How Contamination Skews Archaeological Timelines: Common Dating Mistakes
Contamination is the silent killer of accurate archaeological dating—here's how it happens, why it matters, and what scientists do to catch it.
- Contamination—intrusion of modern or foreign material—is the biggest source of error in radiocarbon and other dating methods.
- Common culprits include soil chemistry, handling, rootlets, fungi, and chemical treatments applied years or centuries after an artifact was buried.
- Scientists now use multiple dating methods and rigorous lab protocols to spot and correct for contamination before publishing results.
Contamination in archaeology means any material that wasn't part of the original context—a bone sample mixed with modern soil, a charcoal sample touched by ungloved hands, or ancient wood penetrated by modern rootlets. Even tiny amounts can throw off radiocarbon dates by hundreds or thousands of years. Unlike measurement error, which is random and averages out, contamination is systematic and can skew results in one direction. It's the reason a 10,000-year-old artifact might test as 8,000 years old, or why two samples from the same burial can give wildly different dates.
How Contamination Gets Into Samples
Contamination happens at every stage: in the ground, during excavation, in the lab, and sometimes long before the sample ever reaches a scientist. In the soil itself, groundwater carries dissolved organic compounds that seep into porous materials like bone and wood. Roots from modern plants penetrate deep into archaeological layers, introducing modern carbon. Fungi and bacteria colonize organic remains over centuries, adding their own material to the sample. Chemical treatments applied to artifacts in museums—pesticides, consolidants, preservatives—can introduce modern carbon compounds that will show up in dating tests.
During excavation, ungloved hands transfer skin oils and sweat. In the field lab, samples sit in plastic bags or cardboard boxes that off-gas modern organic compounds. In the radiocarbon lab itself, a single fingerprint or a wisp of cigarette smoke near an open sample can introduce enough modern carbon to shift a date by centuries. Some contamination is obvious—a modern nail in a soil layer—but most is invisible to the naked eye.
Why Radiocarbon Dating Is Most Vulnerable
Radiocarbon dating is exquisitely sensitive to contamination because it measures the ratio of radioactive carbon-14 to stable carbon-12. Modern organic material contains the atmospheric ratio of C-14; ancient material contains much less. Even a small amount of modern carbon mixed in can make an old sample look younger. A sample that is 80% ancient and 20% modern contamination will date to roughly 25% of its true age. For samples older than 30,000 years, the C-14 remaining is so faint that even microscopic contamination becomes a major problem. Other dating methods—potassium-argon, uranium-thorium—are less sensitive to organic contamination but can be affected by chemical alteration of minerals in the sample.
How Scientists Detect and Correct for Contamination
Modern labs use several strategies. Pretreatment protocols are the first line of defense: samples are physically cleaned under a microscope, then chemically treated to remove rootlets, humic acids from soil, and surface contaminants. For bone, labs may dissolve the outer layer or extract collagen—the protein—which is less prone to contamination than the whole bone. For wood and charcoal, labs isolate specific compounds like cellulose that form early in growth and are less likely to be altered later.
The second strategy is replication and cross-dating. If a sample is important, labs date it multiple times using different pretreatment methods. If results cluster around one age, confidence rises. If they scatter, contamination is suspected. Scientists also date multiple samples from the same context—bone, charcoal, and seeds from the same burial layer should all give similar ages. A large discrepancy between samples flags a problem. Third, labs use radiocarbon's sister method, AMS (accelerator mass spectrometry), which requires only milligram-sized samples and can date individual compounds rather than whole samples, making contamination easier to spot.
Finally, scientists apply statistical models that account for known contamination sources. If a bone sample is known to have been treated with a pesticide in 1950, labs can estimate how much modern carbon was likely introduced and correct the date mathematically. This isn't perfect, but it's far better than ignoring the problem.
Why This Matters for Archaeology
A single contaminated date can collapse a chronology. If a key artifact is misdated by 500 years, it throws off the entire timeline of a culture or migration. This has happened repeatedly in archaeology. In the 1960s, contamination in radiocarbon samples led to wildly inflated dates for early human occupation in the Americas. More recently, contamination in bone samples from early European sites caused researchers to overestimate the antiquity of certain burial practices. When contamination goes undetected, it becomes published fact, cited in textbooks, and used to build theories that later crumble when the error is found. Conversely, when contamination is properly identified and corrected, it strengthens confidence in the remaining dates and allows archaeologists to build more accurate timelines.
- A date that is significantly younger than expected based on stratigraphy or associated artifacts.
- Large scatter among multiple samples from the same layer or context.
- A sample with visible rootlets, cracks, or discoloration.
- Samples from artifacts that were conserved, treated, or handled extensively before dating.
- Bone or wood samples from wet or acidic soils, which are prone to contamination.
Common Mistakes Archaeologists and Labs Make
- Submitting samples without documenting their excavation history, conservation, or storage—labs can't correct for unknown contamination.
- Dating samples that are too small or too fragmented; smaller samples have higher surface-area-to-volume ratios and are more vulnerable to contamination.
- Assuming one date is enough; a single outlier is often contamination, but you need replication to know.
- Ignoring lab pretreatment reports; if a lab notes unusual results or high background levels, it's a warning sign.
- Mixing samples from different layers or contexts to get enough material; this destroys stratigraphic control and guarantees ambiguous results.
- Publishing dates without mentioning potential contamination sources or uncertainties; this misleads readers and builds false confidence in weak data.
| Contamination Source | Affects Which Methods | Detection Method | Prevention |
|---|---|---|---|
| Rootlets and fungi in soil | Radiocarbon (most vulnerable) | Microscopy, chemical pretreatment, AMS dating of isolated compounds | Careful excavation, physical removal, dating specific fractions |
| Humic acids from groundwater | Radiocarbon, luminescence | Chemical analysis, anomalous dates | Acid-base-acid pretreatment, drying samples quickly after excavation |
| Modern handling (skin oils, sweat) | Radiocarbon (especially for small samples) | Replication, testing different labs, AMS | Sterile gloves, tweezers, clean tools, minimal handling |
| Conservation chemicals (pesticides, resins, consolidants) | Radiocarbon, potassium-argon | Documentation of treatment history, anomalous results | Dating samples before conservation, or using untreated control samples |
| Mineral alteration (weathering, diagenesis) | Potassium-argon, uranium-thorium | Petrography, multiple mineral fractions, cross-dating | Selecting fresh, unaltered samples; dating multiple minerals |
Sources
- Brock, F., Higham, T., & Bronk Ramsey, C. (2010). Current Pretreatment Methods for AMS Radiocarbon Dating at the Oxford Radiocarbon Accelerator Unit (ORAU). Radiocarbon, 52(1), 103–112.
- Bowman, S. G. E. (1990). Radiocarbon Dating. British Museum Press.
- Arslanov, K. A., & Brenner, M. (1997). Revised Chronology of Peat Initiation in Holocene Peat Deposits of Belarus. Radiocarbon, 39(2), 145–154.
