In a previous study nitroglycerin failed to dilate coronary collateral vessels during exercise. This study tested the hypothesis that failure of nitroglycerin to increase collateral flow occurred because endogenous nitric oxide (NO) had activated the guanylate cyclase vasodilator pathway so that additional NO from nitroglycerin could have no additional effect. Six dogs were collateralized using intermittent 2-min occlusions of the left anterior descending coronary artery followed by permanent occlusion. One week after permanent coronary occlusion, dogs were exercised on a treadmill (heart rate 202 ± 5 beats/min), while blood flow was measured with radioactive microspheres. Blood flow to the collateral zone during control exercise was 1.90 ± 0.11 ml · min¹ · g¹ compared with 2.28 ± 0.15 ml · min¹ · g¹ in the normal zone (P < 0.05); systolic wall thickening was 23 ± 3% in the collateral zone compared with 27 ± 2% in the normal zone. When N^G-nitro-L-arginine (L-NNA; 20 mg/kg iv) was administered to block NO production, collateral zone flow during exercise decreased to 1.43 ± 0.20 ml · min¹ · g¹ (P < 0.05), and systolic wall thickening decreased to 12 ± 4% (P < 0.05). A subsequent infusion of nitroglycerin (2 µg · kg¹ · min¹ iv) increased collateral zone blood flow to 1.65 ± 0.16 ml · min¹ · g¹ (P < 0.05) and increased systolic wall thickening to 22 ± 5% (P < 0.05). These findings demonstrate that endogenous NO contributes to collateral zone blood flow during exercise. If endogenous NO synthesis is blocked, then nitroglycerin is effective in improving collateral zone blood flow and contractile function during exercise.