224 MAGNUS ON THE EXPANSIVE FORCE OF STEAM, 
clining the tube above the mercury. Upon this the remaining 
water is again poured out of the open arm, and what cannot be 
poured out on account of the form of the tube sucked up by 
a loose piece of string. 
The portion of this arrangement projecting from the case is 
connected at c by means of caoutchouc with a glass tube fg hk, 
fig. 1, leading to an air-pump N M. When the air is rarefied, 
the vapour is able to overcome the pressure to which the water 
is subjected in the closed arm of the short U-shaped tube a 6 d, 
the mercury sinks in this and rises in the open arm. 
The air is then rarefied until the surfaces of the mercury in 
both the arms are situated, at least approximatively, in the same 
horizontal plane. The small difference in height that may still 
exist is measured by means of a telescope, A B, fig. 1. With 
this view the case S P R is furnished with two tubes at R and 
R’ of 12 inch in diameter, which traverse all the cases, and are 
closed at both ends with glass discs, so that it is possible to see 
through them with the telescope. The tubes a 6 d, in which 
the vapour is formed, are so fixed to the inner case that the sur- 
faces of the mercury lie exactly in the direction of these tubes. 
The pressure of the rarefied air is indicated by the barometer 
attached to the air-pump; in order, however, to measure it more 
accurately, I employed a siphon gauge or metre of pressure, 
which is represented in fig. 4. It consists of a tube o pg, curved 
in the form of a U, 3 foot high, and half-filled with mercury ; 
one of its arms is connected by tube mn, fig. 1, with the ex- 
hausted space, while the other is open at 0, or covered with a 
loosely-fitting cork. The difference in the height of the mer- 
cury in both arms was measured by means of the kathetometer, 
AB C, fig. 1. 
To maintain the temperature of the mercury in this siphon 
gauge everywhere equal, it is surrounded with the case H I K L, 
one side of which is of wood, the other three of plate-glass, 
and which is easily closed at the top by a lid. It contains two 
thermometers, one at the bottom, the other at the top, of which 
I took the mean. In the same case was moreover fixed a siphon 
barometer 7 s¢, which consequently had always the same tempe- 
rature as the gauge, and was equally read off by means of the 
kathetometer. 
The barometer had been compared with the normal barometer. 
Both the barometer and the gauge o pg had more than half 
.! 
“ 
