24 Dr. Hare’s improved Eudiometers. 
being introduced to its hilt, and the capillary orifice open, in 
consequence of the pressure of the finger on the lever by which 
it is usually closed. Being thus prepared, let us suppose that 
it were desirable to analyse the atmosphere. Draw out the 
rod 200 measures; a bulk of air, equivalent to the portion of 
the rod thus withdrawn, will of course enter at the capillary 
opening; after which the lever must be allowed to close it. 
Introduce the recipient into a bell-glass of hydrogen, and 
opening the orifice draw out the rod about 100 degrees; close 
the orifice, and withdraw the instrument from the water. 
Apply the projecting wires (WW) severally to the solder in 
the two furrows in the beam (fig. 2) communicating with the 
poles of the four calorimotors ; then raise the jug so as that it 
may receive one of them, and subject it to the acid. By the 
consequent ignition of the wire, the gas will explode. The 
instrument being plunged again into the water of the pneu- 
matic cistern, so that the capillary orifice, duly opened, may 
be just below the surface; the water will enter and fill up 
the vacuity caused by the condensation of the gases. The re- 
sidual air being excluded by the rod, the deficit will be equi- 
valent in bulk to the portion of the rod remaining without; 
and its ratio to the air subjected to analysis may be known by 
inspecting the graduation, 
In the case of the gaseous mixtures above described, the 
deficit has, in my experiments, been 126 measures. Whereas, 
according to the theory of volumes, it ought to be only 120. 
But I have not as yet operated with hydrogen purer than it 
may be obtained from the zinc of commerce; and some allow- 
ance must be made for the carbonic acid of the air, which may 
be condensed with the aqueous vapour produced by the oxy- 
gen and hydrogen. . 
In the invaluable work on the Principles of Chemistry, lately 
published by Dr. Thomson, it is suggested, that in order to 
obtain correct results in analysing the air with the hydro-oxy- 
gen eudiometer, more than 42 per cent of hydrogen should 
not enter into the mixture. I am not as well satisfied of the 
correctness of this impression, as I am generally with the re- 
sults of the wonderful industry and ingenuity displayed in the 
work above mentioned. 
If oxygen is to be examined by hydrogen, or hydrogen by 
oxygen, we must of course have a portion of each in vessels 
over the pneumatic cistern, and successively take the requisite 
portions of them, and proceed as in the case of atmospheric 
air. 
B (fig. 1) represents a glass with wires inserted through small 
tubulures, in the usual mode for passing the electric spark, 
should 
