University of Virginia Library

Search this document 


expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
collapse section 
Chainlines Versus Imposition in Incunabula by Curt F. Bühler
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 

expand section 


Page 141

Chainlines Versus Imposition in Incunabula
Curt F. Bühler

Bibliographers[1] have almost unanimously agreed that the descriptive terms Folio, Quarto and Octavo[2] refer to the number of times the original sheet of paper had been folded. The sheet folded once, with the chainlines vertical and the watermark in the center of one of the resulting two leaves, forms a folio (2°); folded twice, the chainlines horizontal and the "filigrane" in the middle of the fold, is a quarto (4°). Folded a third time (octavo = 8°), the chainlines are again vertical and the watermark will appear at the top of the inner margin in four leaves.

It is also normal to assume that, in a volume with folio leaves, the actual printing (machining) provided two pages on each side of a sheet; in a quarto, four on the same side of each sheet; and in an octavo, eight.[3] There cannot be much doubt that this is true for the vast majority of cases — but it is not always so.[4] Such abnormalities must be taken into consideration whenever the examination of a volume discloses problems which cannot be explained by an analysis based on the norm.

When we come to editions described as "Folio and Quarto." or "Quarto and Octavo" are we to believe that some sheets (as in the first instance, for example) were printed as folios and others as quartos? If this were true and since all leaves must necessarily be of the same size, these quarto leaves must have been printed on a very large press (quite different from that used to print the folio leaves) with paper twice the size of that used for the folios.[5]

Let us examine a few instances. The Sermones quadragesimales de poenitentia by Robertus Caracciolus, printed in two editions by Franciscus Renner at Venice in 1472, has in both cases the collation: [a-i10 k6 l-t10 u12 x-z10 A-C10 D12]. Both editions are described as "2° und 4°" by the Gesamtkatalog


Page 142
der Wiegendrucke.[6] The copy of the earlier edition (GW 6062)[7] in the Bayerische Staatsbibliothek in Munich (2°. Inc. c. a. 106) has the following distribution of paper: [a], folio; [b-g], quarto; [h], quarto = 1, 2, 9, 10 + folio = 3-8; [i-k], folio; [l-q], quarto; [r], quarto = 1, 2, 9, 10 + folio = 3-8; [s-u], folio; [x-A], quarto; and [B-D], folio. The identical distribution of paper occurs in the copies in the British Museum (IB. 19827), Walters Art Gallery and the Cambridge University Library (Oates 1653).[8] Renner's second printing (GW 6063), however, provides an interesting clue. Here the leaves in quires [a-m] are all quarto, those in [p-D] are all folio, with quires [n-o] sometimes mixed. In the Munich example (2°. Inc. c. a. 107), these two quires have leaves 1 and 10 in quarto, the rest are folio. The Cambridge copy (Oates 1654) has quire [n] as the Munich one, while the example in the Pierpont Morgan Library (PML 20486) has the entire quire as folio. Both the Cambridge and the Morgan copies have all quire [o] in folio. In short, certain leaves are found either as folios or quartos,[9] though close examination shows that they are otherwise identical.[10] Surely it seems unlikely that the printer would re-arrange his formes for these leaves, and one may presume that they were all machined in the same manner.

From this analysis, it seems highly probable that the entire book (as well as the earlier edition) was printed as a folio — and that the quarto leaves are merely the result of double-sheets having been cut in half[11] and


Page 143
printed with two pages to each side.[12] Thus, quarto sheets[13] can be used for printing by "folio-imposition".[14] In effect, any large-sized quarto could actually be a folio, when considered from the printer's point of view as opposed to the bibliographer's.

Even more curious cases — and perhaps more enlightening — are those of certain mixed 4° and 8° volumes, in particular two undated publications (probably of the late 1470s) by Johannes Schurener in Rome: Modestus, De re militari, and Solinus, De mirabilibus mundi. So far as the Modestus is concerned, my own copy, which collates: [a8 b12 c10], agrees with that in the British Museum (BMC IV:59 — IA. 17768) in the fact that "the inner sheet of quire [a] and the two outer sheets of quire [b] are quarto". But in quire [a], my copy has a watermark in only the first and second leaves,[15] the wrong two leaves since in full octavo folding one expects to find the watermarks in either of these four leaves: 1, 4, 5 and 8 or 2, 3, 6 and 7.[16] Furthermore, one of the two inner sheets is quarto and the other is octavo. They could never, therefore, have been conjoined. In quire [b], a watermark appears in my copy only on [b]4 (an octavo leaf), showing that this sheet was never attached to any other sheet(s) in this particular signature. In quire [c], a watermark appears only on the sixth and seventh leaves — but not in any other. These facts would seem to preclude the possibility that the machining was carried on in such fashion as to print either four or eight pages on one side of a sheet of paper and then binding these sheets together.


Page 144

The Solinus (collation: [a4 b-i10 k-l8 m-n10])[17] offers even further complications. The British Museum's description (BMC IV:59) of its two copies points out that both have the second and third sheets of quire [l] as quarto and that, in IA. 17771, the outer sheet of quire [a] is also quarto, though octavo in IA. 17772 (G. 8938). So far as concerns quire [l], it is instructive to consider the Cambridge University Library copy (Oates 1410), for in this volume only the second sheet is quarto.[18] Accordingly, only a single sheet of quarto is found in a signature of eight leaves, otherwise octavo. One wonders where the other half of this sheet, if printed, is to be found.

A close comparison of the quarto and octavo leaves of [a]1 and [a]4 indicates that these were printed from the same setting of type.[19] Further, IA. 17771 proves conclusively that the first quire must have been machined as two separate sheets, with only two pages printed on each side. In short, these seem to have been printed on rather a small press with imposition as in folio.

The same "mixed foldings" can also be found north of the Alps. As a characteristic example, one may cite the Psalterium Germanicum, [Strassburg: Heinrich Eggestein, c. 1475 = Goff P-1074]. The collation is somewhat complicated, and BMC I:74[20] suggests the formula: [a-f10 g8+2 h8+1 i-k10]. This does not, however, tell the whole story, since Eggestein apparently had considerable difficulty with the imposition. Actually, in quire [c], leaves 7 through 10 are mounted on stubs in both the Morgan (PML 26106) and the Scheide copies,[21] and so also is [h]9 in the former copy (wanting in the latter). The Scheide (though not the Morgan) example also has [b]10 mounted on a stub.


Page 145

In the Morgan example, quires [a] and [i] are octavo, all the other leaves being quarto except [k]4.7 (octavo).[22] But this sheet is also quarto in the Scheide volume! The watermarks also supply interesting evidence. In the octavo quire [a] of both copies, watermarks appear only on leaves 2 and 4 of PML 26106 and on 3 and 5 of the Scheide copy. This would be impossible if the sheet had been printed with the usual eight pages on one side of the sheet. Again, in the Morgan quarto quire [d], watermarks appear on eight of the ten leaves (only 3.8 being without a mark), an impossibility if printing had been carried on in the normal quarto manner. There seems to be no other explanation available but to assume that the printing proceeded on half-sheets of normal paper and quarter-sheets of the double paper, but in each case with only two pages on each side of every sheet (that is, as in folio printing). It thus became possible for the pressman to print the entire inner or outer forme with a single pull of the lever.[23]

The material set forth above[24] leads to the conclusion that "format" for the fifteenth-century printer occasionally represented something quite different from what it does to the twentieth-century bibliographer. The researcher must needs be wary of traps of this sort which may lie in wait for him; that he cannot, for example, judge the number of pages originally printed on one side of a sheet of paper by the position of the chainlines and the watermarks.



See the following: Catalogue of Books Printed in the XVth Century now in the British Museum (1908-62), I, xviii; Konrad Haebler, Handbuch der Inkunabelkunde (1925), p. 40; Gesamtkatalog der Wiegendrucke (1925-40), III, xxii; Ronald B. McKerrow, An Introduction to Bibliography for Literary Students (1928), p. 165; Fredson Bowers, Principles of Bibliographical Description (1949), pp. 193-194; W. W. Greg, A Bibliography of the English Printed Drama to the Restoration (1939-59), IV, lv-lvi; etc.


These are the only such designations frequent among the incunabula.


McKerrow, p. 29 ff.


See, for example, McKerrow, p. 174.


For such extra large sheets, see Haebler, p. 39; Bowers, p. 194; and Edward Heawood, "The Position on the Sheet of Early Watermarks," The Library, 4th Ser., IX (1929), 47.


So BMC V:191; M.-Louis Polain, Catalogue des livres imprimés au quinzième siècle des bibliothèques de Belgique (1932), no. 982; Indice generale degli incunaboli delle biblioteche d'Italia (1943-65), nos. 2470-71; J. C. T. Oates, A Catalogue of the Fifteenth-Century Printed Books in the University Library Cambridge (1954), nos. 1653-54; but Marie Pellechet, Catalogue général des incunables des bibliothèques publiques de France (1897-1909), no. 3246 simply states "in-fol."


The collation itself suggests that the book was not printed with four pages to a side of the sheet, since a large majority of the quires have ten leaves, which would have required two quarto sheets and a half-sheet. This would have caused awkward and unnecessary problems that would have been avoided with signatures of eight leaves.


"The book was thus printed in three sections, the first two of ten quires each, the last of seven quires" (BMC V:191). But quire [a], with the "tabula," would surely be the last printed, and the fact that the paper is folio (while the other early quires are quarto) suggests that it was actually machined after the folio signatures [B-D]. The distribution in the second edition may well suggest that Renner was using two presses and that two quires were always being printed simultaneously. Further, that the change in paper occurred when each press was printing the seventh quire allotted to it.


See also Pope Pius II, Dialogus de somnio quodam, Rome: Johannes Schurener, 11 Sept. 1475, which has (according to BMC IV:57) the inner sheet of quire [g] as quarto, the rest being folio. However, the Huntington, Morgan and Library of Congress copies are folio throughout.


The abbreviations and spacings are the same throughout, as also broken characters.


As Bowers points out (p. 194), cutting must be "treated as if it were a folding of the sheet."


On such half-sheet imposition, see Dennis E. Rhodes, "Variants in the 1479 Oxford Edition of Aristotle's Ethics," SB, VIII (1956), 209-212; Curt F. Bühler, "The First Edition of Ficino's De christiana religione: a Problem in Bibliographical Description," SB, XVIII (1965), 248-252; etc.


The leaf in the Museum's copy of GW 6062 measures 292 x 211 mm., so that the sheet must have been greater than 422 x 292. The double sheet must then have been greater than 600 x 420 mm. These figures may be compared with the size of the "forma regalis" (700 x 500 mm.) and the "forma mediana" (500 x 300 mm.) as cited by Haebler (p. 39). The average size of the Schurener quarto sheets was 210 x 150 mm., which means that the full sheet measured more than 420 x 300 mm. and the double-sheet in excess of 600 x 420 mm. These very large sheets were used in the printing of the over-size incunables. The sheets which the R-Printer used in Strassburg for the printing of the Specula of Vincent of Beauvais must have been greater than 646 x 475 mm. (BMC I:65), and paper of that size was also used by Peter Schoeffer in Mainz (BMC I:27). Judging from the measurements of the Morgan copy of the Golden Legend (PML 780), one deduces that Caxton used a sheet greater than 520 x 388 mm. William Blades (The Life and Typography of William Caxton [1861-63], II, xvii) asserts that Caxton's largest sheet measured approximately 22 x 16 inches (about 560 x 405 mm.).


Bowers, p. 194, and Greg, p. lvi, suggest the notations: (2°-form) 4° and (4°-form) 8°.


Actually, one-half of the same watermark appears in the upper margin, somewhat nearer the fold than the center.


McKerrow, p. 167.


Again this is a very curious (and suspicious) quiring for either quarto or octavo imposition.


So also in the copy belonging to the Philip H. and A. S. W. Rosenbach Foundation. In this copy, too, only a single watermark appears in signatures [c], [m] and [n], an impossibility with octavo printing. The copy in the Yale University Library now collates [a-v4.8], but this is a quiring which has been imposed on the book by a binder. The leaves in the Yale copy correspond to the following leaves as they left Schurener's press:

  • Original quiring Yale copy
  • a4 a1-4
  • b10 b1-8 c1+2
  • c10 c3+4 d1-8
  • d10 e1-4 f1-6
  • e10 f7+8 g1-4 h1-4
  • f10 h5-8 i1-4 k1+2
  • g10 k3-8 l1-4
  • h10 m1-8 n1+2
  • i10 n3+4 o1-8
  • k8 p1-4 q1-4
  • l8 q5-8 r1-4
  • m10 s1-8 t1+2
  • n10 t3+4 v1-8


Both have the same broken a in "suffragiū" (l. 9) and the b out of line in "ber" (l. 11) of [a]1 recto, and the same misprint "et ‖ et" (a repeat) in lines 4-5 on the recto of the fourth leaf.


The catalogue provides the additional note: "Quires b, c, and g appear to be made up."


I am grateful to Mr. William H. Scheide for permitting me to examine his copy at my leisure.


If printed as a normal quarto with a single sheet inserted, then the make-up would have been either ∣, with 5.6 inserted, or ∣, with 3.8 inserted. With normal printing, the insertion of 4.7 would have been impossible.


Compare McKerrow's observation (p. 61): "I may perhaps surprise some bibliographers by saying that always until about 1800 a normal full-sized forme of type was printed by two pulls of the lever." See also my "Caxton Studies," Gutenberg Jahrbuch 1940, pp. 169-176.


Heawood (p. 45) discusses a further complication in that, on rare occasions, the chain-wires were "placed longitudinally in the mould, giving horizontal lines in a folio book, vertical in a quarto, and so on." Compare also my "Caxton Studies."