﻿Mr. Rollo Appleyard on Dielectrics. 



399 



which conformed precisely, and without undue pressure, to 

 the sheet of dielectric ; and, at the same time, made perfect 

 contact with it, leaving no air-spaces or spark-gaps where 

 ozone or other gases might collect. 



Circular sheets of celluloid, 11| inches in diameter, were 

 employed for these tests. To prevent leakage, a 3-inch 

 margin was allowed all round, so that the working-surface in 

 contact with mercury was 5^ inches in diameter. The method 

 of fixing the dielectric sheet wpjb as follows : — Two flat ebonite 

 rings, 5^ inches internal diameter, and 6^ inches outside, were 

 faced on each side with india-rubber, forming insulating 

 washers. These rings were placed exactly opposite each 

 other, flatwise, one on each side of the sheet to be tested. 

 Circular plates of iron were now laid against the outer faces 

 of the rings, and combined to make two mercury-tight hollow 

 boxes, one on each side of the dielectric. The whole was then 

 gripped vertically in an ebonite vice ; after which mercury 

 was poured in to fill the boxes. This apparatus, with the 

 exception of the insulating rings, is the same as that designed, 

 I believe, by Mr. Evers. The great advantage of the method 

 is that the mercury conforms to the surface of the dielectric, 

 leaving no air-spaces and necessitating no artificial pressure. 

 The surface-leakage can be made as small as desired, by the 

 simple device of leaving enough margin. Experiments made 

 upon dielectrics fixed in this apparatus give results entirely 

 different from those obtained with rigid metallic plates. 



A sheet of celluloid '071 inch in thickness was the first that 

 came under trial. Its behaviour during a cycle of voltage 

 from 150 to 1200 volts is represented in Table II. 



Table II. . 



Volts increasing. 



Volts diminishing. 



Volts. 



Resistance 

 in megohms. 



Volts. 



Resistance 

 in megohms. 



150 

 300 

 450 

 600 

 750 

 900 

 1050 

 1200 



82-7 

 82-3 

 820 

 81-5 

 81-2 

 80-8 

 80-6 

 80-6 



1050 

 900 

 750 

 600 

 450 

 300 

 150 



80-7 

 810 

 81-3 

 81-5 

 82-0 

 823 

 82-7 



