The paper is concerned with assessing the suitability of spark-gaps for the measurement of the amplitude of short-duration recurrent impulses. Measurements are described which were made on a range of gaps between 2-cm diameter spheres, using recurrent impulses of from 4 to about 1/10 microsec duration, at recurrence rates from 3 000 to 100 per second.The breakdown voltage of the gaps tested was found to be independent of both impulse duration and recurrence rate within the limits of experimental error, and the calibration data obtained agree with the corresponding data given in B.S. 358 : 1939 for gap lengths greater than 0.2 cm. In obtaining these results it was necessary to irradiate the gap by means of 0.2 mg of radium.Measurements were also made on a range of gaps between parallel-plate electrodes with 1.0 microsec impulses repeated 400 times per second. For gaps greater than 0.2 cm in length, irradiation by 0.2 mg of radium produced no effect on the breakdown voltage, and did not appear to be essential for the satisfactory use of such gaps.