48 SECTIONAL ADDRESSES. 
theory with the results of conductivity determinations, the assumptions 
underlying it must be remembered, viz. : 
(1) That the electrolyte is completely dissociated into point ions and 
that all interionic forces except Coulomb forces can be neglected. 
(2) That corrections for the overlapping of ionic atmospheres can be 
neglected. 
(3) That the solvent between the ions retains the properties of the 
pure solvent. 
These conditions can only be fulfilled in dilute solutions, and a great 
deal of work has been done recently with uni-univalent electrolytes in a 
number of solvents to test the theory in the dilute range. 
The results show that there is a close approach to a linear relation 
between A, and c?, as required by the theory, for strong electrolytes in 
solvents with a dielectric constant greater than 20. In water, methyl 
and ethyl alcohols, nitromethane and acetonitrile the slopes of the con- 
ductivity curves agree well with theory for a number of electrolytes, and 
any large deviations are such that they can be explained by ionic associa- 
tion. In fact, the body of evidence now available seems sufficient to 
justify the use of the Debye-Hiickel-Onsager equation to represent the 
behaviour of a perfect electrolyte in dilute solution and we have, therefore, 
a new means of judging whether an electrolyte is appreciably associated 
or not. 
The term ionic association marks the contrast between the old outlook 
and the new. Arrhenius thought of the act of solution as separating the 
molecules into ions. We know now that most salts are already ionised 
in the crystalline state, and the question that concerns us is whether the 
condition of complete ionisation persists on solution of the crystal. In 
many cases the conductivity of an electrolyte is less than we should 
expect from the Debye-Hiickel equation, indicating that some modification 
of the state of configuration of the ions has occurred which involves a 
decrease in the conductivity. This may be due either to the formation 
of a covalent linkage between the ions or to a modification of their dis- 
tribution leading in the extreme case to the formation of an ion pair as 
suggested by Bjerrum. The term ionic association is used to cover both 
possibilities. 
The influence of the solvent on the properties of an electrolyte is 
illustrated very clearly by a comparison of the behaviour of uni-univalent 
salts in water and in non-aqueous solvents. In water they are all strong 
electrolytes with a surprisingly uniform behaviour, as shown in 
Kohlrausch’s classic diagram in the Zeitschrift fiir Electrochemie for 1907. 
In non-aqueous solvents, however, Walden’s comprehensive investigations 
have shown that individual differences begin to appear as the interionic 
forces increase and the specific affinities of the ions are brought to light. 
The question naturally arises as to whether the extent of the ionic associa- 
tion is determined entirely by the interionic forces, i.e. by the dielectric 
constant of the solvent. This is clearly not the case, since a number of 
salts are strong electrolytes in methyl alcohol and weak electrolytes in 
nitromethane, which has a higher dielectric constant (37 as against 30-3). 
Walden and Ulich have pointed out that, in general, non-hydroxylic 
solvents such as the nitro-compounds and acetone accentuate the individual 
