562 REGNAULT ON THE 
A dry and a wet barometer were plunged into the same cup 
of mercury, and both kept in contact with a flat metallic rod, 
perfectly vertical and divided into centimetres. The cup of mer- 
cury was placed in a cast-iron vessel 26 centimetres in diameter, 
the bottom of which was covered with a layer of mercury to the 
height of 6 or 7 centimetres, and which could not mix with the 
mercury in the cup containing the barometers. A tube of glass 
23 centimetres in diameter and 1 metre high surrounded the 
two barometers. This tube was filled with water, which dis- 
placing the mercury in the iron vessel, obliged it to rise into the 
annular space between the exterior of the glass tube and the 
interior of the iron vessel. A mercurial thermometer dipped into 
the water of the tube at the height-of the barometrical cham- 
bers; and lastly, an ascending agitator, furnished with flaps, 
allowed of motion being communicated to the water in the tube 
throughout its whole height. The iron vessel was placed upon an 
iron tripod, under which a furnace could be placed, for the pur- 
pose of gently raising the temperature of the water. The ten- 
sion of the aqueous vapour was denoted in every instance by 
the difference in the height of the two columns of mercury, a 
correction being made for the small layer of water in the wet 
barometer, which covered the top of the mercury. 
The water introduced into the barometer should be completely 
deprived of air. To effect this object, the barometer having 
been well-boiled, and whilst still hot at the contracted end, I 
poured out a small quantity of mercury and supplied its place 
by boiling water which had been kept during half an hour in a 
state of ebullition. The finger was immediately placed over the 
opening and the barometer turned round. To be quite certain 
that the condition was actually fulfilled of introducing water free 
from air, the barometer was left upright until the following day, 
then rapidly inclined in order to cause the disappearance of the 
barometric vacuum: if ever so little air remained in the cham- 
ber, a small bubble was observed, which remained for some time, 
until the water had again dissolved it. It was only after many 
fruitless trials that I was enabled to obtain moist barometers 
perfectly free from air. 
The difference in height of the two columns was measured 
by means of a cathetometer; the use of this instrument may 
give rise to considerable errors, if an endeavour is not made to 
correct the deviations which the rays of light are subjected to 
