The likelihood of increasing the number of cigarettes per day (cpd), a common method of compensation, in smokers switching from nonfilter to filter cigarettes, was examined in newly diagnosed and histologically confirmed lung cancer cases including all cell types; 1,242 lung cancer cases and 2,300 sex and age matched hospital controls were interviewed in 20 hospitals from nine United States cities between 1969 and 1984. The mean increase in cpd for lung cancer cases was about twice that of controls. Using switchers who did not increase cpd as the referent group, there was a linear dose-response relationship between the odds of lung cancer and increasing level of compensation. The odds ratio increased from 1.19 to 2.37 in males and from 1.66 to 3.83 in females corresponding to increases of 1-10 to 21+ cpd after switching. Findings from this study suggest that increasing cpd after switching to filter cigarettes is an important risk factor for lung cancer that needs to be emphasized in epidemiologic studies. Proponents of the idea that switching cigarettes is of some benefit, should also advocate that individuals who continue smoking should avoid compensation after switching.