Original article: Hidden library: visualizing fragments of medieval manuscripts in early-modern bookbindings with mobile macro-XRF scanner

A hidden library of medieval manuscript fragments lies deep within the covers and spines of early modern books. When the printing press was invented c. 1450, medieval books and manuscripts became obsolete. The old books were either boiled down to make glue or cut up to make strong bookbindings. We can gather a plethora of historical information from these recycled fragments, such as: locations and dates of certain objects, variations in handwritten texts, and types of manuscripts that were selected for recycling.

But how can we read these libraries if they are hidden from view? By using macro-scale X-ray fluorescence scanning, researchers are able to collect distribution maps of key elements. Incident X-rays cause a core electron to be ejected, leading to a higher energy electron to jump down to fill the open orbital.  The secondary X-rays emitted by this process are characteristic of each element. Mainly, this technique is used to identify elements like iron and calcium, while ignoring organic substrates like parchment and paper. This method is frequently used to study paintings, as can be seen in the previous bites about a Caillebotte painting and Degas Painting.

This article discusses four case studies from the Leiden University Libraries to highlight the use of macro-XRF for studying these hidden fragments. One of these case studies is a large fragment under the endpaper of the book cover. Fig. 1 contains the calcium, iron, and mercury distribution maps. Both the calcium and iron show text from a calendar, and mercury shows the word “decollatio,” likely marking the beheading of John the Baptist. Of these three elements, the calcium distribution is the easiest to read. Iron distribution shows text on both the front and the back of the fragment. Calcium’s secondary X-rays are weaker in energy and could only come from the top of the parchment fragment; however, iron’s X-rays have more energy, meaning they can be emitted from deeper (both sides of the fragment) and still reach the detector. Mercury’s X-rays have the highest energy and are actually from lower layers/deeper fragments. In a way, this is a depth profile of the fragments hidden in the book cover!

Similarly, for a different case study, a thin strip of manuscript used to reinforce the bookbinding of an early-modern book was imaged. The iron distribution showed text from a Book of Hours translated into Middle Dutch of the 14th century. Mercury and copper were also found, likely to add decorative elements, since mercury is associated with red pigments and copper is associated with green pigments. Thus, a colored reconstruction can be seen in Fig 2c.

Unfortunately, one of the biggest challenges for the implementation of these techniques is the geometry. The smallest spot size this instrument can achieve is 40 μm, but it varies based on the distance of the object to the X-ray source as seen in Fig. 3b. For mapping these documents, the smallest spot size possible is needed in order to resolve the fine details. For studying hidden fragments in bookbindings, it is challenging to optimize the focus, because you cannot see the fragments you want to analyze. Also, they are not always flat, meaning the X-ray distance to the object is constantly changing, and the instrument currently lacks fine Z dimension adjustments.

Overall, this research shows how macro-XRF can be used to study hidden medieval text fragments in early-modern bookbindings. Several case studies are discussed to highlight the successes, but it is noted that currently, there are major challenges for this technique with these samples. As this method is already popular for analyzing paintings, this is a good stepping stone to see macro-XRF applied to new challenges, such as objects with complex geometries. However, to open new doors moving forward, different types of data analysis will be needed to deconvolute some of the resulting maps as seen in the images of multiple pages at once.

All figures reproduced/adapted with permission from Duivenvoorden, Käyhkö, Kwakkel, & Dik in Heritage Science 2017.