THE COMBINING VOLUMES OF HYDROGEN AND OXYGEN. 
399 
The last palladium tube used in this work was set up in January, 1914. During 
the following seventeen months, 20 to 30 litres of hydrogen passed through it. On 
June 1st, 1915, it was deprived of hydrogen as completely as possible by heating and 
pumping off the gas. Air was then admitted to the drying train by the T-piece, F, 
and the apparatus was left over a week-end. When examined again not a trace of 
gas could be found in the pump.- The air in the train was next replaced by nitrogen 
and the palladium was heated for five and a half hours continuously by the maximum 
current used in the actual determinations, with the same negative result. Finally, 
after the apparatus had remained for six weeks with nitrogen at atmospheric pressure 
outside the tube, the interior was still vacuous. 
Preparation and Purification of Oxygen: First Method. 
Oxygen was first prepared by electrolysis of barium hydroxide solution in the 
U-tube already described. The gas, after being dried over calcium chloride (fig. 1, 
Q, Q), and phosphoric oxide, R, Id, entered the 2|--litre storage bulb, S, till this was 
filled to atmospheric pressure. The storage bulb had been previously exhausted in 
the same manner as the hydrogen bulb. On immersing the small fractionating 
tube, T, in freshly made liquid air, more than a litre of oxygen could be withdrawn 
from the storage bulb over the phosphoric oxide in the tube, V, and condensed 
as a liquid. If desired, the storage bulb could then be re-charged by continuing 
the electrolysis, and a further quantity of gas could be condensed in T. About 
one-third of the liquid oxygen was allowed to escape into the air through the 
gauge, W, so as to eliminate traces of hydrogen or other low-boiling impurities. The 
next fraction was taken for the determination. A portion of the residue was returned 
to the storage bulb and the last fractions were boiled away into the air. The gas 
finally remaining in the apparatus was removed by the pump. 
Second Method. 
Oxygen as prepared above might possibly contain traces of hydrogen, though the 
solubility of this gas in boiling oxygen is certainly very small.* 
It was deemed advisable, therefore, to make oxygen by some other method which 
could not conceivably yield hydrogen as an impurity. Gas generated by heating 
potassium permanganate should fulfil this condition, and was consequently employed 
to check the results obtained with electrolytic oxygen. 
Each of the four tubes, A, B, C, D (fig. 3), was charged with 250 gr. of finely 
powdered crystals of “ pure ” potassium permanganate in a dust-free atmosphere. The 
permanganate in the fifth tube, E, was three times re-crystallised from boiling water. 
The final crystallisation and subsequent drying were carried out in the decomposition 
tube itself. We may mention here that there was no difference in the results 
