1. A voltage clamp method utilizing a sucrose gap and glass microelectrodes was developed and used to study dog ventricular myocardial fibre bundles. The limitations and the reliability of this method are demonstrated by a series of tests.2. A dynamic sodium current, excited at membrane potentials more positive than -65 mV, was measured. The equilibrium potential for this large, rapid inward current depends directly on [Na](o), shifting 29.0 +/- 2.3 mV (+/- S.E. of mean), as opposed to a theoretically expected value of 30.6 mV, when [Na](o) is reduced to 31% of normal.3. Sodium current is inactivated by conditioning depolarizations. Complete inactivation occurs with conditioning potentials more positive than -45 mV, and 50% inactivation occurs at about -55 mV. The location of the inactivation curve shifts along the voltage axis, when [Ca](o) is varied between 0.2 and 7.2 mM.4. A second, much smaller and slower net inward current, with a threshold around -30 mV, and an equilibrium potential above +40 mV was also observed.5. The ;steady-state' current-voltage relationship (after 300-600 msec) exhibits inward-going (anomalous) rectification with negative slope between -50 and -25 mV.6. A small, very slowly developing component of outward current was observed at inside positive potentials. The equilibrium potential for this current, although slightly dependent on [K](o), is neither identical with the potassium equilibrium potential nor with the resting potential in normal Tyrode solution.7. Anatomical limitations, primarily resistance in the extracellular space within the bundle, prevent complete characterization of the rapid, large sodium current, but do not limit the application of the clamp method to the study of other, smaller and slower currents. The evidence for this is discussed extensively in the Appendix.