Aerobic fitness has been associated with various desirable psychological and physiological characteristics. Recently, attenuation of physiological reactivity during stressful situations was added to this list, although comparison of the stress responses of sportsmen and sedentary subjects has yielded equivocal results. The present study examined cardiovascular patterns rather than single variables, and tried to clarify these matters. Tasks were used that were known to increase blood pressure through different combinations of changes in cardiac output and vascular resistance. Autonomic nervous system dynamics underlying these response patterns were studied using preejection period as an index of beta-adrenergic activity, and respiratory sinus arrhythmia as an index of vagal activity. Pre-existing differences in aerobic fitness in a sample of sedentary subjects were related to their responses during the stressful tasks and the recovery periods afterwards. This approach prevented confounding of the relationship between fitness and stress-reactivity with the psychological effects of regular exercise. Furthermore, it excluded the bias in psychological makeup that is introduced when subjects spontaneously engaged in sports are compared to non-exercising persons. To rule out a third (hereditary?) factor underlying both stress-reactivity and fitness, physiological responses before and after a seven-week training program were compared to those of subjects in a waiting list control group. Substantial individual differences in aerobic fitness were found in spite of the fact that all subjects reported low levels of habitual activity. During two active coping tasks, diastolic blood pressure reactivity and vagal withdrawal were negatively related to these pre-existing differences in fitness. No such relation was seen during a cold pressor test or during recovery from the tasks. Neither beta-adrenergic cardiac reactivity nor heart rate responses were related to fitness, but the absolute heart rate during the tasks was lower in the more fit subjects. Seven weeks of training were not effective in changing either reactivity or recovery of any of the variables. The discrepancy between cross-sectional and longitudinal results in the present study suggests that training of longer duration is necessary to induce the psychological or physiological changes underlying reduced reactivity. The latter may include changes in cardiac vagal/sympathetic balance or in adrenoceptor sensitivity. Alternatively, both psychological and physiological determinants of stress-reactivity may be related to aerobic fitness at a dispositional level.