1. Sodium currents were studied using whole-cell voltage-clamp techniques in neurons acutely isolated from superficial (II/III) and deep (V/VI) layers of guinea pig entorhinal cortex. 2. Sodium currents were larger (peak amplitude) in superficial than in deep layer cells under the same conditions: -1939 +/- 780 (SD) pA (N = 6) versus -307 +/- 257 pA (N = 6). Specific membrane conductance was calculated to be 12.3 +/- 9.6 mS/cm2 for superficial layer cells and 1.4 +/- 0.9 mS/cm2 for deep layer cells. 3. Sodium currents could be activated in superficial layer cells from potentials as depolarized as -20 mV, whereas no significant currents could be activated in deep neurons from potentials more depolarized than about -50 mV. Using a protocol consisting of a 25-ms prepulse and a 20 ms test pulse, the inactivation curves for superficial layer cells were found to be shifted toward more depolarized potentials by an average of 15 mV (V50 = -59.8 +/- 3.8 mV compared with -75.7 +/- 12.0 mV for deep cells). This produced a region of overlap with the activation curves for superficial cells. 4. Over a range of about -50 to -20 mV in superficial layer cells, the region of overlap of the activation and inactivation curves, a sodium current could be activated, which did not fully inactivate during the test pulse (average peak amplitude: -89.5 +/- 48.7 pA; crossover voltage: -39.2 +/- 2.0 mV). Voltage steps to more depolarized potentials, outside the voltage "window", permitted complete inactivation of the sodium current.(ABSTRACT TRUNCATED AT 250 WORDS)