T hree lines of inquiry have recently raised the surprising possibility that very low doses of ionizing radiation may not be harmful after all or may even have net benefits, a phenomenon knownas hormesis. Many studies (but not all) show that laboratory animals exposed to low doses of radiation outlive unexposed animals (1). How could this happen? DNA damage occurs commonly as a result ofnormal metabolic processes as well as from exposure to environmental mutagens. Whether the outcome is harmful depends on the dynamic balance between damage and repair processes. A net benefit can result when protective responses to low-grade exposure more than compensate for the harmful effects of the radiation. For example, a major cause of radiation injury at high doses is thought to result from the production offree radicals. Feinendegen et al. have shown thatfree radical scavengers increase after low-dose radiation, possibly to a greater extent than that necessary to neutralize the radicals produced by the radiation (2). This increased production ofscavengers might increase cell defenses against free radicals that result from exposure to other environmental mutagens or those produced by normal oxidative metabolism. In other work, Wolff and colleagues have found evidence that DNA repair may be enhanced by low doses of radiation (3). This suggests another means ofprotection, namely, that radiationexposed DNA may be more readily repaired after subsequent exposures to mutagens. One study demonstrates that enhanced DNA repair exists in workers occupationally exposed to radiation (4). Third, radiation-induced cell death stimulates cell reproduction as a homeostatic mechanism that maintains cell compartment size. Accordingly, Kondo has suggested another possible responseto low-level stimulation, namely, that immune cell production may be enhanced by low-dose radiation (5). Evidence for increased numbers of lymphocytes in laboratory animals after exposure to low-dose radiation has been presented by several investigators (6-8). Whether this immune enhancement results from direct effects on lymphatic tissues or through stimulation ofcentral neuroendocrine regulatory mechaamsms deserves imvestigation. Epidemiological studies of human populations exposed to rela-tively low doses of ionizing radiation have not shown the existence or absence of low-dose effects. For example, the studies of popula-tions living in areas ofhigh natural background radiation have not shown any increase in adverse health effects (9). In the absence of observable effects, it has nevertheless been assumed that low-level exposures produce the same harmful effects as those seen at high levels of exposure, but with lower frequency. This assumption has become the accepted radiation paradigm, justified on the basis of prudence, andon certain laboratory observations of mutagenic effects of ionizing radiation at relatively low doses. Beginning in the

1950s, fear of genetic effects, together with the associated" target theory" of radiation injury, have continued to dominate radiation protection thinking. As a result, substantial efforts are made to reduce or avoid small exposures, even exceedingly small exposures, to workersand members of the public. In more recent years, accumulated experience has tended to reduce fears ofthe mutagenic effects of low-dose ionizing radiation. Direct observations of mutagenesis in human populations have shown humans to be one-fourth as sensitive as expected from previous indirect estimates based on rodent studies. Furthermore, although some findings are suggestive, genetic studies of survivors of the atomic bombings have …