Gear bulk temperature is the steady temperature of the gear teeth before they enter the contact zone. Generally, it can be appreciably higher than lubricant inlet temperature supplied to the gear mesh. Bulk temperature is one of the main parameters to consider in order to perform a correct prediction of scuffing failures. An experimental method for evaluating bulk temperature of spur gears is described. A power-circulating test rig, able to replicate the loading and environment conditions of a gear pair in a gearbox, is used (speed up to 18000rpm, torque up to 500Nm). In order to measure bulk temperature, a thermocouple is fixed on the middle of the end face of a gear tooth and the measure point is located under the face surface. Signal conditioning is made on board by electronic devices purposely developed and rotating with the gear shaft. Data acquisition is performed on-line while the test rig is running. A comparison between theoretical predictions according to the model proposed by Loewenthal and Anderson and experimental results is presented. A preliminary tuning on the model is suggested, in order to fit experimental data.