Read our COVID-19 research and news.

Mirarr landowners Mark Djandjomerr (center) and May Nango (right) observe researcher Elspeth Hayes sampling stone tool residue near Madjedbebe.


A find in Australia hints at very early human exit from Africa

As archaeologist Chris Clarkson was excavating a rock shelter in northern Australia one day in 2015, May Nango of the aboriginal Mirarr group brought her grandchildren to look at the pit. She pointed to a spot near the back wall of the red sandstone cliff and told the children that it was a wonderful place for their ancestors—the "old people"—to sleep 65,000 years ago, says Clarkson of the University of Queensland in Brisbane, Australia.

Nango's tale was more than an aboriginal "dreamtime" story. She was one of the first to hear from Clarkson's team about new scientific dates for the Madjedbebe rock shelter in Australia's Arnhem Land, a region the Mirarr still call home. The dates, based on new excavations and state-of-the-art methods, push back the earliest solid evidence for humans in Australia by 10,000 to 20,000 years and suggest that modern humans left Africa earlier than had been thought. Published this week in Nature, the findings also hint at when modern humans interacted with other archaic humans.

This early date will force the field to "rethink fundamentally the whole issue of when our species started to colonize Asia," says archaeologist Robin Dennell of the University of Sheffield in the United Kingdom.

The timing of the peopling of Australia has been contentious for decades. Many archaeologists split into two camps, favoring settlement either 60,000 years ago or sometime after 50,000 years ago, depending on whether they trusted the dates from certain sites. Last year, geneticists analyzing DNA from living Aborigines joined the fray, but they came up with a wide range of dates, from 50,000 to 70,000 years ago.

The Madjedbebe rock shelter, formerly known as Malakunanja II, has always been central to the issue. Known for its striking rock art, researchers proposed in 1989 that the shelter was the oldest human occupation in Australia, after they dated sediments containing stone tools to 50,000 to 60,000 years ago using the then-experimental method of thermoluminescence. But skeptics suggested that the 1500 tools and other artifacts could have drifted downward over time in the sandy sediments or that animals or termites had disrupted the layers.

Clarkson had long wanted to re-excavate Madjedbebe to resolve the controversy. Geochronologist Richard "Bert" Roberts, now at the University of Wollongong in Australia, who did the first dates, agreed to redate the site with Wollongong geochronologist Zenobia Jacobs, using optically stimulated luminescence (OSL) dating, a higher resolution form of thermoluminescence dating.

With Aborigine permission, the team reexcavated the site in 2012 and 2015 with painstaking stratigraphic controls. They found hundreds of thousands of new artifacts, including "elaborate" technologies such as the world's oldest ground-edge stone axes, grindstones for pulverizing seeds, and finely made stone points that may have served as spear tips. The earliest people at the site also used "huge quantities of ochre" and are the first humans shown to have used reflective mica to decorate themselves or rock walls.

When humans went down under

Artifacts dated to 65,000 years ago from Madjedbebe support suggested dates of more than 45,000 years for a handful of other Australian archaeological sites.


The team took extensive steps to rule out the migration of artifacts between layers, for example by refitting together broken stone tools found in the same layer. Jacobs dated quartz grains from various layers with OSL, determining when light last struck each grain and thus when it was buried. She dated 28,500 individual grains from 56 samples, checking to be sure that the dates were in proper order, growing older from top to bottom layers. Using a Bayesian statistical technique to narrow the margins of error, she concluded that the oldest human occupation was 65,000 years ago, with a range of about 60,000 to 70,000 with 95% probability. "I think we nailed it," she says.

Other dating experts agree: "I feel really good about the dates," says Sheffield geochronologist Edward Rhodes, calling the resulting chronology "highly robust."

Because people must have traveled across the islands of Southeast Asia to get to Australia, the date suggests humans were moving through Indonesia at the same time as Homo floresiensis, the tiny extinct human nicknamed "the hobbit," was living on the island of Flores; the last date for that species is 60,000 years ago, although so far there's no evidence of encounters between humans and hobbits.

The authors also suggest the new date of 65,000 years for the peopling of Australia pushes back the time when modern humans coming out of Africa mated with archaic species in Asia, such as Neandertals and Denisovans. Living Aborigines carry traces of those two species' DNA, which their ancestors must have acquired by mixing somewhere in Asia before they reached Australia.

But such early mixing with Denisovans and Neandertals is at odds with genetic evidence from living Aborigines and nearby Melanesians, says population geneticist David Reich of Harvard University. Analyses of these people's DNA "confidently" suggests that the interbreeding happened only 45,000 to 53,000 years ago, Reich says. "If these [new] dates are correct, they must be from a human population that was largely replaced by the people who are the primary ancestors of today's Australians and New Guineans," he says. If so, today's Mirarr descend from a later migration.

That makes sense to archaeologist Jim O'Connell of the University of Utah in Salt Lake City, who has favored the later chronology. This is "the only reliable [early] date," he says. "I'd make the argument that the ancestor of [living] Australian Aboriginals and New Guineans got there later."

However, paleoanthropologist Michael Westaway of Griffith University in Brisbane still thinks there was only a single migration. He notes that genomic samples to date don't include Aborigines in northern Australia such as Nango and her family. Their DNA—or that of their ancestors—might help resolve the issue.

In any case, the glimpse of these ancient people's behavior thrills both scientists and Aborigines. "What a wonderfully sophisticated group of first colonists they must have been," says Westaway, a fact not lost on Nango. She proudly noted in a statement: "We are glad to see the old tools and paintings showing that we Mirarr have been caring for this land for so long."