1 Although the endocannabinoids 2-arachidonoylglycerol (2-AG) and anandamide share a similar pharmacology, 2-AG reportedly limits myocardial ischaemia-reperfusion injury whereas anandamide does not. We therefore investigated whether or not anandamide reduces infarct size and which, if any, of the known cannabinoid-signalling pathways are involved. 2 Rat isolated perfused hearts were subjected to global, no-flow ischaemia (30 min) and reperfusion (1 h). Agonists were present from 5 min before ischaemia until the end of reperfusion. Antagonists, where used, were present throughout the protocol. Recovery of left ventricular developed pressure and coronary flow was incomplete in control hearts and not significantly affected by any drug treatment. 3 In vehicle-treated hearts, 2673% (n ¼ 13) of the left ventricle was infarcted at the end of reperfusion. Infarction of the left ventricle was significantly reduced after 1 mM anandamide (1071%, n ¼ 7) or 1 mM methanandamide (1274%, n ¼ 6) but not 1 mM HU210. Neither ACPA (1 mM; CB 1 receptor agonist) nor JWH133 (1 mM; CB 2 receptor agonist), individually or combined significantly affected infarct size. 4 Anandamide (1 mM) did not reduce infarct size in the presence of the CB 1 receptor antagonist rimonabant (SR141716A, 1 mM) or the CB 2 receptor antagonist, SR144528 (1 mM). 5 Despite sensitivity to CB 1 and CB 2 receptor antagonists, the infarct-limiting action of anandamide was not mimicked by agonists selective for CB 1 or CB 2 receptors suggesting the involvement of a novel cannabinoid site of action.