A lost page of the Archimedes Palimpsest rediscovered after a century
A missing page from the famous Archimedes Palimpsest has been rediscovered at the Blois Museum of Fine Arts by researcher Victor Gysembergh, a senior research fellow at the Léon Robin Centre for the Study of Ancient Thought (Sorbonne University/CNRS). Behind this discovery lies a meticulous investigation, one that is reviving the study of an ancient manuscript that still holds many secrets.
From Blois to Archimedes: a most unlikely discovery
Victor Gysembergh did not expect to find a missing page from one of the world’s most famous manuscripts in Blois while browsing through digital archives in search of palimpsests. During a discussion about Blois with colleagues, a town once known for housing royal manuscripts, the researcher recalls making a half-serious remark suggesting they might find a palimpsest there. He notes that the prospect seemed unlikely. Yet only moments later, while browsing digital records, an image immediately caught his eye. On the screen was a 10th-century parchment. The handwriting looked familiar, the geometric recognisable. And above all, a strange detail: an illumination that was clearly a later addition. “It is unusual to add illuminations to manuscripts like this in the 20th century. That proved to be a decisive trigger for my suspicions,” he recalls.
“I deciphered what could be read of the text, and this confirmed that it was a treatise by Archimedes,” recounts the researcher. Comparing it with photographs taken in 1906 by the Danish philologist Johan Ludvig Heiberg, the verdict is clear: this is page 123, missing for decades, from the famous Archimedes palimpsest now held at the Walters Art Museum in Baltimore.
Yet how did this page get to Blois? The page comes from the private collection of André Franck, a theatre director and a leading figure at the Office de Radiodiffusion-Télévision Française in the post-war period. Moving within artistic and intellectual circles, he gradually assembled a highly eclectic art collection, which he bequeathed to the Musée des Beaux-Arts in Blois upon his death in 1971. “A kind of cabinet of curiosities,” as Victor Gysembergh puts it. Alongside the stage costumes and works of art were also pages from illuminated manuscripts, prized by collectors for their aesthetic appeal. Separated from its original manuscript, the page was no doubt acquired for its illumination and its scientific content ignored.
A manuscript recycled during the Middle Ages
The Archimedes Palimpsest is a unique manuscript. Hidden within this medieval prayer book lies one of the most precious ancient scientific texts ever preserved. Like many manuscripts of the time, it was ‘recycled’: the costly parchment was scraped clean so it could be reused. “A book required a herd of animals,” Victor Gysembergh reminds us. Because it was made from animal skin, parchment was a scarce and valuable resource. In the 13th century, Archimedes’ texts, by then illegible or no longer understood, were erased to make way for prayers.
Beneath this later script, however, lie traces of far older knowledge. The original manuscript, copied in 10th-century Constantinople, contained not only the complete works of Archimedes but also a range of other texts, including a speech by the Athenian orator Hyperides, a commentary on Aristotle’s Categories, and additional works that remain to be identified.
A story straight out of a novel
The manuscript’s history spans centuries of upheaval. Likely taken during the sack of Constantinople by the Crusaders in 1204, it was transformed into a prayer book in 1229, before returning to the Byzantine capital. Having become a prayer book, it was then kept at the Metochion of the Holy Sepulchre, a dependency of the Orthodox Church. There, in 1906, the Danish philologist Johan Ludvig Heiberg identified the manuscript and discovered, beneath the religious text, the writings of Archimedes. The find made the front page of the New York Times and immediately attracted the attention of researchers and collectors.
Amid the turmoil of the 20th century, the manuscript disappeared once more, only to resurface in the private collection of a French art dealer Salomon Guerson. During this period, three pages which were clearly visible in the photographs taken by Heiberg in 1906 went missing. The reasons for their disappearance remain largely speculative, though the evidence points to deliberate interference. “It was commercial vandalism,” the researcher explains. As with certain ancient papyri that were sometimes cut up to be sold in fragments, these pages were likely removed from the manuscript to be sold off individually, making them easier to sell.
The interference did not end there. Before being cut up, some of the leaves were artificially embellished with added illuminations to increase their value. The leaf found in Blois bears traces of this type of intervention. However, such alterations proved counterproductive, as altering such a precious manuscript actually diminished the manuscript’s scientific value.
It remains unclear in what context these alterations took place. Salomon Guerson, who owned the manuscript at that time, was almost certainly aware of the object’s significance. However, given the tragic circumstances of the Second World War, during which he was persecuted as a Jew, some researchers suggest that his actions may have been driven by a sense of urgency.
Page 123 of the Archimedes Palimpsest
© Blois, Musée des Beaux-Arts, Inv. 73.7.52. Photographie IRHT-CNRS
One page found, two still missing
The page discovered in Blois contains a passage from the treatise “On the sphere and the cylinder” (Book I, Propositions 39 to 41), dealing with complex geometric demonstrations related to volumes. On one side, the ancient text remains partially legible. On the other, the illumination depicting the prophet Daniel completely obscures the original writing.
Researchers must employ advanced techniques to read what lies beneath these multiple layers. Multispectral imaging, which has been employed successfully since the early 2000s, can reveal inks invisible to the naked eye; however, it remains insufficient to penetrate layers of paint and fully recover the concealed text. Pages covered with illuminations require different techniques, such as X-ray fluorescence, which can detect the chemical composition of the inks beneath the paint layers. These investigations could be carried out at specialist facilities, such as the Stanford Linear Accelerator Laboratory (SLAC) in the United States, where teams have been working on this type of manuscript for twenty years. “They have an experimental setup which has been optimised for palimpsests,” he explains. A valuable asset, especially as these fragile objects require very strict handling conditions.
Analysing these manuscripts also presents a number of IT challenges. At present, each individual image requires specific processing, which is often carried out manually. To expedite and improve this process, researchers are exploring the use of artificial intelligence, particularly through computer vision techniques. Victor Gysembergh is also collaborating with the Sorbonne Cluster for Artificial Intelligence (SCAI) to develop new methods of automated analysis. The goal is to detect traces of invisible writing more quickly and to refine their reconstruction.
A body of work that remains largely unexplored
Despite its significance, the Archimedes Palimpsest has never been studied in depth. Some pages remain largely illegible, and two pages are still missing. Research is ongoing to shed light on these unresolved areas.
The rediscovery of the page in Blois was no mere coincidence. It came about as part of the European ERC PALAI project, led by Victor Gysembergh at the Léon Robin Centre for the Study of Ancient Thought. This programme, dedicated to the study of palimpsests, aims to provide a clearer understanding of why ancient manuscripts were rewritten and the contexts in which they were reused. “The project focuses on the phenomenon of palimpsests in general,” explains the researcher, “particularly in northern Italy during the early Middle Ages, where extraordinary manuscripts were reused.” The Blois page was identified in the course of this systematic investigation.
This rediscovered page serves as a reminder that many treasures remain hidden, sometimes in plain sight. Hundreds of thousands of palimpsests remain unexplored, with countless ancient texts potentially concealed beneath later layers of writing. Museums, libraries, and private collections could therefore play a key role in future discoveries.
“There are treasures just waiting to be discovered,” Gysembergh emphasises - an invitation to continue the investigation far beyond the confines of the laboratory.
Watch the video with Victor Gysembergh and Emanuel Zingg
Comment Victor Gysembergh et Emanuel Zingg vont changer le monde