In vivo studies indicate that blood components, especially leucocytes, contribute to reperfusion injury after myocardial ischaemia. This study was designed to: (1) develop a small animal heart model of ischaemia-reperfusion that demonstrates the contribution of blood to reperfusion injury; (2) determine when the presence of blood in the heart--that is, during ischaemia or during early reperfusion--caused greater dysfunction; and (3) attempt to limit the blood contribution to reperfusion injury by leucocyte depletion. Adult rat hearts were perfused in situ with a Krebs-albumin red cell solution (K2RBC), then isolated. Cardiac pump function was assessed with an intraventricular balloon as left ventricular developed pressure and contractility (dP/dt). Group I served as a non-ischaemic control group. Group II was subjected to global, no flow ischaemia for 30 min followed by 45 min reperfusion. In group III, diluted whole blood replaced the K2RBC for five min immediately before ischaemia. In group IV, diluted whole blood was perfused during the first five min reperfusion. In group V, the hearts were reperfused with leucocyte poor diluted whole blood. Pre-ischaemic pump function values were similar to other blood perfused, isolated heart models. Group I showed no increase in coronary resistance or decrease in pump function with time or in response to diluted whole blood. After 35 min reperfusion, the recovery (% control) of dP/dt in group II was 56(12), in group III it was 39(15) and in group IV it was only 19(6) (p < 0.05). Large increases in coronary vascular resistance, oedema, and contracture during reperfusion were also seen in group IV. When leucocytes were depleted from the diluted whole blood (group V), the recoveries were similar to reperfusion without diluted whole blood (group II). Thirty min of global, normothermic ischaemia caused significant cardiac dysfunction early during reperfusion. Perfusion with unstimulated blood for a limited period further impaired the recovery of function and enhanced myocardial oedema. Dysfunction was particularly evident when diluted whole blood was perfused during the first minutes of reperfusion. The leucocyte depletion studies suggest that leucocytes are necessary, but may not be sufficient, to demonstrate the blood contribution to reperfusion injury.