The state-dependent sodium channel block by mexiletine, tocainide, lidocaine, OPC-88117, aprindine, quinidine, disopyramide and AN-132 was investigated in single ventricular myocytes isolated from guinea-pig hearts. A single conditioning clamp pulse with a duration from 10 ms to 1000 ms was applied from the resting potential (-82 mV) to 0 mV level using a suction pipette whole-cell voltage clamp technique, and the maximum upstroke velocity (Vmax) of a test action potential elicited 100 ms after termination of the clamp pulse was measured as an index of sodium channel availability. In myocytes treated with the eight drugs, such clamp pulse caused a significant decrease in Vmax. With mexiletine, tocainide, lidocaine, OPC-88117 and aprindine, the Vmax reduction was enhanced progressively as the clamp pulse duration was prolonged. With quinidine, disopyramide and AN-132, an appreciable Vmax reduction at the shortest clamp pulse was followed by an additional small enhancement of the Vmax decay. These findings suggest that the former group of drugs may block the sodium channel mainly during the inactivated state (inactivation blockers), whereas the latter one may do so mainly during the activated state (activation blockers). Multiple short clamp pulses caused a greater Vmax reduction than a single prolonged clamp pulse for the activation blockers, and vice-versa for the inactivation blockers. Molecular dimensions of the eight drugs, which were estimated by X-ray diffraction of crystals, did not satisfy a simple size criterion as proposed by Courtney (1988) to explain such different types of sodium channel block by Class-I drugs.