Multi-Spectral Imaging for the Edition

The damage to the fragments described above—which is mostly a result of the glue used when the fragments were pastedowns and the consequent ripping caused when the fragments were lifted and turned into flyleaves—means that large sections of the fragments have been rendered illegible. In an attempt to salvage some of the damaged text for this edition, the project team brought Professor Andrew Beeby (Department of Chemistry, Durham University) to Bristol Central Library to carry out multi-spectral imaging of the Bristol Merlin leaves. He has developed a mobile spectrometer for manuscript imaging as part of his work for Team Pigment (Durham University). This project team is systematically analysing pigments used in medieval manuscripts held in British libraries and special collections using (Raman) spectroscopy.

The multi-spectral images of the Bristol Merlin were recorded using a Canon EOS-60D camera, modified for operation across the full spectrum, and equipped with an apochromatic lens, f = 60 mm (Jenoptik). Spectral images were recorded through a set of bandpass filters centred at 365, 400, 450, 500, 550, 600, 650, 700, 750, and 850 nm with a bandpass of 20 nm (Edmund scientific), with the sample illuminated by a bank of LEDs selected to give full coverage of the filter packs. The images were processed using Adobe Photoshop, with adjustments to blacks, whites, contrast, sharpness, and exposure to bring out any ink residues left in damaged areas.

While large sections of the text remained unreadable even after the images were processed, certain areas became more legible—or at least legible enough to fill in some of the lacunae with reasonable certainty (see Fig. 10 for an example). The various different filters proved both more and less effective depending on the type of damage present. For example, ultra-violet and infra-red filters worked more effectively on sections of parchment where glue had pulled off a relatively thin layer of the surface, meaning that enough ink residue had seeped through to the layer below so as still to be detectable. Areas of text from which the glue had ripped off a thicker layer of the parchment proved irretrievable, since all ink residue had been torn away with the writing surface.