This investigation tested the hypothesis that the effects of nitric oxide synthase on myocardial capillary perfusion were reduced during myocardial stunning. Anaesthetized open-chest rabbits were assigned to either a control group or a group treated with a nitric oxide synthase inhibitor. To induce myocardial stunning, a coronary artery was occluded (for 15 min) and then reperfused (for 15 min) twice. During reperfusion, rabbits were given either saline or N G-nitro-L-arginine methyl ester (L-NAME, 30 mg kg-1) followed by I.V. injection of 150 mg kg-1 fluorescein isothiocyanate (FITC)-labelled dextran (molecular weight, 150 000) for 14 s. Fluorescence microscopy was used to identify the perfused vessels and an alkaline phosphatase stain was used to locate the total microvasculature. The 'closest-individual' method was used to estimate the geometric distribution of capillaries. No significant differences were observed in the total volume fraction (mm3 of capillaries per mm3 of tissue) between the control and stunned regions in either the saline- (0.045 ± 0.008 and 0.042 ± 0.009, respectively) or the L-NAME-treated hearts (0.060 ± 0.010 and 0.049 ± 0.005, respectively). There were no significant differences in the percentage volume fraction of perfused capillaries between the control and the stunned regions (49 ± 4 % and 54 ± 4 %, respectively) in saline-treated hearts. In hearts treated with L-NAME, the percentage of perfused capillaries was significantly reduced. The reduction was significantly greater in the control region (~27 %) than the stunned region (~17 %). Closest-individual analysis of the perfused capillary distribution in both groups demonstrated a similar unchanged distribution. Thus, nitric oxide synthase is an important regulator of basal coronary capillary perfusion, and its effects are significantly reduced by myocardial stunning. Experimental Physiology (2002) 87.3, 335-342.