Understanding Radiocarbon Dating: How It Works and What It Cannot Tell You
When an archaeological report says a site was occupied "around 3400–3200 BCE" the number comes from somewhere. For most prehistoric and early historic archaeology, that somewhere is radiocarbon dating — the measurement of residual carbon-14 in organic material. Understanding what the method actually measures, how it is calibrated, and where it fails helps make sense not just of the numbers in reports but of the public announcements that inevitably oversimplify them.
What carbon-14 is and how it decays
Carbon-14 (C-14) is a radioactive isotope of carbon produced continuously in the upper atmosphere when cosmic-ray neutrons strike nitrogen-14 atoms. Atmospheric carbon dioxide contains a small, roughly constant proportion of C-14 alongside the stable isotopes C-12 and C-13. Living organisms — plants absorbing CO2, animals eating plants — maintain equilibrium with atmospheric C-14 throughout their lives. When an organism dies, it stops exchanging carbon with the atmosphere, and the C-14 it contains begins to decay back to nitrogen-14 at a fixed rate.
The half-life of C-14 is 5,730 years: after 5,730 years, half the original C-14 atoms have decayed; after another 5,730 years, half of the remainder has decayed, and so on. By measuring the ratio of C-14 to C-12 in a sample, laboratories can calculate how long ago the organism died — in theory. The practical range of the method, using conventional technology, is approximately 50,000 years; beyond that, the remaining C-14 is too little to measure accurately.
Willard Libby developed radiocarbon dating at the University of Chicago in 1949 and received the Nobel Prize for Chemistry in 1960.
Pretreatment: acid-base-acid for charcoal
Before a sample reaches the laboratory instrument, it must be pretreated to remove contaminants — principally humic acids (soil-derived organic compounds that infiltrate samples over time) and carbonates (which can carry younger or older carbon). The standard pretreatment for charcoal is the acid-base-acid (ABA) procedure: the sample is soaked successively in hydrochloric acid, sodium hydroxide, and hydrochloric acid again. The acid removes carbonates; the alkali removes humic acids; the final acid neutralises the alkali. Other materials require different protocols: bone collagen extraction for bone, cellulose isolation for wood or seeds.
Contamination is the single most common cause of anomalous radiocarbon dates. A sample contaminated with even 1% modern carbon will give a date centuries younger than the true age. Field collection matters: samples must be collected with clean, uncontaminated tools, stored in foil or laboratory bags, and accompanied by complete context information. A date without a clearly recorded context is a date of limited interpretive value regardless of its precision.
Conventional counting versus AMS
Two types of measurement are in use. Conventional beta counting measures the emission of electrons from C-14 decay in a gas counter or liquid scintillation counter; it requires large samples (typically several grams of charcoal) and measuring times of hours or days. Accelerator Mass Spectrometry (AMS) counts the actual C-14 atoms by separating them by mass in a particle accelerator; it requires samples as small as a milligram and measures in minutes. AMS dating made it possible to date materials that could not previously be sampled — a single grain of wheat, a few strands of hair, ink on a manuscript page — and has transformed the field since its adoption in the 1980s.
AMS is now the dominant technique for research archaeology. The cost per sample has fallen substantially; most university-linked laboratories charge in the range of $300–600 per AMS date. Radiocarbon laboratories accredited for archaeological work include Oxford, Groningen, Beta Analytic (Miami), and the Scottish Universities Environmental Research Centre (SUERC).
Calibration: the IntCal curve and dendrochronology
The foundational assumption of radiocarbon dating — that atmospheric C-14 concentration has been constant over time — is not quite true. Variations in solar activity and the Earth's magnetic field alter the rate of C-14 production in the upper atmosphere, and these variations mean that the raw radiocarbon measurement ("conventional radiocarbon age," expressed in years Before Present with 1950 as the reference year) does not translate directly to calendar years.
Calibration converts the conventional age to a calendar date range using the IntCal calibration curve, published and updated at regular intervals by the IntCal Working Group. The curve is constructed primarily from tree rings: dendrochronology — the counting and pattern-matching of annual growth rings — provides a calendar-year framework going back over 14,000 years in northern hemisphere trees (using the long German oak chronology and Californian bristlecone pine data). Each tree ring is a closed carbon system representing a single year; by measuring C-14 in a sequence of rings from known calendar years, researchers have built the calibration curve directly from real atmospheric C-14 values.
Beyond the range of tree rings, the curve is constructed from corals, cave stalagmites (speleothems), and lake sediment varves, each providing annual or near-annual carbon samples tied to independent dating methods.
What the plus-or-minus means
A calibrated radiocarbon date is always expressed as a probability range, not a point value — for example, "2340–2140 BCE (95% probability)" or "cal BP 1250 ± 30." The range reflects two sources of uncertainty: the measurement error of the laboratory (expressed as ± a given number of years on the conventional radiocarbon age) and the shape of the calibration curve at the relevant point. Some parts of the curve are steep and monotonic, giving tight calibrated ranges; others are plateaus or reversals — sections where different conventional ages map to the same calendar years — giving very broad or bimodal distributions.
The Hallstatt Plateau (roughly 800–400 BCE) is the most archaeologically consequential of these flat sections: radiocarbon dating cannot distinguish within this period with useful precision, which is a serious limitation for the Iron Age archaeology of Europe and the Mediterranean.
The marine reservoir effect
Organisms that live in or rely heavily on marine food sources incorporate carbon from seawater rather than the atmosphere. Seawater has a lower C-14 content than the atmosphere because the deep ocean circulates slowly and mixes old, depleted carbon with surface water. The result is that marine organisms give radiocarbon ages that appear several hundred years older than they actually are — the average global marine reservoir effect is roughly 400 years, expressed as the "R" correction.
The marine reservoir effect varies regionally and temporally: the correction for the North Atlantic is different from the correction for the upwelling-rich waters off the Peruvian coast or the enclosed Baltic. Human populations whose diet was substantially marine — Norse Greenland settlers, Pacific Island cultures, coastal prehistoric peoples — require careful dietary assessment before their bone collagen can be dated reliably. Stable isotope analysis (carbon-13 and nitrogen-15 ratios in bone collagen) provides a dietary signal that can be used to apply appropriate corrections.
What radiocarbon dating cannot tell you
Radiocarbon dates the death of an organism, not an archaeological event. A timber used in a building dates to when the tree was felled, not when the building was constructed — and if the timber is reused from an earlier structure, the date may be centuries older than the building it supports. Charcoal from a long-lived tree like an oak gives the death of the sampled part of the tree, which may be decades older than the death of the tree, which may be decades older than the fire in which the charcoal was deposited.
These "old wood" and "inbuilt age" problems mean that the best samples for radiocarbon dating are short-lived materials: grain, seeds, annual plants, animal bone from a food deposit, or sapwood from a tree ring sequence. A reported date is only as reliable as the reasoning about the relationship between the sample and the event it is intended to date.
The sites on the map each have their own dating histories — some known through radiocarbon, some through dendrochronology, written records, or coin sequences. Understanding how those dates were established is part of understanding what the site actually tells us about the past.