SUMMARY In 10 dogs, we placed a Doppler flow probe, an ameroid constrictor, and a hydraulic cuff occluder around the left circumflex coronary artery (CA). Ultrasonic dimension gauges were implanted in control segments (CS) and ischemic segments (IS) of the left ventricle (LV). A microtransducer was used to measure LV pressure (LVP). Sixteen to 35 days later, when CA occlusion was nearly complete and collaterals had developed, the dogs were studied by telemetry during exercise, while running behind a van (8-11 mph for 3.5 minutes). Control exercise and exercise after propranolol (P) (average dose 0.5 mg/kg, iv) were conducted on the same day 2 hours apart. During control exercise, significant increases occurred in heart rate (HR) (116-249 beats/min), LVP (124-163 mm Hg), end-diastolic pressure (EDP) (8.1-25.9 mm Hg), peak (+)dP/dt (3618-7348 mm Hg/sec), and percent shortening (%AL) of CS (19.7-28.8) (all P < 0.01), but percent systolic wall thickening (%AW) and %AL of IS decreased greatly [22.2 ± 3.8 to 5.1 ± 1.3% (P< 0.01) and 12.6 ± 2.1 to 7.1 ± 2.4% (P< 0.01), respectively]. During matched exercise periods after P, significantly less marked increases occurred in HR (101-182 beats /min), LVP (123-139 mm Hg), EDP (9.4-20.9 mm Hg), peak (+)dP/dt (2801-4312 mm Hg/sec), and %AL of CS (17.1-22.4%) (all P < 0.01). However, %AW and %AL of IS did not deteriorate [19.0 ± 3.6 to 14.3 ± 3.2% (NS) and 11.3 ± 3.1 to 11.5 ± 3.9% (NS)]; both were significantly different from control exercise (P < 0.01). It is concluded that amelioration by propranolol of an exercise-induced myocardial oxygen demand-supply imbalance in a collateral-dependent zone can markedly reduce regional ischemia and dysfunction at the same exercise level.