Scientists have reconstructed a million-year-old skull found in central China, and say they've identified which group of early humans it belonged to.
The researchers suggest their findings, published today in the journal Science, push back when modern humans evolved.
Our species, Homo sapiens, is the only living species in the Homo genus.
But in the past million years, many other Homo species have wandered the Earth, and palaeontologists have have been trying to reconstruct the human family tree.
Previously, researchers thought the common ancestor of all human species, Homo erectus, diverged into several groups about 600,000 years ago.
These include species that would eventually become us, as well as our sister species such as the Denisovans, who have left fossils in caves around eastern Asia, and the "Dragon Man" or Homo longi.
Co-author Xijun Ni, a palaeoanthropologist at Fudan University, said the much older skull in this study was closely related to Denisovans and H. longi.
This suggests modern humans and our sister species diverged from H. erectus some 400,000 years prior to when previously thought.
"Based on our new discovery, we challenged the established timeline of human evolution," Professor Ni said.
But not everyone was convinced by the findings.
The mysterious skulls found in Hubei
The study focused on a fossil found in China's Hubei Province in 1990, in a district then called Yunxian.
It was one of two skulls found in the area, both of which had been squashed underneath one million years of sediment, and distorted out of shape.
Study co-author Chris Stringer, a palaeoanthropologist at London's Natural History Museum, said the Yunxian skull had been classified as the species H. erectus "by default", since other human species were thought not to have evolved yet.
"We doubted that was the correct classification, even from the distorted fossil, so we wanted to reconstruct it properly to test that idea," Professor Stringer said.
Reconstruction prompts evolutionary rethink
The researchers used X-ray scanning to gain a detailed map of how the skull had been crushed, and then used a new digital reconstruction tool to build an idea of what the skull may have looked like in life.
They then analysed the reconstructed cranium, looking for structural clues to suggest which evolutionary groups it belonged to.
"Our work shows that this is a more derived form than H. erectus," Professor Stringer said.
Instead, the researchers suggest, it's a descendant of erectus, belonging to the clade of humans that evolved into the Denisovans and H. longi.
This means that, a million years ago, the H. longi clade of early humans had already diverged from the species that would eventually evolve into our species Homo sapiens.
Previously, genetic research had suggested this divergence only happened 600,000 years ago.
"If Yunxian sits close to the origins of both the H. longi/Denisovan and H. sapiens clades, it may represent one of the most important windows yet into the evolutionary processes that shaped our genus around one million years ago," Professor Stringer said.
He believed the research helped to resolve the "Muddle in the Middle" – the puzzle in human evolution from about 300,000 to one million years ago, where there is much debate on how various human species evolved.
"When I began working in human evolution over 50 years ago the East Asian record was either marginalised, or its fossils were only ever considered as direct ancestors of recent East Asians," Professor Stringer said.
"But what we now see from Yunxian — and from Harbin, Denisova, and many other sites — is that East Asia preserves crucial clues to the later stages of human evolution in general."
Is the reconstruction certain?
Andy Herries, an archaeologist at La Trobe University who wasn't involved in the new research, was not convinced by the analysis.
"It's an amazing fossil at an interesting time period, in an interesting part of the world," Professor Herries said.
"But trying to cram it into H. longi is, I think, unwarranted given all the uncertainties."
Professor Herries said that different tools for learning about these fossils were suggesting very different information on the time period.
He pointed out that the cranium reconstruction had been based on two different skulls, which would influence the researchers' analysis.
He said that genetic testing, or examining the proteins in the fossil, would help to resolve the skull's true identity.
But the researchers said that genetic testing couldn't be done on the skull.
"Given its age of around 1 million years old, it is impossible to get any informative genetic data from the fossil," Professor Ni said.
He added that the researchers had tested proteins in mammal bones discovered at the site, and found they couldn't get useful information out of them.
