To assess the association between exposure to environmental tobacco smoke (ETS) and lung cancer, the authors personally interviewed 292 lifelong nonsmoking lung cancer cases (recruited from 15 hospitals in the study area) and 1, 338 nonsmoking controls (randomly selected by population registries) between 1990 and 1996 in Germany. Subjects were asked by a standardized questionnaire about exposure to ETS in childhood, by spouse, at work, and in transportation and social settings. Several indicators of these different sources of exposure were investigated, using not or low exposed subjects as the reference category. The most informative quantification index was weighted duration of exposure (hours x level of smokiness). No effect of ETS exposure during childhood and no clear effect of spousal ETS were observed. However, for the highest category of exposure, clear effects of ETS at the workplace (odds ratio (OR) = 1.93; 95% confidence interval (Cl): 1.04, 3.58), in vehicles (OR = 2.64; 95% Cl: 1.30, 5.36), and from all sources combined (OR = 1.39; 95% Cl: 0.96, 2.01) were found. Adjustment for occupational carcinogens, radon, and diet did not appreciably change the results. These findings suggest that exposures to high levels of ETS at the workplace and in other public indoor settings appear to be important risk factors for lung cancer risk in nonsmokers. Am J Epidemiol 2000; 151: 241–50.