A two-color optical procedure was used to measure temperatures during combustion in the prechamber of an automotive-type divided-chamber diesel engine. The prechamber design was cylindrical, instead of spherical, with quartz windows at the ends to provide optical access. The temperature was measured for engine conditions in which the combustion timing, overall air-fuel ratio, engine speed, intake-manifold pressure and intake-air temperature were varied. At each condition, the temperature was measured along a single line of sight coincident with the axis of the prechamber and was measured as a function of engine crank angle in at least 92 engine cycles. The ensemble-averaged temperature was determined from these data and reported in this study.The variations in the maximum measured temperatures with changes in the operating conditions were explained by changes in the composition, temperature and pressure of the reacting species. The degree of mixing was a key parameter which affected the properties of the reacting species. The maximum measured temperatures were also compared with the stoichiometric adiabatic flame temperatures calculated assuming either a zero or an infinite mixing rate between previously produced combustion products and the air instantaneously involved in combustion. Better agreement between measured and calculated temperatures was obtained when an infinite mixing rate was assumed.