The method of analysing atmospheric air by means of these 
instruments is as follows. Supply the subsidiary eudiometer 
with its complement of hydrogen gas, by introducing the apex 
of the glass recipient into a bell-glass containing, over mercury, 
the gas in question, and drawing out the sliding-rod, the ori- 
fice being kept open only while above the surface of the mer- 
cury and inside of the bell. 
The gauge-cock (C, fig. 4) of the principal eudiometer being 
closed, and that which opens a communication between the 
recipient and the funnel (I*) open, and the instrument having 
been previously thoroughly filled with mercury, and placed 
over the mercurial cistern, as already mentioned, introduce 
into it, through the funnel, the gas which had been included 
in the subsidiary instrument (fig. 5); next shut off the com- 
munication with the mercurial cistern, re-establish those be- 
tween the recipient and the rod and gauge, and push the rod 
into its tube up to the hilt. The re-entrance of the rod, by 
raising the mercury into the recipient, forces the hydrogen in 
bubbles through the water of the gauge, and displaces all the 
atmospheric air which it previously contained. Now shut the 
passage to the gauge, open that which communicates through 
the funnel with the mercurial cistern, and draw out the rod to 
its utmost extent. Into the eudiometer thus situated and pre- 
pared, introduce successively 100 measures of hydrogen and 
200 measures of atmospheric air, by means of the subsidiary 
eudiometer: then closing the passage to the mercurial cistern, 
and opening the passage to the gauge, push in the rod until 
the water in the gauge indicates that the pressure on the gases 
included is equivalent to that of the external air. The gauge- 
cock being closed, the gases are ready to be exploded. The 
explosion is produced by galvanic ignition, as in the case of 
the eudiometer for water (fig. 1), excepting that instead of car- 
rying the eudiometer to the calorimotor, the circuit is esta- 
plished by lead rods severally attached to the galvanic poles 
by gallows and screws. (See g g, fig. 2.) One of the lead rods 
terminates in a piece of iron immersed in the mercury, the 
other is fastened to the insulated wire of the eudiometer, Un- 
D2 der 
