Objective: The aim was to test the hypothesis that A, receptor mediated cardioprotection by endogenous adenosine is exerted during ischaemia rather than reperfusion. Methods: Anaesthetised open chest rabbits were subjected to 30 min regional ischaemia and 120 min reperfusion, and randomised to one of six groups: group I - saline vehicle (VEH) (n=12) to allow A(1) and A(2) adenosine receptor interactions during ischaemia and reperfusion; group II - both A, and A? receptors were antagonised during ischaemia and reperfusion with 8-p-sulphophenyltheophylline (SPT) (10 mg.kg(-1)) (SPTIR, n = 14); groups III and IV - the selective A, adenosine receptor antagonist 8-(3-noradamantyl)-1,3-dipropylxanthine (KW-3902) was given during ischaemia-reperfusion in low dose (1 mg.kg(-1), LA(1)-IR, n = 11) and higher dose (2 mg.kg(-1), HA(1)-IR, n = 6); group V - KW-3902 (1 mg.kg(-1)) was given only during reperfusion (A(1)-R, n=12); group VI - SPT was given only at reperfusion (SPTR, n=11). Results: In in vitro studies, (1) KW-3902 completely inhibited negative inotropic effects of the A(1) agonist R(-)N-6-(2 phenylisopropyl) adenosine (R-PIA) in catecholamine stimulated papillary muscles, and (2) had no effect on concentration dependent vasorelaxation to adenosine or R-PIA. In in vivo studies, transmural myocardial blood flow in the area at risk (determined using 15 mu m radiolabelled microspheres) was reduced by 98% in all groups from 139(SEM 15.8) to 2.7(1.1) ml.min(-1).100 g(-1) (p<0.001). At 120 min of reperfusion, blood flow in the area of necrosis was significantly less in groups LA-(1)-IR [48.6(6.2)], HA(1)-IR [36.1(7.1)], SPTIR [35.9(6.4)], and SPTR [25.1(5.4)] compared to groups VEH [69.1(15.8])1 and A(1)-R [77.2(11.8)]. The area at risk (Ar) was equivalent among groups. SPT treatment during ischaemia-reperfusion in the SPTIR group increased the area of necrosis (An, assessed by triphenyltetrazolium chloride) relative to Ar (An/Ar) to 51(1.9)% nu 26.0(1.7)% in VEH group. KW-3902 in LA(1)-IR and HA(1)-IR during both ischaemia and reperfusion increased An/Ar to 35.2(2.5)% and 35.2(2.1)% of area at risk, respectively, both of which were significantly less than the SPTIR group. With A(1) blockade at reperfusion (A(1)-R), An/Ar was equivalent to that in VEH [27.0(1.9)%], while an infarct size of 46.7(2.1)% was still observed in SPTR. Conclusions: While adenosine exerts its predominant modulation of infarct size during reperfusion, the cardioprotection mediated by A(1) receptor mechanisms is modest and exerted principally during the ischaemic time period.