It’s a book best judged by its cover.
For decades, the 16th-century reprint of the ancient Greek poet Hesiod’s Works and Days sat inside Northwestern’s Special Collections at Deering Library, waiting for a chance encounter that would reveal the didactic poem’s biggest secret. Written some 2,700 years ago, the piece melds fact and fiction to provide a glimpse into ancient agricultural life. What hadn’t been noticed in Northwestern’s edition was the presence of an unknown manuscript, until exposed by researchers at the Northwestern University/Art Institute of Chicago Center for Scientific Studies in the Arts (NU-ACCESS). Scott Devine, the Marie A. Quinlan Director of Preservation and Conservation, and conservator Tonia Grafakos came across this edition of the seminal poem while researching binding practices for a presentation. The “lost” manuscript was contained in the book’s cover, part of a protective barrier to safeguard the text.
“With this project, it wasn’t the book’s contents that interested us, but the fact that, as the book aged, it became evident that an earlier manuscript was used in the binding process,” says Marc Walton, senior scientist at NU-ACCESS.
After speaking with Devine and Grafakos, Walton began exploring the piece with collaborators Oliver Cossairt, an imaging expert, and Aggelos Katsaggelos, a signal processing analyst, both electrical engineering and computer science.
The researchers developed a visualization technique they hoped might reveal the hidden manuscript in greater detail. Walton was far from certain of the results, but if the effort succeeded, he knew the text could be translated and analyzed to possibly determine why it had been cast aside.
Using hyperspectral imaging and super-resolution x-ray fluorescence, Walton and his colleagues scanned a small area of the book cover, identifying two main text blocks. Where the human eye sees only visible light, hyperspectral imaging divides the spectrum into more detailed bands of light. X-ray fluorescence is a nondestructive technique that provides elemental information that, when applied to a painting (or book cover), can identify the presence of certain pigments used by an artist or author.
“One of this project’s big takeaways is that we could push technology forward,” Walton says. “Working with the Katssagelos and Cossairt labs, we have developed and refined new nondestructive analytical techniques.”
Expanding existing practices to explore cultural heritage artifacts is one of the core missions of NU-ACCESS. Through collaboration between museum professionals and University science and engineering researchers, the center conducts object-based and object-inspired research devoted to problems of interest to museums and cultural institutions.
Postdoctoral fellow Johanna Salvant, a resident expert in painting techniques, is working to enhance how researchers explore the distribution of pigments within a piece.
The history of illuminated manuscripts — text supplemented by imagery — dates back some 1,500 years, and Salvant is collaborating with Walton and the Isabella Stuart Gardner Museum in Boston to investigate materials and medieval techniques used in a 15th-century collection.
“Museum conservators approached us because they were interested in learning what parts of this colorful piece were associated with the original paintings and what parts were associated with 19th-century restoration,” Salvant says.
The first step was to create a super-high-resolution copy of the artwork. This data-rich image serves as a virtual microscope and offers functionality similar to Google Earth.
“We can zoom in close enough to see individual brush strokes and get in-depth information on how the piece is technically constructed,” Walton says.
To discern characteristics no larger than the width of a human hair, the researchers expanded the digital rendering using photometric stereo. The computational imaging technique implemented by Cossairt provides a rendering that merges the high-resolution photograph with a topographic representation.
“Photometric stereo allows researchers to view an object’s topography by separating color from surface shape,” Walton says. “In doing so, we can measure that topography from small changes in brightness that different lighting directions may cause.”
This strategy, combined with Salvant’s use of x-ray fluorescence to examine specific pigment distribution in the paintings, has already led to major discoveries for the team.
“Our research adds to the art history of this piece quite considerably,” Walton says. “Although it was added to the museum as a work by Jean Bourdichon, we can definitively say, because of the painting’s stylistic attributes and quality, that it was in fact completed by another unknown artist.”
Salvant has also revealed specific areas in the French collection of suffrages that are attributable to 19thcentury restoration.
“We’ve made a pseudo–color map that helps us understand the pigments that went into the fabrication of the piece,” she says, pointing to a yellow section. That portion could only be part of the restoration, not the original, since the specific chromate-based pigment did not exist until about 200 years ago.
“Using these nondestructive methods of analysis, we can start to make determinations not only about a piece’s original manufacture but also about what’s happened to it over time,” Walton says. “We can learn how it was made and restored while reconstructing the object’s entire life history.”
Major funding for NU-ACCESS is provided by the Andrew W. Mellon Foundation. Additional support is provided by the Department of Materials Science and Engineering, the Materials Research Science and Engineering Center, and the Office for Research.