The objective of this research was to investigate the long-term relation between body mass index (BMI) and mortality from all causes and from specific causes in the general population. A 29-year follow-up study was conducted in a random sample of white men (n=611) and women (n=697) aged 20–96 years who were residents of Buffalo, New York, in 1960. At baseline, height and weight were determined by self-report. BMI was calculated as weight (kg)/height (m2). During the follow-up period, 295 (48.3 percent) men and 281 (40.3 percent) women died. With the Cox proportional hazards model and adjustment for age, education, and cigarette smoking, a significant linear association was found between BMI and all-cause mortality in men less than age 65 years at baseline (relative risk (RR) = 1.06, 95 percent confidence Interval 1.02–1.09), but not in women (RR = 1.02, 95 percent confidence interval 0.99–1.05). In men age 65 years and older, the relation was quadratic in form (p=0.02), with the lowest risks appearing in the BMI range of 23–27. BMI was most strongly related to cardiovascular disease (CVD) and coronary heart disease mortality in women and younger men. No such associations were observed in older men. BMI was not related to an increased risk of death from non-CVD or cancer in either sex. These findings illustrate the importance of BMI as a risk factor for CVD and coronary heart disease mortality in certain gender-age groups and indicate that the majority of the impact of BMI on overall mortality is due to the strong relation between relative weight and these specific causes of death.