It has been suggested by Kreuger that the point-hemisphere discharge gap should be accepted as a source of known partial discharges in electrical insulation studies. Experiments to evaluate the source are described in which the voltage, point radius, frequency and gap separation were varied. The effect of prolonged use and initial conditioning was also investigated. Results indicate that the discharges were sensibly constant over the frequency range 50-1250 c/s and for voltages less than 1.4 times the inception value. It is concluded that the gap is a satisfactory and flexible source and can be regarded as a secondary standard provided it is used under prescribed conditions.