We studied the relation of the maximal upstroke velocity (Vmax) of action potentials to the peak sodium current (INa) under voltage clamp in single, internally perfused, canine cardiac Purkinje cells under conditions that ensured membrane action potentials due only to INa. Three different methods of altering sodium channel availability were investigated: voltage-dependent inactivation, tetrodotoxin (TTX) block, and use-dependent block by quinidine. Under all three conditions, the relation of Vmax to INa was nonlinear, and no relation was found that would allow prediction of INa results from Vmax measurements. With voltage-dependent inactivation or TTX block, sodium channel availability measured by Vmax was reduced less than availability measured by peak INa, so that Vmax overestimated sodium channel availability. This overestimation of sodium channel availability by Vmax could be attributed to greater sodium channel mobilization during the slowed action potential upstrokes. The overestimation varied with experimental temperature as a consequence of changes in sodium channel kinetics. Vmax also overestimated sodium channel availability during TTX exposure so that the Kd for TTX block was 4.5 micron from Vmax measurements but only 1.6 microM from INa measurements. Use-dependent block of INa by quinidine had a striking voltage-dependent component under voltage clamp that could not be appreciated from action potentials. Consequently, block could be underestimated or overestimated by Vmax measurements. We conclude that Vmax measurements represent a convenient index for INa, but Vmax is not a reliable method for quantitative studies of sodium channel behavior.