Obesity is generally assumed to be an important risk factor for death and morbidity. However, the association between excess body weight and all-cause mortality among younger and older women and the impact of body mass index (BMI) and waist-to-hip ratio (WHR) concurrently is not fully understood. In 1968-1969 we initiated a prospective study comprising a population sample of 1,462 women from Gothenburg, Sweden. During a 24 year period, until 1992-1993, 265 women had died. A multivariable Cox Proportional Hazards Regression model was used to estimate the relative risk of death in relation to BMI and WHR, with age and other covariates of age-specific interest as smoking, physical activity at work and leisure time and serum triglyceride concentration, at start of the study. BMI and WHR were analyzed as independent variables. Younger women (38 and 46 years at baseline) presented a statistically significant non-linear (U-shaped) relation between BMI and mortality. Among older women (50, 54 and 60 years at baseline), a significant negative linear relationship with decreasing mortality in relation to increasing BMI values was seen. For all women a higher WHR was related to an increased risk of death. The lowest risk of death among younger women corresponded to a low WHR and a BMI within the middle range. For older women the highest survival was observed for those with lowest WHR and highest BMI. Thus, in older women a high BMI seems not to be an increased risk as long as adiposity is not centrally located.