | 70 | 71 | 72 | 73 | 74 | 75 | 76 | 77 | 78 | 79 | 80 | 81 | 82 | 83 | 84 | 85 |
D2 | D2 | D2 | D2 | -- | C1 | -- | -- | -- | -- | -- | -- | -- | -- | D1 | C2 | C2 | C2 | C2 |
H | quire: | |||||||||||||||
v | v | v | v | v | r | ? | v | v | [r] | [v] | v | r | r | r | r | r |
86 | 87 | 88 | 89 | 90 | 91 | 92 | 93 | 94 | [95] | [96] | 97 | 98 | 99 | 100 | 101 | 102 |
D2 | -- | C1 | -- | D2 | -- | -- | D2 | -- | [G] | [H] | -- | C2 | -- | D1 | -- | C2 |
J | quire: | |||||||||||||||||
v | v | v | v | v | r | r | r | r | [v] | v | v | v | v | r | r | r | r | r |
103 | 104 | 105 | 106 | 107 | 108 | 109 | 110 | 111 | [112] | 113 | 114 | 115 | 116 | 117 | 118 | 119 | 120 | 121 |
D2 | -- | -- | -- | D2 | -- | D1 | D1 | -- | [--] | -- | C1 | C1 | -- | C2 | -- | -- | -- | C2 |
K | quire: | L | quire: | M | quire: | ||||||||
v | v | r | r | r | v | v | v | r | v | r | r | r | v |
122 | 123 | 124 | 125 | 126 | 127 | 128 | 129 | 130 | 131 | 132 | 133 | 134 | 135 |
-- | D2 | C2 | -- | C2 | -- | C1 | -- | D1 | C1 | -- | D1 | -- | D2 |
N | quire: | O | quire: | ||||||||||||
r | v | r | r | v | v | r | ? | r | v | v | r | v | r | v | r |
136 | 137 | 138 | 139 | 140 | 141 | 142 | [143] | 144 | 145 | 146 | 147 | 148 | 149 | 150 | 151 |
-- | I1 | J1 | -- | I2 | -- | -- | [--] | J2 | -- | I1 | J1 | -- | -- | -- | A1 |
P | quire: | Q | quire: | R | quire: | ||||||
v | r | r | r | v | v | v | r | v | r | r | r |
152 | 153 | 154 | 155 | 156 | 157 | 158 | 159 | 160 | 161 | 162 | 163 |
-- | J2 | J2 | -- | I2 | -- | -- | J1 | I1 | -- | -- | J2 |
S | quire: | ||||||||||||||
r | r | r | r | v | v | v | v | r | r | r | r | v | v | v | v |
164 | 165 | 166 | 167 | 168 | 169 | 170 | 171 | 172 | 173 | 174 | 175 | 176 | 177 | 178 | 179 |
-- | K1 | K1 | -- | K2 | -- | -- | K2 | L2 | -- | -- | L2 | -- | L1 | L1 | -- |
T | quire: | ||||||||
r | r | v | v | [v] | [r] | r | r | v | v |
180 | 181 | 182 | 183 | [184] | [185] | 186 | 187 | 188 | 189 |
K1 | K1 | -- | -- | [--] | [--] | -- | -- | L1 | L1 |
V | quire: | ||||||||||||||||||
v | v | v | v | v | r | r | r | r | r | v | v | v | v | v | r | r | r | r | r |
190 | 191 | 192 | 193 | 194 | 195 | 196 | 197 | 198 | 199 | 200 | 201 | 202 | 203 | 204 | 205 | 206 | 207 | 208 | 209 |
-- | -- | -- | -- | D2 | -- | C2 | C2 | C2 | C2 | D2 | D2 | D2 | D2 | -- | C2 | -- | -- | -- | -- |
W | quire: | "Assumption of Mary" | |||||||||||||
v | r | v | r | v | r | r | ? | v | v | r | v | r | r | v | v |
210 | 211 | 212 | 223 | 224 | 225 | 213 | 214 | 215 | 216 | 217 | 218 | 219 | 220 | 221 | 222 |
D2 | -- | D2 | C2 | -- | C2 | M1 | -- | M2 | M1 | -- | -- | N1 | N2 | -- | N1 |
The first two quires in the codex consist of a variety of Bunch of Grapes paper. I refer to one half of this watermark, the part with most of the grapes, as A. The complementary portion of the mark, containing a leaf and vegetable elaboration, I call B. As Allan Stevenson has instructed us, watermarks normally occur in slightly distinct pairs, as fraternal twins, so to speak, owing to the use of two moulds in paper production.[9] Such twins do appear in the Bunch of Grapes paper, and I have designated these companion marks AB1 (comprising A1 and B1) and AB2 (consisting of A2 and B2).[10]
Of the four possible watermark sequences, one that is immediately apparent in A quire is - A B - in ff. 9-12. These leaves constitute a single sheet, and ff. 10 and 11 are, in fact, a bifolium, and are mounted on a single guard.
As the reader may by now have surmised, the first step in identifying the symmetry in a watermark sequence is to locate the center of the quire. In B quire this is an easy task: ff. 30 and 31 surround the center of the symmetry in this gathering, with the watermark sequence projecting to the left and right of these leaves (i.e. toward the beginning and end of the gathering) in what might loosely be called a mirror-image pattern. As in A quire, the central four leaves seem to comprise a single folded sheet, and here too the innermost two leaves have been left uncut and mounted as a bifolium. But in this quire we find very strong evidence of multi-folding, for this is almost certainly the explanation of the watermark sequence in ff. 23-28 and 33-38. Clearly, three sheets were laid one atop the other and folded, producing a tripled version of the - A B - pattern. Folios 21, 22, 39, and 40 could have resulted from the folding of a single sheet; the gathering would in that event have been quired by folding one sheet, inserting into it three multi-folded sheets, and then inserting one folded sheet into the whole. But once again another arrangement is possible: all five sheets may have been multi-folded. The bottom sheet would in that case have produced ff. 21, 30, 31, and 40, the next sheet ff. 22, 29, 32, and 39, and so on. The mould-side pattern is congruous with either of these methods of quiring, and both are therefore possible.
Quires C, D, F—M (with the exception of ff. 95, 96, and 112), V, and W (except for the interpolated "Assumption of Mary" play on ff. 213-222) consist of paper with the YHS in a Sun watermark. "C" designates the upper portion of this mark, "D" the lower portion. Twins of the mark are indicated by C1 and C2, D1 and D2.[13] The conjugacy of leaves, as indicated by the watermark and mould-side sequences, is corroborated, incidentally, by several tests, including the relative size of the complementary portions of the mark on conjugate leaves and the direction in which the sun's rays point in the watermark design.
C quire is rather small, but it appears to be intact, as evidenced by the watermark and mould-side arrangements. I shall discuss the brevity of this quire later in the present study.
Quires D, E, and F must be considered together since they constitute a single quire which has been disturbed by an interpolation. E quire, ff. 51-52, appears in the midst of the "Marriage of Mary and Joseph" (Play 10), which
An interesting variation on the results of multi-folding appears in G quire. The asymmetrical sequence of watermarks in this gathering may at first seem confusing, but an obvious explanation suggests itself: the sheet which produced ff. 71, 72, 81, and a missing leaf following f. 81 sat head-to-tail with the four other sheets which comprise this quire, and all five sheets were multi-folded (cf. V quire, which was similarly formed). Since there are now nineteen leaves in the gathering, traditional collation would suggest that one leaf was missing somewhere. In order to locate the original position of that leaf, one need only complete the watermark symmetry in the gathering by supplying an unwatermarked leaf after f. 81; this creates a perfect mirror-image pattern extending from the center of the quire between ff. 76 and 77. Other explanations are possible (e.g. interpolation of one or three leaves), but there is no literary or bibliographical evidence to support these ideas. By contrast, the contention that a leaf was excised after f. 81 accords with the textual evidence, for the arrangement of the plays makes it possible that a blank leaf once followed f. 81: "The Trial of Joseph and Mary" (Play 14) ends on f. 81v and the next play, "The Nativity," begins on f. 82r. A blank leaf may well have originally stood between these plays, possibly owing to
As we have seen, once one has identified the symmetry of a watermark sequence, it is possible to locate the violations of that symmetry, and this can in turn help to discover the number of leaves which may have been lost from a gathering. In H quire, this approach may also help to explain one of the cruces of the cycle: the fact that there is no Play 17. In this gathering, C and D are again YHS in a Sun paper whereas G and H in ff. 95 and 96 designate the complementary portions of a Hand watermark.[17] Folios 95-96, along with f. 112 in the "Baptism" play, were interpolated after the text had been transcribed: they were written by a different hand from the main scribe's on paper which does not appear elsewhere in the manuscript. These leaves lack the normal rubrication of the codex, and f. 112, which is the first leaf of the "Baptism," is the only initial leaf of any play which does not have a play number; clearly this leaf replaced the original initial leaf at some point after the rubricator had numbered the plays, and ff. 95-96 were probably interpolated at the same time. The text of these interpolated leaves gives every sign of being faithful copies of lost originals, blending thematically with the material contiguous to it and conforming to the prosodic variety of that material. Moreover, both the "Magi" and the "Baptism" are incomplete without these interpolations.
The most important question about H quire is how many leaves have been lost, and from where? There are now, excluding ff. 95-96, fifteen leaves; one would normally expect either one or five leaves to have been lost. If only one leaf had been lost from the quire, it would have necessarily stood in the present location of ff. 95-96, since both the literary evidence and the watermark and mould-side patterns show that these folios replaced some earlier matter. But the watermark sequence does not square with the loss of only one leaf. It therefore seems probable that five leaves have been lost from the gathering, and this hypothesis is confirmed by the mould-side pattern of the fifteen leaves which have survived from the original quire:
In order to discover the original location of the five lost leaves, we must look first to the position now occupied by ff. 95-96, since, as noted above, these leaves evidently replaced earlier material. The watermark and mould-side sequences rule out the possibility that five or two leaves once stood there, leaving open the question of whether one, three, or four folios were originally present. Any conjecture beyond this must necessarily be uncertain, but one hypothesis fits so well with the evidence in the quire that it deserves mention, viz. four leaves have been removed after f. 91 and one leaf after f. 94. To begin with, it is most probable that one leaf is missing after f. 94 (as opposed to three or four leaves) because only one leaf's worth of material was copied onto ff. 95-96: these two leaves contain eighty-two lines of text and two one-line stage directions as against eighty-four lines of text on f. 94.[18] The loss of three or four leaves from this position is improbable since it would suggest that a significant amount of material had been removed from the middle of the "Magi," a play which seems complete as we have it with respect both to itself and to the other extant English Magi plays. If, therefore, one leaf is missing after f. 91, where could the other four lost leaves have originally stood? Several answers are possible, most of which involve a complicated pattern of excisions at unrelated points in the quire; but excisions of this kind obtain in no other quire, and are therefore unlikely to have occurred here. A more convincing solution is suggested by evidence within H quire that something has been removed after f. 91. Most significantly, the "Shepherds," which ends on f. 91r, is Play 16, and the "Magi," which begins on f. 92r, is Play 18, with no play in between. It seems safe to say that something, presumably a whole play, either followed or was intended to follow the "Shepherds." The watermark and mould-side sequences indicate that if one of the lost leaves originally stood after f. 94, as we have posited, then the other four leaves could have followed f. 91; this is perhaps the only arrangement, in fact, which accords with the literary evidence and does not scatter the missing leaves among different plays in locations which show no sign of textual disturbance. The four missing leaves would have provided space enough for a short play: the "Purification of Mary" (Play 19) takes up about that amount of space, for example. Indeed, if the scribe left the four leaves blank and went on to the "Magi" (and he often did leave space blank in this way, as noted above), the "Purification" may have been the very play which was originally intended to be number 17. Not only would it fit thematically between the "Shepherds" and "Magi," but there is also reason to believe that it was an addition to the original cycle, and that the change of
In J quire f. 112 probably replaced two watermarked conjugate leaves. Folio 111, the two missing leaves, and f. 113 comprised a separately-folded sheet, and ff. 105-110 and 114-119 were multi-folded (ff. 107-108 and 116-117 constitute a sheet which sat head-to-foot with the other two multi-folded sheets). Folios 103-104 and 120-121 appear to contain the first clearcut evidence that the main scribe ever quired with half-sheets, for the mould-side pattern of these four leaves does not fit with any of the ways by which whole
At this point it may be apposite to consider the implications of quiring with half-sheets. G. S. Ivy, in his excellent article "The Bibliography of the Manuscript Book," discusses two of the principal ways in which quires were formed: the "bifolium" method, in which a single sheet was folded once to yield two leaves, and the multi-folded method of simultaneously folding two large sheets.[20] As we have seen, in Vesp. D. viii. there is evidence of further methods of forming quires: the double folding of single whole sheets and the multi-folding of three or more whole sheets; the single folding of a half-sheet can be considered a variety of the bifolium method. As Ivy says (p. 39), once the leaves of a membrane codex have been cut, it is not usually possible to assert whether the bifolium or the multi-folded method has been used. But in Vesp. D. viii. we see that watermark and mould-side sequences in paper manuscripts can indicate that whole sheets were employed, and that many of these were multi-folded. In one sense, the bifolium method was probably always used in this codex, since after the sheets were folded (individually or in piles), they were presumably unfolded and cut along the fold parallel to the shorter side. The resulting two stacks of bifolia were then placed in their original positions. If, however, one understands the bifolium method to mean the formation of quires by piling non-conjugate half-sheets in random arrangement, then it is unlikely that the main scribe used this method in the codex very often. The watermark sequences in quires B, G, T, and V, for example, are unlikely to have been arrived at by chance; rather, they almost certainly resulted from the multi-folding of whole sheets.[21] Also implausible is the idea that unrelated half-sheets were randomly gathered to produce mould-side patterns like v v v v r r r r v v v v r r r [r] in D/F quire, r r r r v v v v r r r r v v v v in S quire, and v v v v v r r r r r v v v v v r r r r r in V quire—all of which are consonant in addition with watermark sequences which result from folding whole sheets. In sum, the main scribe of Vesp. D. viii. appears normally to have folded whole sheets, either individually or in piles. Half-sheets may possibly have been used sometimes, but it seems significant that this question arises with reference only to those quires which have evidently experienced some physical disruption: every gathering the main scribe formed which has escaped such disruption appears to have been quired by folding whole sheets of paper.
Consider, for example, R quire. The mould-side sequence in this gathering indicates textual disturbance or quiring by half-sheet, or both. Before one attempts to choose among these options it is helpful to observe that R
P and R quires, then, may have originally been one gathering which was split by the interpolation of Q quire, just as D/F quire was almost certainly divided by the intrusion of E quire. If P and R were one quire, and if they were formed by folding whole sheets, then the watermark and mould-side patterns indicate that further substantial alteration of the quire must also have taken place. Alternatively, if it was quired with half-sheets, only two leaves following f. 163 (the leaves which would have been conjugate with ff. 152-153) need have been displaced. At the least one can say that these quires have been in some way physically disturbed, and it may therefore be impossible to reconstruct how they were formed. A similar arrangement obtains
Looking briefly at the remainder of the codex, one may observe that the two gatherings which constitute Passion Play 2, S and T quires, appear to comprise four multi-folded sheets and two multi-folded sheets respectively.[25] Folios 184-185, the chain-spaces of which correspond to the Bunch of Grapes or YHS in a Sun paper, were interpolated into T quire as part of the compilation of the cycle.[26] V quire comprises five multi-folded sheets of YHS in a Sun paper (ff. 194-195 and 204-205 sat head-to-foot with the other sheets in the pile). Finally, W quire has lost two leaves, one before and one after the gathering; the latter leaf probably contained the conclusion of "Doomsday" (Play 42). More leaves may, of course, have been lost before and after this quire, but there is no literary or bibliographical evidence that this happened. Folios 213-222, the "Assumption of Mary" (Play 41), were interpolated into W quire: these leaves are a different paper from that found elsewhere in the codex, and the play is written in a different hand.[27] The outer leaves of this play comprise a half-sheet, and since f. 213 is a flyleaf, this scribe may have quired in this instance with a bifolium. But it is also possible that the two leaves which were conjugate with ff. 213 and 222 were also once employed as flyleaves for this play; such an arrangement was quite common in small independent medieval texts, which required multiple flyleaves to protect them until they were bound into larger volumes (Ivy, pp. 53-54).
In sum, bibliographical analysis of the paper in Vesp. D. viii. illumines some of the steps in the compilation of the codex. By examining the overall patterns in the evidence, one may draw reasonably confident inferences about the main scribe's habits of quiring. First of all, it should be said that the present quire-marks appear to be accurate. But one can look back to an earlier stage and reconstruct the physical make-up of the manuscript before the many interpolations and excisions. As we have seen, quires A, B, possibly D/F, H, J, and V each appears originally to have contained twenty leaves. Significantly, all but three of the remaining quires were evidently additions to the original text.[28] It therefore appears that the main scribe's normal practice as he transcribed the bulk of the manuscript was to quire in twenties,
Before concluding I should like to note that analysis of watermark symmetries is not restricted to the quarto format. In folios, for example, one may expect symmetries to obtain, radiating out from the centers of gatherings in inverted and complementary patterns. For instance, if a quire begins with three marked and two unmarked leaves, it should end with two marked and three unmarked leaves. In practice, the symmetries may be fairly complex; e.g. a gathering which begins with three watermarked leaves, then one unmarked leaf, then four marked, one unmarked, and three marked leaves may be expected to conclude with the inverted complementary pattern of three unmarked leaves, then one marked, four unmarked, one marked, and three unmarked leaves. This pattern actually occurs in Harley MS 2013, one of the five cyclic versions of the Chester mystery plays. Transcribed in the year 1600, this folio codex contains no signatures or quire-marks, although there is foliation, presumably by the scribe, George Bellin. For the present I wish only to demonstrate that the quiring of this manuscript can be discovered by reference to watermark and mould-side sequences. I shall therefore represent these patterns in an arbitrary number of leaves, say, ff. 1-70:[32]
I have proposed in this study a method of analyzing symmetries, and violations thereof, in the watermark and mould-side sequences of paper codices. I have tried to show that this kind of analysis can be an essential tool in detecting and locating interpolations and excisions, and in disclosing the history of literary documents. It seems fair to say in view of the findings of the present study that examinations of the material evidence in many paper manuscripts should now be reconsidered.[35]
| ||