My belated education in nuclear energy continues. I just read Power to Save the World: The Truth about Nuclear Energy (Vintage, 2008) by Gwyneth Cravens, a petite, energetic novelist and journalist. Cravens contacted me after seeing my chat with Rod Adams, a nuclear-trained Naval officer, on last May (which I followed up with a post). I recently met Cravens during a tour of the Indian Point nuclear power plant in New York State, which she arranged. I'm feeling a lot better about living near Indian Point, less because of what I learned during my tour (although plant employees were quite informative) than because of Power to Save the World.

The 2007 book describes how Cravens morphed from a nuke-fearing greenie who in the 1980s opposed the Shoreham nuclear plant on Long Island, where she lives, into a proponent who believes that we need nuclear power to save us from global warming and other adverse effects of fossil fuels. Cravens repeats the refrain that the risks of nuclear energy have been exaggerated; nuclear power, both civilian and military, hasn't killed a single person in the U.S. over the past half century. But she fleshes out these statements with surprising (to me) details.

—Day in and day out we are all bombarded by radiation, including alpha and beta particles, x-rays and gamma rays. Americans receive an average of 360 millirem (a rem is a measure of radiation dosage) a year from radon gas and other background sources, cosmic rays (doses rise at higher elevations, where there is less atmospheric protection), consumer products (such as smoke detectors), and medical procedures. (I learned at Indian Point that the 360 estimate has recently been revised upward to 620). A set of dental x-rays delivers 39 millirem, a flight from New York City to Los Angeles 1.5 millirem. Radiation treatments for cancer can deliver millions of millirem to a specific organ and tens of thousands to the whole body. Federal regulations allow nuclear workers to receive up to 5,000 millirem annually, but they receive less than 240 on average. U.S. nuclear plants increase the radiation in their neighborhoods by less than one millirem a year, on average.

—There is no clear-cut evidence of adverse health effects from radiation at levels below 100,000 millirem a year. The health effects of radiation have been calculated from people who received extremely high doses, including survivors of the atomic attacks on Hiroshima and Nagasaki, the Chernobyl accident, and "radium girls" who painted radium on watches and other instruments in the early 20th century. The cancer epidemiologist Charles Key told Cravens that "compared to tobacco, gasoline, drunk drivers or being a couch potato, radiation is of very little risk to most of the public."

—Claims about adverse effects from low levels of radiation are often based on a so-called linear non-threshold model. The model assumes, for example, that if an exposure of n millirem kills 50 percent of a population, then 0.1 n will kill 5 percent, 0.01 will kill 0.5 percent and so on. There is no evidence for this model. Background radiation from natural sources varies around the world from an annual dosage of less than 100 to over 10,000 millirem. (Residents of Ramsar, Iran, receive up to 26,000 millirem a year!) Studies have not found increased cancer or other illnesses in areas with naturally high radiation.

—Fifty plant and emergency workers died of acute radiation exposure in the 1986 Chernobyl disaster in the U.S.S.R., the worst nuclear accident in history. The explosion contaminated more than 200,000 square kilometers with radioactive fallout, but radiation in parts of this zone is now lower than in Finland and other regions of the world with naturally high radiation. The International Agency for Research on Cancer estimates that radiation releases from Chernobyl caused a slight increase in thyroid cancer but adds that "smoking will cause several thousand times more cancers in the same population." So far, there have been no excess deaths among the 200,000 "liquidators" who helped clean up the mess from Chernobyl compared with controls.

—The worst nuclear accident in U.S. history was the 1979 partial meltdown at Three Mile Island in Pennsylvania, which led to venting of radioactive gas. The highest dose received by plant workers was 4,000 millirem, 1,000 less than the annual dose permitted for U.S. nuclear workers. The highest dose for people living near the plant was 100 millirem. There is no credible evidence of increased cancer or birth defects among plant workers or residents near Three Mile Island.

—According to the U.S. Centers for Disease Control, so-called depleted uranium, which consists primarily of the isotope U 238 and not the more fissionable U 235, "has never caused a case of cancer in animals or humans." The dense metal is used to make tank armor, armor-piercing projectiles, shielding for x-ray machines, boat keels and other applications.

—Nuclear power in the U.S. has grown steadily more efficient and cheaper. Plants now operate at 90 percent of peak capacity (up from about 50 percent a few decades ago) compared with 73 percent for coal, 29 percent for hydroelectric, 16 to 38 percent for natural gas, 27 percent for wind and 19 percent for solar. In 2005 nuclear power was cheaper per kilowatt than any alternative.

—The waste from coal-burning plants is much greater in volume and more harmful than from nuclear generators. If you, as an average American, got all your electricity from nuclear plants, you'd generate one kilogram of nuclear waste during your lifetime, enough to fit in a soda can. If you got all your electricity from coal, you'd generate almost 70 tons of waste. Coal plants emit far more radioactive materials than nuclear plants do; each year a 1,000-megawatt coal plant disperses about 27 metric tons of uranium, thorium and other radioactive substances. Coals plants also emit mercury and other toxins, in addition of course to carbon dioxide and other greenhouse gases. An estimated 24,000 Americans die prematurely per annum because of pollution from coal plants; in China, the number is 400,000.

—Hydropower has killed many more people than nuclear power. About 1,000 Americans have died in dam collapses in the past 100 years. Dam collapses caused by a typhoon in China in 1975 killed 26,000 people immediately; another 145,000 people later died of disease and famine. The output of hydroelectric plants is decreasing because of droughts, possibly brought on by global warming.

—The footprint of nuclear power is much smaller than that of solar and wind. A 1,000-megawatt nuclear plant like Indian Point requires less than two square kilometers of land. Comparable solar and wind plants would require, respectively, 130 and 500 square kilometers of land, and they cannot produce a steady supply of power, as nuclear plants do.

I've always had a knee-jerk distrust of nuclear advocates, just as I have of right-wing Congressmen, psychiatric-drug shills and string theorists. But I trust Cravens and the experts she interviewed—including physicists, engineers and epidemiologists—over many years of reporting. If you're agonizing over whether to support nuclear energy, read Cravens's book and see if you find it as persuasive as I do. I also welcome (and expect) challenges to the assertions above.