290 BELL SYSTEM TECHNICAL JOURNAL 



which had a resistance of about 3.5 milliohms. The inductance L consisted 

 of two coils connected in series, each containing a magnetic link. The 

 larger of these coils had 187 turns of copper wire, the smaller 50 turns. 

 The larger coil provided the greater sensitivity, on account of the more 

 intense magnetization of the link. The relation between the current / 

 in the coil and the deflection on the magnetic link meter^ used to measure 

 the intensity of magnetization was obtained by calibration with direct 

 current. 



The inductance of the two coils in series was about 700 microhenries and 

 the d-c resistance about .39 ohms. The time constant of the coils L/R 

 was thus about 1800 microseconds, which is large compared to the duration 

 of the main surge of a lightning stroke, which may last for 100 to 500 mi- 

 croseconds. The instrument will thus effectively integrate the main surge, 

 but will not record the charge which may be caused by a small tail current 

 of much longer duration. 



The measurements of charge were made mainly to determine the time 

 to half-value of the currents. The theoretical curves in Part I and else- 

 where in this paper are based on a current of the form: 



J{t) = 1.157 (e~°' -r'') 



where a = .013 • 10*, 6 = .5 • 10^ for a current reaching its crest value / 

 in about 5 microseconds and decaying to its half-value in 65 microseconds. 

 If a = R/L = .00056-10® for the surge integrator, the total charge re- 

 corded for a current of the above wave shape is 



Q = /1. 15 r^— - r^ = /•83-10-« 



for a current decaying to its half-value in 65 microseconds. The relation- 

 ship between Q/J and the time to half-value, /1/2, is as follows for currents 

 reaching their half- values in 65, 130, 260, and 520 microseconds: 



Q/J: 83 160 295 540 microseconds 



/1/2 : 65 130 260 520 microseconds 



From a curve of Q/J versus /1/2, the time to half-value may be obtained 

 from the observed ratio of charge to crest current. The values given later 

 on, in Table I, were obtained in this manner. If a triangular wave shape 

 had been assumed, the times to half-value would have been Q/J and there- 

 fore somewhat longer. 



2.3 Measurements oj Voltage Between Sheath and Copper Shield 



These measurements were made by means of a magnetic link voltmeter 

 consisting of a solenoid of inductance L containing the magnetic link and a 



