The effect of i.v. ergonovine tartrate infusions (0.05-20 micrograms/kg/min, 12 minutes duration) on coronary arteries was studied in 14 conscious dogs instrumented to continuously measure vascular diameter by an ultrasonic dimension gauge using 10-MHz piezoelectric crystals. Ergonovine induced a biphasic coronary response: small, transient dilation during the first minutes of infusion, followed by slowly developing constriction reaching its maximum 5 to 15 minutes after the end of the infusion and persisting at this level for at least 10 minutes. The threshold dosage for significant constriction was 0.05 microgram/kg/min. A dosage of 5 micrograms/kg/min (cumulative 60 micrograms/kg, corresponding to 35 micrograms/kg ergonovine maleate) caused a decline in mean left circumflex artery diameter by 137 +/- 15 micrometers (= 4.6%) without significantly altering heart rate, plasma catecholamines or plasma renin activity. Coronary venous O2 saturation did not decline, indicating the absence of coronary resistance vessel constriction. The epicardial artery constriction was not attenuated by a vasopressin antagonist. Under adrenergic blockade (2 mg/kg phentolamine and 2 mg/kg nadolol) or under ganglionic blockade (5 mg/kg pentolinium tartrate), ergonovine (5 micrograms/kg /min) caused substantial elevation in mean arterial pressure, while the decline in coronary artery diameter was attenuated. When this increase in arterial pressure was prevented by appropriate bleeding, the ergonovine-induced coronary constriction was not diminished by adrenergic or ganglionic blockade. The serotonin antagonist methysergide (0.5 mg/kg) completely abolished the ergonovine-induced coronary artery vasomotion. It is concluded that ergonovine in dogs causes an epicardial coronary artery constriction comparable to the diffuse coronary artery narrowing in men not suffering from variant angina pectoris. These constrictions are not mediated by an adrenergic mechanism.