Humanity reached Australia even earlier than previously thought, according to new dating of tools from a rock shelter in Arnhem Land, Northern Territory. The arrival was so early it throws doubt on ideas of when modern humans first left Africa, and the timing of interbreeding with Neanderthals and Denisovans.

The question of when people arrived in Australia has attracted considerable debate since isotope dating was invented. Estimates have varied between 40 and 60,000 years. The earlier dates have created some puzzlement, since they bump up so closely against when the first modern humans are thought to have left Africa.

Now that confusion could increase further, with the announcement in Nature of a dig at the Madjedbede rock shelter. Previous efforts to date items from the same site produced ages between 50,000 and 60,000 for the oldest discoveries, but doubts were cast over these, leading Dr Chris Clarkson of the University of Queensland to try again.

Clarkson told IFLScience that when the previous study of the Madjedbede site used optical stimulated luminescence to measure the ages of sediments in which the tools were found, it was the technique's first application in Australia. Moreover, previous researchers had “only taken a small series of samples from one wall of the site” and each sample had been large enough to mix together layers, increasing uncertainty. Consequently, many anthropologists were skeptical of the findings, but doubters were more likely to think the estimates overstated than understated the artifacts' ages.

Yet when Clarkson had the layers re-tested he not only produced an estimate of 65,000 years for tools found there, but had reasons to feel confident in the figure.

The lowest excavated layers are exceptionally dense with relics of human occupation. In the subsection of the site that has been examined, an astonishing 11,000 items indicative of human occupation have been found, from stone flakes to the world's oldest edge ground hatchets. Many of these fit together in ways that make Clarkson confident they have not moved to lower layers, a problem which often bedevils attempts to estimate the age of objects from their strata.

At the front of the site, items were encased in rock rubble from a roof-fall, preventing movement. Indeed, there are even items at greater depth, suggesting humans may have visited the shelter even earlier than 65,000 years ago, but Clarkson told IFLScience there was sufficient doubt over these items ages for them to be excluded from the paper.

Madjedbede hosts signs of human habitation at every level from 65,000 years ago on, but three periods are particularly abundant. These eras (65-53,00 years, 27-13,000 years and since 7,000 years ago, respectively) coincide with times when the region was wet, suggesting the shelter was more heavily used in those periods. The reasons for this are unclear. “Maybe the population was greater at those times, or maybe people used the shelter more to get out of the rain,” Clarkson said to IFLScience.

 

The results appear to lay to rest theories that humans wiped out much of Australia's megafauna soon after arriving. Major extinctions occurred around 40,000 years ago, and the long co-existence between humans and great beasts makes it unlikely we were the primary cause of their demise.

Another puzzle concerns why Clarkson's dates conflict with the earliest evidence for human occupation elsewhere in Australia. Deaf Adder Gorge, 70 kilometers (40 miles) south of Madjedbede also contains relics estimated at 50-60,000 years old, but there is a dearth of known sites of similar age elsewhere. Clarkson told IFLScience he thinks it is more likely we simply haven't found the right locations, but it is also possible the first arrivals lingered in the area for a long time before dispersing.

One the other hand, in the Ice Age conditions of the day Madjedbede was approximately 300 kilometers (200 miles) from the coast. There were almost certainly earlier stopping places that are now under water, and therefore almost impossible to find.

Broader questions concern humanity's great migration out of Africa. As Clarkson pointed out, the site houses complex technologies, as would be expected of people who made successful ocean crossings. Our oldest remnants from South East Asia were much younger, but Clarkson says what can be seen here; “Tells us these people were innovating, adapting to a different country with new flora and fauna.”

Indigenous Australians have DNA from both Neanderthals and Denisovans. Some measurements with biological clocks have placed the times at which modern humans interbred with these peoples as recently as 50-60,000 years ago. However, since we seen no evidence either group made it to Australia, Clarkson's results suggest that either their DNA was brought to the continent in subsequent waves of settlement, or the interbreeding happened earlier, as more recent studies suggest.

While Clarkson acknowledges later arrivals may have mixed with the people who first inhabited Madjedbede, he told IFLScience there is too much continuity in the technology and culture seen there for the first inhabitants to have been displaced by a subsequent wave.

Many questions might be answered if DNA from Madjedbede's early residents can be retrieved. This has yet to be done, but Clarkson told IFLScience some samples are undergoing further study that might provide this holy grail.