124 ON THE CALCULATION OF THE CONDUCTIVITY OF 
by recrystallization. Solutions of them were found to be neutral 
and free from sulphates and magnesia. Neither potassium nor 
other metals could be detected in the sodium chloride with the 
spectroscope. Sodium, but no other metal, could be detected in 
a flame coloured by the potassium chloride. The hydrochloric acid 
was obtained as chemically pure, and gave no residue on evapora- 
tion. It was free from sulphates. 
The Water Used. 
The water was doubly distilled, with addition of sodium 
hydrate, in a tin-lined retort, and condensed in a block-tin pipe, 
the first part of the distillate being rejected. It was stored in 
bottles which had been used for this purpose for several years. 
It gave no residue when evaporated, was neutral, and gave no 
colour with Nessler’s reagent. 
Preparation and Analysis of the Simple Solutions. 
The simple solutions were prepared by dissolving about the 
amount of salt required for the strongest solution, and subsequent 
diluting. The concentration in each case was determined by 
volumetric analysis. A solution of silver nitrate was used in 
estimating the chlorine in the potassium and sodium chlorides, 
and the amounts of salt present were calculated from the data 
thus obtained. 
In making an analysis 1 ¢. ¢. (or 5 © ©) of the solution at 
18°C was drawn off by a pipette, placed in a flask, diluted, and 
coloured distinctly with neutral potassium chromate. Silver 
nitrate standardised at 18°C was run in from a burette, and a 
glass bulb filled with potassium chromate of the same shade as 
the solution being analysed, was held before the eye. The end 
point by this means could be seen quite sharply. 
A solution of ammonia was employed for estimating the 
hydrochloric acid, with cochineal as an indicator. 
The pipettes and burettes used were tested by weighing the 
water which they delivered. They were found to be accurate 
to 0.1 per cent. 
