The left anterior descending coronary artery was ligated in 58 open-chest anaesthetised dogs; 23 were controls, 15 were given intravenous propranolol 1 mg·kg−1 before and at 6 h intervals after coronary ligation, nine had bilateral cervical vagal nerve stimulation (VS) before and for 4 to 6 h after coronary ligation, and 11 had both VS and propranolol. None of the 20 dogs undergoing VS developed ventricular fibrillation within the first hour after coronary ligation compared to nine of the remaining 38 (P<0.05). Compared to controls, myocardial creatine kinase (CK) depletion in the epicardial layer of the infarct centre measured 24 h after coronary ligation was significantly less in the groups treated separately with vagal nerve stimulation and propranolol. Myocardial blood flow (MBF) measured at 15 min after coronary ligation was reduced to the normal myocardium by the interventions, but was unchanged at the infarct centre. Severely ischaemic myocardium (MBF ≤ 20% of normal) was better protected by the interventions than was moderately ischaemic myocardium. At 15 min after coronary ligation, the heart rate — blood pressure product (RPP) was reduced compared with controls by propranolol (18% reduction, P<0.05), reduced more by vagal stimulation (by 37%, P<0.001) and still more by vagal stimulation with propranolol (by 43%, P<0.001). Preservation of CK in myocardium with MBF ≤20% of normal was improved by VS and propranolol given separately roughly in proportion to reduction in RPP, but further reduction in RPP by VS and propranolol together did not improve CK levels further. We conclude that there may be an optimum level of indices of oxygen demand for preservation of very ischaemic myocardium in experimental infarction.