To examine the relationships between casual, ambulatory and home blood pressure measurements in the general population, these measurements were obtained in 1, 695 of 3, 744 subjects aged 20yr or older in Ohasama, Japan. Of these 1, 695 subjects, 1, 207 measured their home blood pressure more than 14 times in each of the morning and evening (881 untreated subjects including normotensives and untreated hypertensives, 56.4±11.5yr of age; 326 treated subjects, 66.0±9.2 yr of age). We analyzed data in these 1, 207 subjects, examining the distribution of each measurement, the relationships among measurements, and the factors affecting the blood pressure differences among the measurements. For systolic pressure, the casual measurement was the highest among the methods examined. The daytime ambulatory measurement was significantly higher than morning and evening home measurements. Morning home measurements were significantly higher than those in the evening. For diastolic pressure, however, the morning home measurement was the highest among the methods examined. Short-term pressure variability (standard deviation and variation coefficient of ambulatory measurements) was greater than long-term pressure variability (standard deviation and variation coefficient of home measurements). The pressure variability in treated subjects was greater than that in untreated subjects. The correlation between casual pressure and the other pressures was not as strong (r<0.567). Among the relationships between ambulatory and home measurements, the strongest correlation was observed between the 24-h ambulatory measurement and the morning home measurement (r=0.738) in untreated subjects. The morning home measurement was highly correlated with the evening home measurement (r>0.814). The differences among the methods examined were affected by blood pressure level and age. It should be noted that in elderly and treated subjects, blood pressure measurement using one method does not necessarily correlate with that obtained using the other methods. This information is useful for the estimation of the value of one type of blood pressure measurement from values obtained with other methods. (Hypertens Res 1999; 22: 261-272)