Selected artifacts, including non-hominin mammal bones, found at the new site Zhaoyu Zhu et al / Nature

For hundreds of thousands of years, small bands of ancient humans ranged across a sandy, hilly grassland. They survived on the mammals around them—perhaps hunting them, perhaps scavenging for their carcasses—and their tools were rudimentary, razor-sharp blades formed from chipped stone. They lived in fear of the big cats and large predators that stalked their children.

And they were isolated. They likely rarely encountered other creatures who looked like them, beyond the 25 or 50 with whom they lived.

This was the strange existence of early-human ancestors on the savannah. It’s a classic scene—the image of hunter-gatherers roaming grasslands is just as at home in a biology textbook as it was in 2001: A Space Odyssey. But according to new research, this scene did not play out only in Africa.

Our ancient ancestors lived in China, too.

Ancient humans appear to have reached northwestern China about 2.1 million years ago, and they lived there for hundreds of thousands of years, according to a new study published Wednesday in Nature. It suggests that hominins migrated out of Africa much earlier, and spread much farther east, than once thought.

Previously, the earliest ancient-human presence outside of Africa had been a Homo erectus fossil found in a cave in Dmanisi, Georgia. It was dated to 1.85 million years ago. This newly discovered community of early humans lived roughly 250,000 years earlier than that group, and did so 3,500 miles to the east.

“We need to rethink when hominins first left Africa,” said Robin Dennell, an archeologist at the University of Sheffield and one of the authors of the paper. “We have shown that the earliest evidence from outside Africa is at least 2.1 million years old, and therefore 250,000 years—or 10,000 generations—older than Dmanisi in Georgia.”

“It’s so old that the earliest members of our own genus, the genus Homo, may have migrated out of Africa,” said Michael Petraglia, a professor of anthropology at the Max Planck Institute who was not involved in the new study. These creatures would have likely been Homo erectus; or possibly even Homo habilis, the first ancient primate to be called Homo.

Petraglia emphasized that the discovery also changes our understanding of Ice Age China. “They’ve added on something like 400,000 years of prehistory [in East Asia]—and it’s not every day you get to do that,” he told me. “Some of the oldest sites in China were previously only 1.6 or 1.7 million years old. They’re now saying the oldest sites are 2.1 million years old.

The research was led by Zhaoyu Zhu, an archeologist and climatologist at the Chinese Academy of Sciences. Zhu and his team have spent the past 13 years excavating a unique site on China’s Loess Plateau, a rare spot protected from erosion and glaciation, and continually buried in wind-blown sand over the past several million years.

While the new paper identifies a human presence, the researchers have not yet found any early-human fossils at the site. They have unearthed a wealth of early stone tools left behind by our ancestors, buried under many layers of solidified sand. These artifacts are mostly chipped flakes of stone, a type of primitive blade created by smashing two river-smoothed cobbles together. Hominins in Africa are known to have used this technology during the same period.

Claiming an ancient-human presence from a bunch of stone flakes could prove controversial among some researchers, who only feel confident dating an ancient-human presence when they find the remains of an early human, like teeth, a jaw bone, or ancient DNA.

But both the paper’s authors and outside experts told me they felt comfortable asserting that ancient humans lived on this site.

“It comes down to two general points,” said John Kappelman, a professor of anthropology and geology at the University of Texas at Austin who was not connected to the research. First, the stone tools “look like they were produced by humans,” he told me. They also show evidence of manufacture and maintenance. Some flakes have an almost serrated edge, suggesting that their creator smashed them against a cobble multiple times in order to improve them. Others “appear to show resharpening or sharpening,” he said, meaning their users attended to their tools and tried to improve them.

Second, the flakes are the most substantial feature found in this geological layer. “These are the largest pieces of stones—either the flakes themselves or the cores from which flakes were removed—in a very fine-grained sedimentary sequence,” Kappelman said. This supports the idea that they have a human, not a natural geological, origin.

Dennell, one of the authors of the paper, agreed. “Very importantly here, there are no geological processes that could have flaked these stones. The Loess Plateau is a stone free landscape—it is basically an enormous deposit of wind-blown dust, deposited year upon year by the winter monsoonal winds for the last 2.5 million years,” he told me in an email.

“Hominin remains are incredibly scarce,” he continued. “Their skeletons are very fragile, preservation is very rare, and they were not very common. In contrast, a single hominin can generate thousands of stone tools in a lifetime. Additionally, fossils will never indicate the first actual appearance of an animal—the first recorded appearance is always later than the first actual appearance, no matter whether it’s a hominin or a hippopotamus.”

“As an archaeologist, I have always found it strange that some refuse to accept artifacts unless found with hominin remains when the antiquity of humankind was first established in the early 19th century from stone artifacts in France and Britain, and not from skeletal evidence,” Dennell added.

Petraglia, the Max Planck professor who was not connected to the study, told me that the paper’s findings were “very strong.”

“I think it’s excellent,” he said. “I’m very excited about the work. It’s such a major finding, and it changes a lot of our views about the migration of early humans out of Africa, and when they got into Asia.”

But the study’s importance relies on another claim, too: that the authors know the stones are 2.1 million years old. To reach this conclusion, researchers deployed a technique known as paleomagnetism.

Paleomagnetism relies on a simple planetary fact: The Earth’s magnetic field seems to reverse every 800,000 years. As rocks harden, they record the planet’s magnetic field in their structure: Lava now cooling on the surface of Hawaii, for instance, will retain the Earth’s current North Pole–facing orientation for millions of years. Researchers can sample the surface of a vertical rock face and track these changes in magnetic polarity, matching the reversals they detect with well-dated reversal events recorded elsewhere in the world. The Loess Plateau rock turned out to align to two reversals captured elsewhere in the world, including one switch, 2.1 million years ago, first observed on distant Reunion Island.

Over the past decade, a more cosmopolitan notion of human evolution has emerged. Homo sapiens do not seem to have been the only advanced hominin to stalk the Earth, and they did not evolve in a single spot in Africa. (My colleague Ed Yong recently detailed an aspect of this idea.) Instead, Homo sapiens shared Earth with at least two other types of ancient human: Neanderthals, who occupied much of Eurasia; and the more mysterious Denisovans, who may have ranged throughout Southeast Asia.

Starting around 100,000 years ago—when genetic evidence suggests that Homo sapiens permanently departed Africa—these three forms of ancient human inhabited much of the same territory, interacting and occasionally mating with each other. Today, only Homo sapiens survives, though some modern humans retain Neanderthal or Denisovan DNA.

But this new find suggests that some early humans ranged much of the Earth well before that late date. “This is way earlier than Neanderthals—way, way earlier,” Petraglia told me.

The finding suggests that ancient humans migrated out of Africa many times, though these migrations were not always successful, he argued. “Some populations got all the way over to eastern Asia, but we have to imagine that these were small, sort of hunting-and-gathering populations. And while they may have mated across East Asia, it doesn’t mean they survived for a long period of time. Some populations might have become isolated, and some might have become extinct.”

These migrations may even have predated the worst of the modern Ice Ages. While the Earth endured some large swings in temperature about 2 million years ago, the mighty, continent-churning glaciations of the past million years had yet to parade down from the North Pole.

The Loess Plateau likely alternated between arid steppe and moist grassland every 40,000 years. Tools also appear to grow more scarce in the site during these cold and dry intervals, suggesting that ancient humans could only adapt so much to life out of the tropics. And it’s worth noting that the new study does not claim that hominins have continuously inhabited Asia for the past 2 million years.

Kappelman, the University of Texas professor, took a more robust view of this ancient-human community. He noted that different types of australopithecus—a hominin that briefly shared Africa with Homo habilis for 500,000 yearslived for millions of years without ever appearing to leave Africa. But 2.1 million years ago, as the first fossils from our genus, Homo, appear—scientists suddenly find evidence of an ancient-human presence across much of the Old World.

“What this paper suggests is: Boom! You get this dispersal, all the way across what was then the known Earth. The pieces were being filled in there very early on,” Kappelman told me. “It’s the kind of thing where, if we saw this for some other species, it would be remarkable.”

He could recall only one other mammal that moved into a new territory and immediately dispersed across it: ancient horses. “They originated in North America, then migrated into the Old World about 11 million years ago. And then, boom, it’s like gangbusters. They’re everywhere,” he said.

Ancient humans might have gone through a similar expansion, even before they gained all the traits that define a modern Homo sapiens. We just have less evidence of this explosion: Since ancient humans were carnivores, they never would have been as abundant as horses, which are lower on the food chain.

“The world is overrun with people today,” Kappelman said. “There was a period in time when it wasn’t.” The challenge is trying to piece together how humans went from scarce to ubiquitous. “What was it about their behavior? Because we don’t see australopithecus going out, even though they are around at the same time. We just don’t find other hominins in Eurasia until the rise of Homo.

We want to hear what you think. Submit a letter to the editor or write to