To determine whether body weight is a risk factor for mortality among older persons, we analyzed body mass index (weight [kilogram]/height [square meter]) data for 4710 white, National Health and Nutrition Examination Survey respondents who were aged 55 to 74 years during 1971 through 1975, in relation to their survival over an average of 8.7 years of follow-up. In a multivariate analysis that controlled for elevated blood pressure, smoking, and poverty, we found no additional risk associated with weight among women and a statistically significant, but moderate, additional risk (relative risk, 1.1 to 1.2) among men in the upper decile (body mass index, greater than or equal to 30 kg/m2). In contrast, low weight (body mass index, less than 22 kg/m2) was associated with increased mortality (relative risk, 1.3 to 1.6) except for women aged 55 to 64 years. We conclude that the accepted definition of overweight (body mass index, greater than or equal to 27.8 kg/m2 [men] or greater than or equal to 27.3 kg/m2 [women]) lacks specificity and may be inappropriate for older persons who do not have weight-related medical conditions. The low-weight mortality association, consistently demonstrated, deserves serious scrutiny.