120 
MR, W. R. BOUSFIELR: TONIC SIZE IN RELATION TO 
and if av g took a concentrated solution of either salt at a suitable pressure and 
a temperature above the melting-point of the salt, and were gradually to drive off 
the water, we should come gradually to a condition of things in which the conduction 
was entirely autolytic. Hence we cannot expect the near equality of ionisation which 
we find in the case of KC1 and NaCl at the dilute end of the range to persist at the 
concentrated end where the idiosyncrasies of self-ionisation may produce a marked 
effect. 
(d) Mixtures of KC1 and NaCl Normal Solutions.— If two solutions have the 
same ionisation, there should be no change of molecular aggregation upon mixture, 
since, according to our theory, change of ionisation and change of ionic hydration 
are necessarily concomitant. Density measurements afford a very sensitive test 
of changes of molecular aggregation, and the results in this instance show that 
the change of aggregation upon mixing normal KC1 and NaCl solutions in various 
proportions is very small. 
The question of the conductivities of such mixtures has a bearing upon the 
validity of our main hypothesis. If the conductivities of such mixtures are calculated 
according to the so-called “law of mixtures,” they all come out about 0*0003 too 
high ; but calculated as below they come within 0 * 000 ] of the correct values. The 
residual difference, both with the densities and the conductivities, may be due to the 
slight difference in ionisation in normal solutions which we attribute mainly to a 
difference in autolytic ionisation, 
A.ccoi dm^ to our hypothesis we have, for the conductivity of the solutions of KC1 
and NaCl respectively, 
_ ma 1 _ mot l 
Kl_ 1000 ' Vl rf K ‘ 2 ~ Ibbo ' ffr 2 ’ 
where — stands for the sum of the reciprocals of the radions in each case. 
If we mix a volume V x of the KC1 solution with a volume V 2 of the NaCl solution, 
and there is no contraction, we have for the concentrations of KC1 and NaCl. 
respectively, in the mixture 
m 
mYo 
Vi + V nis ~ Vj + Yo ‘ 
Hence, if 17 be the viscosity of the mixture, we get for the portions of the 
conductivity of the mixture due to the KCl and NaCl respectively, assuming the 
coefficients of ionisation to be equal, 
a inYi 
1 
and a 
mV, 
1000 ' Vj + V, 
rjr 1 
1000 ‘ 
Vi + Vo 
v, 
Ki^h 
and V * 
K 2 r)2 
Vi + V,- 
V 
' Vi+V. 
2 ’ V 
