1. The roles of adenosine 5'-triphosphate (ATP)-sensitive potassium channels (KATP) and endogenous adenosine in the regulation of coronary flow have been assessed in the isolated, buffer-perfused heart of the rat. 2. In the presence of glibenclamide 10 microM there was a significant (P < 0.001) reduction in coronary flow from a baseline value of 8.78 +/- 0.76 ml min-1 g-1 to 3.89 +/- 0.59 ml min-1 g-1. This change was accompanied by a significant (P < 0.01) reduction in cardiac mechanical performance as shown by the decrease in the pressure-rate product from 21,487 +/- 2,577 mmHg min-1 to 6,950 +/- 1,104 mmHg min-1. 3. The non-selective adenosine antagonist 8-phenyltheophylline (10 microM) also caused a significant (P < 0.001) reduction in coronary flow from a basal value of 10.4 +/- 0.6 ml min-1 g-1 to 6.32 +/- 0.60 ml min-1 g-1. The subsequent addition of glibenclamide, in the presence of 8-phenyltheophylline, brought about a further significant (P < 0.001) reduction in coronary flow to 3.05 +/- 0.55 ml min-1 g-1 and this value was similar to that in the presence of glibenclamide alone. 4. In hearts perfused under constant flow conditions, exogenous adenosine caused dose-related reductions in coronary perfusion pressure described by a maximum reduction in pressure of 30.7 +/- 3.9 mmHg and an ED50 of 977 +/- 813 pmol. Addition of glibenclamide caused a significant (P < 0.01) increase in coronary perfusion pressure of 44.7 +/- 7.2 mmHg and a significant (P < 0.05) rightward shift of the dose-response curve for the depressor effects of adenosine (ED50 = 13.5 +/- 3.8 nmol), with a depression (P < 0.05) of the maximum (16.3 +/- 2.4 mmHg). 5. In conclusion, both KATP and endogenous adenosine make major contributions towards coronary vascular tone and the regulation of coronary flow in the rat isolated heart. Furthermore, in the coronary vasculature a significant proportion of the vasodilator action of adenosine is mediated through the activation of KATP.