An exhibit at the Neanderthal Museum in Croatia Nikola Solic / Reuters

The first step to re-creating 50,000-year-old technology is to collect a bunch of rocks. So began Andrew Sorensen’s plan to study a great mystery in archaeology: how Neanderthals controlled fire.

Sorensen, an archeologist at Leiden University, collected a special kind of rock called flint off the beaches of England. If you hit it in just the right ways, flint will break to expose sharp edges that can be used to butcher meat, scrape hides, and cut wood. And if you strike it against a mineral called pyrite, sparks will fly. Flint plus pyrite plus tinder equals fire.

Archeologists have found evidence of Neanderthal fire pits. They have even found tar that Neanderthals likely made by deliberately heating birch bark. What they have never found are tools that Neanderthals could have used to start fires on demand. Without it, Neanderthals would have needed to collect fire from natural sources such as lightning strikes, which would have required walking long distances to find fuel to keep fires going and enduring cold spells with raw food when they went out. The mastery of fire would have made life much easier. Many think it was a key turning point in human evolution.

Sorensen suspected that flint tools called bifaces may hold the answer. Bifaces are essentially hand axes used in all sorts of cutting—a “Neanderthal Swiss army knife,” as Sorensen put it. So he took a bunch of flint home, shaped it into bifaces, and tried to create fires in an indoor lab. Through trial and error, he found that striking pyrite against the flat side of the biface produced sparks that could ignite tinder. “I wasn’t setting off any fire alarms or anything,” he says. He extinguished the tinder instead of blowing on it and feeding it progressively larger pieces of fuel.

Biface and pyrite that can be used to make fire (Andrew Sorensen)

Now Sorensen had a hypothesis to test. When he repeatedly struck the pyrite against the bifaces, the scraping left marks on the rocks. If Neanderthals did this, presumably their pyrite left marks on their bifaces, too. So he and his co-authors went looking for Neanderthal bifaces in museums to study marks called microwear. In a new paper in Scientific Reports, Sorensen and his co-authors suggest that Neanderthals used bifaces and pyrite to start fires, based on the similar microwear patterns on real Neanderthal stone tools and on the tools he re-created in a lab. Sarah Hlubik, an anthropologist at Rutgers University who also studies the early origins of fire and was not involved in the study, says the paper is “really exciting.” It’s the first physical evidence of Neanderthals starting fires.

But it’s not easy to look for such subtle hints in old rock, especially as many things can leave similar marks. “This question of how you tell apart what sandstone versus quartzite versus limestone versus pyrite looks like, that’s really new,” says Rebecca Wragg Sykes, a researcher at the University of Bordeaux who was not involved in the study. The authors “know this is at the edge of knowledge,” she added. The technique could open new doors for studying stone technology.

Dennis Sandgathe, an archaeologist at Simon Fraser University and a skeptic of the idea that Neanderthals could start fires, remains unconvinced. His skepticism comes from 40,000-to-100,000-year-old Neanderthal caves that he’s excavated in France. His team has found plenty of evidence of fires during relatively warm periods at these sites, but none in relatively cold periods. He thinks it’s because Neanderthals were collecting fire from lightning strikes, and lightning is more common when it’s warm than when it’s cold. Sandgathe also says pyrites and stone bifaces should be in the same layer in the archeological record if they were being used to make fire, but he has never found them together.

Still, he admits, the archeological record is a patchy one. “It’s a pretty crappy source of data. It’s, unfortunately, the only one we really have,” he says. And this absence of data makes it hard to figure out 50,000-year-old technology—even with all of our modern tools and know-how.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.