﻿WO Mr. E. H. Kennarcl on 



cylinder. Now while the magnet is rotating let the electro- 

 meter be disconnected from earth and let the magnetization 

 be then reversed. N and <£ then change sign, so that part 

 of the electrification on the outer cylinder will be driven on 

 to the electrometer and a deflexion will result. 



In practice the electrometer was left insulated while the 

 position of the spot was read every 60 seconds and the 

 direction of magnetization reversed immediately after each 

 reading. By rotating the magnet first in one direction and 

 then in the other, which changes the sign of the effect due 

 to electromagnetic induction, and then combining the de- 

 flexions obtained in the usual manner, the effect of dis- 

 turbances was largely eliminated from the final result. 



The only results obtained with the apparatus entirely in 

 its final form consist of 3 series aggregating 70 observations. 

 The greatest deflexion during any period of 60 seconds was 

 + 11 divisions of the scale, while the mean numerical 

 deflexion was less than 4 divisions for each series. For the 

 effect apparently due to electromagnetic induction when the 

 magnetization was reversed, these observations yield the 

 values -f-0'65 + 0'34 (probable error) and — 0*7 + 0'35; mean, 

 O'O + 0*24. The value zero is thus a very probable value for 

 the true deflexion. 



This conclusion is supported by all the inferior series of 

 observations, which show much larger irregular disturbances. 

 That this negative result was not due to defective insulation 

 is evident from the fact that the spot always drifted slowly 

 across the scale when the electrometer was not earthed ; in 

 addition, the insulation was always carefully tested at the 

 start, and on several occasions artificial deflexions were 

 produced while the magnet was in rotation, either by altering 

 the electrostatic potential of the inner cylinder or by 

 charging up the outer cylinder and then connecting it to 

 the electrometer. 



In order to determine the magnitude of the effect to be 

 expected according to the Moving Force Line Theory, N 

 was determined with a ballistic galvanometer at various 

 points on both cylinders. Its maximum value was found to 

 be 350,000 C.G.S. units. The speed was about 72 turns per 

 second, so that the maximum value of </> would be +0'25 

 volt ; on the outer cylinder (f> was about 1/10 less than on 

 the inner. 



From these results and from observations on the capacity 

 of the electrometer and connexions a rough calculation 

 of the theoretical deflexion was made. Corrections are 

 necessary on account of the charge bound inside the outer 



