Asian and African studies blog

27 May 2021

Fragments of Abbasid Sciences: From Desert Monastery to Digital Reunion

As the Qatar Digital Library (QDL) uploads its two millionth image this week, we’d like to celebrate the nearly 80,000 images of British Library Arabic scientific manuscripts that contribute to this achievement.

One of the most fascinating of these manuscripts and one of the oldest is a thousand-year-old fragment of a Christian Arabic miscellany in Or. 8857. Enhanced cataloguing facilitated by the British Library/Qatar Foundation Partnership has provided a glimpse of the scientific interests and texts available to readers in the monasteries of the Near East around AD1000 and also of the diverse communities that produced these manuscripts in monastic scriptoria. Creating a digital surrogate of this fragment for the QDL has also allowed us to virtually reorder its folios and even remotely reunite it with another, larger fragment from the same manuscript held in another collection.


Acquisition and condition

On 30 May 1921, the British Museum acquired five folios of a Syriac manuscript along with thirty-three folios of a very ancient Arabic manuscript from F.W. Bickel, an antiquities dealer in Zürich specialising in Christian oriental manuscripts.

Off-white paper with two lines of cursive text in the Latin alphabet
Acquisition note: ‘Bought of F. W. Bickel. 30 May, 1921.’ (British Library, Or. 8857, endleaf verso [ii-v])
CC Public Domain Image

When this purchase was recorded in the British Museum acquisition register, the fragmentary Arabic manuscript was given the shelfmark Or. 8857 along with a typically brief description: ‘Or. 8857. A fragment of a work on the calendar, followed by some prescriptions. 33ff. XIth. cent. 8o Arabic’. Clearly the manuscript was old – 5thAH/11th AD century according to the acquisition register. But details about its contents were scanty, and nothing was said about its provenance.

Off-white paper divided into three with small boxes on left and right and large one in the centre, all of which are filled with cursive text in the Latin alphabet in black ink
Entry for Or. 8857 in the British Museum acquisition register ( List of Oriental Manuscripts 1909–1921. Or. 6948–9034, p. 275 [British Library, ORC GEN MSS 7])
CC Public Domain Image

When the thirty-three paper folios that comprise Or. 8857 entered the British Museum, they were evidently in disarray. Not only is there no evidence to suggest that the folios arrived with a binding, but worse – the sewing that held the quires had disintegrated and the loose bifolia had broken apart along their spine-folds to become individual folios. At some point, probably shortly after their acquisition, all thirty-three folios were mounted on paper guards and sewn into a new binding with little regard to their original order but perhaps preserving the order in which they had arrived at the British Museum.

Off white manuscript folio with two columns of text in black ink in the Arabic script and red stamp with British Museum seal at bottom
The first folio, according to the manuscript’s present arrangement, is not what it seems. Its layout suggests either poetry or two columns of prose but, in fact, it is a list of the planets that rule each hour of the day, and it runs horizontally across the page despite the columns. What appears to be an eastern Arabic five (٥) in the upper left corner – perhaps explaining the western Arabic five in the lower right-hand corner – is actually a Coptic seventy (𐋰), which indicate that this is really not the first but the penultimate folio (British Library, Or. 8857, f. 1r)
CC Public Domain Image

After this conservation work, the manuscript seems to have rested unnoticed until a more complete list of its contents was prepared for the Subject Guide to the Arabic Manuscripts in the British Library (pp. 353, 357, 385 and 389). But it was not catalogued in detail until it was selected for digitisation for the QDL.


Date and context of production

The manuscript is written in a squat and angular script that has been described as ‘Kufic’. This script is now considered one of a loosely defined group of scripts generically called Abbasid Bookhand because they were developed in the early Abbasid chancery and employed for copying books on both sacred and secular topics from roughly the 3rd/9th to 5th/11th century. They were then replaced by the maghribī script in the extreme west of the Islamic world and by the naskh script almost everywhere else.

Apart from the manuscript’s archaic script and paper, other features help to define the time and place it was copied. Chief amongst these are its quire signatures, numbers that tell the bookbinder the correct order in which to bind the quires that make up a manuscript. In this manuscript two sets of quire signatures are found on the first and last folio of each quire. These quire signatures are written using two separate systems of alphabetic numerical notation: Greek and Georgian. The use of these two numeral systems alongside an Arabic text written in Abbasid Bookhand and featuring the distinctive punctuation marks displayed in this manuscript all attest to the collaboration of multi-ethnic and multilingual artisans in the Syrian, Palestinian and Egyptian monastic scriptoria of the early Abbasid period. The particular combination of quire signatures found here, however, is most typical of the scriptorium of the Monastery of St Catherine, Sinai, especially during the late-4th/10th and early-5th/11th century.

Double page spread of manuscript on off-white paper with writing in Arabic script in black ink, with several features highlighted by red, green and blue circles placed over the text
Opening from the Book of Seasons (Kitāb al-azmina), which displays a variety of punctuation and space-filling marks as well as Greek quire signatures (circled in green, Η = 8 right and Θ = 9 left), Georgian quire signatures (circled in red, Ⴆ = 7 lower right and Ⴇ = 9 upper right) and Coptic folio number (circled in blue, 𐋯𐋩 = 69) (British Library, Or. 8857, ff. 10v and 17r)

Or. 8857, ff. 10v:
Or. 8857, ff. 17r:
CC Public Domain Image


Reordering the folios

The quire marks demonstrate that Or. 8857 is a fragment containing the remains of quires 5–9 of a larger original manuscript. But without putting the folios back in their original order, it would be impossible to know how much of each quire has survived. Luckily, each folio also has a number in its head margin. Although these folio numbers are likely to have been added somewhat after the quire signatures, they are early and they also attest to the multilingual context in which the manuscript was produced and consumed since they are written using the Coptic epact alphabetic numerals. The use of these numerals was not restricted to the Coptic community, and they are commonly referred to as ‘register letters’ (ḥurūf al-zimām) since they were favoured by merchants and administrators for use in their registers and account books.

Like the Arabic alphabetic numerals (ḥurūf al-jumal, commonly called abjad) – the numerical values of which happen to be explained on ff. 1v–2r of this manuscript – the Greek, Georgian and Coptic (zimām) alphabetic numeral systems all have a base of ten (unlike Roman numerals, which also have a sub-base of five) and they are additive (unlike Roman numerals, which also subtractive) rather than positional (like Arabic numerals). This means that to write the number 123 in alphabetic numerals, one does not write the letter representing 1 in the hundreds place, 2 in the tens place and 3 in the ones place as done with Arabic numerals. Rather, one writes the letters representing 100 (+) 20 (+) 3.

Table with first column and row in grey background with Greek letters in the central cells
Greek majuscule alphabetic numerals 1–900

Table with first column and row in grey background with Georgian letters in the central cells
Georgian majuscule (Asomtavruli) alphabetic numerals 1–900

Table with first column and row in grey background with Coptic letters in the central cells
Coptic epact or zimām numerals 1–900

Reading the Coptic (zimām) foliation along with the quire signatures, it becomes clear that Or. 8857 is a fragment of five quaternions (quires 5–9) comprising folios 37–71 of a larger manuscript of unknown extent. Quires 5, 6 and 8 are still complete with eight folios each, while quire 7 is missing the two folios of its inner bifolium, and only the first three folios from quire 9 are preserved.

Five schematic diagrams of thick or hatched blue lines forming concentric c-shaped items flipped so that they are open to the left
Visualisation of the original quire arrangement of the folios in Or. 8857. Historic Coptic (zimām) foliation at left and modern British Museum foliation in brackets at right. Note that the Georgian signatures for quires 7 and 8 are erroneously reversed. (Visualisation produced with Viscodex)
CC Public Domain Image


Diverse monastic reading material

Once we know the original order of the folios, we can see that Or. 8857 contains a variety of texts on subjects more or less obviously suited to the monks of Monastery of St Catherine.

1) Fragment of a Christian prayer (f. 37r–37v [British Museum f. 18r–18v]);

2) Prayer Taken from the Book of the Prophet David (Duʿā mustakhraj min Kitāb Dāwūd al-nabī, ff. 37v–41r [BM ff. 18v–22r]);

3) Prayer Composed by One of the Righteous Christian Believers (Duʿā allafahu baʿḍ al-muʾminīn al-muḥiqqīn min al-Naṣārá, ff. 41r–47v [BM ff. 22r–28v]);

4) Three recipes for incense (ff. 47v–49v [BM ff. 28v–30v]);

5) The Book of Seasons (Kitāb al-azminah, ff. 49v–70v, ff. 56–57 missing [BM ff. 30v–33v, 11r–16v, 3r–10v, 17r–17v, 1r and 1v]);

6) Fragment of an astrological text (ff. 70v-71v [BM ff. 1v-2v]).

The prayers that occupy the first eleven folios are clearly appropriate in a monastic context although certain features may seem jarring to the modern eye. One prayer ends with the invocation ‘O Lord of the Worlds!’ ( yā Rabb al-ʿĀlamīn, f. 18v), for example, and another is preceded by the basmala ( bi-sm Allāh al-Raḥmān al-Raḥīm, f. 22r), both phrases which occur in the Qurʿān and appear distinctly Islamic today. But during this early period, and for centuries after Or. 8857 was copied, these phrases were used in common by the Arabic-speaking adherents of all the Abrahamic faiths. On the other hand, although incense does not necessarily imply church ritual, the Trinitarian formula ‘In the name of the Father, the Son and the Holy Spirit’ (bi-sm al-Ab wa-al-Ibn wa-rūḥ al-qudus, f. 28v) at the beginning of the incense recipes attests to their Christian context.

The last two texts in the fragment, however, seem less typical of a monastic library. The Book of Seasons is a sort of almanac containing information about the calendar, the heavens, weather phenomena, human illness and health and agricultural matters as they pertain to the twelve months of the year. This genre of literature, in which titles like the Book of Seasons or the Book of Asterisms ( Kitāb al-anwāʾ) are common, provided important guides for living in harmony with the natural rhythms of the year – especially useful for monastic communities surviving in often harsh and semi-isolated conditions. Indeed, one of the earliest authors of this genre was Abū Zakarīyā Yūḥannā ibn Māsawayh (d. 243/857), a Nestorian Christian hospital director at Baghdad, personal physician to the Abbasid caliphs and teacher of the Nestorian physician and translator Ḥunayn ibn Isḥāq (d. 260/873).

Single page of Arabic-script text in black ink with several words in red ink on off-white paper
Information on the names of the months in Syriac, Greek and Persian from the beginning of the Book of Seasons, preceded by the basmala (British Library, Or. 8857, f. 30v)
CC Public Domain Image

The fragment ends with an anonymous introductory text on astrology, which includes an unusual method for determining a person’s ascendant not by observing their natal horoscope chart, but through numerological analysis of their name and that of their mother. While this text may seem the least appropriate in a monastery, there was considerable legal and theological disagreement about which of the various astrological practices were licit or illicit, and knowledge of the planets' influences on the environment and the human body was generally considered an important part of maintaining good health and wellbeing.


Fragments reunited

A much larger fragment of the same manuscript of which Or. 8857 is also a fragment is now held at the Biblioteca Ambrosiana in Milan under the shelfmark X 201 sup. According to a note by Mons. Enrico Rodolfo Galbiati (Doctor of the Ambrosiana 1953–84, Prefect of the Ambrosiana 1984–89) written in the margin of the Ambrosiana’s copy of Löfgren and Traini’s Catalogue of the Arabic Manuscripts in the Biblioteca Ambrosiana (vol. 1, p. 33), X 201 sup. was amongst a lot purchased in 1910 from an unknown dealer in Munich by Ambrogio Damiano Achille Ratti, then Prefect of the Biblioteca Ambrosiana (1907-14), but soon to lead the Catholic Church as Pope Pius XI (1922–39).

X 201 sup. has also been digitised and is now available on the Ambrosiana’s Biblioteca Digitale, where I stumbled upon it, immediately recognising its similarity to Or. 8857. Like Or. 8857, the Milan manuscript is a miscellany combining Christian material with texts on herbal remedies, medicine, astrology and related topics. Likewise, the same Abbasid Bookhand and number of lines per page are found in both manuscripts. But it is the Greek and Georgian quire signatures alongside Coptic foliation found in both manuscripts that prove they are two pieces of the same puzzle.

According to the Coptic foliation and bilingual quire signatures, Or. 8857 contains ff. 37–71 (ff. 56 and 57 are missing) of the original manuscript, and its last quire signature is 9 on f. 17r (f. 69r of the Coptic foliation). The Milan manuscript contains 227 folios (beginning at ff. 97 and ending at f. 337 of the Coptic foliation, with some gaps), and its first quire signature is 13 on f. 101r (f. 5r of the modern Ambrosiana foliation).

We know that the quires in Or. 8857 were quaternions, which have eight folios each, so we would expect the Milan manuscript to be composed of quaternions too – although it should be pointed out that irregular quires are not unusual in manuscripts. Between the beginning of quire 9 (the last in Or. 8857) and the beginning of quire 13 (the first to begin in the Milan manuscript) there were four quires, which if they were all regular quaternions, should equal thirty-two folios (4 quires x 8 folios in each quire = 32 folios). When we count from the beginning of quire 9 on f. 69 of the Coptic foliation and to the end of quire 12 on f. 100 there are, indeed, exactly thirty-two folios, confirming that the two manuscripts are fragments from the same original manuscript.

Even though 77 folios have been lost from the original manuscript (ff. 1–36, 56–57 and 72–96 of the Coptic foliation, plus another 14 within the body of X 201 sup.), a very substantial 260 folios have now been identified, and this will no doubt form the basis for future studies into Abbasid scientific traditions amongst Christian monastic communities.

Thanks to international digitisation projects, the magic of IIIF, and the Mirador viewer there are fewer barriers than ever before to studies of this kind. In fact, anyone with a computer and access to the internet can virtually reunite the two fragments of this manuscript by following the steps below.

1) Navigate to X 201 sup. on the Ambrosiana’s Biblioteca Digitale, and click on the words ‘Visualizza la copia digitale’. The images will open in the Mirador viewer via your web browser. Open the dropdown menu at the top left corner of the viewer window and choose a location in the viewer window at which to display Or. 8857.

A screen shot showing the cover of a book with a red binding, with thumbnails of pages on the bottom, and the file menu in the top left hand corner dropped down and highlighted in a red box with rounded edges

2) You will now see that a blank canvas has opened at your chosen location.

Screen shot with a book with a red binding atop thumbnails of pages on the left-hand side and a dark grey area with a red-outlined oval on the right-hand side

3) In another browser window, navigate to any page on the QDL displaying images of Or. 8857 and expand the tab marked ‘Use and Share this Record’.

Screen shot showing a white page with thumbnails of book spines in the centre and text on the bottom third, some of which is on a grey background. The lowest grey background is inside an oval outlined in red

4) Under the heading ‘IIIF details’, locate the IIIF logo next to the IIIF manifest for Or. 8857, drag the logo to the Mirador window in your web browser and drop it anywhere on the blank canvas (see step 2).

Screen shot with a black banner at the top and text with a grey background in the middle, with some of the text highlighted by lines and hollow boxes in red and light blue

5) Alternatively, you can copy the IIIF manifest ( located next to the IIIF logo on the screen in step 4 and click on the blank canvas in the Mirador viewer (see step 2). This will open the screen below, where you can paste the IIIF manifest into the field marked ‘Add new object from URL’ and click ‘Load’.

Screen shot showing a primarily white screen with a series of thumbnails of manuscript pages on the top third of the screen

6) You can now use the dropdown menus to choose how you would like to view each of the manuscripts and even repeat the steps above to add more canvases and view other IIIF compliant objects at the same time.

Screen shot of a black background with a matrix of thumbnails showing various pages of manuscripts

Bink Hallum, Arabic Scientific Manuscripts Curator, British Library Qatar Foundation Partnership
CCBY Image

Thanks to Dr Adrien de Fouchier, OP (Biblioteca Apostolica Vaticana) and Dr Stefano Serventi (Biblioteca Ambrosiana) for their generous help and advice with my research for this blog.


Chrisomalis, Stephen, Numerical Notation: A Comparative History (New York: Cambridge University Press, 2010), pp. 135–37, table 5.1 (Greek and Georgian); 139, table 5.3 (Greek); 150, table 5.5 (Coptic/ zimām, the numerals for 600 [𐋸] and 700 [𐋹] are erroneously reversed) and 178, table 5.20 (Georgian)

Ifrah, Georges, The Universal History of Numbers from Prehistory to the Invention of the Computer , trans. by D. Bellos, E.F. Harding, S. Wood and I. Monk (New York–Chichester–Weinheim–Brisbane–Singapore–Toronto: John Wiley & Sons, 2000), pp. 220 (Greek), 225 (Georgian), and 545 (Coptic/zimām),

Kawatoko, Mutsuo, ‘On the Use of Coptic Numerals in Egypt in the 16th Century’, Orient 28 (1992) 71, fig. 3 (helpfully, gives variant forms for most numerals)

List of Oriental Manuscripts 1909–1921. Or. 6948–9034 (British Library, ORC GEN MSS 7)

Löfgren, Oscar and Renato Traini,Catalogue of the Arabic Manuscripts in the Biblioteca Ambrosiana, 3 vols, Fontes Ambrosiani LI, LXVI and Nuova Serie II (Vicenza: Neri Pozza Editore, 1975–95) vol. 1, item 33, pp. 33–35

Pataridze, Tamara, ‘Les Signatures des cahiers unilingues et bilingues dans les manuscrits Sinaïtiques (Georgiens, Arabes et Syriaques)’, Manuscripta Orientalia 18.1 (2012) 15–35

Subject Guide to the Arabic Manuscripts in the British Library, compiled by Peter Stocks, ed. by Colin Baker (London: British Library, 2001)

Varisco, Daniel, ‘The Origin of the Anwāʾ in Arab Tradition’, Studia Islamica 74 (1991) 5–28