The intrinsic cardiotoxicities of bupivacaine and lidocaine were examined in the isolated, perfused rat heart. The perfusates contained no protein and were equilibrated with a gas mixture of 95 per cent O2 and 5 per cent CO2. Autonomic activity, competitive binding, and postseizure hypoxia and acidosis were absent in this experimental model. The effects of the two local anesthetics were evaluated at normokalemia (5.9 mEq/l) and hyperkalemia (9.0 mEq/l). For normokalemia, the ratio of the potency of bupivacaine to that of lidocaine was 14 for slowing ventricular rate to 50 per cent of control, 6 for slowing atrial rate to 50 per cent of control, and 17 for doubling of the PR interval. The action of bupivacaine to slow ventricular rate was due to an inhibitory effect on both AV conduction and atrial rate. For lidocaine, ventricular slowing was mediated mainly by an inhibition of atrial rate with decreased AV conduction playing a minor role. Hyperkalemia of 9.0 mEq/l had little effect on heart rate or AV conduction in the absence of bupivacaine or lidocaine. It did, however, greatly potentiate the effect of both local anesthetics to slow ventricular rate. For bupivacaine, ventricular slowing to 50 per cent of control during hyperkalemia was accomplished almost entirely via an inhibition of AV conduction, while for lidocaine it occurred because of inhibition of both AV conduction and atrial rate. Regardless of the mechanism, hyperkalemia of this degree increased the ventricular slowing effect of both bupivacaine and lidocaine.