TEMPERATURE COEFFICIENT OF ICE. JOHNSTONS 135 



platinum wires, d and d^ sealed in the end as shown in the 

 iigure. Thesa four tubes are fitted carefully in corks, and 

 then in the "U" tube as shown. 



Now suppose we fill the tube with an electrolyte and pass 

 a current through it, by way of the electrodes b and b r There 

 will result a polarization at these electrodes and the value of 

 the current will vary somewhat with the time. If this current 

 is measured by a sensitive galvanometer and if the difference 

 in potential of the two electrodes is measured by a voltmeter, 

 a value of the resistance of the electrolyte could be calculated 

 .at any particular time; but this is a varying quantity, and not 

 the true resistance of the electrolyte. 



Now if two other electrodes a and a (see Fig. 5), are placed 

 in the position shown in the figure, their ends c and c x being 

 lower somewhat than the extremities of the pjatinum wires 

 d and d v and then a current is parsed through the electrolyte 

 "by way of the electrodes d and d^ then if we measured the dif- 

 ference of potential between the points c and c r by some electro- 

 static instrument and knew, the value of current, we could 

 calculate the value of the resistance of the electrolyte between 

 the points c and c,, and the resistance so calculated would be 

 constant in value and unaffected by polarization. This would 

 be so, because when there is a variation in the current due to 

 polarization or any other causes, there will be a proportional 

 change in the potential difference between the two potential 

 electrodes, so that the ratio of the potential difference to the 

 current will be constant (with a const, temp.). Therefore the 

 resistance determined in this way will have a constant value. 

 Thus, polarization effects will be eliminated. 



The current passing through the electrolyte, (which was 

 ice in this case), was measured by a Dolezalek electrometer, 

 A, (see Fig. 5). The potential of ea?h potential electrode 

 was measured by a Wilson Tilting electroscope, D, the plate 

 of the electroscope being kept at a potential of 320 volts, from 

 small storage cells. A caMbration curve for this instrument, 

 as it was used in these experiments is shown in Fig. 0. It 



