Direct myocardial depression plays a role in the cardiovascular toxicity of local anesthetic agents, but this role is obscured by concomitant cardiac, systemic, and CNS events: seizures, hypoxia, acidosis, sympathetic activation, bradycardia, and A-V heart block. Direct injection of small bolus doses of lidocaine and bupivacaine into a branch of the left coronary artery was used to minimize these systemic effects. Regional contraction in the zone supplied by the coronary artery was measured with a sonomicrometer. Both agents caused a dose-dependent reduction in the extent of systolic contraction, and a 4.9:1 (lidocaine:bupivacaine) dose ratio produced a 50% depression of contraction. The duration of depression, taken as the time for 95% recovery of systolic contraction, was about 25% (P less than 0.05) longer with bupivacaine for an equal degree of depression. Coronary blood flow was reduced modestly by both agents. These results suggest that differences in the magnitude or duration of direct myocardial depression cannot explain the clinical perception that the cardiovascular toxicity of bupivacaine is greater than that of lidocaine.