261 
1909-10.] On an Electrically Controlled Thermostat. 
absorption from the, air. The concentration of this stock solution was 
determined by precipitation with barium chloride. All precautions were 
taken in the determinations, and a platinum Gooch crucible was used for 
the filtrations. 
About 0*5 grammes of the solution were rapidly pipetted with a special 
pipette into a small weighed glass bottle containing about 10 c.c. of water. 
The increase in weight of the bottle gave the weight of solution taken. 
In this way any increase in weight during the weighing was avoided. 
The volume of the filtrate and washings was measured, and a correction 
applied for the weight of BaS0 4 dissolved. The solubility was taken as 
1 mg. of BaS0 4 in 400 c.c. of filtrate. 
Determination I. 
Weight of solution taken (in air) . 0*5306 grms. 
Weight of BaS0 4 obtained (in air) 0*7674 grms. 
Yolume of filtrate . . . 400 c.c. 
Determination II. 
0*5331 grm. 
0*7710 grm. 
390 c.c. 
The molecular weight of BaS0 4 was taken as 233*46, and of H 2 S0 4 as 
98*076. 
The weights used were of platinum, and were carefully calibrated. To 
allow for the buoyancy of the air the uncorrected percentages were 
multiplied by 0*9991. The percentages of H 2 S0 4 , as calculated from these 
data, are : — 
(1) 60*78 5 ; (2) 60*78 3 . 
The above details are recorded in order that the true percentages of 
H 2 S0 4 may be calculated should the figures assumed above for the 
molecular weight and solubility of barium sulphate prove to he incorrect. 
The capacity of the cell was determined with a normal solution of 
Merck’s purest potassium chloride, which by previous trials had been found 
to give the same conductivity as specially pure chloride prepared from 
potassium perchlorate with all precautions. The conductivity for this 
concentration of potassium chloride is given by Kohlrausch and Maltby 
as 0*09828. 
The molecular weight of barium sulphate is only known to four signifi- 
cant figures. The concentrations can therefore only be vouched for to 1 in 
2000 at most. 
For the higher concentrations the conductivity varies rapidly with the 
concentration, and the accuracy of the conductivity determination is greater 
than is required. Near the maximum, however, the conductivity varies 
slowly with concentration, and the full accuracy is required. The limit of 
error of the conductivity determination is less than 3 in 10,000, as can be 
seen from Table I. 
