592 



REPORT — 1897. 



In the case of the insulating oils it was found impracticahle to make direct 

 readino-s, on account of the dense deposit of carbon on the electrodes, which 

 materially altered the conditions of the discharge. 



The method of Macfarlane and Pierce was therefore adopted, and a pair of 

 spheres of 1 cm. diam. were used as a subsidiary electrometer, the results previously 

 obtained being used for the reduction of the results now obtained. At high 

 potentials it was found impossible to get a consistent set of readings, on account of 

 the violent agitation of the liquid, especially in the case of the lighter oils, like 

 kerosene, in which the convection effects were very pronounced. 



It was found that, while a smaller difference of potential is necessary to pro- 

 duce a discharge through a given distance for large spheres than for small ones, 

 when they are close together, for longer distances the air is dielectrically stronger 

 for large than for small spheres. For spark-lengths of more than 3 mm. the curves 

 are practically straight, and the dielectric strength is therefore constant. 



The values of the dielectric strength of air, at ordinary pressures, are aa follows : — 



Dielectric Strength. 

 Diam. of Spheres 

 in Cm. 



•5 

 10 

 20 

 30 



All of these values are considerably higher than that obtained by Macfarlane 

 for planes, viz. 23'8 kilovolts per cm. 



The results for the insulating oils are not as uniform as those for air, but the 

 same general characteristics were found, except in a few cases. In all cases the 

 curves for the spheres of 3 cm. diam. are fairly straight, and it would appear that 

 Macfarlane's conclusion that the dielectric strength of liquids is constant for plane 

 electrodes is warranted. 



The following estimates for the dielectric strength of the oils experimented 

 upon are given : — • 



large induction coil, through the primary of which an alternating current of 

 E.M.F, 50 volts and frequency 125 was passed, the electrometer and the spark 

 micrometer being connected in parallel with the secondary of the coil, A variable 

 resistance was used to regulate the potential. The results were not very satis- 

 factory, but the values of the spark-length for the largest spheres were situated 

 between those obtained by Steinmetz and Siemens, the frequencies of the alter- 

 nating currents used by them being respectively 150 and 100. In fact the results 

 appear to bear out Taumann's contention that the more rapidly the potential ia 

 changed the less will be the potential required to spark across any given distance. 



