The present paper reports two studies which address the extent to which cardiovascular responses to behavioral stress can be conceptualised as a stable individual difference variable in children. In the first study, eighth grade girls (mean age = 14.2 yrs) who had participated in a study of the familial aggregation of cardiovascular responses to behavioral stress were asked to repeat the assessment protocol when they were in (he ninth grade (mean age = 15.1 yrs). Results showed that mean task-induced increases in diastolic blood pressure and heart rate during serial subtraction, mirror image tracing, and isometric handgrip tasks were reliable across test sessions, particularly due to the reliable increases exhibited during the mirror image tracing task. In the second study, fifth grade boys (mean age = 10.4 yrs) who had participated in a previous experiment (Matthews & Jennings, 1984) of cardiovascular responses during a video game were tested in the ninth grade (mean age - 14.0 yrs) in the protocol used in Study 1. Results showed that mean tusk-induced increases in diastolic blood pressure and heart rate during video games were related to mean task-induced increases during serial subtraction, mirror image tracing, and isometric handgrip tasks over three years Liter. The association for heart rate increases during the video game and isometric handgrip task was particularly impressive. These results suggest that task-induced diastolic blood pressure and heart rate responses of children are stable over substantial periods of time and amiss important developmental transitions, and generalize across tasks requiring qualitatively different behavioral responses. Nonetheless, future research is needed in adults and children to further identify factors which influence the stability and generalizability of cardiovascular responses to behavioral stress.