Font-de-Gaume Cave:
First direct radiocarbon dating of a Paleolithic painting in the Dordogne

On March 9, 2026, a study was published in the journal Proceedings of the National Academy of Sciences (PNAS) that marked the first direct radiocarbon dating of cave paintings in the Dordogne region. An international team led by Ina Reiche of the French National Center for Scientific Research (CNRS) determined the age of two images in the Grotto of Font-de-Gaume — a bison and a mask. Prior to this work, no Paleolithic art site in southwestern France, including Lascaux, had received direct chronological data: the age of the paintings was estimated solely by style, indirect analogies, and the context of the finds.

The cave that remained silent for a century

Font-de-Gaume is located in Les Eyzies-de-Tayac, a Dordogne commune known as the "capital of prehistoric man." Within a radius of several kilometers, there are approximately 200 decorated caves from the Paleolithic era. Formally opened to science in 1901, Font-de-Gaume is one of the few such sites that still allows visitors, albeit in strictly limited numbers.

The approximately 120-meter-long cave walls are covered with over 200 images: bison, horses, mammoths, woolly rhinoceroses, and figures traditionally referred to as anthropomorphic masks. The paintings are executed in a polychrome technique — with red and yellow ochre and black pigment. The grotto is a UNESCO World Heritage Site, part of the "Prehistoric Monuments and Rock Art of the Véser Valley," along with Lascaux, which was closed to the public in 1963 due to microbial growth on the walls.

The majority of the paintings were traditionally attributed to the Magdalenian culture . Dating them by style gave a range of several millennia — and without direct measurements, remained speculative.

A barrier that no one has tested

The radiocarbon method works with organic carbon. In a living organism, the ratio of the isotope C-14 to the stable C-12 is maintained at a constant level, and after death, it slowly decreases. The half-life of C-14 is approximately 5,730 years, allowing dating of objects up to approximately 50,000 years old.

Cave paintings lend themselves to this method only if the artists used charcoal. In Spanish caves and at Chauvet, charcoal pigments have been discovered and dated long ago. In the Dordogne, however, it was taken for granted that local artists worked exclusively with mineral pigments: iron oxides and manganese oxides. These substances contain no organic matter and are inaccessible to C-14 analysis.

The fundamental problem was that this assumption hadn’t been systematically tested. No study had confirmed the absence of carbon in Dordogne pigments — it’s just that no one had deliberately looked for it.

How to see what "shouldn’t be"

Ina Reiche’s team from the Chimie ParisTech-PSL/CNRS laboratory built a methodology on two non-invasive techniques that allow the chemical composition of pigments to be studied without physical contact with the wall.

The first is Raman microspectrometry. A laser beam hits a surface, the molecules scatter it with characteristic frequency shifts, and the shape of the spectrum allows the substance to be unambiguously identified. Carbon and manganese oxide produce fundamentally different Raman "fingerprints," although both appear dull black to the naked eye.

The second is hyperspectral imaging, or reflectance imaging spectroscopy. The camera records spectra in hundreds of wavelength ranges — from ultraviolet to near-infrared — and creates a chemical map of the entire panel. In these images, coal and manganese patterns appear in different colors, although they are indistinguishable to the naked eye.

It was the hyperspectral imaging of the Carrefour "Cross Panel" that revealed something unexpected: the bison HB15 and the anthropomorphic mask contain carbon, while the neighboring bison HB14 is painted with manganese oxide. The hyperspectral analysis was conducted by specialists from TU Delft; their images of the panel clearly demonstrate the chemical distinction between the two types of pigments.

Permission for minimal intervention

The uniform distribution of charcoal across the entire outline eliminated contamination. If charcoal had landed on the walls by accident — from tourists’ torches or pencil marks — it would have settled chaotically, without a clear connection to the drawing’s lines. The systematic alignment of the charcoal signal with the outlines indicated the original composition of the pigment.

This provided the basis for the next step. A special permit was required to collect microsamples — as a World Heritage site, the grotto is protected by the strictest procedures. The permit was granted as an exception.

Accelerator mass spectrometry (AMS) was used for dating — the most sensitive radiocarbon dating method, capable of processing samples weighing only a few milligrams. The analysis was performed by the specialized platform Laboratoire de mesure du carbone 14 (CEA/CNRS/IRD/ASNR/French Ministry of Culture).

Bison: A little later than thought

Bison HB15 was dated between 13,461 and 13,162 years ago (in calibrated years before present, calBP — the counting starts from 1950, adjusted for fluctuations in atmospheric C-14 concentrations). This is somewhat later than stylistic estimates, but still falls within the Magdalenian period of the Upper Paleolithic.

The very fact of a measured date is already a result. Previously, all attributions to the Madeleine were based on comparative analysis. Now, the date is not derived by analogy, but obtained by direct measurement.

Mask: Seven Thousand Years of One Image

The anthropomorphic mask yielded significantly more complex data. Three separate sections of the figure yielded three distinct chronological ranges. The main parts of the mask were dated: 15,981 to 15,121 years ago and 15,297 to 14,246 years ago — the Magdalenian period, preceding the bison’s dating.

The third result was unexpected. The mask’s left eye was dated to between 8,993 and 8,590 years ago — the Mesolithic, a fundamentally different culture. The gap between the mask’s first lines and this final stroke is approximately 6,000–7,000 years. This means that people living in Europe after the end of the last Ice Age, in a world without mammoths, came to the same grotto and added a detail to the drawing created by their distant predecessors.

When a wall is not a monument, but a living place

Similar additions made at different times were also recorded in other caves — Altamir , Pech Merle . However, these were observations based on stylistic analysis and overlapping layers of pigment. Direct dating of individual strokes of the same figure, even across millennia, represents a qualitatively different level of precision.

These findings rethink the concept of cave art as a recorded event. The images on the walls of Font-de-Gaume were created over several working sessions, separated not by years, but by millennia. People didn’t simply contemplate the old drawings; they worked with them.

For people who lived 9,000 years ago, ancient paintings retained some meaning — otherwise, it’s hard to explain why they would venture into the depths of a dark grotto and add an eye to a six-thousand-year-old figure. Science has yet to answer this question.

Madeleine’s: who drew it originally?

The creators of the original Font-de-Gaume images belonged to the Magdalenian culture of the Late Paleolithic . The Magdalenians lived approximately 17,000 to 12,000 years ago in Western and Central Europe. Hunter-gatherers, they produced a variety of bone and antler tools, including composite harpoons, eyed needles, and carving tools. Small mammoth ivory sculptures found at their sites depict the same animals as those found on the cave walls.

The climate of the Magdalenian Périgord was strikingly different from today’s. The Last Glacial Maximum had passed about 20,000 years ago, but Europe had not yet warmed up. Open steppes with mammoths, woolly rhinoceroses, and wild horses gradually gave way to sparse forests. It was these animals — some of which were already beginning to disappear — that the artists of Font-de-Gaume depicted.

The question of why they drew deep underground by the light of tallow lamps and torches remains unanswered. Hypotheses include ritual hunting practices, shamanic rites, and the marking of collective memory. New chronological data does not resolve this debate, but it does provide a previously unexisting framework.

How they dated before: from intuition to isotopes

The history of prehistoric art chronology begins with Henri Breuil, an abbot and scholar who dominated the field in the first half of the 20th century. Breuil laid out a stylistic evolution of cave painting: early drawings — simple linear silhouettes; later — detailed, three-dimensional figures. His system was influential, but it relied on analogies rather than direct measurements.

In the 1960s, André Leroi-Gourhan proposed an alternative stylistic periodization, dividing the development of cave art into four stages. The first direct radiocarbon dates for cave paintings appeared in the late 1980s and early 1990s, when the AMS method achieved the necessary sensitivity. The Chauvet Grotto, discovered in 1994, yielded dates of approximately 36,000 years ago — twice as old as predicted by stylistic schemes, undermining the idea of a linear progression of style.

Uranium-thorium dating (U-Th) developed in parallel, analyzing the decay of uranium in the carbonate crusts that grow over the drawings. This method provides a minimum age for the image — the lower chronological limit. In 2012, U-Th dating showed that several Cantabrian drawings were over 40,000 years old, sparking discussion about the possible involvement of Neanderthals in the creation of the cave art. However, the uranium-thorium method is of little use in the Dordogne region, where carbonate crusts are poorly developed.

Pigments under laser sight

Black in Paleolithic painting was achieved using two main materials: manganese oxide and coal. They are indistinguishable in appearance. Paleolithic manganese mines have been found in the Dordogne — an obvious source of mineral pigment. But coal was ubiquitous: any hearth yields it in abundance.

A key step for Reiche’s team was to measure the pigment’s composition rather than assume it. In 2023, the same research group published the first data on the detection of carbon pigments in a number of Font de Gaume figures — a 2026 paper translated this discovery into specific, absolute dates.

Two figures on one panel – and different chemistry

The "Cross Panel" features three images side by side. Bison HB14 is painted with manganese oxide and cannot be radiocarbon dated. Bison HB15 and the mask contain charcoal and have now been dated. They are from the same place, on the same surface, but with different pigments.

Why manganese was chosen in one case and coal in the other is unknown. The difference in technique may reflect different artistic traditions, different sources of raw materials, or the individual preferences of a particular artist. The very existence of both pigments in the same gallery refutes the previous image of the Dordogne as a zone of homogeneous, purely mineral painting.

A region that remained without absolute dates

The Dordogne and Périgord region is one of the world’s largest centers of Paleolithic art. Hundreds of decorated caves cover several thousand square kilometers — and not a single one had direct radiocarbon dates until 2026. Lascaux, Pech Merle, Rouffignac, and Cap Blanc — all were dated indirectly.

The 2026 study shattered the long-held belief that charcoal exists in Dordogne pigments, and it can be measured. For the first time in a century of studying this region, researchers had numbers, not just analogies. The chronological map of one of the richest prehistoric art zones in the world began to fill with measured data.