302 Profs. J. A. Fleming and J. Dewar. On the 



show if the rapid series of charges of the condenser gives a current 

 exactly equal to the rapid series of discharges. When this equality- 

 exists we know that there is no sensible leakage current passing 

 through the condenser dielectric, or loss of charge during the interval 

 of time which elapses between two successive contacts of the vibrating 

 switch. 



This disposition of apparatus enabled us therefore very quickly — 



(1) To measure the capacity of the condenser when having any 



selected frozen liquid as dielectric at any temperature. 



(2) To measure the resistance of the above dielectric. 



(3) To determine the temperature of the dielectric layer. 



(4) To heat up the whole condenser slowly to any temperature 



above that of liquid air, 



(5) To determine whether there was any sensible leakage current 



through the dielectric. 



The usual method of procedure, after having filled the condenser 

 with the electrolyte to be examined, was to cool down the whole con- 

 denser to the temperature of liquid air by slowly immersing it in 

 several litres of liquid air contained in the large vacuum tube. Then 

 it was raised out into the cold gaseous air lying above the surface of 

 the liquid air, in which position the whole condenser rose very slowly 

 in temperature from —185° C. up to the melting point of the dielectric, 

 and during this time observations for capacity, resistance, and tem- 

 perature were successively and quickly taken at various temperatures. 



The great mass of the condenser favoured a very slow tempera- 

 ture change, and thus assisted to keep all parts of the condenser at 

 the same temperature. 



The operation of measuring the dielectric constant of the frozen 

 electrolyte placed in the condenser consisted then in measuring the 

 current flowing through the galvanometer with a constant switch- 

 frequency, and charging voltage, when the condenser had the 

 selected electrolyte as its dielectric ; and then again when the 

 frozen electrolyte was melted out and the condenser put together 

 with the cones the same distance apart and at the same temperature 

 but having gaseous air as the dielectric. As a matter of fact, each 

 time of experimenting one careful measurement was made of the 

 charge and discharge current through the condenser set up with the 

 cones a fixed distance apart and the cones cooled to the temperature 

 of liquid air, the dielectric being gaseous air of the same tempera- 

 ture. The capacity of the condenser, when the vertical distance 

 separating the cones was 1*5 mm., was very nearly l/8000th of a 

 microfarad, with air as dielectric. Under these circumstances the 

 multiplication of the observed galvanometer deflection by a factor, 

 after applying a correction for the capacity of the leads and vibrator, 



