1. Ca2+ currents were recorded using the whole-cell mode of the patch-clamp technique from mouse pancreatic B-cells kept in culture for 1-4 days. B-cells were identified in the cell-attached mode by their response to a change in the glucose concentration from 3 to 15 or 20 mM or by their inward currents. 2. Only one component of Ca2+ current was observed in these cells, which activated at potentials greater than -50 mV and was blocked by nitrendipine (5 microM), and increased in amplitude by CGP 28392 (5 microM). 3. During maintained depolarizations the Ca2+ current inactivated considerably but not completely. Inactivation was most marked at potentials where the Ca2+ currents were large, but in general was slower for currents at potentials greater than 0 mV than at more negative potentials. 4. Two-pulse experiments showed that the inactivation curve for the Ca2+ current was U-shaped, returning to unity at potentials approaching the Ca2+ equilibrium potential. Measurements of Ca2+ entry showed that inactivation was dependent on the amount of Ca2+ entering during the pre-pulse, independent of the pre-pulse potential. 5. Ca2+ currents were not appreciably slowed when BAPTA, a faster buffer of Ca2+, replaced EGTA in the pipette solution. 6. Replacement of Ca2+ in the external solution by Ba2+ increased the amplitude of the inward current and largely abolished inactivation. Large inward currents through Ca2+ channels were observed in the absence of divalent cations in the external solution (+EGTA), which were presumably carried by Na+. These currents did not inactivate during 150 ms depolarizations, but were increased in amplitude by CGP 28392 (5 microM) and blocked by D600 (30 microM). 7. The observations suggest that normal mouse pancreatic B-cells have only one type of Ca2+ channel which is dihydropyridine sensitive and inactivates by a mechanism which is almost purely Ca2+ dependent. Inactivation of the Ca2+ current will probably be important in the control of Ca2+ entry during glucose-induced electrical activity.