Section 17. Effect of Concentration on Conductance. 



49 



It seemed therefore to be of especial interest to test the applicability of 

 these formulas at the widely different temperatures employed in our 

 experiments. When such a test is made by direct substitution the results 

 are in a high degree dependent on the values of A employed, since in 

 dilute solutions A A is a relatively small quantity; yet in several 

 instances authors have not given sufficient consideration to this matter. 

 The most satisfactory method of procedure seems to us to be the elimina- 

 tion of the A value, which can not be determined with sufficient accuracy 

 by extrapolation, by writing the functions in the following form : 



A = A K O (Kohlrausch) A = A K A> O (Barmwater) 



A = A K A* O ( van't Hoff) A = A K A 2 O (Rudolphi) 



and then plotting the values of A along one coordinate axis and those of 

 the C-A function constituting the last term (that is, C h , A* C h , etc.) along 

 the other axis. If the function in question holds, the points will of course 

 lie upon a straight line ; and by comparing, in the case of the different 

 functions, the deviations of the separate points from the best representa- 

 tive straight line that can be drawn, a measure of the degree of applica- 

 bility of each function is obtained. All our complete series of measure- 

 ments and those of Kohlrausch and Maltby on the same salts at 18 have 

 been studied in this way, a plot on a very large scale being employed. 

 The straight lines were drawn in every case so as to represent most 

 closely the points for the concentrations 100 or 80, 10, and 2 milli- 

 normal, and the average of the percentage deviations of the observed A 

 values at these three points taken. These averages for the two functions 

 are given in the following table under C l and (CA)*, respectively. 



Table 10. Mean percentage deviations of the observed values of the equivalent 

 conductance from those calculated by the cube-root functions. 



It will be seen that the deviations from either function are insignificant 

 up to 156, but that they become considerable at the higher temperatures. 

 It may be of interest to state also the percentage deviations of our straight 

 line corresponding to the Kohlrausch function from the points repre- 



