Air-Blast upon the Spark Discharge of a Condenser. 681 



curve is not symmetrical with respect to its maximum 

 ordinate, the best mode of procedure is as follows : — Having 

 taken a series of values of the cymometer current a without 

 the extra damping resistance, and those of the current % 

 when the resistance is inserted in the cymometer circuit, 

 and observed the maximum values A and A x of the quantities, 

 for various values of n we plot two curves, each to abscissae 

 m/N, but one having ordinates a/A and the other ordinates 

 tfi/Ax. The latter curve should lie outside the former, but 

 should have the same value for its maximum ordinate, viz., 

 unity. We then draw a number of horizontal lines across 

 the hump of the resonance curves, as shown in fig. 1, with 



C/7P / 777.777.. 



G/1P 2 77277V 



C/^P 5 777772,. 



§0- 







/ 



\ 







1-0 



0-9 

 0-8 



07 

 06 



£0-4 

 ^ 0-3 



|o-i 



















14) 



0-9 



OS 



0-7 



06 



£0-5 



*04 

 «: 



kO-3 



!° ,? 













\ 



































































































































































/ 









































































































































































■97 -98 -99 1-00 1-01 1-0? 1-03 

 Frequency /f/ir/o y/v 



•97 



97 -98 -99 1-00 1-01 1-02 1-03 



fffEQUEA/Cr ft/ir/o n /^ 



Fig-. 1. — Resonance Curves taken with air-blast on spark-gap. 



•99 1-00 1-01 1-0? 1-03 

 fffEQUE/vcr ff/ir/o y/y 



such ordinates that the greatest length intercepted by the 

 curves is not more than O06 on the same scale that N is taken 

 as unity. Let the half length of any such horizontal inter- 

 cept be called x and the corresponding ordinate a/A be 

 denoted by y, we then calculate the value of 



D = 7TX 



\ 





and this gives us the value of the sum of the decrements of 

 the spark and cymometer circuits. It is well to calulate D 

 from several, say four or five, measurements of x and y for 

 intercepts of different lengths, and then take the mean value 

 for D. In the same way from the outer resonance curve we 

 can calculate the value of Dj. The difference between D and 

 Dj is the increment in the decrement due to the added 

 resistance wire, and if II is its high frequency resistance and 



