THE COMBINING VOLUMES OF HYDROGEN AND OXYGEN. 
423 
with great care. They were mounted on capillary tube, and, in the case of the 3-way 
taps, had the two bores drilled diagonally and at right angles to each other. 
Experience has shown this to be the safest type for preventing leak. All taps through 
which gas passed were mercury cupped. The plugs were so finely ground that a very 
small quantity of grease sufficed to lubricate them, and they were kept well “ home ” 
in their barrels by means of rubber bands. When in use, the plug was turned through 
the smallest possible arc, a precaution which tends to prevent the appearance of striae, 
particularly when the tap is in service for a long time. The grease used was the 
well-known rubber-paraffin-vaseline mixture, and, contrary to the experience of 
GermannC 13 ), we found that oxygen had no detectable effect on its appearance or 
properties. No leak between the interior of the apparatus and the air was ever 
observed, though the point was frequently tested. Nitrogen was tested for directly 
by exploding down about 1200 c.c., of hydrogen and 600 c.c. of oxygen, the quantities 
being adjusted so that hydrogen was in slight excess. The hydrogen residue, of 
about 3 c.c., was transferred to a small explosion eudiometer and mixed with 2 c.c. of 
oxygen. After passing a spark and introducing a small quantity of alkaline pyrogallol 
solution, about 1 c.mm. of gas remained unabsorbed. In other experiments hydrogen 
and oxygen residues were sparked over water or potash ; in no case was any diminution 
in volume noticed. 
The presence of nitrogen in either gas might give rise to traces of oxides of 
nitrogen, and eventually nitric acid, as a result of the explosions, or to ammonia as a 
result of the sparking. Oxides of nitrogen (except nitric oxide), nitric acid, and 
ammonia would remain with the ice in the explosion vessel. The water formed left 
no residue on the glass or mercury surfaces when evaporated; was odourless, tasteless, 
and neutral in re-action, and, when tested for nitric acid and ammonia, gave negative 
results. Traces of nitrogen, or other gases, in the hydrogen, provided that they were 
left unaltered by the explosions and sparking, would introduce no error, since they 
would be measured as part of the hydrogen residue. Any such impurity in the 
oxygen, on the other hand, would make the hydrogen residue too great by an amount 
equal to three times the volume of the original impurity. 
The presence of carbon compounds in either gas would almost certainly lead to the 
formation of carbon dioxide in the explosion vessel. Since the vapour pressure of 
carbon dioxide, at the temperature of the acetone, solid carbon dioxide mixture used 
for freezing out the water, is more than 40 mm. of mercury, this gas would escape 
with the hydrogen from the explosion vessel to be retained and detected in the liquid- 
air-cooled spiral. 
We have no direct proof of the completeness of synthesis, but the hydrogen 
residues were sparked at different pressures and for varying periods of time without 
affecting the results. The measured volume of hydrogen left in the pipette at the 
end of a determination was usually removed by the mercury pump, but in 
experiments 7, 8 and 9 of series 4, it was transferred to the explosion vessel and 
