It has recently been suggested that the inactivation of calcium currents in molluscan neurones may depend on calcium entry, rather than membrane potential as such (Tillotson, 1979). We have studied calcium current in identified, voltageclamped neurones of the snail bathed in sodium-free saline containing 25 mm CaCl2, 10 KCl, 5 MgCl2, 5 Tris-HCl, with 75 mm TEA-Cl and 2 mM-4-aminopyridine to reduce potassium currents as much as possible.Under these conditions inward calcium currents turn offexponentially over the first 100 msec at potentials up to + 20 mV. At these potentials the currents do not inactivate completely, even during pulses up to 1 see in duration. With potentials more positive than + 20 mV the current becomes outward at the end of a 400-msec pulse. Cadmium ions (1-2 mM) completely abolished the inward current and, for potentials above +30 mV, also reduced the net outward current at the end of a 400-msec pulse, suggesting that some of this outward current was calcium-activated.The rate of decline of the calcium current was dependent on its size. When the current was reduced by reducing [Ca]o from 25 to 5 mm the decline was slowed. (For a pulse to 0 mV the time constant changed from 46 to 88 msec.) The inactivation was also studied in two-pulse experiments. A 250 msec pre-pulse of variable voltage was followed after a 100-msec interval by a test pulse to 0 mV. The inward current during the test pulse was reduced as the pre-pulse became more positive in the range -30 to + 30 mV, but then increased with further depolarization, approaching the value for no pre-pulse when the pre-pulse was + 110 to + 120 mV.This result may be interpreted in terms of calcium entry causing calcium inactivation, or alternatively in terms of the calcium entry leading to a Ca-activated potassium current instantaneously activated by depolarization during the test pulse. Such Ca-activated K currents have been shown to be voltage-dependent (Gorman & Hermann, 1979), and it might therefore be expected that such an effect would increase with increasing test pulse potential. This second explanation is less likely as the relationship between test pulse current relative to that with no pre-pulse