B.—CHEMISTRY. 49 
differences between electrolytes, while the hydroxylic solvents suppress 
them. The reason for this is probably to be found in a difference in the 
nature of the solvation of the ions in the two classes of solvents. There 
is abundant evidence that the ions in solution have a number of solvent 
molecules attached to them either by co-ordinate linkages or as a result 
of the dipole character of the solvent, and these solvent atmospheres 
exert an important influence on the behaviour of the ions. For instance, 
the evidence of the ionic mobilities of the alkali metals and of Washburn’s 
transference experiments leaves no doubt that the effective size of the 
ion is determined by its diameter, including any envelope of solvent which 
it carries with it. 
Bjerrum has considered the effect of ionic size on the probability of 
the formation of ion pairs, which would contribute nothing to the 
conductivity of a solution, and has shown that if the sum of the radii of 
the two ions is below a certain value, the number of ion pairs will increase 
rapidly. Hence, the solvation of ions may have an important effect in 
preventing the ions from coming near enough together to form ion pairs. 
Sidgwick has considered solvation from the electronic standpoint, and 
has emphasised the importance of the donor and acceptor properties of 
hydroxylic solvents in enabling them to form co-ordinate links with both 
anions and kations. Non-hydroxylic solvents, however, like nitromethane 
and acetone, can only form co-ordinate links with kations, leaving the 
anions chemically unsolvated, and it is very significant that in such 
solvents lithium salts are weak electrolytes, while in hydroxylic solvents 
they are less associated than the salts of the other alkali metals. It 
would thus appear that the existence of a chemical link between the 
solvent molecules and both ions is of cardinal importance in preventing 
ionic association. This view of the protective action of hydroxylic solvent 
molecules is confirmed by the effects of small quantities of water on the 
conductivity of solutions in nitromethane and other non-hydroxylic 
solvents. For example, the conductivity of lithium thiocyanate, a weak 
salt, is increased 60 per cent. by the addition of 0-1 per cent. of water, 
while that of tetra-ethylammonium iodide, a strong electrolyte, is only 
increased 0-22 per cent. by a similar addition. 
Until recently, most of our ideas about electrolytic solutions were 
based on experience in water, and this gave quite a false impression of the 
simplicity of the problem, since in water, thanks to its high dielectric 
constant and to the protection afforded by the chemical solvation of 
both ions, all uni-univalent salts exhibit almost ideal behaviour. But a 
survey of non-aqueous solutions reveals at once a much more complex 
situation, in which the chemical nature of the solvent and the affinities 
of the ions are often the predominant factors. Thus, a purely physical 
theory (pace Faraday), like that of Debye and Hiickel, while invaluable 
in explaining and predicting the behaviour of an ideal electrolyte, is far 
from giving a complete picture of electrolytic solutions even in the dilute 
range, since it leaves out of account the chemical nature both of the ions 
and of the solvent molecules. 
Looking back over the century we see how the mechanism of electrolytic 
conduction has gradually been disclosed to us, and how in recent years 
the many-sided influence of the solvent has come more and more into 
1931 E 
