586 REGNAULT ON THE 
absolutely stationary ; the position which the meniscus then occu- 
pies is exactly the mean of those extreme positions which it took 
during its very small oscillations. The barometer dipping into 
the same cup does not exhibit sensible oscillations, so that it can 
be conveniently read off after the observations on the meniscus. 
A second observer reads at a distance with a telescope the 
thermometers immersed in the retort; errors of parallax are 
thus avoided in the reading. 
In this manner a certain number of determinations were made 
at the same pressure, allowing between two consecutive deter- 
minations an interval of eight or ten minutes to elapse ; it was 
thus easy to perceive the perfect steadiness of the temperatures 
indicated by the thermometers for the same pressure, and to be 
convinced, that the least change in the pressure was indicated 
immediately by a corresponding variation in the temperature. 
For obtaining a still greater pressure in the apparatus, the 
stop-cock 7 was opened with caution, and a quantity of air ad- 
mitted sufficient to establish the desired pressure. 
Thus the temperatures at which water boils were obtained 
successively from the lowest pressures up to that of the atmo- 
sphere. This latter was measured by the barometer attached to 
the apparatus, by the operation described at page 570. 
For ascertaining the temperatures at which water boils under 
higher pressures than those of the atmosphere, the apparatus 
was arranged in the manner shown at fig. 6; that is ‘to say, the 
measuring apparatus of fig. 2 was replaced by that of fig. 8, and 
the tube ¢¢' was brought into connection with a forcing-pump. 
The tube mn is about. 4 metres in length and 14 millimetres in- 
terior diameter ; it is fixed to a long board kept perfectly ver- 
tical. Three thermometers with large bulbs, placed at convenient 
distances, show the mean temperature of the column of mercury 
in the tube mn. 
By means of the forcing-pump the requisite pressure is esta- 
blished in the apparatus. This pressure is measured by the 
height of the barometer, added to the height of mercury repre- 
senting the difference between the mercurial levels in the tubes 
mn and pg. The determination of this last height requires 
some precautions if great exactitude is aimed at. In the actual 
position of the apparatus both columns of mercury oscillate in 
the tubes mn and p q. These oscillations take place only within 
very narrow limits, and they are often only perceptible by means 
of the telescope of the cathetometer, which magnifies consi- 
