They came out wearing sunglasses.
Fifteen people, unusually pale, blinking at a spring afternoon in the foothills of the Pyrenees, with a crowd and a bank of cameras waiting for them at the entrance.
None of them knew the date.
They had gone under with no watches, no phones and no way to be told anything at all about the world they had left.
What they had was a guess.
Asked how long they had been down there, most of the group landed on a number that fell well short of the real one, and not one guessed high.
Why the count comes up short
The human clock does not run at exactly 24 hours.
Left to itself, with no sunrise to reset it every morning, it drifts slightly long, and daylight is the signal that drags it back into line each day.
Take the light away and the drift compounds.
Each waking stretch runs a little further, each sleep starts a little later, and the cycle the body is actually living can stretch well past 24 hours without the person noticing anything odd.
Then the arithmetic does the rest.
Someone counting in subjective days fits fewer of them into the same real week, so the tally they carry out is short by roughly the amount their cycle got stretched by.
The room they were counting in
The site was a limestone system in the Ariege, near the Spanish border.
Cold and wet, holding around 50 degrees year round with the air close to fully saturated, which is why everything they owned stayed damp for six weeks.
Nothing was piped in.
Electricity came off a pedal bike, and drinking water had to be hauled up from a well roughly 148 feet below the camp.
The camp itself was spread out.
Cooking and gathering happened in one chamber, the tents stood about a third of a mile away, and a third space held the scientific gear.
Who went down and what was measured
Fifteen volunteers, eight men and seven women, aged 27 to 50.
They went in during the middle of March and came back up on the twenty fourth of April, which made the stay 40 days exactly.
The instrumentation was not casual.
Sleep sensors, behavioral logging, cognitive and brain activity tests before and after, and swallowed capsules that reported core body temperature from inside each person.
The bill ran to roughly 1.4 million dollars.
The first week was chaos, by the leader’s own account, because everyone was on a different rhythm and half the group was asleep whenever the other half wanted to work.
By the second week they had organized.
The group built a rough collective rhythm around shared goals, and by the end two thirds of them said they would rather have stayed down there longer.
The answer took five years to arrive
The mission ended in the spring of 2021.
The peer reviewed analysis of what happened to those fifteen body clocks did not appear for another five years, and when it did, in a chronobiology journal, it carried a formal correction almost at once.
Everything in between came from interviews.
That includes the figure most people remember, the claim that the group averaged something like 32 hour days, which was the expedition leader talking to reporters rather than a measured result.
The distinction matters more than it sounds.
A number from a press conference and a number from the paper are different objects, and for five years the world only had the first kind.
Drifting clocks are not a harmless curiosity either, as two pilots proved when they both fell asleep at 36,000 feet with 153 people behind them.
What it cannot settle
There was no control group.
A neuroscientist at a major French institute said so publicly at the time, and it is the objection that has stuck to the project ever since.
Fifteen people went in and nobody comparable stayed out.
So the changes measured afterward have nothing to be measured against, which is the difference between a rich observation and a result.
Confinement research can be done the other way.
Volunteers were paid to lie flat for 60 days in a study where the untreated group lost 48 percent of their calf strength, a figure that exists only because somebody was left alone as a comparison.
The precedent underground is older than any of this.
A geologist came out of an Alpine cave in the sixties insisting it was the twentieth of August, when it was already the middle of September.
He went back down a decade later, for six months in a Texas cave, and concluded he had spent it living on 48 hour cycles.
One man, then, and fifteen people now, all drifting the same direction.
Which is the part nobody disputes.
Put a person in a French cave or any other place the sun cannot reach, and they will lose time rather than gain it, every time.
Read the whole thing?
Get the week's signal, not the noise
Our sharpest reporting on energy, climate and nature — free, once a week.
