STRESSES IN THE DIELECTRIC 



139 



from a hole in the centre. This was connected to an india-rubber 

 tube provided with a stop-cock and to a bisulphide of carbon 

 manometer. The plates being immersed in the liquid near 

 together, and to begin with to earth, air was blown into the space 

 between them through the india-rubber tube until a large central 

 space extending from plate to plate was cleared of liquid. The 

 cock was then turned off* and the manometer read. The lower 

 plate was then brought to the same potential as in the experiment 

 above described, and at once the manometer showed an increase of 

 pressure, this increase being needed to balance the difference between 

 the electrical pressures at right angles to the lines of force in the 

 air and in the liquid. If h is the increase of height in the mano- 

 meter and (T the specific gravity of the liquid used, we have, since E 

 is the same in the liquid and in the air, 



If the difference of potential and distance apart are the same 

 as in the previous experiment we may eliminate E by the result 

 of that experiment. If we write K p for the value of K as involved 

 in the pressure at right angles, and K T for its value as involved in 

 the tension along the lines of force, the two experiments give us 

 respectively : 



<\c i\ E2 7 



( K P - A) -5 = g' 1 ^ 



and 



whence 



A' 



Kp = 



The values obtained for K p and K T were nearly equal to each 

 other and to the value obtained by the ballistic galvanometer, 

 which we may denote by K. 



The following Table shows the results in a few cases obtained in 

 two series of experiments, the earlier comparing K p and K T , the 

 later comparing K T and K. 



