Dogs were subjected to exercise on a treadmill, using a protocol in which the speed and slope were increased every 3 min, and which elevated both heart rate (to a mean of 198+/-14 beats.min(-1)) and mean arterial blood pressure (to 150+/-4 mmHg). Then, 24 or 48 h later, the dogs were anaesthetized with a mixture of alpha-chloralose and urethane and subjected to a 25 min occlusion of the left anterior descending coronary artery. The control dogs (instrumented but not exercised) were subjected to the same procedure. In some dogs the nitric oxide synthase inhibitor aminoguanidine (50 mg.kg(-1); intravenous) was administered 30 min before occlusion. Baroreflex sensitivity (BRS) was determined by the rapid bolus injection of phenylephrine 60 min before, and again 3 min after, the onset of occlusion. Exercise markedly reduced the consequences of coronary artery occlusion 24 h (but not 48 h) later, without modifying myocardial tissue blood flow. In the exercised dogs there were reductions in arrhythmia severity [ventricular fibrillation (VF) during occlusion, 0%; survival from the combined ischaemia/reperfusion insult, 70%] compared with controls (VF during occlusion, 36%; survival, 9%). BRS was preserved during occlusion in the exercised dogs (before occlusion, 1.60+/-0.54 ms.mmHg(-1); 3 min after occlusion, 1.37+/-0.4 ms.mmHg(-1)), but not in controls (before occlusion, 1.28+/-0.29 ms.mmHg(-1); 3 min after occlusion, 0.45+/-0.12 ms.mmHg(-1); P<0.05), and other ischaemic changes (inhomogeneity of electrical activation and changes in the ST-segment, recorded over the ischaemic region) were also less marked in the exercised dogs. Exercise-induced cardioprotection was abolished by aminoguanidine (VF during occlusion, 25%; survival, 0%). The results show that even a single period of exercise affords delayed protection against ischaemia/reperfusion-induced VF and other ischaemic changes. Since this protection is abolished by aminoguanidine, and since (inducible) NO synthase activity was elevated 3-fold in left ventricular samples 24 h after exercise, we suggest that this protection is mediated by nitric oxide.