Somewhere in a Beijing collection sits a fragment of a child’s lower jaw that has spent the past three years quietly annoying everyone who studies human origins.
It surfaced at Hualongdong, a collapsed cave in Anhui province in eastern China, alongside the remains of around 16 people who died roughly 300,000 years ago. The jaw fits a partial skull catalogued as HLD 6, a youngster of about 12 or 13. Xiujie Wu of the Institute of Vertebrate Paleontology and Paleoanthropology in Beijing led a team that set the bone against 83 other fossil hominins and a sample of living humans. They published the results in the Journal of Human Evolution.
What came back was a mess, in the productive sense. Heavy bone through the body of the jaw, the build of an archaic human, sat beside a slimmer, more modern-looking hinge, with faint suggestions of a chin that never quite arrives.
Why a missing chin matters
A chin seems a thin peg to hang a taxonomy on, and yet it’s one of a handful of skeletal features unique to Homo sapiens. Neanderthals did without. Denisovans, as far as their sparse remains suggest, did too. María Martinón-Torres, director of Spain’s National Research Center on Human Evolution and a co-author on the study, told CNN that HLD 6 has no true chin, only weakly expressed traits that seem to anticipate one.
Speaking to Scientific American, she likened the blend to fossils from Jebel Irhoud in Morocco, similar in age and usually read as early members of the sapiens lineage. If that comparison holds, a population in eastern China was wearing modern facial hardware at the same time as populations in north Africa, which is not how the textbook version goes.
One juvenile jaw is still one juvenile jaw, and the team stopped well short of naming a new species.
A crushed skull from Hubei
The older problem is considerably bigger. In 1990, a cranium came out of a river terrace above the Hanjiang in Hubei province, flattened by something like a million years of sediment. Its geological age runs between 0.94 and 1.10 million years. Since nothing more advanced was thought to exist that early, Yunxian 2 was filed under Homo erectus by default and largely left there.
A team led by Xijun Ni of Fudan University and the Chinese Academy of Sciences, working with Chris Stringer of the Natural History Museum in London, put it through CT scanning and structured light imaging, unpicked the fractures digitally and rebuilt the thing. Writing in Science in September 2025, they place the skull inside the Asian longi clade, the group that probably includes the Denisovans and forms the sister lineage to Homo sapiens.
The swing is substantial. A Natural History Museum release frames it in round numbers: the origin of Homo sapiens pushed back by some 400,000 years, to just over a million years ago. The tree behind that claim is more exact. It puts the split between the sapiens lineage and the longi and Denisovan lineage at roughly 1.32 million years ago, with the Neanderthal lineage splitting off earlier still, near 1.38 million years ago. Prior genetic estimates had put that same divergence nearer 600,000 years, a gap closer to 700,000 years than the headline figure above suggests. Stringer is quoted in that release calling Yunxian 2 “an early member of the longi clade and linked to the Denisovans”.
Where other researchers dig in
“Unwarranted given all the uncertainties.” That’s how Andy Herries, an archaeologist at La Trobe University with no role in the study, summed up the idea of filing the fossil into H. longi, in comments to ABC Science. His procedural objection lands harder. The reconstruction borrowed elements from Yunxian 1, the more heavily crushed of the two skulls from the site, which risks feeding assumptions into the very shape under analysis.
In the same report, Ni conceded that a fossil this old yields no usable genetic data, and that protein testing on animal bones from the site came up empty too. The whole case then rests on comparative anatomy and phylogenetic modelling, with no genetic evidence to weigh against it. Speaking separately to CBS News, Herries put the deeper doubt more bluntly: fossil shape is “not always a perfect indicator for human evolution”. Two lineages under similar pressures can arrive at similar proportions without being close kin, which is why morphology and molecular clocks so often talk past each other.
A single paper, however carefully argued, is a proposal rather than a consensus, and both of these findings sit in a field where the sample size runs to dozens of individuals across a million years.
That leaves the two fossils doing the same work from opposite ends of the last million years, pulling the tidy diagram out of true. Populations meeting, parting and meeting again look less like a branching tree and more like a braided river, silt shifting between channels. Nothing pulled out of a Chinese cave so far has lined up neatly for the family photograph, and the ones still in the ground are unlikely to start.
