128 Mr. E. L. Nichols on the Alternating 



between copper poles (approximately 112,000 ohms), on the 

 one hand, and the parallel circuit between the ball and point 

 on the other. The resistance of the latter path was infinite 



Fig. 2. 



/ 

 / 



\ 

 \ 



-. / \ 



V 



\l 



A- 



i/*X \ \ 



/ /y v * \\ 



;f //'to i _v-\ 



/ n la 13 \tj is 10 17 la \ ; \?o 



II 





12^4 S 07// 



/ 

 I 



/ 



J* * S 7/," 9! / 11 12 13 \V» 15 10 17 lfi ° v • 



•/ // i ■ \ i - 



ii \ y 



v V/ 1 / 



whenever the arc was interrupted, falling to finite values 

 during each discharge. Increase of current through 11 

 indicates, therefore, the formation of the arc. Such increase 

 is found to exist during the second half of each cycle, that is 

 to say, during that interval in which the ball is positive ; and 

 it might be inferred from these curves alone that the dis- 

 charge was an intermittent one taking place always from ball 

 to point. Other curves, taken simultaneously with A and B, 



the Thomson galvanometer being shunted around the indicator 

 © © 



galvanometer, lead to the same conclusion. The curves 

 marked C and D (fig. 2) show the results obtained. It was 

 thought that they would give the fluctuations in electromotive 

 force between ball and point, corresponding in time to the 

 current fluctuations in the resistance R. The indicating 

 galvanometer, however, owing to the very rapid alternations 

 to which the circuit was subjected, was found to possess such 

 high self-induction as to materially influence the result. 

 Strictly speaking, the curves C and I), therefore, give the 

 periodic changes of electromotive force at the terminals of the 

 indicator and not those occurring at the ball and point. 

 These curves are nevertheless of considerable interest. The 



