It has recently been shown that a sodium background current, i(b,Na), exists in cardiac muscle cells whose effect is to depolarize the membrane so that the resting potential, V-m, is positive to the potassium equilibrium potential, E-K. In ventricular cells, where i(b,Na) is smallest, V-m, is about 10 mV positive to E-K (E-K = -87 mV at 37 degrees C). Yet, replacement of Na+ ions by large impermeant cations does not cause the expected hyperpolarization. We have studied this problem in guinea-pig myocytes using a single microelectrode recording technique in combination with a rapid external solution switch. Cells depolarized less than or equal to 0.5 mV from potentials between -80 and -73 mV and hyperpolarized up to 5 mV from potentials between -73 and -64 mV when 70 mM choline chloride or N-methyl-D-glucamine chloride were used to replace 70 mM Na+ in the bathing solution. Replacement by 70 mM lithium chloride, however, only caused hyperpolarization in very depolarized cells when the voltage change was much smaller. The changes were complete almost as soon as the solution change, i.e. within 250 ms, indicating that the actions are attributable to the external solution change rather than to secondary changes in intracellular concentrations. Patch clamp recording was used to investigate the mechanism involved. These experiments showed that the presence or absence of the inward rectifier current i(K1) determines in which direction Na+ removal acts. In the absence of i(K1) the changes are attributable to removal of i(b,Na), whereas in the presence of i(K1) the changes resemble the i(V) relation for i(K1), implying that Na+ regulates i(K1) in a way that can mask the changes in i(b,Na). These results explain why removal of Na+ does not lead to hyperpolarization in ventricular cells as would be expected if changes in i(b,Na) were solely responsible. Computer reconstruction shows that the effects may be attributed to actions of sodium removal on the conductance and gating of i(K1).