The bet Libby had to make
Willard Libby proposed radiocarbon dating in 1949 on an assumption he could not test: that the amount of carbon 14 in the atmosphere had been the same for the whole period he wanted to measure.
The logic was sound. Cosmic rays strike nitrogen in the upper atmosphere and produce carbon 14 at a steady rate; living things take it in and hold the atmospheric proportion; at death the intake stops and the clock starts. If the atmospheric level was constant, the amount remaining gives the time elapsed.
Libby tested the method against objects of known age — wood from an Egyptian tomb, a fragment of the Dead Sea Scrolls — and the dates came out close enough to convince. He received the Nobel Prize in Chemistry in 1960.
⚠️ The assumption turned out to be wrong, and it took two decades to find out by how much. Radiocarbon ages drift from calendar dates by up to several centuries in some periods.
How tree rings fixed it
The correction came from an unlikely direction: counting rings.
A tree lays down one ring a year, and each ring locks in the carbon 14 level of the year it grew. A wood sample of known ring count therefore gives a direct reading of the atmosphere at a known date. Overlapping sequences from living trees, dead trunks and bog oaks extend the record back more than 12,000 years.
Matching radiocarbon measurements against those rings produced the calibration curve. The current version, IntCal20, is built from tree rings, corals, lake sediments and marine records.
The reason the atmosphere varied is threefold: solar activity modulates cosmic ray flux, the Earth’s magnetic field deflects it, and the oceans exchange carbon with the air on a delay of centuries.
⚠️ This is why a radiocarbon result is properly written as a range, not a date. “1260-1390” is the honest form; “1325” is not.
The forger who was dated by his paint
Han van Meegeren sold a Vermeer to Hermann Göring during the German occupation of the Netherlands. After the war he was charged with collaboration for selling national heritage to the enemy — a capital offence.
His defence was that the painting was his own forgery. He proved it by painting another one in custody.
Carbon 14 did not feature in that trial, but it settled the argument later. Scientific analysis of his other alleged Vermeers, including radiocarbon dating of the canvas and lead isotope work on the pigments, confirmed twentieth-century material.
The technique has since become routine in art authentication. A canvas grown as flax in 1935 cannot carry a seventeenth-century painting, however convincing the brushwork.
The Turin Shroud
In 1988 the Vatican permitted three laboratories — Oxford, Zurich and Arizona — to date a sample from the cloth. All three returned dates between 1260 and 1390.
That range coincides with the first documented appearance of the shroud, in Lirey in France in the 1350s, where a bishop reported it as a painted forgery whose maker had confessed.
The result has been contested. The sampled corner may have been rewoven after the 1532 fire, which would mix newer thread into the measurement; a 2005 paper argued this, others have disputed the evidence for it. No full retest has been permitted since.
⚠️ Whatever the resolution, the measurement itself has never been faulted. The dispute is about what was sampled, not about the dating.
The bomb spike
Atmospheric nuclear testing in the 1950s and early 1960s nearly doubled the carbon 14 in the atmosphere. The peak came in 1963, and the level has fallen back since the Partial Test Ban Treaty stopped the tests.
This ruined the method for anything formed after 1955 — and created a far more precise one.
Because the curve rises and falls steeply, a sample from that period can be placed to within a year or two rather than a century. Forensic scientists use it to establish year of birth from tooth enamel and year of death from other tissues. Art authenticators use it to catch forgeries painted after the war.
⚠️ The spike is fading. As the excess carbon 14 works its way into the oceans and biosphere, the atmospheric level is approaching its pre-war value, and the technique will expire with it — some time around 2030.