Blood pressure (BP) usually peaks in the morning. The circadian variation of the onset of cardiovascular disease mimics this circadian BP variation. To examine the determinants of the BP difference between the self-recorded BP in the morning (home BP) and daytime average ambulatory BP a cross sectional study was done in the general population of Ohasama, Japan. 1207 subjects > or = 20 years measured both home (more than 14 times) and ambulatory BPs (326 treated for hypertension and 881 untreated subjects), The prevalence of subjects with the systolic BP difference (home BP in the morning - daytime ambulatory BP) of > or = 10 mmHg (high morning BP) was 5.6% in untreated normotensives, 2.9% in untreated hypertensives, and 25.8% in treated hypertensives. This trend was also observed for diastolic pressure. Multiple regression analysis demonstrated that age, male sex, and use of antihypertensive drugs were positively associated and day-night difference of BP was negatively associated with the high morning BP, respectively. These results suggest an insufficient duration of antihypertensive action of widely used antihypertensive drugs in Japan from the 1980s to 1990s. The amplitude of the day-night difference of ambulatory BP in subjects with a high morning BP was lower (non-dipping) than that without high morning BP. The high morning BP is not necessarily accompanied by hypertension but might be mediated, at least in part, by an insufficient duration of action of antihypertensive drugs. The high morning BP accompanies so-called non-dipper pattern of circadian BP variation. An insufficient duration of action of drugs may partly mediate non-dipping in subjects with antihypertensive medication.