304 Profs. J. A. Fleming and J. Dewar. On the 



rected value is given of the dielectric constant of the various 

 electrolytes used, the dielectric constant of air at ordinary pressure 

 and temperature being taken as unity. Temperatures are all given 

 in terms of our standard platinum thermometer P x .* 



The temperature at which the observations are interrupted in the 

 case of each electrolyte depends upon the temperature at which 

 sensible electric conductivity makes its appearance. One of the 

 most remarkable facts which has attracted our attention in these 

 experiments is that electric conductivity in electrolytes which are 

 frozen and then reduced to very low temperatures extends far down 

 into the solid condition, but there is for each body apparently a 

 temperature at which electric conductivity begins to increase very 

 rapidly, so much so, that in some cases a rise of temperature of only 

 two or three degrees makes all the difference between being able to 

 measure the dielectric constant well by the above-described method 

 and not being able to measure it at all. 



At the lowest temperatures reached, all the frozen electrolytes we 

 have used appear to be practically perfect insulators, or, at any rate,, 

 to have enormously high electric resistivity, and then, as i^hey are 

 gradually heated up, long before their melting point is reached, quite 

 suddenly they begin to exhibit very sensible conductivity. We have 

 not attempted to interpret this result into the language of any 

 theory of electrical conduction, but it is evidently a very important 

 fact in connection with electrolytic conduction. 



Hence, in the above described observations, a point of temperature 

 is reached for each frozen electrolyte as it warms up, at which, for 

 the frequency of charge and discharge (viz., 120 per second) we were 

 using, further observation on the dielectric constant became impos- 

 sible, at least by the method employed. Cases in which the true 

 dielectric measurement is doubtful, owing to such incipient conduc- 

 tion, are indicated by a query. 



The following measurements have, then, been made on certain well 

 known electrolytes frozen as described : — 



* See "Thermo-electric Powers of Metals and Alloys," Dewar and Fleming, 

 < Phil. Mag.,' July, 1895, p. 100. 



