The highly selective adenosine A1 receptor agonist, 2-chloro-N6-cyclopentyl-adenosine (CCPA), has been shown to be as cardioprotective as ischaemic preconditioning when evaluated with an early staining method using tetrazolium. However, tetrazolium-positive tissue measured 3 h after reperfusion may still overestimate the long term salvage. To test for this possible artefact, a 72 h reperfusion rabbit model of myocardial infarction was used, and infarct size was assessed by histology. Myocardial infarction was induced by a 30 min coronary occlusion. Rabbits were assigned to a control group receiving no treatment, pretreatment with 0.125 mg.kg-1 CCPA, or 0.25 mg.kg-1 pretreatment with CCPA (0.25 mg.kg-1) followed by an A1 selective antagonist, 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) 30 min after reperfusion to reverse the haemodynamic side effects. In the 0.125 mg.kg-1 CCPA group, 30.8(SEM 4.2)% of the ischaemic zone was infarcted, which was significantly less than that seen in the control group [46.5(3.0)%; p < 0.01]. Reversing the side effects of CCPA by giving DPCPX soon after reperfusion did not block the protective effects [26.2(1.9)% infarction; p < 0.01 v control]. This finding confirms a genuine anti-infarct effect of adenosine A1 receptor stimulation when given prior to the onset of ischaemia. Furthermore blocking the A1 receptors soon after reperfusion reverses the side effects but does not block protection.