Extensive preparative surgery and lengthy experimentation may lead to high rate of complications and mortality in myocardial ischaemia studies. These problems are particularly common when pigs are used as the subject as they are prone to develop lethal ventricular arrhythmias. Here, a closed-chest model is presented, in which the trauma of major preparative surgery is avoided. One-hundred and twelve pentobarbital-anaesthetized, mechanically ventilated pigs were used. Coronary occlusion was produced by injection of a 2 mm diameter ball via a modified coronary angiography catheter. Reperfusion was induced by retraction of the ball via a thin filament attached to the ball. The amount of the myocardium at risk (MAR) was8·23± 2·41% (mean ± SD) of the left plus right ventricular weight. It was possible to carry out scheduled 24 h experiments in 87 out of 93 animals (93·5%). Preparative mortality was 1·8% and 24 h mortality 6·5%. Ventricular fibrillation ( VF) occurred during preparation in 36%, during coronary occlusion in 7·3% and during reperfusion in 5·0% of the animals. VF was significantly related to a large zone of MAR and insufficient premedication. Catheter- or ball-induced complications were found in 10·7%. Mortality and incidence of VF are considerably lower in this closed-chest model than in a previously reported open-chest pig preparation.