Holman Jenkins Jr. writes in today’s WSJ:

For the first time, Germany’s Green Party will control one of the country’s state governments, just in time to decide the fate of several nuclear plants temporarily idled in response to the Japanese nuclear mess — the same mess that catapulted the Greens to unfamiliar success in state elections two weeks ago.
Were they to surrender to their inner Al Gore, their answer to the resulting electricity shortage would be for consumers to make do with much higher prices. But Green Party leader Winfried Kretschmann is unlikely to seek immediate return to the political wilderness. Forget the fantasy talk of wind and solar. The choice then is nuclear or coal.
Countries all over the world are making similar choices in light of Japan’s disaster, whose severity the government now puts on a par with Chernobyl (even if the radiation releases haven’t been remotely comparable). Thanks to Fukushima, officials everywhere will have to reacquaint themselves with one of the great noise-to-signal puzzles of modern science: How much harm does low-level radiation exposure do?
For 60 years, the hunt has been on for handfuls of “excess” cancers in populations subjected to unnatural doses of radiation. The findings have been more politically vexing than scientifically satisfying.
The once-golden Hiroshima and Nagasaki studies, jointly funded by the U.S. and Japanese governments, showed little or no low-dose cancer risk, and even a boost to longevity from low-dose exposure in the form of reduced deaths from “non-cancer” diseases. But the A-bomb studies have had their scientific crown knocked askew in recent decades. One reason is “survivor bias” — having survived not only the bombings but homelessness, hunger and a typhoon in the immediate aftermath, the subjects may be a hardier lot than the Japanese population at large.
Investigations of U.K. infants exposed to X-rays in the womb and workers in U.S. weapons plants gradually began nudging the bomb studies aside in the 1980s, for reasons due partly to legal and regulatory convenience: Their findings seemed to confirm the simple and intuitive “linear, no-threshold” hypothesis, which holds that radiation is always dangerous in direct proportion to dose.

The rest here.