Blood pressure and pulse were monitored throughout 24 hours in 13 middle-aged men using a sphygmomanometer with an automatically inflatable cuff. These volunteers were in part mistakenly believed to have a "mildly elevated blood pressure" (in the light of a conventional non-chronobiologic approach, involving time-unspecified spot checks of blood pressure). Within about 24 hours, blood pressures and pulse varied remarkably: the coefficient of variation extended from 6.9 to 14.9% for systolic blood pressure (SBP), from 8.5 to 15.4% for diastolic blood pressure (DBP) and from 8.0 to 17.9% for pulse. By applying the single cosinor method (involving the least-squares fit of a 24-hour cosine curve), statistically significant circadian rhythms were observed in SBP, DBP, and pulse in all but 2 cases. In most cases, the timing of peak values was estimated to be in the evening for all 3 variables. Circadian rhythmicity accounted for up to 46, 30, and 42% of the total variability of SBP, DBP, and pulse, respectively. This marked circadian rhythmic variability is an important aspect of cardiovascular function.