Time-Lag in the Spark Discharge. 221 



and different in origin, there is certainly a connexion between 

 them. For it was found that in a sphere gap, as indicated 

 by Peek's experiments, there is in general no regular lag, 

 and the limiting peak potential of a magneto which will 

 sometimes cause a spark is the same .as the steady sparking 

 potential. (Some exceptions to this statement will be noted 

 later.) And when the limiting peak potential and the steady 

 spark potential are thus identical, there is also little or no 

 irregular lag; the sparking potential with the alternating- 

 source is definite, and the limiting potential which will 

 sometimes produce a spark is also the peak potential which 

 will produce regular sparking. In such gaps if a spark 

 sometimes passes, it always passes, and no effect is produced 

 on the sparking potential, either with steady or alternating- 

 sources, by increasing the initial ionization of the gap. The 

 two kinds of lag appear together and vanish together, when 

 the change is made from point to sphere electrodes. 



The actual sparking potential. 



Some curious results were obtained in experiments directed 

 to determine what is the potential between the electrodes 

 when the spark actually passes. If there is a time-lag, 

 whether regular or irregular, this potential will not in 

 general be equal either to the steady sparking potential or 

 to the peak potential of the alternating source. It can hardly 

 be less than the former, but it will usually be less than the 

 latter. If the time-lag is regular, the actual sparking 

 potential will be constant so long as the wave-form is 

 constant ; but if the time-lag is irregular and due to the 

 casual ionization, the potential when the spark actually 

 passes may presumably have any value greater than the 

 steady sparking potential ; it will simply be the value which 

 happens to obtain when the casual ions are present at the 

 right place*. Accordingly a determination of this actual 

 sparking potential should indicate whether the lag is regular 

 or irregular. 



The determination is made possible and easy by the results 

 described in a recent paper by Mr. C. 0. Paterson and 

 myselff- It is shown there that if a spark gap is connected 



* The i; right place 1 ' will depend to some extent on the potential 

 prevailing. For the ion has to be in a field of given intensity in order 

 to start the discharge ; and the region in which such a. field exists will 

 be the larger the greater the potential difference between the electrodes. 

 But this consideration does not alter the conclusion that the actual 

 sparking potential should be determined by chance. 



t " Some Characteristics of the Spark Discharge and its effect in 

 igniting explosive mixtures," Phys. Soc. Proc. March 14, 1919, (The 

 paper is not published yet.) 



