Voiv. 6, 1920 CHEMISTRY: NO YES AND MacINNES 23 
Noyes and Falk's^ summary of the experimentally determined transfer- 
ence-numbers. These numbers, which are equal to the ratio u+/(u+ + 
u~) of the mobility of the cation-constituent to the sum of the mobilities 
of the two ion-constituents, show variations that correspond to changes 
in the ratio u+/u~ of the mobilities of the two ion-constituents between 
zero concentration and 0.3 normal, of 4}/ 2% in the case of sodium chloride, 
7V2% in the case of hydrochloric acid, and 24% in the case of lithium 
chloride. Moreover, Maclnnes^ has shown that the chloride-ion constit- 
uent has the same equivalent conductance in 0.1 normal solutions of lithium 
chloride, potassium chloride, and hydrochloric acid, although the conduc- 
tance-ratios A/Ao for these substances are 0.833, 0.862 and 0.925; so that, 
if we account for the constancy of the chloride-ion conductance by the 
probable assumptions that the three substances are equally ionized and 
that the chloride-ion has the same mobility in the three solutions, we 
must conclude that the hydrogen-ion and lithium-ion decrease in mobili- 
ties between zero concentration and 0.1 normal by amounts that differ 
from each other by 10% — a result that makes it not unreasonable to sup- 
pose that the whole decrease in equivalent conductance (of 1^/2% and 16% 
in the two cases) may be due entirely to decrease in mobility of the ions, 
and not at all to decrease in ionization. 
When, indeed, in addition to these conclusions that neither the activity- 
coefficients nor the conductance-ratio is determined primarily by the de- 
gree of ionization, we take into consideration the fact that there is no 
property which affords any direct evidence of the existence of un-ionized 
molecules in solutions of most of the largely ionized inorganic substances 
up to moderate concentrations, it seems advisable to adopt for the present 
the hypothesis that such substances are completely ionized, and to at- 
tribute the decrease in the conductance-ratio wholly to decrease of ion- 
mobility, and the change in activity-coefficient entirely to some unknown 
effect of a physical nature. 
It would lead far beyond the scope of this paper to discuss the many 
classes of phenomena that seem to substantiate this assumption. A sum- 
marized description of them was given many years ago by one of the authors 
of this paper, who at that time, however, suggested that they might be 
explained more fully by the hypothesis that the ions are partially united 
as a result solely of their electrical attraction into loosely bound molecules, 
which differ fundamentally from the stable molecules formed as a result 
of chemical affinity in accordance with the law of mass-action. The 
known facts may, however, prove to be better accounted for by the simple 
hypothesis of complete ionization, supplemented by some other, purely 
physical, explanation of the cause of the decrease of ion-mobility and of 
ion-activity with increasing concentration; and this now seems the most 
promising method of treatment, as has recently been urged by various 
authors. As said above, we cannot here discuss in detail the hypothesis 
