The feasibility of substituting a shorter duration of ambulatory blood pressure monitoring (ABPM) for 24h ambulatory monitoring to evaluate increased office BP measurements was investigated by analysing the records of 126 normal volunteers and 168 mildly hypertensive patients receiving placebo, including 22 studied on two occasions two months apart. The value to be predicted was the long-term awake BP mean for an individual (usual BP). Among the subjects studied on two occasions, there were no systematic differences between the two occasions and there was reasonable correspondence between the hourly patterns. Estimated variance components were substituted in a formula for the standard error of prediction (SEP) of the usual awake mean BP, for various durations and frequencies of monitoring. The lower limit of the SEP was determined by the patient-by-occasion component of variance. As few as six hours of monitoring with two to three readings/hour achieved most of the gain in precision obtainable by going from single BP readings toward continuous measurement during an entire awake period. In the remaining subjects, the influence of the day rhythm of BP on short-term monitoring was examined by measuring the differences in means of 6h periods from the full awake mean BP for starting times from 7 am to 3 pm. Among normotensives and hypertensive patients with near normal office BP (< 144/< 96 mmHg), there was little difference due to starting time. There was, however, a slight positive bias in the more hypertensive subjects.(ABSTRACT TRUNCATED AT 250 WORDS)