How Ancient Cave Paintings Were Made

On December 18, 1994, three explorers working a cliff in the Ardèche region of southern France squeezed through a narrow passage and entered a chamber that no one had seen for roughly 29,000 years. On the walls were lions, rhinoceroses, bears, and horses, drawn with confident lines and shaded for depth. This article explains how Ice Age people made such images: which minerals became paint, how it was put on stone, how artists saw in total darkness, and how researchers date work this old. It also covers what the evidence cannot yet tell us.

Key takeaways

• Cave painters used iron-oxide ochre for reds and yellows, manganese-bearing minerals and charcoal for blacks, and probably mixed them with binders such as animal fat or water.
• Paint was applied by hand, by swabbing and blotting, and by blowing pigment over a surface, which is how hand stencils were made.
• Artists lit caves with torches and stone fat lamps. Experiments show torches burn for roughly 20 to 60 minutes each.
• Chauvet Cave in France was visited between about 37,000 and 28,000 years ago. Lascaux is usually dated to 17,000 to 22,000 years ago. Paintings on Sulawesi are at least 51,200 years old.
• Dating relies on radiocarbon for charcoal and uranium-series methods for the mineral crusts that grow over the art, and the second method gives only minimum ages.

What the first painters had to work with

Upper Paleolithic artists had no manufactured paints and no brushes from a shop. They had rocks, charred wood, animal products, and a great deal of patience. Analyses of the pigments at Lascaux show that the palette was limited, and that limit is part of why the images look so graphic: reds, yellows, and blacks, with the dark outlines doing most of the work.

Ochre is the workhorse. It is clay or rock colored by iron oxides, and the same family of compounds as rust. The mineral hematite gives red, and goethite gives yellow to brown. At Lascaux, researchers have identified iron oxide, hematite, and goethite in the pigments, along with manganese-containing minerals for darker tones (Lascaux, Wikipedia summary of the pigment studies).

Black came from two sources. Manganese-bearing minerals give a dense, matte black or dark brown. Charcoal, made from burned wood, gives a softer black that can be used like a drawing stick or ground into powder. At Lascaux charcoal may have been used only sparingly. At Chauvet the black drawings are usually described as charcoal, although some researchers have questioned the source of the black pigment in places (Chauvet Cave, Wikipedia).

Pigment	Raw material	Color	What the evidence shows
Red ochre	Hematite, an iron oxide	Red to red-brown	Used in Chauvet hand stencils and across Lascaux
Yellow ochre	Goethite, a hydrated iron oxide	Yellow to brown	Identified in Lascaux pigment analyses
Manganese black	Manganese-bearing minerals	Black to dark brown	Identified at Lascaux
Charcoal	Burned wood	Black	Dominant black at Chauvet; used sparingly at Lascaux

A pigment is only a powder. To stay on a wall, it needs a medium. The leading proposal for Lascaux is that the powder was suspended in animal fat, groundwater, or clay and then swabbed, blotted, or blown onto the rock. Nobody has recovered a recipe. The honest summary is that researchers know the colorants well and the binders only by inference.

From yellow rock to red paint: heating and grinding

Raw ochre does not come in every shade, and Ice Age people seem to have known how to change it. When goethite, a hydrated iron oxide, is heated, it loses water and converts to hematite, shifting from yellow toward red. Archaeologists have argued that some Upper Paleolithic groups in Europe heat-treated ochre on purpose, though how often this was deliberate is disputed.

A campfire could also have taught the trick by accident. Either way, the chemistry is simple, like toasting bread: heat drives out water and the color shifts.

The same toolkit appears far earlier. At Blombos Cave in South Africa, a team led by Christopher Henshilwood reported in Science in 2011 an ochre-processing workshop about 100,000 years old. It included ochre, bone, charcoal, grindstones, hammerstones, and two abalone shells that held a liquefied ochre mixture. The purpose of the mixture is unknown, but the find shows that people were grinding, mixing, and storing pigment tens of thousands of years before the famous French caves were painted.

Grinding matters because particle size sets how a pigment behaves. Finer powder spreads more evenly and sticks better, while coarse grains leave a gritty, broken line. A grindstone with a patch of red on it is a small piece of evidence for a surprisingly deliberate industry.

How the paint went on: hands, daubs, and blown pigment

Surviving marks suggest more than one application method, and some of them are easy to read. A line can be a stick of charcoal or manganese used like a crayon. A patch of fill can be a daub pressed on with a pad or a hand. Wikipedia's summary of the Lascaux studies lists swabbing, blotting, and blowing through a tube as the leading ideas for how the paint reached the stone (Lascaux, Wikipedia).

The most direct evidence is the hand stencil. To make one, a person pressed a hand against the wall and blew pigment around it, leaving a pale silhouette inside a halo of color. Chauvet has red hand prints and stencils made this way. The technique works like a stencil from a hardware store, except that the template is the artist's own body.

Brushes are harder to prove. Stone does not preserve the soft tools that might have been used, and researchers are cautious about naming the material from the marks alone. Fur, moss, fibers, and chewed twigs have been proposed in various places, but a firm identification requires the tool itself or its residue. That caution is worth keeping in mind whenever a documentary describes a "prehistoric brush" as established fact.

Preparing the wall: scraping, engraving, and the rock's own shape

Paint was not always the first step. In Chauvet, artists cleared loose debris and mineral concretions from parts of the wall before drawing, which left a lighter, smoother surface that makes dark lines easier to see. In some places they also incised lines around outlines, creating grooves that catch light and can suggest depth or movement (Chauvet Cave, Wikipedia).

At Lascaux, some designs are engraved where the rock is soft enough to cut. Elsewhere the wall's own shape appears to have been used: a bulge can become a shoulder, and a ridge can become a back. This is one of the reasons caves are better understood as three-dimensional sites than as flat galleries. A photograph of a panel removes the contour that the artist exploited.

An analogy helps here. Think of a sculptor who finds a figure in a block of stone. Cave artists did not carve, but they did read the surface, and the wall participated in the picture. Lascaux's artists working on curved ceilings and Chauvet's artists working on cleaned panels were solving the same problem from different directions: how to make a mark that the light would reveal.

Lighting the dark: torches, fat lamps, and a test cave

Painted chambers can be hundreds of meters from the entrance, so the artists carried light. Lascaux has produced a sandstone lamp about 17,000 years old, 224 millimeters long and weighing 560 grams, with a bowl that held black combustion residue and a handle engraved with signs. Analysis of the remains identified juniper, apparently a wick used for ignition (Stone Lamps of the Palaeolithic). Sophie de Beaune and Randall White reported in 1993 that about 70 lamps had been found at Lascaux in groups, although the archaeologist André Glory argued that the Lascaux object may have been a ritual incense or smoke burner rather than a lamp.

A 2021 study in PLOS ONE by Medina-Alcaide and colleagues tested what such light would have been like. They built replicas of torches and fat lamps and burned them in a Spanish cave. Wooden multi-stick torches burned for an average of 41 minutes, with a range of 21 to 61 minutes, and lit up to almost six meters around them. They gave about five times the light of a double-wicked grease lamp but made a lot of smoke (Phys.org summary of the study).

Grease lamps fueled by bone marrow from cow and deer were dimmer, comparable to a candle, and lit a space up to about three meters across. They burned steadily with little smoke for over an hour. The authors saw the two as complementary: torches for getting around and crossing wide spaces, lamps for lighting a small area for a long time.

That balance gives a rough sense of the work schedule. A painter at a wall might need a steady lamp, a fresh torch for the walk, and a supply of fuel. It also hints at why many panels are in smaller side chambers, although no study has shown that the lighting dictated where artists chose to paint. Claims that flickering flames made animals "move" are interesting but speculative, and no experiment has demonstrated that the images were designed for that effect.

Why the paint survived tens of thousands of years

Survival is partly a matter of chemistry. Iron oxides such as hematite are already the end product of oxidation, which is why rust-colored soil does not rust further, and they hold their color under light and air. Charcoal is nearly pure carbon, which resists decay and does not bleach. By contrast, many plant dyes and organic colorants used in later centuries fade within decades, a problem we cover in our article on why colors fade.

The setting matters as much as the pigment. A deep cave holds a steady temperature, little wind, no rain, and no direct sunlight. At Chauvet, a rockfall sealed the entrance about 29,000 years ago, closing the cave like a jar with the lid on. Lascaux shows what happens when that seal is broken: within a few decades of heavy visitation, fungi and lichens were spreading across the walls, and the cave was closed to the public.

Mineral crusts add a second layer of protection. Where calcite grew over a painting, it can act as a clear glaze that locks the pigment in place, and it also makes the painting datable. Fresco painters later discovered a similar effect on purpose, because lime plaster absorbs carbon dioxide and hardens over the pigment. Our explainer on why fresco painting is so durable covers that process.

Not every Ice Age painting lasted. Open rock shelters, exposed cliffs, and wet caves have lost almost everything. The images that survive are a sample biased toward deep, dry, stable places, which means the record overrepresents certain kinds of caves and probably underrepresents work that was made outdoors, on skin, on wood, or on the body itself. The cave art we study is a fraction of a larger tradition.

Chauvet: a gallery sealed for about 29,000 years

The cave now called Chauvet-Pont d'Arc was explored on December 18, 1994 by Eliette Brunel-Deschamps, Christian Hillaire, and Jean-Marie Chauvet. Hundreds of animal paintings have been catalogued, showing at least 13 species, and the cave also holds cave bear skulls, paw prints in the clay floor, and rounded hollows thought to be bear nests. UNESCO inscribed it as a World Heritage site on June 22, 2014 (Chauvet Cave, Wikipedia).

Chauvet became famous in part because of an argument about its age. Early radiocarbon dates on charcoal drawings came back remarkably old, with one sample reported in 1999 at about 32,900 years, which was striking for work this skilled. A 2012 study using chlorine-36 to date rockfalls placed the sealing of the entrance about 29,000 years ago, which helped explain why the art survived.

The most complete chronology so far is a 2016 PNAS study led by Anita Quiles. It used 259 radiocarbon dates and concluded that people visited the cave in two phases: from about 37,000 to 33,500 years ago and from about 31,000 to 28,000 years ago. Most of the dated black drawings fall in the earlier phase. A 2020 recalibration put the oldest painting at about 36,500 years ago.

The lesson from Chauvet is that sophistication does not track age. An older cave with confident perspective, overlapping animals, and shaded fills overturns the picture of a slow climb from clumsy to refined. Skills were passed down inside communities, and a good tradition can arrive fully formed in a new place.

Lascaux: a cave closed by its own visitors

Lascaux was found on September 12, 1940, when a teenager named Marcel Ravidat followed his dog into a hole near Montignac in the Dordogne. He returned with friends Jacques Marsal, Georges Agnel, and Simon Coencas, and the priest and prehistorian Henri Breuil visited nine days later. The cave holds nearly 6,000 figures, of which over 600 are painted, and the paintings are usually estimated at 17,000 to 22,000 years old (Lascaux, Wikipedia).

Of the more than 900 figures identified as animals, 605 can be identified precisely, and 364 of those are equines (horses). The Hall of the Bulls is the most famous space. Its four large black aurochs bulls (extinct wild cattle) dominate 36 animals, and one bull is about 5.2 meters long, the largest animal discovered in cave art so far.

The story of how Lascaux was nearly destroyed is a chemistry lesson. Opened to visitors, it attracted about 1,200 people a day, and their breath, heat, and humidity changed the cave's air. Carbon dioxide levels rose, fungi and lichens spread across the walls, and the cave was closed to the public in 1963. A black mold outbreak led to another closure in January 2008. A full-scale replica, Lascaux IV, was inaugurated in December 2016. UNESCO inscribed the Vézère Valley decorated caves, including Lascaux, in 1979.

A cave wall is a delicate system. The paintings survived tens of thousands of years because the cave's temperature, humidity, and gas levels were stable. Visitors broke that stability within a few decades.

How anyone knows how old a painting is

Three dating approaches do most of the work, and each has a limit.

Radiocarbon dating measures carbon-14 in organic material such as charcoal. Living things absorb carbon-14 from the atmosphere, and after death it decays with a half-life of about 5,730 years, so the remaining fraction tells how long ago the wood was burned. Results must be calibrated against a curve built from tree rings and other records, and the calibration curves have been revised, which explains why Chauvet's oldest date has shifted slightly. The method works only on carbon-bearing pigment, and it only dates when the wood was burned, not necessarily when it was drawn on the wall. See our explainer on radiocarbon dating for the details.

Uranium-series dating is used when the pigment is mineral and cannot be radiocarbon dated. Over centuries, dripping water can deposit a thin film of calcite over a painting. That film traps uranium but starts with essentially no thorium, and uranium decays to thorium-230 at a known rate, so the ratio gives the age of the film. Because the film formed on top of the painting, the painting must be older, so the result is a minimum age. This is the method behind most claims about the oldest art. It is one case of the broader family covered in our guide to radiometric dating.

Rockfall and surface studies, such as chlorine-36, date when a rock face was exposed or sealed, and they help to fix when a cave entrance collapsed. They do not date paint directly, but they bracket the period in which the art must have been made.

Think of a dated mineral crust as a postmark on a letter sealed inside an envelope. It proves the letter existed by that date. It cannot say how much earlier it was written.

Site	Region	Age evidence	Dating basis
Chauvet	Ardèche, France	Visited about 37,000 to 33,500 and 31,000 to 28,000 years ago	259 radiocarbon dates, Quiles et al. 2016
Lascaux	Dordogne, France	Usually 17,000 to 22,000 years	Archaeological estimate, early Magdalenian
Maros caves	Sulawesi, Indonesia	Hand stencil, at least 39,900 years	Uranium series, Aubert et al. 2014
Leang Tedongnge	Sulawesi, Indonesia	Warty pig, at least 45,500 years	Uranium series
Leang Karampuang	Sulawesi, Indonesia	Three figures with a pig, at least 51,200 years	Laser uranium series, 2024
La Pasiega, Maltravieso, Ardales	Spain	Red marks and stencils, over 64,800 years	Uranium-thorium, Hoffmann et al. 2018

Older than Chauvet: Sulawesi, Spain, and who painted first

For a long time Europe seemed the birthplace of figurative art, partly because Europe's caves had been studied the longest. That view shifted in 2014, when Maxime Aubert and colleagues reported uranium-series dates from Sulawesi in Indonesia. A hand stencil in the Maros caves was at least 39,900 years old, and a babirusa (a "pig-deer") at Leang Timpuseng was at least 35,400 years old. The authors concluded that people were making rock art about 40,000 years ago at opposite ends of the Eurasian world.

The record has been moved back further since. In 2024, a team reported a painting at Leang Karampuang, in the Maros-Pangkep region of South Sulawesi, of three people gathered around a large pig, measuring about 92 by 38 centimeters in a single dark red pigment. Laser dating of the calcite on top gave a minimum age of 51,200 years. Co-author Maxime Aubert called it the oldest evidence of storytelling. Chris Stringer of the Natural History Museum, who was not part of the study, called the date "quite provocative," said the work looked sound but needed further dating, and suggested that representational art began in Africa (Al Jazeera report).

The Iberian Peninsula adds another twist. A 2018 Science study by Dirk Hoffmann and colleagues dated calcite crusts over red marks, a hand stencil, and painted formations at La Pasiega, Maltravieso, and Ardales. The crusts were older than 64,800 years, which is at least 20,000 years before modern humans reached Europe by the authors' reckoning. They inferred that Neanderthals made the art. That conclusion rests on timing rather than any direct sign of who did the painting, and it remains debated.

Note what these dates do and do not show. Minimum ages cannot rank cultures against each other, and a handful of preserved panels is a tiny sample of what once existed. What the results suggest is that image-making is old, widespread, and probably older still than any painting that has survived.

What the choice of animals suggests

Cave art is not a snapshot of daily life. Horses make up most of the identified animals at Lascaux, followed by deer and aurochs. Yet at many Ice Age sites the animals people actually ate most, such as reindeer, are rare in the art. Researchers generally read this as a sign that animals were chosen for symbolic reasons rather than their importance as food, although preservation and changing climates complicate the comparison.

Chauvet pushes the point further. Its paintings include lions, rhinoceroses, and bears, species that were dangerous rather than dinner. The cave floor itself is covered with cave bear traces, which might have shaped how people thought about the place. Beyond the animals there are hand prints and stencils, so the repertoire is not purely zoological.

The cognitive archaeologist David Lewis-Williams proposed that some images reflect altered states of consciousness, with the dark, enclosed spaces of caves acting as sites for ritual. This is one of several interpretations, including hunting magic, storytelling, and territorial marking, and none of them has been demonstrated. Because each fits some sites and fails at others, the careful position is that there was probably no single purpose behind a tradition that lasted tens of thousands of years.

What the paintings do show is a set of mental abilities that are fully modern. To draw a lion in a cave, an artist needs to keep an absent animal in mind, plan a trip with materials and light, and share conventions with a viewer. If this sounds like language, it is not language, but it is another form of symbolic communication. Our piece on the psychology of creativity looks at how the brain does this kind of work today.

What people get wrong about cave paintings

"Cave men lived in these caves." Painted chambers deep underground are generally not living spaces. Chauvet shows only two phases of human visits, and the floor is covered with cave bear traces, including hollows thought to be bear nests.

"The art got steadily better over time." The Chauvet-versus-Lascaux comparison undermines this, since the older cave is not the cruder one. Styles shift with place and tradition, not along a straight line toward realism.

"Prehistoric artists used brushes and a standard paint recipe." The best-supported methods are hand application, swabbing, and blowing, and the binder recipes remain inferred. Be wary of any source that gives you exact ingredients.

"Radiocarbon dates the paint." It dates carbon in the pigment, and only where the pigment is charcoal. For mineral paint, researchers rely on crusts formed over the painting, which give minimum ages.

"We know what the paintings mean." We do not. We know many details of how they were made, which is a very different kind of knowledge from understanding why.

For comparison with how later artists understood pigments, see our articles on how pigments create color in paintings and how ancient Egyptians made their pigments. The Egyptian palette includes synthetic colors, which makes a useful contrast with the mined earth colors of the caves.

The takeaway

A cave painting is a chain of choices. Someone found the right rock, heated or ground it, mixed it with something that would stick, brought fire into the dark, prepared the wall, and made a mark that could be read by torchlight. The materials were simple, but the planning was not, and the fact that the same pigments recur from southern France to Sulawesi shows how old and widespread the habit of picture-making is. Each new dating result has pushed that history back, and the oldest paintings found so far are very unlikely to be the first ones made.

FAQ

Q: What paint did cave painters use? A: They used mineral and carbon pigments: iron-oxide ochre for red and yellow, manganese-bearing minerals and charcoal for black. The binder is less certain. Animal fat, groundwater, and clay have been proposed for Lascaux, but no recipe has been recovered.

Q: How were hand stencils made? A: An artist pressed a hand to the wall and blew pigment around it, which left a pale silhouette surrounded by color. Chauvet has red hands made this way, and similar stencils at Sulawesi have been dated to at least 39,900 years ago.

Q: How do scientists date cave paintings? A: Charcoal can be radiocarbon dated directly. Mineral paint cannot, so researchers use uranium-series dating on calcite that grew over the painting, which gives a minimum age. Chlorine-36 dating of rockfalls adds context by showing when entrances were sealed.

Q: How did the artists see inside the caves? A: They used fire. Lascaux has yielded stone lamps, one with a juniper wick, and a 2021 experiment found that wooden torches burned for about 41 minutes on average, while bone-marrow grease lamps burned for over an hour with little smoke.

Q: What is the oldest known cave painting? A: The oldest dated figurative painting is at Leang Karampuang in Sulawesi, at least 51,200 years old. Red marks in Spain have been dated to more than 64,800 years ago, but they are not figurative, and their attribution to Neanderthals is debated.

Q: Why is Lascaux closed to visitors? A: Crowds of around 1,200 visitors a day altered the cave's carbon dioxide, heat, and humidity, and fungi and lichens spread on the walls. It was closed to the public in 1963, and visitors now see a replica, Lascaux IV, which was inaugurated in December 2016.

Q: Why did people paint animals they did not hunt? A: Nobody knows for certain. At many sites the commonest foods, such as reindeer, are rare in the art, which suggests symbolic or ritual reasons. Proposed explanations include altered states of consciousness, storytelling, and hunting magic, and none is proven.

---

Read next:

• How Impressionism Changed Art History
• The Psychology of Color in Art and Design
• Street Art and Graffiti: The Science and History Behind the World's Wildest Art Form