There are about a dozen studies of the incidence of skin cancer among irradiated populations with known skin doses that are available for estimating the risk of radiation-induced skin cancer. It is of note that they provide no evidence for a dose threshold and are compatible with a linear dose-response relationship, at least for ultraviolet radiation exposed skin. The studies also provide varying amounts of evidence concerning a number of other important issues in assessing skin cancer risk: types of skin cancer induced by ionizing radiation, the appropriateness of relative risk vs absolute risk models, combined effects of ionizing and UV radiations, and variations in sensitivity to skin cancer induction among demographic and genetic subgroups. Little epidemiological information is available on several factors, such as the RBE for high-LET radiation, the effects of dose protraction or fractionation, or variations in risk by age at irradiation. A reasonable estimate of skin cancer lethality was 0.2 per cent when weighted for the relative proportions of squamous cell and basal cell skin cancers. Average risk estimates of radiation-induced skin cancer incidence were: absolute risk (AR) of 8.5 X 10(-4) person-year-Sv and excess relative risk (RR) of 52 per cent/Sv. Lifetime skin cancer risk was calculated by life-table methods for males from exposures spread out over ages 20-60 years. The estimates for excess skin cancer incidence were 2 per cent and 11 per cent per Sv under the AR and RR models, respectively, while the corresponding mortality risks were 4 X 10(-5) and 2 X 10(-4) per Sv.