We investigated the utility of home blood pressure measurements for determining the risk of stroke. We also analyzed the relationship between home blood pressure and the incidence of stroke. Home blood pressure and screening blood pressure measurements were obtained from 1, 789 residents (aged 40yr or older) of a rural Japanese community. Blood pressure was measured at home with a semiautomatic device. A mean (±SD) of 23.0±7.5 measurements were made for each subject. Subjects without a history of stroke and who were not receiving medication for hypertension (n=1, 256) were prospectively followed up for 4.4±2.1yr. Subjects were subdivided into quintiles according to their baseline blood pressure. The association between the baseline blood pressure and the incidence of the first-ever stroke was examined with the Cox proportional hazards regression model, adjusted for age and sex. The lowest risk of stroke morbidity occurred in the subjects in the third quintile for home systolic blood pressure (117-123mmHg) and in those in the second quintile for home diastolic blood pressure (66-70mmHg). The subjects in the fifth quintiles for home systolic (_??_133mmHg) and diastolic blood pressure (_??_81 mmHg) had a significantly increased risk of stroke morbidity. The subjects in the first and the second quintiles for home systolic blood pressure and those in the first quintile ffor home diastolic blood pressure tended to have an increased risk as compared with subjects in the lowest risk groups, although this increase was not statistically significant, indicating two possibilities: a trend toward a J-shaped relationship or no decrease in risk of the first-ever stroke in subjects with home blood pressure level less than 123/70mmHg. This relationship was not observed for screening blood pressure. We conclude that home blood pressure measurements can provide additional prognostic information to that obtained from blood pressure measurement in a medical environment. (Hypertens Res 1997; 20: 167-174)