ERC PALAI "A page of the Archimedes Palimpsest rediscovered after more than a century"

From Blois to Archimedes: an unlikely discovery

While browsing digital archives in search of palimpsests, Victor Gysembergh never expected to find in Blois a missing piece of one of the most famous manuscripts in the world. "I was talking about Blois with colleagues — a city that had once housed royal manuscripts. I half-jokingly said: 'let's see if there isn't a palimpsest over there.' It was quite unlikely," the researcher recalls. A few clicks later, an image catches his eye. On the screen: a 10th-century parchment. A handwriting that seems familiar. Geometric figures. And above all, a strange detail: a clearly later illumination. "It's not common to add illuminations in the 20th century to manuscripts like that. That's what really tipped me off," he recalls.

"I deciphered what could be read of the text, and that confirmed it was a treatise by Archimedes," the researcher explains. By comparing it with photographs taken in 1906 by Danish philologist Johan Ludvig Heiberg, the verdict was in: this was folio 123, missing for decades, from the famous Archimedes Palimpsest now held at the Walters Art Museum in Baltimore.

But how did this page end up in Blois? The folio comes from the collection of André Franck, a theatre figure and key personality at the ORTF in the post-war years. Close to artistic and intellectual circles, he assembled over time a highly eclectic art collection, which he bequeathed to the Museum of Fine Arts in Blois upon his death in 1971. "A kind of cabinet of curiosities," summarises Victor Gysembergh. Among the stage costumes and art objects were also pages from illuminated manuscripts, prized by collectors for their aesthetic value. Separated from its original manuscript, the folio was most likely acquired for its illumination, without its scientific content being identified.

A manuscript recycled in the Middle Ages

The Archimedes Palimpsest is a singular manuscript. Behind this medieval prayer book lies in reality one of the most precious ancient scientific texts ever preserved. Like many manuscripts of the time, it was "recycled": the costly parchment was scraped down for reuse. "A book is a flock," Victor Gysembergh notes. Made from animal skins, parchment was a rare resource. In the 13th century, the texts of Archimedes — having become illegible or incomprehensible — were erased to make way for prayers.

But beneath this new writing, traces of ancient knowledge survived. The original manuscript, copied in the 10th century in Constantinople, contained the complete works of Archimedes, but also other texts — among them a speech by the Athenian orator Hyperides, a commentary on Aristotle's Categories, and other texts not yet identified.

A story worthy of a novel

The manuscript's fate spans centuries and historical upheavals. Probably carried off during the Crusaders' sack of Constantinople in 1204, it was turned into a prayer book in 1229 before returning to the Byzantine capital. As a prayer book, it was kept at the Metochion of the Holy Sepulchre, a dependency of the Orthodox Church. It was there, in 1906, that Danish philologist Johan Ludvig Heiberg identified it and discovered, beneath the religious text, the writings of Archimedes. The event made the front page of the New York Times and immediately sparked the interest of researchers and collectors.

But the manuscript disappeared again amid the turmoil of the 20th century, before resurfacing in the collection of French art dealer Salomon Guerson. It was during this period that three folios — clearly visible in the photographs taken by Heiberg in 1906 — vanished. The reasons for this disappearance remain largely a matter of hypothesis. However, all signs point to a deliberate act. "It was mercenary vandalism," the researcher explains. As with certain ancient papyri sometimes cut up to be sold in fragments, these pages may have been separated from the manuscript to be sold individually, as more easily marketable pieces.

The process didn't stop there: before being cut out, some folios were decorated with artificially added illuminations to increase their value. The folio found in Blois bears the mark of this. A counter-productive practice, since altering such a precious manuscript actually diminishes its scientific value.

It remains to be understood in what context these alterations took place. Salomon Guerson, the manuscript's owner at the time, was most likely aware of the object's importance. But given the tragic circumstances of the Second World War — as he was being persecuted as a Jew — some researchers put forward the hypothesis of an act driven by urgency.

One page found, two still missing

The folio discovered in Blois contains a passage from the treatise On the Sphere and Cylinder (Book I, propositions 39 to 41), devoted to complex geometric demonstrations on volumes. On one side, the ancient text remains partially legible. On the other, the illumination depicting the prophet Daniel completely conceals the original writing.

To read what lies hidden beneath these successive layers, researchers will need to resort to advanced techniques. Multispectral imaging can already reveal inks invisible to the naked eye, but it is not enough to penetrate pictorial layers. This technique, already used successfully in the early 2000s, should make it possible to reveal a large portion of the still-hidden text. But in the case of pages covered with illuminations, other techniques are required — such as X-ray fluorescence, capable of detecting the chemical composition of inks beneath the painted layers. These analyses could be carried out at specialised facilities, such as the Stanford Linear Accelerator (SLAC) in the United States, where teams have been working on this type of manuscript for twenty years. "They have an experimental setup optimised for palimpsests," he notes. A valuable asset, especially since these fragile objects require very strict handling conditions.

Analysing these manuscripts also poses computational challenges. Today, each image requires specific processing, often carried out manually. To speed up and improve this process, researchers are exploring the use of artificial intelligence, particularly through computer vision techniques. Victor Gysembergh is thus collaborating with SCAI (Sorbonne Cluster for Artificial Intelligence) to develop new automated analysis methods. The goal: to detect traces of invisible writing more quickly and refine their reconstruction.

A corpus still largely unexplored

Despite its importance, the Archimedes Palimpsest has never been studied exhaustively. Some pages remain largely illegible, and two folios are still unaccounted for. It is precisely to shed light on these remaining unknowns that the research continues.

The rediscovery of the Blois folio was no accident. It is part of the European ERC PALAI project, led by Victor Gysembergh at the Centre Léon Robin. Dedicated to the study of palimpsests, this programme aims to better understand the logic behind the rewriting of ancient manuscripts and the contexts in which they were recycled. "The project concerns the phenomenon of palimpsests in general," the researcher explains, "particularly in northern Italy in the early Middle Ages, where extraordinary manuscripts were recycled." It is within this work of systematic identification that the discovery of the Blois folio fits.

This rediscovered page is a reminder that many treasures remain hidden in plain sight. Hundreds of thousands of palimpsests, still largely unexplored — so many ancient texts potentially concealed beneath layers of more recent writing. Museums, libraries, and even private collections could thus play a key role in future discoveries. "There are treasures just waiting to be looked at," the researcher emphasises — an invitation to continue the investigation well beyond the laboratory.