Between 1945 and 1963 the atmospheric weapons tests roughly doubled the carbon-14 in the air, and the isotope spread evenly around the globe within a few years of being made. The Test Ban Treaty stopped the production in 1963. What happened next was not decay — carbon-14 has a half-life of 5,730 years and had barely begun. The oceans and the biosphere simply started absorbing the excess, and the concentration in the air fell by half roughly every eleven years.

That falling line is an accident. For about sixty years it has also been the most precise instrument anyone has for dating a single human cell.

The logic is short. A cell copies its DNA when it divides, using whatever carbon is in circulation that year. After the final division the DNA is left alone. So the carbon-14 in a cell's genome is a stamp of the year that cell was born, and reading it against the atmospheric curve gives a date.

In 2005 Kirsty Spalding and Jonas Frisén used this to settle a question that could not be asked any other way. Whether the adult human brain makes new neurons had been argued for decades, and the experiment that would resolve it — label a living person's dividing cells, come back years later, look — is not one anybody may run. The tests had already run it, on everyone, without consent or design. The answer in the occipital cortex was no: those neurons are as old as the person carrying them, while the non-neuronal cells packed around them turn over steadily. The same stamp later showed that heart muscle does renew into adulthood, and that about a tenth of a person's fat cells are replaced every year — at every adult age, at every body weight.

None of it was measurable before. None of it will be measurable for much longer.

The method has never depended on the size of the spike. It depends on the steepness of the fall. A date can only be read off a curve dropping fast enough to tell one year from the next, which is why the researchers note, flatly, that resolution before the bomb tests is very poor. A flat line has no years in it.

The line is flattening now, and it is flattening because of a second contamination running the other way. Fossil fuels are old enough to have lost all their carbon-14, so burning them pours carbon into the air that dilutes what remains. Atmospheric levels have already come back down through the pre-industrial value. Heather Graven modelled where they go next: on current emissions, fresh organic material in 2050 will carry the same radiocarbon signature as material from 1050, and will be indistinguishable from it by dating.

So the two futures are not a choice between keeping the instrument and losing it.

On the high-emissions path the curve keeps falling, and every new leaf, bone and forged canvas reads as medieval. On the successful path emissions come down, the curve settles near its natural value, and the years stop being distinguishable at all. One future makes the record lie. The other makes it silent.

The instrument was never the mark the bombs left. It was the speed at which the world was clearing it. A body can only be dated while the evidence is still leaving.