Conductivity produced by Heating Salts. 731 



are not obtained in the absence of water vapour in some 

 cases. I£ this is always so then dry KI would not give 

 any leak.) 



We hence conclude that in most cases definite chemical 

 changes take place when increased conductivity is brought 

 about. 



Second Series. 



I. Hie Effect of Low Temperature on the Conductivity 

 produced by Zinc Chloride. 



The zinc chloride was heated to 360° C. and the tube in 

 which it was heated could be put into communication with 

 the testing apparatus in two distinct ways, by means of 

 glass tubes. 



One of these glass tubes A passed direct from the part 

 containing the heated zinc chloride into the testing vessel, 

 the other B was bent, in between the two parts of the 

 apparatus, to form a flat spiral. This spiral was immersed 

 in solid C0 2 . The air was drawn over the zinc chloride and 

 through A or B into the testing apparatus by means of a 

 large glass aspirator so arranged as to produce an air current 

 of about 10 cms. per second in the tube containing the 

 electrode C (fig. 1). 



(a) Effect upon the positive ions. — The rate of leak w T as first 

 tested without aspirating, then when aspirating through A, 

 and lastly when aspirating through B which was roughly 

 equal in length to A. When the air was drawn through the 

 flat spiral of glass immersed in the solid C0 2 contained in a 

 beaker, the rate of leak was reduced to one-eighth the value 

 it had when drawn through A. When drawn through A, a 

 fog was produced in the aspirator like that previously noted 

 by Dr. Willows and myself *, but when the air was drawn 

 from the heating tube through B hardly any fog could 

 be noted. 



(b) Effect upon negative ions. — In this case the leak was 

 reduced to § the value when aspirating through B that it had 

 when aspirating through A. The fog was much reduced. 



The electrode C (fig. 1) was charged to 190 volts in both 

 cases. 



It is probable that the reason why more effect is apparently 

 produced by the low temperatures on the positive ions is that 

 the rate of leak obtained depended largely upon the velocity 

 of the ions, as it was not possible under the conditions of the 

 experiment to obtain a saturation-current. 



* Phil. Mag. Oct. 1904, pp. 452-3. 



