TEMPERATURE COEFFICIENT OF ICE. — JOHNSTONE. 131 



It was observed that this current p'oduced a deflection of 

 51 scale divisions. Therefore the current necessary to produce 

 a deflection of one scale division, will be 



2 

 1000 = 3.29 X 10~^ amp. 



(10001 + 1930)51 



The specimen of ice was prepared as follows : Two brass 

 electrodes, — circular discs, were made ; a copper rod was 

 soldered to one of them and a copper wire was soldered to the 

 edge of the other one. A cylinder of ice, 3 cm. in height, was 

 cut from ice, obtained from the Dartmouth Lakes. The 

 electrodes were then frozen to this cylinder of ice by warming 

 them slightly and then pressing them to the upper and lower 

 surfaces of the ice. This conductivity cell, so to speak, was 

 placed on a plate of parafine wax, and the whole thing was 

 placed in a box, which was kept in the open air, shaded from 

 the sun. Of course experiments could only be performed when 

 the air was below the temperature of 0°C, which was quite 

 frequent at this period of the year. Several sets of readings 

 are given below, together with a set of readings from Ayrton 

 and Perry's papers.-^ The apparatus was connected u]:) as in 

 Fig. 2. 



In the actual experiment, the current was made to pass 

 through the ice for a considerable period of time, in some cases 

 the circuit being unbroken for 48 hours. 



When the current was suddenly reversed after flowing for 

 quite a length of time in one direction, a very much greater 

 deflection of the galvanometer was obtained than at first. This 

 deflection decreased somewhat with time. Thus for instance, 

 the deflection changes from 9 divisions, on one side of the zero, 

 to 13 divisions on the other side when the current is reversed 

 through the ice. This is an increase of 40^ of the current, which 

 passed through the ice in the initial case. If this is all due to 



1. Loc cit. 



