The most compelling evidence for radiation hormesis—for the idea that a certain amount of radiation over time can be good for you—comes from a much-cited 2004 study of the residents of certain apartments in Taiwan.  The construction steel for their buildings had been accidentally contaminated with radioactive cobalt-60 (half-life 5.3 years), and 10,000 people were subjected to up to an average of 1,300 millirem/year of radiation over a period of up to 20 years.  (The maximum was 91,000 millirem in the first year, 1983.)  The aggregate dose per person ranged from 40 to 600 rem (40,000 to 600,000 millirem) over the time they occupied the contaminated buildings.

In an acute radiation event, 400 rem causes death.  That’s what killed the 47 workers at Chernobyl.

So, how high was the cancer rate among the contaminated apartment dwellers compared to the general population in Taiwan during the years 1983-2003?  Here’s the chart:

The mean cancer mortality in Taiwan from 1983 to 2003 was 116 deaths per 100,000 person-years.  (It went up over that time due to the aging of the population.)  The mean cancer mortality of the 10,000 radiated apartment dwellers was 3.5 deaths per 100,000 person years—3% of the normal cancer death rate.  There “should” have been 232 cancer deaths among the apartment residents, but there was only 7.  If the cobalt-60 radiation was good for them, it was very good for them.

The original paper, “Is Chronic Radiation an Effective Prophylaxis Against Cancer?”, was published in the Spring 2004 Journal of American Physicians and Surgeons and may be found here.

Two books examine hormesis.  One for a popular audience is Underexposed: What If Radiation is Actually Good for You? (2005), Ed Hiserodt; Laisser Faire.  A technical one is Radiation Hormesis and the Linear-No-Threshold Assumption (2009), Charles Sanders; Springer.