1 Anaesthetized dogs were subjected to 1 h occlusion of the left circum¯ex coronary artery followed by 2 h of reperfusion. Relaxant responses were examined in coronary artery rings removed proximal (nonischaemic) or distal (ischaemic) to the site of occlusion. 2 Relaxant responses to acetylcholine (ACh) were similar in nonischaemic and ischaemic artery rings. In addition ACh-induced relaxation of nonischaemic and ischaemic artery rings was equally susceptible to inhibition of nitric oxide (NO) synthase using L-N G -nitroarginine (L-NOARG, 10 74 M), or to inhibition of soluble guanylate cyclase using 1H-[1,2,4]oxadiazolo[4,3-a]quinoxaline-1-one (ODQ, 10 75 M). 3 In nonischaemic arteries, the relaxation to ACh was una ected by high K + (67 mM) but in ischaemic arteries, the maximum relaxation to ACh was signi®cantly reduced from 113+6 to 60+2% (ANOVA, P50.05). Tetraethylammonium (TEA, 10 73 M), an inhibitor of large conductance calcium activated potassium (BK Ca ) channels did not inhibit the response to ACh in nonischaemic arteries but in ischaemic arteries TEA signi®cantly shifted the concentration response curve to ACh to the right (pEC 50 ; nonischaemic, 7.07+0.25; ischaemic, 6.54+0.21, P50.01, ANOVA) without decreasing the maximum relaxation. TEA did not a ect the responses to sodium nitroprusside in either nonischaemic or ischaemic arteries. 4 In conclusion, ischaemia/reperfusion did not change the sensitivity of endothelium-dependent relaxation to L-NOARG or ODQ indicating that ischaemia did not a ect the contribution of NO or cyclic GMP to ACh-induced relaxation. However, in ischaemic arteries the opening of the BK Ca channels contributed to relaxation caused by ACh whereas TEA had no e ect in nonischaemic arteries. The factor responsible for the opening of this potassium channel was a factor other than NO and may be endothelium derived hyperpolarizing factor (EDHF).