a Dielectric in an Electrostatic Field. 205 



this potential I still got no deflexion although the lowest calcu- 

 lation from Righi's results should have given me 38 divisions. 

 The time of charging was varied from almost immediate 

 sparking to a gradual increase of potential that lasted 

 occasionally five minutes or more. Just at the time the spark 

 passed there was often a sudden jar of the image, sometimes 

 up and at other times downward. This could not have been 

 mistaken for the phenomenon expected. A calculation from 

 Quincke's results and formula show that this potential should 

 have given at least 20 divisions. Thus by the observations 

 of both experimenters I should have obtained a large deflexion. 

 After a series of experiments with these high potentials the 

 glass was ruptured, so that the limit of possible strain was 

 reached. During the course of the experiments four tubes, 

 of the best German Thuringian glass, were used to be sure 

 that it was not a special case of a single specimen. 



I next tried a hard rubber tube in the same manner as the 

 glass. The outer cylinder was put in place and the spaces 

 filled with kerosene. The specific inductive capacities of the 

 two are about the same, so that the potential gradient between 

 the charged plates was practically a straight line. The sensi- 

 tiveness of the apparatus was so great that gradual changes 

 of temperature caused the image to move slowly across the 

 field. With sparks up to seven millimetres no change in 

 length was noted, except the jar before mentioned at the time 

 the spark passed. This motion was too rapid to be measured, 

 but I guessed it to be about seven divisions. The time of 

 charging was varied as in the experiments with glass. It is 

 stated that this substance elongates about ten times as much 

 as glass*. 



These results, so far at variance with the statements of 

 former investigators, have caused me carefully to study their 

 methods and results, and one cannot but be impressed with 

 their discrepancies. Consider first the work on change of 

 volume. The method employed was to charge a conducting 

 liquid in a glass thermometer and to observe the change of 

 volume of the glass by the rise or fall of the liquid in the 

 capillary tube. 



The experiments give the following results. Govi found 

 that when the internal liquid was water the height fell, show- 

 ing an increase in the volume of the glass. With nitric acid 

 a greater change was observed than with water, and with 

 alcohol a much greater change than with nitric acid. Ether, 

 when charged, rose in the tube and then fell ; mercury and 

 olive-oil showed no effect. Duter, experimenting later, states 

 * Quincke, Wied. Ann. x. p. 200. 



