This study examined the relative impact of five behavioral factors--activity, posture, location, social involvement, and tension--on the 24-hr blood pressure (BP) variability of 21 normotensives, 18 borderline hypertensives, and 18 sustained essential hypertensives. Multiple-regression analyses indicated that within each diagnostic group, activity accounted for more variance in BP variability than any other behavioral dimension. For each behavioral dimension examined, the magnitude of the relationship with BP was generally greater for the normotensives than for both hypertensive groups. In contrast, variation due to individuals was a better predictor of BP variability for the two hypertensive groups than for the normotensive group. Number of months hypertensive and use of antihypertensive medication were related to BP variability for the sustained hypertensive group. Findings have implications for the use of ambulatory BP monitors and interpretation of resulting data.