The association between marital history at midlife (in 1950) and mortality (as of 1991) was studied in a group of intelligent, educated men and women (N = 1,077) who participated in the Terman Life-Cycle Study initiated by Lewis Terman in 1921. Results confirm that consistently married people live longer than those who have experienced marital breakup but suggest that this is not necessarily due to the protective effects of marriage itself. Individuals who were currently married, but had previously experienced a divorce, were at significantly higher mortality risk compared with consistently married individuals. Furthermore, individuals who had not married by midlife were not at higher mortality risk compared with consistently married individuals. Part of the relationship between marital history and mortality risk may be explained by childhood psychosocial variables, which were associated with both future marital history and mortality risk.