834: Mr. R. T. Lattey on the Dielectric 



This is illustrated by some calculated values in Tables II. 

 and III. 



Table II. 

 Calculated values of I/i. 



K 2 . ?'/E = l. 10. 100. 1000. 



1 1-30 1-08 1-01 T002 



2 1-56 1-16 1-02 1-005 



10 300 178 1-10 1-011 



100 ..... 9'6 6-62 1-97 1101 



Table III. 

 Values of C 2 + ^— corresponding to Table II. 



K 2 . r/R=l. 10. 100. 1000. 



1 2 30 0-21 002 0-002 



2 2-56 0-22 0-02 0-002 



10 4-0 0-28 0-02 0-002 



100 10-6 0-77 003 0002 



To make r/R large, R must be small; but this will 

 increase K 2 (= (h ^-Wc o R 2> ) unless L-p^Cf 1 is 



small and C is large. 



This illustrates the fallacy of the general assumption that 

 exceedingly high frequencies are necessary for all methods 

 of determining the dielectric constant of an electrolyte. 

 The assumption is correct when the method referred to 

 in section I. A is used. The results given by methods B 

 and C appear at first sight to be independent of frequency ; 

 but it must be borne in mind that R is not independent 

 of p. As a first approximation, it is generally stated that 

 Bp~* is constant for a straight wire. Some experiments 

 made by the author seem to show that for circuits and 

 frequencies such as were used in these experiments, E^ -1 is 

 more nearly constant. The circuit is most free from errors 

 due to leakage when it is symmetrical, i. e. when Ci = . 

 In this case, K becomes equal to phj2R } and is practically 

 constant so lono- as L is constant. 



II. Estimation of the Factoe K. 



The electrolytic condenser is removed, and the current 

 estimated for various values of C ; if I is the maximum 

 current when C = Co and I is the current when C = C, then 

 equation (1 b) leads to 



*('-§) - (¥-')' IM 



