The Copenhagen Male Study is a prospective cohort study initiated in 1970/71 comprising 5249 employed men between the ages of 40 and 59 years. Included in a registry follow up of lung cancer were 4931 men who responded sufficiently to a number of questions on tobacco habits and who could be classified into social classes. During the 17 years of follow up, lung cancer was diagnosed in 144 men. By the end of the follow up period, 135 had died. Substantial social inequalities in the risk of lung cancer were found with a gradually increased risk with low social class, Kendall's tau B = 0.07, p less than 0.001. In multivariate analysis, compared with the highest social class (highly educated, administrators), the lowest social class (unskilled workers), had a highly increased risk, relative risk (with 95% confidence limits), RR = 3.7 (1.9-7.3). If in the analysis, adjustments were made for form of smoking, amount smoked, whether inhalation took place, number of pack-years and age, the increased risk dropped to RR = 2.9 (1.5-5.9). We conclude that the substantial social inequalities in lung cancer risk are only to a minor degree explained by social class differences in tobacco smoking habits.