This population study included 399 subjects, of whom 370 (93%) showed a significant diurnal blood pressure (BP) rhythm. The nocturnal BP fall was normally distributed and averaged 16 +/- 9 mm Hg systolic and 14 +/- 7 mm Hg diastolic (mean +/- SD). The amplitude of the diurnal BP curve followed a positively skewed distribution, with a mean of 16 +/- 5 mm Hg for systolic BP and 14 +/- 4 mm Hg for diastolic BP. The daily BP maximum occurred at 15:54 +/- 4:47 for systolic BP and at 15:11 +/- 4:20 for diastolic BP. Thirty-four subjects were reexamined after a median interval of 350 days. The test for the presence of a significant diurnal rhythm was discordant in only two subjects. Repeatability (twice the standard deviation of the differences between paired recordings expressed as a percentage of the mean) varied from 11 to 25% for the 24 h, daytime, and overnight BP, and from 76 to 138% for the parameters describing the diurnal BP rhythm. In nine subjects with an initial night/day ratio of mean BP less than 0.78, the nighttime BP was significantly increased at the repeat examination, whereas the opposite tendency was observed in nine subjects with an initial ratio greater than 0.87. In conclusion, the distribution of the nocturnal BP fall is unimodal. The reproducibility of the ambulatory BP is satisfactory for the level of BP and for the presence of a diurnal BP rhythm, but not for the parameters of the diurnal BP curve. Thus, one 24 h recording is insufficient to fully characterize an individual's diurnal BP profile.