STUDY OBJECTIVE--To determine the effect of marital status on mortality for men. In particular, to examine whether subgroups of unmarried men (widowed, single, and divorced/separated men) have a similar mortality to married men. DESIGN--Cohort study. SETTING--Whitehall civil service, London, between 1967 and 1969. PARTICIPANTS--A total of 18,403 men aged 40-64 years with 18 years' follow up. MEASUREMENTS AND MAIN RESULTS--Cause-specific mortality rates and risk factors at baseline were determined. Overall mortality was greater for all groups of unmarried men. Patterns of mortality were different in the subgroups of unmarried men. Widowed men had a significantly greater risk of dying from ischaemic heart disease (relative risk (RR) 1.46, 95% confidence interval (CI) 1.08, 1.97) which persisted after exclusion of deaths that occurred in the first two years. Divorced men had greater cancer mortality (RR 1.49; 95% CI 1.06, 2.10) that could not be explained simply by their greater consumption of cigarettes. The initial increased mortality for single men was no longer evident after adjustment for other risk factors, suggesting that single status in itself may not increase the risk. The risk for single men may have been underestimated, however, by over adjustment for possible intermediary factors. CONCLUSIONS--Previous studies, which have examined total mortality only or have grouped all unmarried men, have masked interesting differences in the cause of death between subgroups of unmarried men. The extent to which the findings are explicable by psychosocial factors or the role of other environmental factors, which may also differ in relation to marital status, is unclear. Future work should not assume that all unmarried men have similar mortality risks and must examine the life course of each subgroup to advance our understanding of the possible causal role of marital status in disease aetiology.