Cadbury Research Library, University of Birmingham

The Cadbury Research Library holds the University of Birmingham's special collections: more than four million manuscripts, photographs, archives and associated objects, and about 250,000 rare books, the earliest of them printed in 1471. The reading room on the Edgbaston campus is open to members of the university and to members of the public alike.

The part that concerns this atlas is one collection inside it. In 1924 Alphonse Mingana, an Assyrian scholar and priest, made the first of three buying journeys to the Middle East, funded by Edward Cadbury. What he brought back is now more than three thousand Middle Eastern manuscripts in over twenty languages, among them Arabic, Syriac, Ethiopic, Georgian, Hebrew, Samaritan and Armenian.

A collection formed that way carries its own problems into a catalogue, and it is worth being direct about them. Manuscripts bought on the market in the 1920s arrive with no find context, sometimes bound together in combinations that a dealer or a previous owner made rather than a scribe, and the library inherits both the books and the errors attached to them. Part of the collection is now digitised and it is described in the FIHRIST union catalogue, which is what makes those errors findable by somebody other than the person who happens to open the volume.

Two parchment leaves in the Mingana Collection carry Quranic text in an early Hijazi hand. They came to Birmingham with the rest of the collection in the 1920s and sat there for some ninety years without being catalogued as anything in particular, because they had been bound together with leaves of a different seventh century Quran and the volume had always been read as one manuscript.

Alba Fedeli, working on the collection for a doctorate on early Quranic manuscripts, recognised that the two sets of leaves were not the same book. The university submitted a sample for radiocarbon dating at Oxford and announced the result in July 2015: the parchment dated to between 568 and 645 CE at 95.4 per cent probability. The other manuscript the leaves had been bound with is now catalogued separately as Mingana Islamic Arabic 1572b.

What the date settles is narrower than much of the coverage suggested, and the distinction is the reason the find matters rather than a qualification on it. Radiocarbon dates the death of the animal whose skin the parchment is, not the moment a text was written on it, and parchment could be stored or scraped and reused. The leaves are early on any reading and the range overlaps the lifetime of the Prophet, but the measurement is a terminus for the material and not a date for the writing. The specialist literature on early Quranic manuscripts says so plainly, and so did the university's own announcement.