Pentobarbital and alpha-chloralose are widely used for experimental anesthesia, but their direct electrophysiological actions at anesthetic concentrations are unknown. Trough and peak concentrations measured by high-performance liquid chromatography averaged 27 +/- 3 and 45 +/- 13 mg/l (means +/- SD) for pentobarbital and 41 +/- 15 and 103 +/- 13 mg/l for alpha-chloralose in dogs receiving them for general anesthesia. The direct effects of each agent on papillary muscle action potentials obtained from guinea pigs killed by decapitation were studied in vitro. Pentobarbital increased action potential duration to 95% by 24 +/- 6 and 33 +/- 4% at 25 and 50 mg/l (P less than 0.001 for each), respectively, and caused corresponding increases in effective refractory period. Furthermore, pentobarbital reduced maximum rate of voltage change (Vmax) of phase 0 in a voltage-, rate-, and concentration-dependent fashion, suggesting use-dependent sodium channel blocking actions. The voltage dependence of Vmax was shifted by 3.7 +/- 1.7 (P less than 0.01) and 6.5 +/- 1.8 mV (P less than 0.001) by 25 and 50 mg/l pentobarbital, respectively. In canine ventricular muscle, pentobarbital caused rate- and concentration-dependent decreases in Vmax and increases in action potential duration and refractory period over a concentration range of 5-100 mg/l. alpha-Chloralose was devoid of direct electrophysiological effects in both species. We conclude that pentobarbital has potentially important electrophysiological actions on ventricular tissues at concentrations required for general anesthesia and may confound the results of in vivo studies.