Study objective - To determine associations between physical activity at age 35-45 years with peripheral arterial disease and cardiovascular risk factors at age 55-74 years. Design - Cross sectional survey of the general population - Edinburgh Artery Study. The presence of peripheral arterial disease was determined using the WHO/Rose questionnaire on intermittent claudication, and the ankle brachial pressure index at rest and during reactive hyperaemia. Levels of physical activity undertaken at the time of the survey and at the times the subjects were aged 35-45 years were measured by self administered recall questionnaire. Setting - City of Edinburgh, Scotland. Participants - Altogether 1592 men and women aged 55 to 74 years, selected from the age-sex registers of 10 general practices spread geographically and socioeconomically throughout the city. Main results - Participation in moderate or strenuous activity when aged 35-45 years was reported by 66% of men and 40% of women. In men, but not in women, less peripheral arterial disease (measured by an increasing trend in the ankle brachial pressure index) was found with increasing amounts of exercise at age 35-45 years (p < 0.001). Higher levels of exercise at age 35-45 years were associated with lower blood viscosity (p < 0.05) and plasma fibrinogen levels (p < 0.05) in men and women aged 55-74 years, and also with higher current alcohol intake (p < 0.001) and high density lipoprotein cholesterol concentrations (p < 0.01) in women aged 55-74 years. After adjustment for age, sex, life-time smoking, social class, body mass index, and alcohol intake, the association between leisure activity aged 35-45 years and the ankle brachial pressure index aged 55-74 years remained highly significant in men who had at some time smoked (p < 0.001) but not in men or women who had never smoked (p > 0.05). Conclusion - The risk of peripheral arterial disease, particularly among male smokers, is inversely related to previous physical activity in early middle age, suggesting a protective effect of exercise.