To investigate the relation between nocturnal decline in blood pressure and mortality, we obtained ambulatory blood pressures in 1542 residents aged 40 years or over of a rural Japanese community. Subjects were followed-up for a mean of 5.1 years and were then subdivided into four groups according to the percent decline in nocturnal blood pressure: 1) extreme dippers: percent decline in nocturnal blood pressure ≥ 20% of the daytime blood pressure; 2) dippers: decline of ≥ 10% but < 20%; 3) nondippers: decline of ≥ 0% but < 10%; and 4) inverted dippers: no decline. The relationship between the decline in nocturnal blood pressure and mortality was examined by the Cox proportional hazards regression model adjusted for age, sex, smoking status, previous history of cardiovascular disease, and the use of antihypertensive medication. The mortality risk was highest in inverted dippers, followed by nondippers. There was no difference in mortality between extreme dippers and dippers. This relationship was observed for both treated and untreated subjects, was more pronounced for cardiovascular than for noncardiovascular mortality, and did not change after the data were adjusted for 24-h, daytime, and nighttime blood pressure levels. Am J Hypertens 1997;10:1201–1207 © 1997 American Journal of Hypertension, Ltd.