Cold case: the first animals may have arrived before the ice

A new study argues that the animal kingdom may be about 200 million years older than a widely used method has made it look. The paper, published in Science Advances on October 2, 2026, estimates that animals could have originated between about 800 million and 700 million years ago. That window reaches back to before the Cryogenian Period, and possibly into it. The Cryogenian is the age of the severe glaciations known as Snowball Earth. ^[1][2][3]

That is the headline, and here is the limit on it. Nobody has found an animal fossil that old. The date comes from a molecular clock, a statistical model of genetic change, and the authors say plainly that their analysis does not prove animals existed then. What it does is show that the older date is plausible once a questionable assumption is removed. ^[1][2]

The study was led by Orin M. Lole Durbin and Ross P. Anderson, with authors from Oxford, Virginia Tech, ETH Zurich, Yale and the University of California, Berkeley. ^[1][3] Anderson is the senior author. The journal's full text was not available to this newspaper, so what follows rests on the University of Oxford's release, an independent report in Sci.News, and the archived data and documentation the team posted with the paper. ^[1][2][3]

How a clock gets its age

A molecular clock works like this. Researchers compare the genetic differences among living species, then use fossils to pin the tree of life to real dates, and a model of how fast genes change turns the differences into elapsed time. The fossils supply the limits. A minimum age comes from the oldest firm fossil of a group, which says the lineage had to exist by then. A maximum age is harder, because it says a lineage cannot be older than a certain point, and that rests on an absence. ^[1][3]

The absence that has done the most work is a small site in Weng'an, China. The deposit is roughly 590 million years old and preserves microscopic life in extraordinary detail, yet it holds no definitive animal. The reasoning followed naturally. If animals had existed when Weng'an formed, surely such a good window would have caught one, so animals must be younger. Estimates constrained that way came out roughly 200 million years later than the new ones. ^[1][2]

The new paper, titled "Re-evaluating molecular clock maximum age calibrations revives pre-Ediacaran divergence estimates for animals" in its data archive, tests how much the answer depends on that choice. The team ran Bayesian analyses in a program called MCMCTree under several alternative calibration scenarios. ^[3] The title says the result "revives" earlier estimates. The paper is arguing that the older numbers were not so much wrong as wrongly overruled.

A second window, and nothing in it

The strongest part of the argument is not statistical. It is a fresh look at a different deposit. The researchers examined more than 140 samples from the Kheseen Biota in Mongolia using scanning electron microscopy. They found well-preserved microfossils, among them acritarchs, which are the organic-walled remains of single-celled organisms, and embryo-like fossils with internal cells. None could confidently be identified as an animal. ^[1]

That matters because of when Kheseen formed. It is younger than Weng'an, and by its time animal fossils are known from other places. So here is a site that preserves tiny things beautifully, sits in a period when animals demonstrably existed, and still shows no definitive animals. ^[1][2] If Weng'an's silence proves animals had not yet evolved, Kheseen's silence should prove they had not evolved when it formed either, and that is plainly wrong.

Anderson put the point this way, as the Oxford release reports it. The Kheseen fossils undercut the argument that animals would necessarily turn up at Weng'an if they already existed. Animals may have lived in environments these sites do not represent, or the local chemistry may simply not have preserved them. ^[1]

I find this the most persuasive piece of the paper, and it is worth dwelling on why. A fossil record is a memory, and like any memory it keeps what happened to be stored in the right conditions and loses the rest. A gap in it can mean nothing happened. It can also mean nothing was kept. The first animals, according to the research team's account, may have been small and soft-bodied, exactly the kind of creature that leaves nothing behind. ^[1] Treating the absence of their remains as the absence of the animals was always a bet. Kheseen is a documented case of the bet losing.

Older anchors

Having set aside Weng'an as a maximum-age limit, the team turned to older rock. It used deposits dating to roughly 850 million to 730 million years ago, including the Svanbergfjellet Formation in Norway, the Bitter Springs Group in Australia and the Chuar Group in Arizona. ^[1] These are the sites that supply the new ceiling. The clock was then re-run with them, and the origin of animals slid back into the 800 to 700 million year range. ^[1][3]

This is not the only hint that the animal story starts earlier than the body fossils say. Chemical fossils, or biomarkers, in ancient rocks have been interpreted as consistent with sponges living at least 650 million years ago. That is earlier than the oldest definitive animal fossils visible to the naked eye. ^[1][2][3] The oldest undisputed animal body fossils are reported at about 574 million years old. ^[2] Between that figure and the new window lies a gap of well over 100 million years, and closing or explaining it is the problem the study sets out to address. ^[2]

Into the ice

Now the climate. The Cryogenian began around 720 million years ago and included glacial episodes so severe that they are commonly called Snowball Earth. The proposed origin window therefore runs from before that period to the start of it. ^[1][2]

The usual picture of the dawn of animals is a warming, oxygenating world in which complex bodies became possible. If the new dates are right, that picture needs a rewrite. Animals, or their immediate ancestors, would have been there before the ice came, and some lineage would have had to get through it. ^[1]

The authors do not claim this, and neither should anyone else yet. The study does not show that ice caused animals to evolve. It does not show that animals survived any particular glacial episode, and it does not settle whether they arose before the glaciations, during them or after. ^[1][2] What it does is make the timing question consequential. A world of ice is a very different place to look for a first animal than a warm shallow sea, and a hypothesis that tied the origin of animals to the planet's wildest climate swings can now be tested against rocks rather than argued in the abstract.

That is my reading, and it is an opinion, not a finding. The date is where the paper's weight lies, and the ice is where its consequences begin. Anyone who has ever tried to date something from a handful of surviving objects will recognise the pleasure and the danger of the situation. A clue can point somewhere exciting and still fail to arrive.

What would change the picture

The uncertainty is substantial, and the study's own materials say so. The results are sensitive to calibration choices, which the archived dataset documents across multiple scenarios. ^[3] The 800 to 700 million year range is a model-supported possibility. It is not a settled birthday for the animal kingdom. ^[1][3] The full sampling details, the exact ranges each scenario produces and the sensitivity tests beyond the data summaries are questions that the paper's text and independent groups will have to answer. The Dryad repository holds input files, model configurations and outputs, so others can rerun the analyses and check. ^[3]

There is also the matter of whether older fossil limits are themselves safe. The new constraints come from deposits that also lack definitive animals. A critic might fairly ask whether the paper has swapped one inference from absence for another. The authors' answer, by the account in their release, is that the point is not to prove animals were present at those sites but to show that the absence at Weng'an cannot serve as a ceiling. ^[1] Whether that reply satisfies the field is the open question.

The team's own next step is practical. They say future work should search deposits from more places, more environments and more kinds of preservation, and should combine body fossils, traces of animal activity and chemical biomarkers. ^[1][2] Nobody has yet found body fossils or traces of animal activity in the older deposits the new clock relies on, and that is the hunt to watch. A single convincing animal trace in rocks from before the ice would do more than any recalculation. Equally, further study of Kheseen could say whether animals were truly absent from its living community or merely missing from what was kept.

If the clock is right, the first animals have been waiting for us in rock we have not yet learned to read.