Archaeology: A Clock That Needs Calibrating

When Willard Libby developed radiocarbon dating in the 1940s, archaeology gained its first reliable clock for the recent past. The method rests on a chain of facts: cosmic rays striking the upper atmosphere convert nitrogen into carbon-14, a radioactive isotope that mixes through the air and enters every living thing. While an organism lives, it constantly exchanges carbon with its environment, so its tissues match the atmosphere's carbon-14 proportion. The moment it dies, exchange stops. From then on, the carbon-14 already present decays back into nitrogen at a fixed rate, halving every 5,730 years. Measuring the carbon-14 left in wood, bone, or cloth tells scientists how long ago the organism died — when it last breathed or photosynthesized.

The clock has hard limits. After roughly eight or nine half-lives — about 50,000 years — the surviving carbon-14 is so scarce it drowns in background noise, and older material cannot be dated. Sensitivity also matters: early instruments needed grams of material and counted rare decay clicks; modern accelerator mass spectrometry counts individual carbon-14 atoms and works with milligrams. But precision alone does not guarantee an accurate calendar date, because the calculation hides an assumption — that the atmosphere's starting proportion of carbon-14 has always been the same. It has not.

Atmospheric carbon-14 levels have drifted over time, pushed up and down by changes in solar activity, Earth's magnetic field, and the carbon cycle itself. A plant that grew in a carbon-14-rich century starts with more isotope than the standard calculation assumes, and so appears younger than it is; one from a poor century appears older. The solution is calibration. Tree rings provide an independent year-by-year chronology: matching overlapping ring patterns from living and ancient trees, dendrochronologists have built sequences back more than 12,000 years, and each ring's carbon-14 can be measured directly. Plotting those measurements yields a calibration curve converting a raw radiocarbon age into a corrected calendar range. The curve is not smooth — it has wiggles and flat stretches — so one radiocarbon age sometimes maps to several possible calendar ranges, an ambiguity no amount of careful laboratory work can remove.

Two further distortions follow. The first is the reservoir effect: ocean water mixes slowly, so marine organisms build tissue from carbon long out of contact with the atmosphere, and a freshly caught fish can yield a radiocarbon age hundreds of years too old; freshwater creatures near limestone suffer a smaller version of the same problem. The second distortion is industrial: since the 1800s, fossil-fuel burning has poured ancient, carbon-14-free carbon dioxide into the air, diluting the atmospheric proportion — the Suess effect — while nuclear testing in the mid-twentieth century briefly doubled it. Any conversion that ignores which reservoir a sample fed on, or which century's air it breathed, risks systematic error.

Finally, the sample itself may lie. Groundwater carries dissolved carbonates into buried bone; roots penetrate old wood; humic acids from soil coat charcoal. Such contamination is dangerous in proportion to a sample's age, because the arithmetic is unforgiving: just one percent of modern carbon mixed into a truly ancient, effectively dead sample will make it appear roughly 37,000 years old. Laboratories therefore pretreat samples aggressively — washing, acid and alkali baths, extracting only the most trustworthy fraction, such as collagen from bone. None of this invalidates the method: radiocarbon dating is strongest where its weaknesses are respected, with dates published alongside error ranges, and important chronologies cross-checked against tree rings, lake sediments, and historical records.