A historical cohort study was conducted in Misasa town, Tottori prefecture, Japan, where radon spas have been operating for a long time. Misasa town was divided into an elevated radon level area and a control area, with mean indoor radon levels of about 60 and 20 Bq/m3, respectively. In total, 3,083 subjects in the elevated radon level area and 1,248 in the control area, all aged 40 or older on January 1, 1976, were followed up until December 31, 1993, for a mean period of 14 years. The mortality rates from all causes exhibited no difference between the elevated radon level area and the control area for both sexes. No difference was observed in the incidence of all-site cancers (age, period-adjusted rate ratios by Poisson regression, RR = 1.06, 95% confidence interval (CI) 0.79-1.42 for males, RR = 0.90, 95% CI 0.65-1.24 for females), while stomach cancer incidence seemed to decrease for both sexes (RR = 0.70, 95% CI 0.44-1.11 for male, RR = 0.58, 95% CI 0.34-1.00 for female) and lung cancer incidence for males only seemed to increase (RR = 1.65, 95% CI 0.83-3.30 for male, RR = 1.07, 95% CI 0.28-4.14 for female) in the elevated radon level area. Caution is needed in the interpretation of these findings, however, since the individual exposure level was not measured and major confounding factors, such as smoking and diet, could not be controlled in this study.