7 OCT 2026 — The oldest fossils that are unmistakably animals are about 574 million years old. A study published on 2 October in Science Advances argues that animals themselves are probably far older, arising somewhere between about 800 and 700 million years ago. The argument does not rest on a new fossil. It rests on showing why a much-used piece of evidence for a younger date does not hold.
The team, led from the University of Oxford with colleagues at ETH Zürich, Yale and the University of California, Berkeley, tested that argument against Mongolian fossils.
Two records that disagree
Scientists have long had two answers to when animals appeared. Body fossils, from the soft-bodied Ediacaran organisms onwards, reach back only to about 574 million years ago. Chemical traces in ancient rocks, lipid biomarkers linked to sponges, point to animals before 635 million years ago, the paper says.
Molecular clocks are meant to settle such disputes. They use the genetic differences between living species, plus fossil dates, to estimate when lineages split. But a clock needs both a minimum age (animals must be older than their oldest fossil) and a maximum age, a point before which animals had not yet appeared. The maximum is harder to defend, and the paper shows how much the answer depends on it.
Testing a fossil site in China
Recent studies used the Weng'an Biota in China, about 590 million years old, as a maximum. Its microscopic fossils are preserved in exceptional detail, yet none is definitely an animal. The logic was that if animals had existed then, Weng'an would have caught them.
The same gap in Mongolia
The Oxford team checked that logic against the Kheseen Biota in Mongolia, which is preserved the same way, in phosphate rock, and shares some of the same microfossils. It formed about 550 to 526 million years ago, when animals were already living elsewhere, including ones with mineral skeletons. Yet the Kheseen rocks contain no definite animals.
"The Kheseen Biota breaks the argument that the exceptional microfossils of Weng'an mean we would have seen animal fossils in the assemblage had they existed at the time," Ross Anderson, the senior author, said in the university's release. If animals can be missing from a site formed when they certainly existed, their absence from Weng'an proves little.
A new clock, and a new window
With Weng'an set aside, the authors argue that older deposits make better maximum ages. These include the Svanbergfjellet Formation in Svalbard and the Bitter Springs Group in Australia. Both are well preserved and lack animals, and both are dated to between about 830 and 790 million years ago.
Using the Svanbergfjellet limit, their clock puts the last common ancestor of all animals at about 772 million years ago, give or take about 40 million years. The Bitter Springs limit gives about 791 million. Either way, the origin falls before or during the Cryogenian, the period of the Snowball Earth glaciations, and in all their analyses most of the later diversification still comes after it.
What remains uncertain
The study does not find an older animal. It argues that one key calibration was unreliable and that the result depends on which maximum is chosen, with wide error margins either way. The authors say this uncertainty limits what can be said about why animals arose, whether through rising oxygen, more nutrients in the ocean or the ice ages themselves.
"The causes and trajectory of animal evolution will remain uncertain," they write, until fossil and chemical evidence gives "sufficient confidence in maximum calibrations".
View author profile → · Editorial policy
