Dr. Hare's Improved Eudiometers. 73 
Let us suppose that this Eudiometer has been thoroughly 
filled with mercury, the sliding rod being drawn out to its 
greatest extent; and that it is firmly fixed over a mercurial 
eistern in the position in which it is represented in the draw- 
ing, the little funnel-shaped part at the bottom, descending 
into the fluid to the depth of half an inch. Above this part 
is seen a cock (C), the key of which, in addition to the per- 
foration usual in cocks, has another, at right angles to, and 
terminating in, the ordinary perforation. When the lever 
(L) attached to the key of this cock is situated as it 
is seen in the drawing, the tube containing the sliding 
rod communicates with the recipient, but not with the mer- 
cury of the reservoir. Supposing the lever moved through a 
quarter of a circle. to the other side of the glass, the tube in 
which the rod slides, will communieate at the same time with 
the recipient and the reservoir. By means of the gauge- 
eock (C), the passage between the gauge and the recipient is 
opened or shut at pleasure. 
As subsidiary to this Eudiometer, another is provided with 
a rod and graduation exactly similar,* excepting its being 
shorter. (See fig. 5.) 
* In order to ensure accuracy in the measures of gas, made by the 
subsidiary Eudiometer, it is necessary to attend to the following pre- 
cautions. In the first place, the instrument must be proved air tight, 
and free of air bubbles, by the means prescribed already in the case of 
the Eudiometer for water. (See note, page 70.) The presence of air 
bubbles is always indicated by the extent of the vacuity which appears, 
when the glass recipientis held uppermost, and which disappears when it 
is held lowermost, the weight of the mercury acting upon the elasticity 
of the tubes, always causes a minute change; but by the smallest bub- 
ble of air, the effect is very much augmented. The Eudiometer should 
be introduced into the vessel whence the gas is to be taken, about ten 
per cent. more than is necessary being drawn in by opening the orifice 
and duly drawing out the red. The Eudiometer being lifted from the 
mercury, with as little change of position as possible, the rod may be 
adjusted accurately to the point desired. A momentary opening: of the 
orifice causes the excess to escape. The gas thus measured and in- 
cluded, is then easily transferred to the principal Eudiometer, by in- 
troducing the apex of the subsidiary instrument under the funnel, (see 
P, fig. 4.) opening the orifice, and forcing the sliding-rod home. 
Vou X.—No. 1. i6 
