We used microelectrodes to monitor the recovery (i.e., decrease) of intracellular pH (pHi) after using internal dialysis to load squid giant axons with alkali to pHi values of 7.7, 8.0, or 8.3. The dialysis fluid (DF) contained 400 mM K+ but was free of Na+ and Cl-. The artificial seawater (ASW) lacked Na+, K+, and Cl-, thereby eliminating effects of known acid-base transporters on pHi. Under these conditions, halting dialysis unmasked a slow pHi decrease caused at least in part by acid-base transport we refer to as "base efflux." Replacing K+ in the DF with either NMDG+ or TEA+ significantly reduced base efflux and made membrane voltage (Vm) more positive. Base efflux in K(+)-dialyzed axons was stimulated by decreasing the pH of the ASW (pHo) from 8 to 7, implicating transport of acid or base. Although postdialysis acidifications also occurred in axons in which we replaced the K+ in the DF with Li+, Na+, Rb+, or Cs+, only with Rb+ was base efflux stimulated by low pHo. Thus, the base effluxes supported by K+ and Rb+ appear to be unrelated mechanistically to those observed with Li+, Na+, or Cs+. The combination of 437 mM K+ and 12 mM HCO3- in the ASW, which eliminates the gradient favoring a hypothetical K+/HCO3- efflux, blocked pHi recovery in K(+)-dialyzed axons. However, the pHi recovery was not blocked by the combination of 437 mM Na+, veratridine, and CO2/HCO3- in the ASW, a treatment that inverts electrochemical gradients for H+ and HCO3- and would favor passive H+ and HCO3- fluxes that would have alkalinized the axon. Similarly, the recovery was not blocked by K+ alone or HCO3- alone in the ASW, nor was it inhibited by the K-H pump blocker Sch28080 nor by the Na-H exchange inhibitors amiloride and hexamethyleneamiloride. Our data suggest that a major component of base efflux in alkali-loaded axons cannot be explained by metabolism, a H+ or HCO3- conductance, or by a K-H exchanger. However, this component could be mediated by a novel K/HCO3- cotransporter.