To determine the effects of an exercise-training program on coronary collateral development, we instrumented 13 1-yr-old beagles with left circumflex (LCf) coronary artery flow probes and balloon occluders and left atrial and aortic catheters. The LCf artery was constricted resulting in a 58 +/- 4% reduction of the peak reactive hyperemia response following release of a 154-s LCf occlusion. All dogs were studied during the first week of the study protocol. Resting heart rate, cardiac output, and left atrial and aortic pressures were evaluated before and during a 1-min LCf occlusion. Myocardial blood flow was also measured with radioactively-labeled microspheres injected into the left atrium during the LCf occlusion. Subsequently the dogs were exercised at 6.4 km/h and 12% grade, and all hemodynamic and blood flow measurements were repeated. The animals were then randomized to either a sedentary or exercising group. The six sedentary animals were confined to their cages, while the seven training beagles did sprint and endurance running for 75 min/day, 5 days/wk. After 12 wk the hemodynamic and blood flow studies were repeated at rest and during treadmill running. There were no statistically significant differences in resting or exercise hemodynamics, response to LCf occlusion, or myocardial blood flow in the two groups in wk 1. Hemodynamics and blood flow data were virtually unchanged in the sedentary animals after 12 wk. However, the trained dogs demonstrated less evidence of left ventricular failure following LCf occlusion while running and had significantly higher myocardial blood flows. Their resting collateral flow increased from 0.52 +/- 0.16 ml . min-1 . g-1 at wk 1 to 0.90 +/- 0.28 ml . min-1 . g-1 at wk 12 (P less than 0.05), while the ratio of collateral to normal myocardial blood flow increased from 0.46 +/- 0.12 to 0.64 +/- 0.16 (P less than 0.05). Thus chronic exercise can stimulate coronary collateral development, and the enhanced collateral flow has a salutary functional effect.