OBJECTIVE--To evaluate Holter and treadmill responses in patients with stable angina or recent myocardial infarction in order to compare the mechanisms of ischaemia and its symptomatic expression in these two groups. PATIENTS--75 patients with ischaemic ST segment depression on both a treadmill stress test and ambulatory Holter monitoring. Group A comprised 35 patients with stable angina, and group B comprised 40 patients in the early period after infarction. SETTING--The coronary care unit and cardiology department of a district general hospital. DESIGN--A prospective, between group, comparative study. RESULTS--Treadmill test showed demand driven ischaemia in both groups. Although ST depression occurred at comparable rate-pressure products and workloads, it was associated with angina in 80% of group A compared with only 40% of group B (p < 0.005). During Holter monitoring, ST depression was associated with an attenuated increase in rate in group A and almost no increase in rate in group B (18.2% v 3.7%; p < 0.005), suggesting that reductions in myocardial oxygen delivery were contributing to the ischaemic episodes, particularly in group B. Ischaemic episodes were more commonly silent during Holter monitoring, particularly patients in group B, only two of whom experienced angina in association with ST depression. Spectral and non-spectral measures of heart rate variability were significantly reduced in group B compared with group A. Patients with silent exertional ischaemia in group A had significantly less heart rate variability than patients who experienced angina but this difference was not seen in group B. CONCLUSION--In stable angina, myocardial ischaemia is usually painful and demand driven, whereas in the early period after infarction silent, supply driven ischaemia predominates. The failure of myocardial ischaemia to provoke symptoms in some patients with stable angina may be related to autonomic dysfunction affecting the sensory supply to the heart. In the early period after infarction despite clear evidence of autonomic dysfunction, other mechanisms must also be important as there was no tendency for the reduction in heart rate variability to be exaggerated in the subgroup with silent exertional ischaemia.