] 04 
MR, W. R. BOUSFIELD: IONIC SIZE IN RELATION TO 
ionic sizes upon rates of transference would theoretically lead to the result that the 
reciprocals of the Hittorf migration numbers should be expressible as a linear 
function of the ratio of the radions. This turns out in fact to he the case, and we 
incidentally arrive at a useful method of extrapolation to determine the value of the 
Hittorf number for an electrolyte at “ infinite” dilution. 
These considerations enable us to determine the coefficients B for the separate ions 
in the expressions for the radions. These coefficients B we refer to as the “ hydration 
numbers,” the relation between the hydration numbers and the migration numbers at 
infinite dilution being' of the form 
B = BiNj + BsNo. 
Turning now to a consideration of the viscosities of the KC1 and NaCl solutions, it 
is shown that the viscosity of these solutions can be represented approximately as a 
linear function of the radions, as can also the viscosity of mixtures of normal KOI and 
NaCl solutions. 
Passing to a consideration of the general relation of viscosity to ionic size, the 
extended conception of the radion is introduced, and an approximate value is given to 
the radion of water, which expresses the average radius of the water molecules 
reckoned upon the same scale as the radions of the solute. Using this value of the 
water radion in conjunction with the values of the radions of the solute determined 
from the conductivities, it is shown that the viscosity of the solutions can be expressed 
with a fair approach to accuracy by the expression 
V = C l/3r, 
where r stands for the radion of a given species of molecules, and /3 for the fraction of 
the total volume occupied by such species. Since is, upon the extended concep¬ 
tion of the radion, the average molecular radius of the whole solution, we may express 
this result by saying that the viscosity of an aqueous solution is proportional to its 
radion. 
In order to correlate ionic sizes with osmotic pressure, a prolonged attempt was 
made to measure the vapour pressure of dilute KC1 solutions at 18° C. A large 
barometer tube was used, closed by a small tap at the bottom, so that minute 
differences of level could be determined by removing and weighing the mercury 
cistern. A similar arrangement was used to determine simultaneously the variations 
of atmospheric pressure during each observation. It was found, however, that the 
variations of atmospheric pressure were often larger than the differences of vapour 
pressure to be measured, and no sufficiently accurate results could be obtained. 
Recourse was therefore had to the freezing-point determinations of Jahn with KC1 and 
NaCl solutions. The variations of ionic size with temperature are probablv serious 
