While factory-made cigarettes dominate the market in most countries, the use of handrolled cigarettes accounts for a substantial proportion of the tobacco consumption in Norway. In the present study, we examined the impact of tobacco smoking on lung cancer in general, and the effect of handrolled cigarettes in particular. The data used was from a self-administered mailed questionnaire which included questions about smoking habits and which was completed by about 26,000 men and women in 1964-65. During the follow-up from 1966 to 1993, 333 lung cancers in men and 102 in women were registered. The analysis was performed by use of the Cox proportional hazards regression models. A clear dose-response relationship was found both for cigarette smoking, and for pipe smoking (in men). The dose-response relationship of cigarette smoking was seen in all the three histologic groups considered-squamous cell carcinoma, adenocarcinoma, and small cell carcinoma. The highest relative risks were noted in squamous cell and small cell carcinoma. A higher risk of lung cancer was found for cigarette-smoking women who started cigarette smoking before the age of 30 compared with similar groups of men. In a combined analysis of men and women, an elevated relative risk of 1.9 (95 percent confidence interval = 1.2-3.3) was found for those smoking only handrolled cigarettes compared with those smoking factory-made filter cigarettes only.