Time-Lao in the Spark Discharge. 227 



potentials, have to be fulfilled. The auxiliary gap will be 

 more effective the greater is its sparking potential and the 

 greater is the ratio of the capacity in parallel with both gaps 

 in series To the capacity in parallel with the leaky gap. 

 These relations were investigated and found to be in 

 accordance with the obvious theory of the action of such 

 an auxiliary gap. When the spark passes across the auxiliary 

 gap, its equivalent resistance becomes small compared with 

 that of the leak, and the whole potential to which the con- 

 denser in parallel with both gaps has been charged by the 

 magneto is thrown on the leaky gap. If the capacity of this 

 condenser is sufficiently large to maintain this potential while 

 the smaller capacity in parallel with the leaky gap is charged 

 up, a spark will pass across the leaky gap, so long as its spark 

 potential is less than that given by the magneto. 



In the experiments on the hard gaps the sparking distance 

 of the main gap was always set at the same value, so that the 

 sparking potential for steady potential was the same ; for this 

 potential depends only on the geometrical form of the 

 electrodes and not on the material*. The capacities in the 

 circuit were also always unchanged ; but the sparking dis- 

 tance of the auxiliary gap could be changed. A steady 

 source of potential was used in place of a magneto to cause 

 the spark to pass the auxiliary gap, and the length of this 

 gap was increased until this spark was accompanied by a 

 spark across the main shunted gap. This length was thus 

 a measure of the sparking potential of the main gap when 

 the potential was applied for the very short time (probably 

 less than a millionth of a second) between the start of the 

 spark in the auxiliary gap and the dissipation of the charge 

 conveyed by it through the shunt. Accordingly the harder 

 the main gap, the greater was the necessary sparking distance 

 of the auxiliary gap. The method was extremely sensitive : 

 a degree of hardness barely perceptible in the difference 

 between the steady sparking potential and that with a 

 magneto would increase the limiting value of the auxiliary 

 gap by 50 percent. Hardness, therefore, like the difference 

 between point and sphere gaps, if measured by impulse ratio, 

 increases greatly with the frequency. 



By this method the conclusions already established were 

 confirmed and it was shown that hardness (measured again 

 by impulse ratio) decreased, instead of increasing, it' the 

 spark potential was increased either by increase of pressure 



* The sparking potential is probably slightly smaller for aluminium 

 electrodes than for those of less easily oxidizable metals, but such differ- 

 ences were too small to be appreciable here. 



