Turn a manuscript page and the lines seem to run in two directions. One hand has written across the page you are holding. A fainter hand belongs to a book whose original shape has disappeared. The same piece of parchment has served different readers, and neither history can be understood by looking only at its darkest words.
A manuscript made from erased and reused writing material is called a palimpsest. The word can suggest a secret waiting for one magical camera. The actual investigation is more absorbing: a page has to be conserved, its material responses recorded, its old arrangement reconstructed, and its marks read by people who know what to look for.
The prayer book is real, too
The Archimedes Palimpsest Project’s account begins with a useful correction to the manuscript’s famous name. The object is a Greek prayer book completed in 1229. Its makers reused parchment from several earlier manuscripts, including a tenth-century copy containing works by Archimedes. These are medieval copies of ancient mathematical writing, not pages penned by Archimedes himself.
The reused material also carries works associated with the orator Hyperides and a commentary on Aristotle’s Categories. A single later binding gathered fragments of several earlier reading worlds. “The book underneath” is therefore not always one complete lost volume waiting to be restored.
The Walters Art Museum’s account describes leaves being cut, rotated, and refolded. Reusing parchment changed the orientation and order of writing as well as its visibility. A researcher needs to distinguish the later page sequence from the arrangement of the older leaves. That physical problem matters to the reading: adjoining marks on a photograph may have belonged to different pages in an earlier book.
It is tempting to cast the prayer-book makers as people who failed to recognize a treasure. A more useful question is what their material and devotional needs made possible at that moment. The prayers deserve study in their own right. Their survival and the survival of the undertext are now intertwined.
First preserve the page that carries the question
A conservation paper by the project team describes mold-damaged parchment, adhesive residues, and the work needed to prepare leaves for imaging. Some pages had also acquired modern painted images over the manuscript writing. The investigators were dealing with several kinds of interference, not one uniform layer of erasable ink.
This gives conservation a central role. A dramatic enhancement is of little use if the surviving source is damaged in obtaining it. The page holds information that a future technique or a different scholar may read more successfully. Caring for its material condition keeps those possibilities open.
Before trying to identify letters in a real image, notice ordinary physical clues: an edge, a hole, a stain, or a direction of writing. Such features can help you recognize the same place across different photographs. A convincing discovery begins with knowing exactly what surface you are comparing.
Twelve photographs still need an interpretation
In the project’s account of the August 2007 imaging session, each leaf was photographed at twelve wavelengths from ultraviolet through infrared. The capture system recorded whole leaves at a resolution the scholars judged sufficient for detailed reading. Yet the account stresses that the separate photographs did not immediately yield easily legible text.
Different materials respond differently to illumination. Comparing and combining those responses can make a weak trace stand out. A displayed color may therefore represent a useful contrast between recorded signals rather than the color a medieval reader would have seen.
The project’s discussion of early imaging experiments identifies another difficulty: photographs have to align accurately. Slight changes between images can confuse the tiny details of a character. Processing also produced unwanted features that hindered reading. Improving contrast and establishing what a mark means are connected tasks, but they are not the same task.
Think of a familiar example from a faded photograph. Increasing contrast may help you notice an outline; it does not establish who the person is. Reading a palimpsest similarly requires evidence about script, language, layout, and context after a mark becomes easier to see.
When light meets a layer it cannot read through
Some obscured pages required a different approach. SLAC’s account of the 2006 investigations describes X-ray fluorescence used to detect iron associated with ink. The researchers also examined signals from elements present in later paint. Mapping material signals helped distinguish text-related traces from interference.
The distinction matters because “use a stronger camera” does not describe every imaging problem. Visible-light photography, photographs made under selected illumination, and elemental mapping collect different kinds of information. Choosing among them depends on what is obscuring the writing and what physical evidence survives.
Compare a page without pretending to read Greek
The photographs accompanying Sorbonne University’s report on a rediscovered leaf offer a useful starting point. They show the two sides of a real object, including a later painted image. Open the photographs and choose one small area to examine before reading the full account. Understanding Greek is not required for this first exercise.
| Look for | Record |
|---|---|
| The physical surface | Find an edge, stain, or hole. Where does it interrupt the area available for writing? |
| Direction | Can you distinguish groups of traces running along different lines? |
| A later addition | What does the painted side allow you to see, and what does it conceal? |
| A remaining uncertainty | What still cannot be identified from appearance alone? |
Keep “I can see a curved trace” separate from “this is a particular letter.” The first is an observation you can point to. The second needs an argument that may draw on neighboring marks and specialist knowledge. Writing both statements separately makes the image more interesting, because it shows where seeing ends and interpretation begins.
A missing page can turn up in plain sight
The investigation is still changing. In a report published in March 2026, Sorbonne University described Victor Gysembergh’s identification of a missing leaf in the Blois Museum of Fine Arts. Recognizable writing and a later added image prompted investigation. Comparison with Johan Ludvig Heiberg’s photographs from 1906 confirmed the connection.
The rediscovered leaf contains material from On the Sphere and the Cylinder. Its identification illustrates a kind of research that a spectacular laboratory image can overshadow: examining catalogued objects, recognizing details, and testing a match against an older record. The university’s report also explains that further imaging would be needed to read writing hidden beneath the added painting.
A palimpsest’s wonder is not that technology supplies words where none survive. It is that surviving material can answer a better question when conservation, imaging, old photographs, and scholarly reading are brought together. Every visible trace has a location and a history. Recovering that history can be as satisfying as recovering the sentence.
Read another material clue
Compare a different way of reading concealed writing in Herculaneum Scrolls: Reading a Library Without Opening Its Pages. Then consider what a manuscript’s materials can establish in The Bakhshali Manuscript: A Tiny Dot and a Big Dating Lesson.
Explore more art, ideas, and independent projects at Incoprea.
