Carbon-14 Dating Held `Inaccurate'

Serious problems with the accuracy of carbon-14 testing - archaeologists' favorite technique for measuring time - could mean the ages of many ancient artifacts and the dates of important climatic changes are probably underestimated, scientists have reported.

Carbon-14 dating, which relies on the steady decay of radioactive carbon atoms to calculate elapsed time, can be as much as 3,000 years off for objects 20,000 years old, according to scientists from Columbia University's Lamont-Doherty Geological Observatory, at Palisades, N.Y.

In a paper published in a recent edition of Nature, Edouard Bard, Richard Fairbanks, Bruno Hamelin and Alan Zindler report that their corrections for the carbon-14 dating method make it agree with more precise time-estimating methods.

``It will stretch the time scale. It will change everything,'' Bard said in a telephone interview. In the future, he said, researchers ``will have to change a little bit what they are saying'' about age.

``It's a very interesting result, no question about that,'' said R.E. Taylor, director of the radiocarbon dating laboratory at the University of California, Riverside.

``From an anthropological and archeological perspective,'' he said, ``we're going to be sensitized. We're going to see that dates out in that range (beyond 20,000 years) have more anomalies than we thought.''

The inaccuracy was discovered as a result of cores being drilled from ancient coral reefs near the island of Barbados. The ages of coral at various depths differed significantly when measured by carbon-14 dating, compared with a highly reliable dating method based on uranium decaying to become thorium. This suggested the widely used carbon-14 method seriously underestimates age.

A sample dated at 10,000 years by carbon-14, for example, is actually 11,500 years old. And a sample dated at 20,000 years is actually 23,000 years old, Bard said.