The duration per week of exposure to others' tobacco smoke in different locations was tabulated from the questionnaire responses of 37,881 non-smokers and ex-smokers who received multiphasic health checkups in 1979 and 1980. Altogether 63.3 per cent of subjects reported some exposure, 34.5 per cent were exposed at least 10 hours per week, and 15.9 per cent at least 40 hours per week. Duration of exposure per week was weakly correlated with serum thiocyanate level in a small test group. Exposure was strongly related to age with a peak of 78.2 per cent in the twenties decade and an accelerating decline thereafter to 13.9 per cent in those age 80 and over. Sex and race were related to passive smoking only to a relatively small degree. Persons reporting longer exposures were more apt to report greater use of alcohol and marijuana, exposure to occupational hazards, and being currently not married. Lack of college education was most frequent among those denying exposure but was directly related to exposure duration among those reporting some passive smoking. Studies of the health effects of passive smoking should take into account these associated factors. Although the reported passive smoking of married persons was strongly related to their spouses' habits, categorization by spouses' smoking resulted in considerable misclassification.