THE COMBINING VOLUMES OF HYDROGEN AND OXYGEN. 
401 
When the washing and drying train had been set up it was exhausted by a 
water pump from three T-pieces simultaneously at N, O and P. Each of the 
permanganate tubes was then heated in turn, the train being exhausted between 
each oxygen filling. When all air had been removed the T-pieces were sealed off. 
After leaving the drying train the gas passed through the tap, Q, to the fractionating 
tube, R, where it was liquefied. When a sufficient quantity of liquid had been condensed 
the tap, Q, was closed, and a portion of the liquid was allowed to boil off and escape 
through the gauge, S. The second fractionating tube, T, which had been previously 
exhausted, was now cooled in liquid air, and the 3-way tap, Y, was opened. On 
lowering the liquid air vessel from It a small quantity of liquid oxygen soon condensed 
in T. The 3-way tap was then turned so that the gas bubbled through its liquid. 
The object of this was to prevent superheating so that any higher-boiling impurities 
would be more effectively retained. A quantity of gas, washed by its own liquid, was 
allowed to escape by the gauge, W, or was removed by the pump. The next fraction, 
bubbled in the same way through its own liquid, was taken for the determination, and 
the residue in both fractionating tubes was rejected. 
The Measuring Apparatus. 
It may be taken as established by earlier workers that the ratio of the combining 
volumes is slightly greater than 2 when the gases are measured at normal temperature 
and pressure. Our procedure, in brief, has been to explode one volume of oxygen 
with rather more than two volumes of hydrogen, the variable and carefully measured 
excess over two volumes being sufficient in quantity to ensure, a residue of hydrogen 
being left after synthesis. The gases were measured consecutively in the same 
apparatus. By working with hydrogen in excess, the possibility of the formation of 
such substances as ozone, hydrogen peroxide, or oxide of mercury, should be diminished 
or eliminated. 
The final operation was to estimate the hydrogen residue, when the combining- 
volumes could be calculated. 
For measuring the gases at 0° C., and 760 mm. pressure, the method was the same 
as that used by Gray and Burt( 10 ) for the volumetric analysis of hydrogen chloride. 
In fact, some parts of the original apparatus w'ere kindly lent to us by Dr. Gray. 
The measuring pipette (fig. 4) consisted of a thick-walled glass bulb, A, of about 
300 c.c. capacity, sealed to capillary tubes at either end of its vertical diameter. 
The lower capillary expanded into the dead-space, B, which was furnished with a 
glass point. The upper capillary, after a right-angle turn, led to the 3-way tap, C. 
The pressure of gas in the bulb was registered by the vertical distance between the 
mercury surface in the dead-space and the mercury surface in the upper chamber, D, 
of the manometer, which communicated with the dead-space as shown in the figure. 
The upper chamber of the manometer, of the same diameter as the dead-space* 
VOL. ccxvi.— a. 3 i 
