﻿398 



Mr. .Hollo Appleyard on Dielectrics. 



insulation has been computed from the deflexions obtained 

 after the corresponding voltage had been applied for one 

 minute. The upper readings correspond with increasing- 

 voltages ; the lower curve completes the cycle with diminish- 

 ing voltages, and shows the falling-off of insulation with 

 current, in a manner curiously like the hysteresis curve of 

 The observed values are given in Table I. 



iron. 



Table I.— Celluloid Sheet. 





volts 

 150 



volts 

 300 



volts 

 600 



volts 

 750 



volts 

 600 



voHs 

 300 



volts 

 150 



r i m 



Galvanometer J 

 Headings. j 2m 



19-8 

 20 



85-8 

 89-3 



533 

 554 



1310 

 1338 



605 

 663 



152 

 158 



45 

 45 



Corresponding 1 

 Eesistance j- = 

 in megohms. J 



4040 



1840 



597 



303 



526 



1039 



1778 



This table includes the actual galvanometer-deflexions from 

 which the insulation-resistance at each voltage was calculated. 

 First and second minute readings are recorded, and exhibit 

 very strikingly the apparently negative " electrification/' — the 

 first-minute readings being less than the second-minute 

 readings. 



A series of tests was made upon celluloid sheets pressed, 

 in the above manner, between two plane surfaces of metal, and 

 the same general features were observed in each case. As no 

 definite fault could be discovered in any sample, I was led to 

 suppose that the abnormal change in resistance, and the 

 corresponding negative " electrification,*'' were to be accounted 

 for by some complex molecular rearrangement within the 

 substance of the celluloid itself. Leakage over the edges 

 might obviously be urged in explanation, but the conditions 

 did not seem to allow of any serious loss in this manner. 

 While employing unyielding metallic surfaces against the 

 comparatively hard surface of the celluloid, the chances of 

 uneven pressure upon the sheet were very great : which would 

 imply uneven electrical contact between the dielectric and the 

 plates. 



The next experiments were made with a view to test the 

 celluloid in perfect contact with the opposed plates ; to remove 

 all chances of leakage-error ; and to provide, at the same time, 

 a more evenly distributed pressure over the surface. The hard 

 metallic plates were therefore replaced by mercury surfaces 



