François Déroche, with contributions by Annie Berthier, Marie-Geneviève Guesdon, Bernard Guineau, Francis Richard, Annie Vernay-Nouri, Jean Vezin, Muhammad Isa Waley
As a specific form of handwritten book, the codex is directly linked to the concept of ‘gatherings’, defined as ‘group[s] of leaves or sheets brought together to form part of a book.’[1] More commonly used today is the term ‘quire’, which originally meant four sheets of folded parchment or paper but now refers to a gathering of any number of sheets, folded into leaves set one within another and sewn together. Other key terms that require definition here include ‘bifolium’, a rectangular piece of writing material folded down the middle so that it forms two leaves or ‘folios’ (abbreviated as ‘f.’ or ‘fol.’; plural ‘ff.’ or ‘foll.’), each folio having two sides, called ‘pages’ (‘p.’, plural ‘pp.’). In order to distinguish the two sides of a folio, the first side to be read is called the ‘recto’ (abbreviated ‘r’, ‘ro’, or ‘a’) while the second is known as the ‘verso’ (‘v’, ‘vo’or ‘b’) illus. 16). Often, only ‘verso’ is explicitly indicated, the absence of specification meaning that the recto is intended; this is the system that will be followed in this book. Lastly, the term ‘double page’ refers to the verso of one folio opposite the recto of the next in a manuscript lying open. In manuscripts written in Arabic or other Semitic characters, the left hand page is the recto.
Basic concepts
As mentioned above, a quire is composed of bifolia, that is to say rectangular sheets of parchment or paper folded in two, forming two leaves of equal size.[2] The folding operation represents a crucial stage in the making of bifolia, and can be performed on a single piece of writing material or on a stack of pieces cut to the same dimensions. In the Middle Ages, the West took this technique further, developing a method that was used frequently - though not exclusively - to produce quires from a single sheet of parchment or paper.[3]
Depending on the desired format, the sheet would be folded only once (folio format), or twice (quarto), or three times (octavo), and so on. The number of folds determined the format and number of leaves in the quire. With one fold, a gathering comprised just one pair of leaves; two folds produced two pairs (termed a binion); three folds produced four pairs (a quaternion); and so on. All comprised an even number of bifolia, since they were obtained by folding a single sheet. The paper or parchment had then to be cut in all those places - apart from the central fold - where two leaves remained attached to one another.
Types of quires
It is entirely possible to assemble quires containing two, four, or eight leaves, etc., in various formats, without employing the folding method on a sheet of parchment or paper; likewise, other types of quires exist which contain an odd number of bifolia. To clarify matters, the various combinations have been set out in the chart below, which includes the traditional names for the various types of fold. It is important, however, to note that such gatherings can also be produced by assembling several sheets.
| Number of bifolia | Type of fold | Name given to quire type | Number of folios |
| 1 | folio | 2 | |
| 2 | quarto | binion | 4 |
| 3 | ternion | 6 | |
| 4 | Octavo | quaternion | 8 |
| 5 | quinion | 10 | |
| 6 | duodecimo | senion | 12 |
| 7 | septenion | 14 | |
| 8 | extodecimo | octonion | 16 |
Beyond eight, we simply refer to gathering of nine, ten, etc., bifolio. Some of the quires listed in the chart above cannot obtained simply folding – a point to be discussed later. On the other hand, all these quires can be called ‘regular’ because they are made up of complete bifolia.
Anomalies
It sometimes happens that a gathering has had one or several leaves added or removed, in such a way that these leaves have no counterpart and are therefore ‘independent.’[4] To add a single leaf - or ‘singleton’ - to a quire, a scribe could employ one of two techniques. The first involved taking an independent leaf whose width was slightly greater than that of the other leaves in the quire, so that folding it down to the exact dimension left a narrow strip called the ‘stub’ (illus. 17). The leaf could then be added to the quire simply by sewing the overlapping stub into the gathering by the normal method, thus making it an integral part of the quire. Examination of the manuscript will reveal this overlapping tab, which is not the remnant of a leaf that has been torn out but the tab of a singleton: for example, in the second quire of a copy of the Ɉāmi‘ al-fuṣulayn (Paris BNF arabe 69o5), folio ɪ5 was glued by its stub to the bifolium of ff. ɪ4-ɪ6.[5]
On occasion, several stubs appear side by side where a number of leaves have been added to a quire. Furthermore, one or several singletons can be stitched together into one complete gathering.[6] The use of parchment involved almost routine recourse to these techniques, as will be discussed later in greater detail.
Another method for inserting an independent folio entailed use of a ‘guard’ (illus. 18). This term refers to a narrow strip of paper or parchment folded down the middle and slipped into a gathering, a separate leaf having been already glued to one (or both) wing (s) of the guard, thereby incorporating the leaf or leaves into the quire. Guards are also used to restore manuscripts when the gutter of a quire is badly damaged: the folios are attached to guards in order to reconstruct quires.
When one leaf of a bifolium is removed (without being torn out), the survival of its conjugate leaf is threatened; for that reason a stub is left by cutting the leaf along the gutter margin near the spine-fold. Hence whenever a stub is detected in a manuscript, it is essential to make sure that the text contains no gaps. If a lacuna appears in the spot corresponding to the stub, it must be concluded that the stub is the vestige of a now-missing folio.
Quires within a manuscript
The number of quires making up a manuscript naturally varies considerably. In some cases there may be only one, often somewhat larger than an ordinary gathering.[7] The term ‘monobible’ might be applied to manuscripts of this type.[8] It is nevertheless extremely rare, in the field under discussion, to find single quires in which the number of folios greatly exceeds the norm. One manuscript, Berlin Staatsbibliothek (henceforth SB) Sprenger 5ɪ7, datable to shortly before 459/ɪo66-ɪo67[9], comprises a single gathering of no fewer than forty folios; it may be of Indian provenance. This brings to mind the unusual case of ‘monobibles’ that seem to have been produced in northwestern India following contact with Islam; one manuscript dating from ɪ77o, Munich Bayerische Staatsbibliothek, (henceforth BSB) cod. hind. 6, is composed of 28o bifolia in a single gathering.[10]
The fold constitutes a weak point in the architecture of a codex. Their concern to prevent it from tearing sometimes led makers of books to reinforce it by inserting a strip of paper or parchment in the fold of the quire either in the centre fold or around the spine - which was then stitched along with it. This strip is also called a ‘guard’.[11] The above-mentioned MS. Sprenger 5ɪ7 provides one of the few known examples of the use of this technique in an Islamic manuscript, perhaps explained by the thickness of the one and only quire composing it. Mourad Rammah mentioned to the present writer finding a quire that employs the same system in a juridical manuscript dated Jumādā II 4o4/ December ɪoɪ3 (Museum of Islamic Art, Raqqada, Tunisia, MS. Rutbī 247).
Examining quires
Medieval scholars knew how important it was to observe closely the state of the gatherings in a manuscript.[12] When examining a manuscript, codicologists must inspect its composition meticulously. Any anomalies may indicate that text has been added, shuffled, or removed.[13] It is worth pointing out once again that this examination must be conducted with great care to avoid damaging the book, especially if a tight binding makes it impossible to inspect the centre folds.
If any irregularities are found -for example, a quire of four folios in a manuscript where all other quires are quinions, or a quire composed of an odd number of folios - the text must be checked for possible lacunae and/or transpositions of folios. An anomaly may also reflect a deliberate act on the part of the copyist. He may have noticed, when nearing completion of the task, that the remaining leaves of the quire on which he was working were insufficient to contain the rest of the text (which did not, however, call for an entire additional gathering); one or two leaves might therefore be inserted in the quire in order to gain the space required. Irregularities at the end of a manuscript often stem from situations of this kind.
The type of quire in a manuscript can be identified from the numbering, if any, to be found in it. Such indications, sometimes located on the recto of the first folio of each quire, will help determine whether the gatherings are quaternions or quinions, etc. When there are no such clues, the first task is to locate the stitches of the sewing, which are found in the fold of the central bifolium of each quire. Once the stitching for one quire has been located, the search should be extended to neighbouring quires. By counting the number of leaves between two contiguous sets of stitches, the number of folios in each quire can, in theory, be easily determined if the number remains constant across several intervals of this type. For example, if sewing is noted between ff. 36-37, 46-47, 56-57, 66-67 and 76-77, then there are ten leaves between each set of stitches: five belonging to the second half of the one gathering, and five to the first half of the next. If the binding permits, thorough verification of this pattern must always be done, examining eυery quire for the number of folios it contains. This process is essential to discovering manuscripts - admittedly rare - in which quires of different types alternate regularly. Mid-quire signatures can also be used as a guide to counting, along the same lines as stitching. In manuscripts that have been restored in the West, it sometimes happens that two sets of stitching are found within a single quire, as for example in Paris manuscript BNF arabe ɪ544 (between ff. 7 and 8, then again between 8 and 9, as well as between ɪ6 and ɪ7 and then again between ɪ8 and ɪ9).[14] Up to now, no other examples of this practice have been found in manuscripts handled by Oriental binders of the early period. In manuscripts re-bound in the West, it is not unusual to find that the original structure has been entirely lost, the only remaining evidence being the old sewing holes.
When examining parchment manuscripts, there are additional details that must be noted, as will be explained in a later section
Describing quires
There are simple formulas which make it possible to describe quires and to draw attention to significant details. In one formula,[15] which can be applied very widely, gatherings are described as a sequence of numbers: the first, in Arabic numerals, indicates the number of gatherings with a regular number of bifolia, the number of bifolia being indicated by a Roman numeral; these two numerals are followed by the number of the last folio of this homogeneous sequence, in parentheses. A volume of ɪoo folios comprising exclusively quinions will thus be described in the following way: ɪo V (ɪoo). Another manuscript, composed of six quinions followed by a final quaternion, will be described as 6 V (6o), IV (68); if a leaf is lacking, for example, from this last quire, it will be indicated as follows: 6 V (60), IV-ɪ (67). If, on the other hand, an addition has been made to that final quaternion, it will be indicated thus: 6 V (60), IV+ɪ (69). In the case of a sole quire, it is not necessary to put the number ɪ before the Roman numeral.
Once again, when manuscripts are made of parchment, the specificity of the medium sometimes requires a more detailed description of quires, as will be explained in the next section.
Quires within parchment manuscripts
Examining quires of parchment
In addition to the examination procedures described above, in manuscripts written on parchment it is important to note the sequence of hair sides and flesh sides (see illus. 10a and 10b). The side that forms the recto of the first folio of a quire is known as the outermost side. If f. ɪ ro of a parchment manuscript is the hair side - assuming it is indeed the original first folio - then the quire would be described as having the hair side outermost.
Before taking a closer look at the way in which Islamic scribes composed quires of parchment, it may be useful to mention again, briefly, the folding technique used by Western scribes in the Middle Ages. To form a gathering, they generally folded the skin in half either once, twice, three, or four times, thereby producing quires of two (folio format), four (quarto), eight (octavo), or sixteen sheets (sextodecimo).[16] But this system of folding had another consequence, one defined by Gregory’s Law, so called after a German scholar who was the first to notice that, within a given quire, two facing folios always displayed the same type of side, hair or flesh.[17] Thus, if folio x vo is the hair side of the parchment, the recto opposite that page (f. x + ɪ ɪo) will also be the hair side.
The distinction between the two sides can be annotated in a simplified way according to the following system for describing gatherings of parchment folios. Quires are numbered in Roman numerals; the folios are given the number they bear in the manuscript in Arabic numerals, flanked by small capital letters ‘H’ (hair) and ‘F’ (flesh) - or conversely - to indicate the nature of the recto and verso sides. The middle of the gathering is marked by a slash (/), the presence of a stub by the letter S. Other letters may be used to provide additional information, such as L for a folio that is lacking. The description of the fifth quire of a manuscript might then be noted as follows:
V: H39F H40F H41F H42F S H43F/F44H F45H S F46H F47H F48H
If a manuscript is paginated, rather than foliated, this descriptive method need merely be adapted by using a hyphen, for example, to link the pages comprising the recto and verso of a folio, thus giving:
V: H77-78F H79-80F H81-82F H83-84F S H85-86F/F87-88H F89-90H S F91-92H F93-94H F95-96H
Early Qur'ān manuscripts
Copies from the first/seventh and second/ eighth centuries
The oldest surviving Arabic manuscripts are Qur'āns, and date from the second half of the first/seventh century; most of them are fragments written in Ḥijāzī-style script, which provides the basis for the dating. Few of these copies contain continuous sequences of folios, which are essential to understanding how parchment was used to make up quires in those early days. Qur'ānic fragment Paris BNF arabe 328a, apparently copied in the late first/seventh century,[18] nevertheless features several groups of leaves bearing uninterrupted text: folios 4 to 22, 23 to 40, and 4ɪ to 48, to which might be added ff. 57 to 64: these latter form Fragment 328b, which is a distinct item from the palaeographic standpoint but may conceivably have been part of the same Qur’ān.[19] In the absence of extensive codicological examination, owing to the present physical condition of the manuscript, the following analysis is offered on a conditional basis: the fragment contains four quaternions, namely ff. 7-ɪ4, 24-3ɪ, 32-39, and 57-64, to which might be added the quire of ff. 42–48 (that is to say, seven leaves), from which the original last folio is missing. Folios ɪ5-2ɪ, in contrast, simply represent an irregularity. The sequence of the sides of parchment is also interesting, as for example in ff. 7-ɪ4:
F7H H8F F9H H10F/F11H H12F F13H H14F (illus. 19)
Since this arrangement occurs elsewhere, it might be thought that this manuscript represents an example of quires produced by the folding method. But two of the quaternions tend to refute this hypothesis: the bifolia of conjugate pairs 43-48 and 44-47, as well as those of 59-62 and 6o-6ɪ, present flesh sides opposite hair sides, which constitutes a strong argument against folding. In any event, the flesh side is regularly used as the outermost side of the various gatherings.
It should not be assumed, however, that quaternions of this type were the only model followed at the time - another fragment in Ḥijāzī style, Paris BNF arabe 328c, is composed of quinions arranged in the normal way (that is to say, with the flesh side on the recto of all the folios in the first half). The general approach will only be understood once Ḥijāzī Qur’ān manuscripts are better known. For the moment, we can merely note the relative diversity in the use of parchment during this period, the late first/seventh and early second/eighth century. The number of folios per quire in no way seems unusual; in contrast, the use of the flesh side as outermost side would appear to deviate from the norm observed among the already published collections. It should nevertheless be noted that the flesh side is also outermost in quires in Greek manuscripts.[20]
From the early second/eighth century, a Qur’ānic fragment that displays palaeographic evolution from the Ḥijāzī style is noteworthy for its quires of ten bifolia, assuming that the reconstruction proposed here is accurate. The arrangement of the sides of the parchment is as follows:
HɪF H2F H3F H4F H5F F6H F7H F8H F9H F10H/H11F F12H F13H F14H F15H F16H F17H F18H F19H F20H.
Qur’āns from the third/ninth century...
*The remainder of this article is exclusively available in the printed version of the related book. The book is available in both electronic and printed formats within Our Publications in the following link:
http://doi.org/10.56656/100099
[1] The Ne𝓌 Shorter Oxford Dictionary (Oxford: OUP, 1993).
[2] A leaf is defined as one half of a bifolium, and should be distinguished from a page, which is one of the two sides of each leaf.
[3] As M. Maniaci has pointed out, the model suggested by L. Gilissen (Prolégomènes à la codicologie) does not invariably apply to Byzantine manuscripts. Maniaci, ‘L’art de ne pas couper les peaux en quatre: les techniques de découpage des bifeuillets dans les manuscrits byzantins’, Gazette du liυre médiéυal 34 (Spring 1999), 1–12.
[4] Muzerelle, in Vocabulaire, p. 92, defines a feuillet indépendant as ‘a folio lacking its conjugate pair, i..e., the other half of the bifolium to which a given folio belongs.’ But the hypothesis advanced here is that the presence of a singleton does not necessarily indicate the loss of half a bifolium, since the leaf may have been separate from the start.
[5] FiMMOD 274. See the chapter ‘The Writing Surface: Paper’.
[6] See below.
[7] An excellent example is provided by MS. Paris BNF suppl. turc 986, a miscellany of fascicles (juz’), each composed of a single gathering. As a rule, these quires are larger than those usually found in manuscripts composed of several gatherings: the second quire numbers twenty-two folios (ff. 19–41), not counting a bifolium of parchment that is in fact a cover. See G. Vajda, ‘Trois manuscrits de la bibliothèque du savant damascain Yūsuf ibn ‘Abd al-Hādī’, ɈA 270 (1982), pp. 229–256 (reprinted in La transmission du saυoir en Islam, 1983). See also P. S. Van Koningsveld, The Latin-Arabic Glossary of the Leiden Uniυersity Library (Leiden, 1976) p. 69, note 89, which mentions Jerusalem manuscript JNUL Yahuda MS Ar. 409; a comment in this direction in The Qur’an, Scholarship and the Islamic Arts of the Book (London, [1999]), p. 36.
[8] Muzerelle, Vocabulaire, p. 60, suggests monobible as the French term.
[9] W. Ahlwardt, Verzeichnis der arabischen Handschriften, vol. II, pp. 249–250, no. 1557.
[10] MUNICH 1982, p. 234 and fig. 48.
[11] See the entry under ‘guard’ in Michelle Brown, Understanding illuminated manuscripts: a guide to technical terms (Malibu, CA/London, 1994). In French, this type of guard is known either as a préserυateur (Lemaire, Introduction, p. 43) or fond de cahier (Muzerelle, Vocabulaire, p. 98).
[12] F. Rosenthal, The Technique and approach of Muslim scholarship (Rome, 1947), p. 12.
[13] An illustration of this effect on parchment manuscripts will be given below.
[14] Sauvan and Guesdon, Cat. 5, p. 88.
[15] Witkam, Cat. 1, pp. 13–14
[16] Gilissen, op. cit., pp. 26–35; Lemaire, Introduction, pp. 69–94; see also the discussion in Maniaci, op. cit. The following comments partly reproduce a published article: François Déroche, ‘L’emploi du parchemin dans les manuscrits islamiques: quelques remarques liminaires,’ Codicology, pp. 27–40.
[17] G. R. Gregory, ‘Les cahiers des manuscrits grecs’, Comptes rendus des séances de l’Académie des inscriptions et belles-lettres (1885), pp. 261–268.
[18] M. Amari, ‘Bibliographie primitive du Coran’, H. Dérenbourg (ed.), in Centenario della nascita di M. Amari (Palermo, 1910), vol. I, pp. 18–19; E. Tisserant, Specimina codicum orientalium, p. XXXII and pl. 41a; G. Bergsträsser and O. Pretzl, Die Geschichte des Korantexts, GdQ, vol. III (Leipzig, 1938), p. 255 and fig. 9; N. Abbott, The Rise of the North Arabic script and its Ḳur’anic deυelopment . , p. 24; A. Grohmann, ‘The problem of dating early Qur’ans’, Der Islam 33 (1958), pp. 216, 222, 226 and n. 48; Grohmann, AP I, p. 42, n. 1; Déroche, Cat. I/1, pp. 59-60, n° 2; F. Déroche and S. Noja Noseda (eds.), Le manuscrit Arabe 328 (a) de la Bibliothèque nationale de France (Lesa, 1998).
[19] Déroche, Cat I/1, p. 60, n° 3.
[20] P. Ladner, Lexicon des Mittelalters, s. v. ‘Pergament’, vol. VI, col. 1886.
| Source note: This article was published in the following book: Islamic Codicology: An Introduction to the Study of Manuscripts in Arabic Script _ English version, 2005, Al-Furqan Islamic Heritage Foundation, London, UK, pp. 65-102. |
