Constant of Ice and Alcohol at very Loxv Temperatures. 7 



this will show itself by making the galvanometer readings in the two 

 cases nneqnal. Mr. Petavel, who assisted ns in these observations, 

 and to whom our thanks are due, arranged a convenient switch- 

 ing device which enabled the galvanometer to have its position 

 in the circuits instantly changed to take either the charge 

 currents or the discharge currents of the condenser, and the 

 equality of these readings is taken as an indication that no sen- 

 sible leakage takes place across the dielectric during the passage 

 of the contact maker from one stop to the other. This method of 

 exchanging the position of the galvanometer also eliminates errors 

 due to the setting of the scale, as the deflections are on opposite sides 

 of the zero. The above- described arrangements having been made, 

 the ice condenser w r as cooled down to the temperature of liquid air 

 by immersing it in the liquefied gas contained in a large vacuum 

 vessel. 



In order to take the temperature of the condenser a platinum wire re- 

 sistance thermometer was placed in the inside of the inner cylinder and 

 in close contact with it. The ice having been reduced in temperature 

 to —185° C, or —198° platinum temperature, the capacity of the con- 

 denser was measured. The condenser was then raised out of the liquid 

 air and allowed to warm up very slowly, and its capacity taken at 

 various stages as the temperature rose. Before and after the experi- 

 ment with the ice the capacity of the condenser was taken when the 

 metal cylinders were at the temperature of —185°, but the dielectric 

 was gaseous air at that temperature instead of ice, and the results so 

 obtained enabled the dielectric constant of the ice to be calculated. 



These experiments being to a considerable extent preliminary 

 experiments, and intended merely to explore the ground, we do not 

 make any particular claim for the accuracy of the numbers as deter- 

 minations of a physical constant. We are arranging improved 

 methods for repeating the whole of these measurements. All we 

 desired in the first instance to do was to examine the mode in which 

 the dielectric constant varied with temperature and its approximate 

 magnitude at each temperature. The following Table III shows the 

 observed value of the dielectric constant of ice between — 185° C. 

 (-198° pt.) and about -120° (-130° pt.). The voltage used on the 

 condenser was 24'1 volts, and kept perfectly constant during the 

 whole time. In order to ascertain if the galvanometer deflection was 

 really due to the sequence of capacity charges or discharges only, 

 and not to any meas arable admixture of conduction current through 

 the ice, a resistance of 1,000 to 90,000 ohms was occasionally 

 inserted in the galvanometer circuit, and the absence of any observed 

 change in the galvanometer deflection was taken as a proof that 

 sensible conduction did not interfere with the true capacity effect. 

 The galvanometer used had a resistance of 500 ohms. 



