into the tube until it is about one fourth filled. The volume of
the gas is then read off on the scale and reduced to standard
conditions. Approximately an equal volume of pure oxygen is then
introduced and the volume again read off and reduced to standard
conditions. This gives the total volume of the two gases. From this the
volume of the oxygen introduced may be determined by subtracting from
it the volume of the hydrogen. The combination of the two gases is now
brought about by connecting the two platinum wires with an induction
coil and passing a spark from one wire to the other. Immediately a
slight explosion occurs. The mercury in the tube is at first depressed
because of the expansion of the gases due to the heat generated, but at
once rebounds, taking the place of the gases which have combined to form
water. The volume of the water in the liquid state is so small that it
may be disregarded in the calculations. In order that the temperature of
the residual gas and the mercury may become uniform, the apparatus is
allowed to stand for a few minutes. The volume of the gas is then read
off and reduced to standard conditions, so that it may be compared with
the volumes of the hydrogen and oxygen originally taken. The residual
gas is then tested in order to ascertain whether it is hydrogen or
oxygen, experiments having proved that it is never a mixture of the two.
From the information thus obtained the composition of the water may be
calculated. Thus, suppose the readings were as follows:

Volume of hydrogen taken 20.3 cc.
Volume of hydrogen and oxygen 38.7
Volume of oxygen 18.4
Volume of gas left after combination has taken place (oxygen) 8.3

The 20.3 cc. of hydrogen have combined with 18.4 cc. minus 8.3 cc. (or
10.1 cc.) of oxygen; or approximately 2 volumes of hydrogen have
combined with 1 of oxygen. Since oxygen is 15.88 times as heavy as
hydrogen, the proportion by weight in which the two gases combine is 1
part of hydrogen to 7.94 of oxygen.

~Precaution.~ If the two gases are introduced into the eudiometer in the
exact proportions in which they combine, after the combination has taken
place the liquid will rise and completely fill the tube. Under these
conditions, however, the tube is very likely to be broken by the sudden
upward rush of the liquid. Hence in performing the experim