Central to the psychophysiologic reactivity hypothesis of the etiology of coronary artery disease is the assumption that reactivity is an individual characteristic that is stable over time. Although heart rate (HR) and blood pressure reactivity appear to meet this criterion, temporal stability of cardiac autonomic control as measured by analysis of heart period variability (HPV) has not been assessed. In this study, we tested the stability of HPV, measured in both the time and frequency domain, during a quiet, resting baseline and in response to 5-min mental arithmetic and reaction time tasks, in 20 normal subjects measured in three testing sessions during a 9-month period. Stability, assessed by the intraclass correlation coefficient (ICC), was excellent for resting baseline measures of HR and HPV, with ICCs of 0.68-0.86. However, HR and HPV reactivity to either arithmetic or reaction time tasks generally was less stable, with ICCs of 0.17-0.73, in contrast to results of previous studies demonstrating long-term stability of HR responses to psychological challenge. Stability of aggregated reactivity scores was only slightly improved. Whether for individual tasks or aggregated measures, reactivity of total and low-frequency measures of HPV was moderately stable but stability of high-frequency HPV reactivity was poor.