Although it has been hypothesized that depressive persons have an excess risk of cancer, few prospective data are available. The association between depressiveness and subsequent incidence of lung cancer was studied in the nationally representative Mini-Finland Health Survey. The study population comprised 7,018 adult men and women, free from cancer at the baseline, carried out in 1978-1980. During a 14-year follow-up, 605 cancer cases occurred, of which 70 were male lung cancer patients. Mental problems and disorders were assessed at the baseline examination using standard interview techniques (General Health Questionnaire and Present State Examination). The relative risk of lung cancer between depressive persons and individuals with a normal depressiveness score was 3.32 (95% confidence interval 1.53-7.20). Neither adjustment for the potential confounding factors of age, education, geographic area, smoking, alcohol consumption, body mass index, serum cholesterol, leisure-time exercise, general health, and use of antidepressant medication nor exclusion of cancer cases occurring during the first 4 years of follow-up notably altered the results. There was a strong interaction between depressiveness and smoking. The relative risks of lung cancer between smokers and nonsmokers were 3.38 (95% confidence interval 1.09-10.52) at normal depressiveness score levels and 19.67 (95% confidence interval 2.57-150.7) at strongly elevated levels, respectively. It is possible that depressiveness modifies the effect of smoking on lung cancer risk either by biologic mechanisms or by affecting smoking behavior.