General Discussion of Results. 
427 
by experiment, while in other cases they have been extrapolat¬ 
ed. This will subsequently be shown to be unjustifiable. 
In Plate II is platted the molecular conductivity of various 
salts in ethyl alcoholic solutions, from the determinations of 
Vollmer at 18°. The molecular conductivity is represented by 
ordinates, and the cube root of the volume, in which one gram- 
molecule of the substance was dissolved, is represented by abscis¬ 
sas. The cube root of the volumes was employed in order to 
represent the values for the more dilute solutions in the figure. 
It will be noticed that the salts of the alkalies yield limiting 
values for ^ while in the case of CaCl 2 there seems to be no 
tendency for the curve to become asymptotic. The same is true 
for Ca (No 3 ) 2 , and from my determinations for PeCl 3 and SbCl 3 . 
In Plate III is represented the molecular conductivity of 
salts in methyl alcohol from Vollmer’s determinations at 18° C. 
and in Plate IV the results of Zelinsky and Krapiwin deter¬ 
mined at 25° C. 
In most cases it will be seen limiting values are reached, 
Cdl 2 being, however, an exception. Zelinsky and Krapiwin’s de¬ 
termination for KI indicates that the curve would not become 
asymptotic, and therefore no limiting value for n would be 
reached. Nevertheless, Carrara assigns 97 as the value of /x,co, 
and Cohen 1 has pointed out that the difference in the determina¬ 
tions of Carrara and of Zelinsky and Krapiwin may be attributed 
to the effect of the platinum black electrodes, which they used, 
on the alcohol. 
Carrara calculated the values of /i <» for a number of salts in 
methylic alcohol solutions and found them to agree very closely 
with his own experimental results. In the following table are 
given these results and also the value of ^ in water. 
1 Loc. cit. 
