Being male constitutes a risk factor for early death, and this may be connected to concepts of masculinity. The objective of the present study was to investigate the association between masculinity rank during late adolescence and mortality patterns in a cohort of 49 321 Swedish men tested for compulsory military training in 1969. The measure of masculinity consisted of a five-grade ranking from the conscription information that comprised leisure interests and occupational preferences. Information on all-cause mortality, alcohol-related mortality, suicide, mortality from other violent causes and mortality from cardiovascular disease was collected from national registers for 1970-2003. The analyses were performed using Cox proportional hazard models with hazard ratios as estimates of relative risk. For all-cause mortality, the crude relative risks versus ordinary masculinity were: lowest masculinity 1.98 (95% CI 1.71 to 2.31), low masculinity 1.38 (95% CI 1.24 to 1.53), high masculinity 0.90 (95% CI 0.81 to 1.01) and highest masculinity 0.78 (95% CI 0.62 to 0.97). After adjustments for childhood class, smoking, alcohol drinking, blood pressure, short stature, psychiatric disorder, low intelligence and contact with police or childcare authorities, conscripts who were ranked lowest on masculinity had a remaining increased relative risk of all-cause mortality (1.49; 95% CI 1.28 to 1.75) and suicide (1. 79; 95% CI 1.31 to 2.43). For mortality from violent causes other than suicide, no statistically significant associations were demonstrated. Earlier research has proposed that masculinity may hold both positive and negative aspects regarding lifetime health. The major conclusion here is that being ranked less masculine is associated with higher mortality.