aoe ON THE CALCULATION OF THE CONDUCTIVITY OF 
tions. The method of procedure in calculation will be more 
exactly described after the experimental determinations have 
been dealt with. 
To obtain the experimental data required in the calculation, 
it is necessary to make up a long series of solutions of each salt 
and measure their conductivity. The volumes of the constituent 
solutions before mixing must be known, and in cases where there 
is an appreciable change of volume on mixing the densities of 
the single solutions and of the mixture must be found. These 
determinations will be taken up in order. 
Purity of Salts. 
Of the salts used the sodium chloride had been furnished 
by Merck, the barium chloride by Eimer & Amend of New 
York. The former was tested for metals of the alkali and 
alkaline earth groups, as well as for iron. No indication of 
these was found. The barium chloride was tested by heating 
in a Bunsen flame. No impurities were discovered in this way. 
A further indication that the salts were sufficiently pure for 
the purpose in view is given by the comparison of the conduc- 
tivities with those of Kohlrausch, as shown farther on in the 
paper. All the salt, however, used in the experiments described 
here, was recrystallized once by the writer. 
Purity of Water. 
The water used in making up the solutions was obtained by 
adding potassium hydroxide, about 1 gramme to a litre, to ordi- 
nary distilled water, and redistilling by means of a copper boiler 
lined with tin, the distillate being collected in a block-tin worm. 
The conductivity of the water before the potassium hydroxide 
was added was about 1.87 in terms of the conductivity of 
mercury at 0°Cx10°. After distillation the conductivity 
ranged from 1.1 to 0.95, the latter being that of the water used 
for the very dilute solutions. 
At an earlier stage of the experiments water was purified by 
methods derived from a paper by G. A. Hulett.* Instead of the 
* Journal of Physical Chemistry, Vol. 1, No. 2. 
