BACKGROUND The results of twin and family studies suggest that heredity has a small influence on smoking behavior. METHODS We conducted a genetic analysis of several aspects of smoking behavior among subjects in the National Academy of Sciences-National Research Council Twin Registry. The registry includes male twins who were born in the United States between 1917 and 1927 and who were members of the armed services during World War II. Information on smoking history was available for 4775 pairs of twins, who were first surveyed in 1967 through 1969, when they were 40 to 50 years old, and then re-surveyed in 1983 through 1985, when they were 56 to 66. Eighty percent of the subjects in this cohort had smoked at some time in their lives, 60 percent were smokers in 1967 through 1969, and 39 percent were smoking in 1983 through 1985. Similarities between twins in smoking habits at base line and at the second follow-up 16 years later were examined. The comparison of concordance for smoking between monozygotic and dizygotic twins was used to assess the relative contribution of familial and genetic factors. RESULTS In 1967-1969 survey the ratio of observed to expected concordance for smoking was higher among the monozygotic twins than among the dizygotic twins for those who had never smoked (overall rate ratio, 1.38; 95 percent confidence interval, 1.25 to 1.54), for former smokers (overall rate ratio, 1.59; 95 percent confidence interval, 1.35 to 1.85), for current cigarette smokers (overall rate ratio, 1.18; 95 percent confidence interval, 1.11 to 1.26), and for current cigar or pipe smokers (overall rate ratio, 1.60; 95 percent confidence interval, 1.22 to 2.06). The data also suggest genetic influences on quitting smoking. Monozygotic twins were more likely than dizygotic twins to be concordant for quitting smoking (overall rate ratio, 1.24; 95 percent confidence interval, 1.06 to 1.45). CONCLUSIONS In this cohort of adult male twins, there were moderate genetic influences on lifetime smoking practices.