1. The mechanisms by which cooling inhibits insulin secretion were investigated by capacitance measurements of exocytosis in single mouse pancreatic B-cells maintained in short-term tissue culture. 2. A reduction of the bath temperature from 34 to 24 degrees C produced a gradual inhibition of exocytosis. Inhibition of exocytosis was use dependent rather than time dependent. The steady-state inhibition amounted to 90%, which was paralleled by a 30% reduction of the peak Ca2+ current. 3. The Q10 values (between 27 and 37 degrees C) for inhibition of exocytosis and the peak Ca2+ current amplitude were determined as > 5 and 1.6, respectively. From the temperature dependence of exocytosis, an energy of activation was estimated as 145 kJ mol-1. 4. Suppression of exocytosis was not the result of a reduction of Ca2+ influx. When the Ca2+ currents were blocked by 30% (comparable to that produced by cooling) by using a low concentration of Co2+, exocytosis was reduced by < 25%. 5. Elevation of cytoplasmic free Ca2+, by photorelease of 'caged' Ca2+ from Ca(2+)-nitrophenyl-EGTA preloaded into the B-cell, was equally effective at eliciting exocytosis at 24 and 34 degrees C. 6. Cooling produced 70% inhibition of exocytosis evoked by infusion of Ca2+ through the recording electrode. Omission of either MgATP or cAMP from the electrode solution resulted in a comparable reduction of exocytosis. Cooling had no additional inhibitory effect when exocytosis was already suppressed by removal of cytoplasmic MgATP. 7. Our data indicate that exocytosis of granules already docked beneath the membrane is little affected by cooling in the B-cell. Instead, the high overall temperature sensitivity of insulin secretion arises because the replenishment of the readily releasable pool is temperature dependent.