The risks of developing lung cancer from exposure to radon are about three times higher than the most commonly used estimate, and smokers are at dramatically higher risk than nonsmokers, the National Research Council reported yesterday.

The risks are about the same as those cited by the Environmental Protection Agency in its effort to induce homeowners to test for radon, but they are sharply higher than the estimates in a 1984 study widely used by U.S. radiation experts.

The council, an arm of the National Academies of Sciences and Engineering, also said ending exposure to radon can reduce risk of cancer over time but apparently not eliminate it. According to its analysis, the effects of radon decline, "but not to zero, regardless of the number of years since exposure."

Radon is an odorless, colorless gas formed by radioactive decay of uranium in the soil. The gas is ubiquitous and apparently harmless in the environment but can accumulate to dangerous levels in underground mines and in buildings. As a cause of lung cancer, the EPA considers it second only to smoking.

The gas has become a major health concern in many parts of the nation, including Maryland and Virginia. Studies in Montgomery and Fairfax counties have found elevated radon levels in as many as 50 percent of houses tested.

In West Springfield, children were removed from several elementary school classrooms this week after officials found radon levels more than quadruple the federal safety guideline.

The EPA and the Nuclear Regulatory Commission asked the council to study the problem in 1984, after the National Council on Radiation Protection issued a study that, according to some scientists, underestimated the risk of radon. The EPA subsequently developed its own risk estimates.

The National Research Council essentially validated the EPA's risk figures. It concluded that someone exposed each year to the equivalent of 100 picocuries of radon -- twice the EPA's safety guideline -- would face a lifetime lung cancer risk 1.5 times that of someone exposed to background levels of radon.

That is about three times higher than the estimates from the 1984 study. The report warned, however, that uncertainties remain.

The council used a new statistical technique to analyze data from four epidemiological studies involving more than 22,000 underground miners. As expected, the analysis found that the risk of lung cancer increases with cumulative exposure.

It also found that the risk to smokers is "10 or more times greater than in nonsmokers." According to the report, smoking does not appear to be simply an added risk but can multiply the risks of radon. Smoking also has been found to multiply health risks from other pollutants, such as asbestos.

Many health officials have expressed concern that children may be at greater risk from radon because their rapidly dividing cells are more susceptible to radiation damage. The council said it could not document increased risk for children, however.

"We think that it's true but, when we tried to model it, the statistics weren't significant," project leader William H. Ellett said. "The committee members think there is something there."

Radon carries health risks long after exposure has ended, the council concluded. "After 15 years, the risk drops by about half," said Dr. Jonathan Samet, a physician at the University of New Mexico Cancer Center and member of the research panel. "Nobody has ever said that there's a threshold for radon, and this model would project some risk for any level of exposure."

Evidence that radon risks drop at all is "a departure from usual thinking," he said. Long-term studies of atomic bomb survivors in Japan have found no decline in relative risk from radiation exposures more than 40 years ago.

The finding may mean that radon, which emits alpha radiation, may affect the body differently than gamma radiation, the type to which bomb victims were exposed at Hiroshima and Nagasaki in 1945.

The EPA has recommended that homeowners take action if radon levels reach 4 picocuries per liter of air.

Jerry Puskin, chief of the agency's bioeffects analysis branch, said that level would still leave a homeowner with an extra 2-in-100 lifetime risk of developing lung cancer. The lifetime risk without radon exposure is 4 to 5 in 100.