The National Heart, Lung, and Blood Institute Twin Study is a multicenter, longitudinal study of 514 white, male twin pairs examined during military induction at the mean age of 20 years (1943) and by the National Heart, Lung, and Blood Institute Twin Study at mean ages 48 (1971), 58 (1981), and 63 years (1986). Of these, 121 of 254 monozygotic pairs and 113 of 260 dizygotic pairs had complete data for all examinations. Using these data, genetic influences on maximum body mass index (BMI) and changes in BMI during this 43-year interval were estimated. BMI (kg/m2) was calculated for each examination and for the maximum weight as recalled by the participant at the last examination. Regression equations were fitted to each person's four-examination measurements to estimate the trend in weight change over adulthood. Twins gained an average of 0.11 kg/m2 per year from mean ages 20-63 years. Maximum BMI averaged 28.4 kg/m2 for monozygotic twins and 28.3 kg/m2 for dizygotic twins. The distribution of the ages at maximum BMI appeared to be bimodal for each zygosity, with modes around ages 20 and 60 years. Heritability (variation attributable to genetic factors) was estimated to be 0.71 (95% confidence interval 0.55-0.87) for maximum BMI and 0.70 (95% confidence interval 0.55-0.84) for trend in adult weight gain. In contrast, variability of BMI around the trend line showed no evidence of significant genetic determination.