Slowing of the recovery of the Na channel from local anesthetic block at low external pH has been well described. In contrast the Ca channel has been reported not to show such an effect. Because of the absence of this response, a two-site model for the interaction of Ca antagonists with the Ca channel has been proposed. We sought to determine whether these results were a consequence of utilizing the poorly lipid-soluble Ca antagonist diltiazem as the test drug. We have measured the time constants of recovery (tau r) of Na (INa) and Ca currents (ICa) recorded from rabbit atrial myocytes during control and exposure to the lipid-soluble Ca antagonist nicardipine as external pH was reduced from 7.8 to 6.9. In the absence of drug, tau r for INa decreased from 19 +/- 1.3 to 12.1 +/- 0.6 ms for the pH reduction. Similarly, tau r for ICa decreased from 53 +/- 7 to 33.6 +/- 7 ms. During exposure to nicardipine tau r for INa increased from 2.24 +/- 0.8 to 5.96 +/- 0.8 s as external pH was reduced. For ICa, tau r also increased from 1.43 +/- 0.55 to 2 +/- 0.32 s. For each channel type, the results are well explained by a single blocking site model with hydrophobic and hydrophilic pathways to a single membrane binding site as proposed by Hille (J. Gen. Physiol. 69: 497-517, 1977).