Abstract. The follow-up of the French cohort of uranium miners has been extended to 1994, and a new source of information for causes of death has been used. The paper presents the new results regarding the risk of death among the cohort, and analyses the impact of the methodological changes on these results. The extension of the follow-up results in a substantial increase in statistical power compared with previous analysis (þ25% for person-years and þ74% for the number of deaths). The use of the National Mortality Database as the principal source for causes of death allows to reduce the potential bias in the calculation of standardized mortality ratios (SMR). As a consequence, an excess risk of deaths from laryngeal cancer, suggested in the .rst analysis, is not con.rmed. The analysis shows the existence of an excess risk of deaths from lung cancer among French uranium miners (85 observed deaths, SMR ¼ 1.9, 95% con.dence interval CI: 1.5–2.3), and an increase of this risk with cumulative exposure to radon (excess relative risk per 100 working level month ¼ 0.6, 95% CI: 0.1–1.2). These results con.rm he existence of a risk of death from lung cancer in a population chronically exposed to relatively low levels of radon.