from Medieval Academy News

Seeking the Invisible: Forensic Science at the Parker Library
by Mark Clarke

Just before Christmas 2001, a one-day workshop on the examination of medieval manuscripts by Video Spectral Comparator was held at the Parker Library, Corpus Christi College, Cambridge, through the kindness of the Fellow Librarian Christopher de Hamel, of the Assistant Librarian Gill Cannell, and of Trevor Emmett, Department of Forensic Science, Anglia Polytechnic University, Cambridge. During the day participants were invited to drop by to see the instrument demonstrated. A large number did—resulting in one of the best collections of palaeographers and codicologists one might hope to assemble—in a suitably seasonal and festive atmosphere.

The Video Spectral Comparator instrument was originally developed for the decipherment of altered inscriptions and the detection of forgery, in particular of erasures and substituted writing (for example, an increased the amount of money on a check). It became apparent that it might be used for reading effaced or obscured writing on manuscripts and for revealing underdrawings, and thus also be useful to the codicologist and conservator. Early versions are owned, for example, by The British Library, and by the paper conservators of Camberwell College of Art, London.

The Comparator is essentially a box, with a variety of light sources, connected to a video camera and thence to a computer. It provides special illumination: ultraviolet, colored visible light, or infrared. (It may be noted that the illumination and heat is no greater than that of a short spell in a display cabinet, and this equipment has been found to pose no risk to manuscripts.) The image of the manuscript under this special illumination is recorded by a special video camera inside the instrument and displayed immediately on a computer screen in real-time. This image can be captured by the computer for printing out or publication. The instrument may be used to increase the contrast between ink and background, allowing faded text, erased text, or text on blackened pages to be read. In certain cases it can also identify later additions to the original script by distinguishing different inks.

Inks may fade considerably, or their supports (papyrus, leather, parchment, or paper) may darken through age, dirt, or fire damage, such that the writing is all but invisible to the naked eye, so that even viewing in transmitted light will not help. A simple, widely-used technique for ink on parchment is to use a hand-held ultraviolet lamp. The parchment fluoresces, but any residual ink (especially iron-gall ink) will quench this fluorescence, and so appear darker against a lighter background. This does not, however, always work. A more sophisticated approach is the use of a band-pass filter reflectography apparatus, of which the Comparator is a superior example.

In this technique, the instrument varies the illumination on the manuscript until an improved contrast between the text and the background becomes apparent. The writing is illuminated with light in a series of narrow spectral bands (i.e., a series of illuminations of a single, precise, almost unmixed, color) in sequence, ranging from ultraviolet to near infrared. The image is viewed through a video camera that is sensitive throughout this spectral region. At certain wavelengths the contrast between the support and the ink will be greater than it is under normal viewing conditions, due to the relative differences in their light absorption in the ultraviolet, visible, and near-infrared regions. This phenomenon is similar to that known as “metamerism,” the phenomenon by which objects of different colors can appear to match in color under certain conditions of illumination, but appear unmatched if the spectral composition (color) of the illumination is changed. This is why one often takes clothes out of a shop to examine them in daylight, to check that they match under different conditions of illumination.

Two samples of writing may be compared: the instrument can repeatedly change the illumination on two selected images, in an attempt to find a setting at which they appear different. If it succeeds, the inks are indeed different, and if it fails, they are at least similar inks, if not necessarily the same.

The infrared region of the spectrum can also be used to reveal hidden layers and underdrawing. Many paints are transparent to infrared radiation. Infrared illumination can be shone on a manuscript, pass through the layers of paint, bounce back off the parchment, and be detected as a bright “white” patch by the infrared-sensitive camera of the Comparator. Carbon ink, however, strongly absorbs infrared, so if it is present—either as a faded ink or as a drawing under the other paints—the infrared is absorbed by it, does not bounce back, is not detected by the camera, and appears as a dark area. This works surprisingly well. Ink that has faded in the visible region of its spectrum may still be visible in the near-infrared region. Underdrawing may be photographed using an infrared “light” source and infrared-sensitive film, but the Comparator is vastly easier to use and provides immediate results, both on a computer screen and thence in printed form.

Band-pass reflectography will also to some extent discriminate between inks. Vegetable inks are only visible at wavelengths shorter than 700–750 nanometers, and iron-gall inks shorter than 1200–1400 nm, while carbon inks are visible to the limits of infrared visualization, around 1900 nm. Detecting intrusive words is thus often possible by noting the location of non-matching or anomalous inks.

Many pigments that appear the same to the eye may, like inks, also appear very different in the near-infrared region. The two important medieval illuminators’ blue pigments, ultramarine and azurite, look very similar to the eye, that is in the visible region of the spectrum, but are totally different when viewed under infrared illumination with an infrared-sensitive camera. Ultramarine becomes transparent to the point of invisibility, while azurite appears black. Where several craftsmen have worked on one manuscript, different hands and ateliers may be determined from differences in their use of these two pigments to achieve similar pictorial effects.

Results: Participants in the workshop were invited to bring questions and problems and to suggest manuscripts in the Parker Library which they would like examined, or alternatively to bring manuscripts along for inspection. Corpus Christi manuscripts examined included four Anglo-Saxon manuscripts: The Parker Chronicle (MS 173), The Canterbury-Ely Pontifical (MS 44), The Winchester Troper, (MS 473) and The Red Book of Darley (MS 422).

The Red Book seemed a good place to start, as a photographic facsimile is in preparation; the first few pages contain much illegible black text, and on later pages the red writing has become very faded and abraded. The faded red text—almost white to the eye—immediately became legible. (The black was less successful; perhaps it is lost forever, as the parchment is darkened dramatically, apparently as a result of an ancient treatment with galls to refresh the text.)

The Winchester Troper, an early eleventh-century mansucript, was examined by Susan Rankin with exciting results, revealing as it did, on the first folio, several lines of previously illegible music. She wrote to me later that “some parts were erased . . . leaving remnants of melodies for which concordances often do not exist. Further, the musical notation is so small that work with the naked eye on anything erased is well-nigh impossible. The computer images have not recovered everything, but they have made more original writing visible than before.”

Readers may be aware of the spectacular results that similar techniques have achieved reading the lower text of the privately-owned Archimedes Codex. Natalie Tchernetska, who works on the Codex, was present, and it had been hoped to examine the detached leaves present in the Cambridge University Library, but in the event it was not possible to move them.

Stella Panayotova of the Fitzwilliam Museum Cambridge brought a manuscript with her from the museum, and writes of her experiences of the day: “The manuscript I brought . . . was manuscript 300, known as The Isabella Psalter and Hours, which suffered from a flood in the mid 19th century, and I suspected that some of the miniatures had been ‘restored’ after that. The machine showed significant variations when we compared the pigments of seemingly untouched images and those that looked as if they have been touched up.”

Readers who would like further information on the Comparator or its uses are welcome to contact me at markey@gn.apc.org.

Editor’s note. The author, who trained as a paper conservator and conservation scientist (with a doctorate on Anglo-Saxon manuscript pigments) at Cambridge, is the author of The Art of All Colours: Mediaeval Recipe Books for Painters and Illuminators (London: Archetype, 2001). He currently works at the Foundation for Fundamental Research on Matter in Amsterdam.