Two Homo erectus skulls from Yunxian, central China, have been redated to 1.77 million years ago using cosmogenic nuclide burial dating — the decay ratio of aluminum-26 to beryllium-10 in quartz grains, which functions as a clock that starts ticking when sediment is buried beyond the reach of cosmic rays. The previous estimate placed them at roughly one million years. The revision adds 770,000 years.
At 1.77 million years, the Yunxian skulls are contemporaneous with the Dmanisi fossils from Georgia — the earliest widely accepted Homo erectus outside Africa. But the two populations look different. The Dmanisi individuals have small brains, roughly 600 cubic centimeters, comparable to late Homo habilis. The Yunxian skulls are substantially larger, exceeding 1,000 cubic centimeters. Same species, same age, separated by thousands of kilometers, with brains nearly twice the size.
The standard dispersal narrative accommodated this by placing Yunxian later — a descendant population that had time to evolve larger brains after the initial migration. The redating eliminates that cushion. If both populations existed simultaneously, the brain size difference cannot be the product of gradual change along a single lineage after dispersal. Either the variation was already present in the source population before any group left Africa, or differentiation happened extraordinarily fast after splitting.
The first possibility implies that early Homo was far more variable than the fossil record suggests — that Africa contained populations spanning the full range from 600 to 1,000 cc, and different subsets migrated in different directions. The second implies that brain size can change dramatically within a few thousand generations under strong selection, which is fast but not impossible given what we know about cranial plasticity in mammals.
The deeper point is methodological. The cosmogenic nuclide technique dates the sediment, not the fossil. It measures when grains of quartz were last exposed to cosmic radiation — when they were last at or near the surface. Once buried deep enough, the cosmogenic isotopes begin to decay at known rates, and the ratio between two isotopes with different half-lives gives the burial time. The technique is independent of the fossil itself, which means it cannot be biased by expectations about what age the fossil “should” be. Previous dating methods were more susceptible to contamination, reworking, and the subtle influence of theoretical priors on which dates were accepted and which were questioned.
The revision changes the question. The old question was: how did Homo erectus brains grow over time? The new question is: why were they already different at the start?