TEMPERATURE COEFFICIENT OF ICE. JOHNSTON E. 137 



will be seen that this method of measuring high resistances of 

 any kind, insulators for example, is very advantageous, for 

 currents as small as 10~ 14 amperes, can he measured with ease 

 by the Dolezalek electrometer, while the Wilton Tilting electro- 

 scope can be made very sensitive. For electrolytes with low 

 resistances, the current could be measured by a galvanometer 

 and the potential by the electrometer, since the latter as a 

 current measure would be too sensitive for use in this case. 



PKEPARATION OF THE ICE. 



Pure water was obtained with a resistivity of about 1x10 

 ohms. 



After cleaning the U U" tube very carefully first, in a 

 solution of potassium bichromate and sulphuric acid, then in 

 an alcohol-ether solution and then washing several times in 

 distilled water, both hot and cold the pure water was put in 

 the tube, its resistance carefully determined and temperature 

 noted. The specific resistance of a sample of this water was 

 then determined by the KohlrauscJi method, the temperatures 

 being the same in the two cases. From these results the a cell 

 constant" of the apparatus can be calculated. 



The "U" tube was then placed inside a cylindrical glass 

 vessel, about 15 cm. in diameter and 45 cm. in height. This 

 was then placed in an earthenware- jar, which in turn was sur- 

 rounded by an ice-salt mixture, contained in a bucket. The 

 while apparatus was kept in a refrigerator. 



A thermo-couple (Fig 7) consisting of a German-silver- 

 iron junction, was used to measure the temperature of the ice, 

 which was formed in the "U" tube. The junction was enclosed 

 in a capillary tube, which could be slipped in and out of one 

 of the electrode tubes, "a ", (Fig. 5). For a diagram of 'the 

 connections of the thermo-couple see Fig. 7. A very careful 

 calibration of this instrument was made over the range of 



