Epidemiological evidence of lung cancer risk from radon is based mainly on studies of men employed underground in mines where exposures are relatively high in comparison to indoor exposure. Risk from residential radon can be estimated from occupational studies. Nevertheless, as such extrapolations depend on a number of assumptions, direct estimation of the risk is needed. The present study of lung cancer mortality was designed as a follow-up of a population (N = 12,004) in a radon prone area of the Czech Republic covering the period 1960-1999. Information on vital status and causes of death were obtained mostly from local authorities and from the national population registry. Exposure estimates were based on one year measurements of radon progeny in most houses of the study area (74%). Exposures outside the area (16%) were based on country radon mapping. Mean concentration of 509 Bq/m3 is higher than the country estimate by a factor of 5. By 1999, a total of 210 lung cancers were observed, somewhat more than the nationally expected number (O/E = 1.10) in comparison to generally low numbers corresponding to cancers other than lung (O/E = 0.81). The excess relative risk per standard radon concentration (100 Bq/m3) was 0.087 (90% CI: 0.017-0.208). This value is consistent with risk coefficients derived in other indoor studies. The present follow-up demonstrated that increased incidence of lung cancer depends linearly on exposure in terms of average radon concentration in the course of previous 5-34 years. Adjustment for smoking did not substantially change this estimate, although the risk coefficient for non-smokers (0.130) was higher in comparison to that for ever smokers (0.069), but not statistically different.