576 REGNAULT ON THE 
hilk, were read off simultaneously by means of two cathetome- 
ters and two observers. This precaution is essential, particularly 
when the tensions are rather considerable, because the air con- 
tained in the tube hi of the manometer acts as an air-thermo- 
meter; if the observations were made successively, notable er- 
rors might arise from the displacement of the menisci produced 
by slight changes of temperature occurring in the surrounding 
air. 
In the greater number of the foregoing experiments, the co- 
lumn of mercury which serves to measure the pressure is heated 
to the same temperature as the vapour of which the tension is 
to be measured, and sometimes the space in which the vapour 
is eliminated is in easy communication with the tube containing 
the mercury ; that is especially the case with the apparatus de- 
scribed at page 563. Now, if within the limits of temperature 
at which we are operating, the tension of the vapour of mercury 
is a notable quantity, this tension will be added to that of the 
liquid under examination and render the determinations inexact. 
It therefore becomes necessary to determine in a direct manner 
the elastic force of mercurial vapour between 0° and 100°. I 
have shown at page 572 that sufficiently exact values could be 
obtained for the tension of aqueous vapour, even when the ba- 
rometric tube which measures the pressure is at a temperature 
8° or 10° below that of the globe in which the vapour is formed. 
I have had occasion to prove the same fact with essence of tur- 
pentine, even for still greater excesses of temperature. It is 
probable, that in this case, it is the air remaining in the appa- 
ratus which, pressed back into the barometric space, transmits 
the pressure exhibited in the globe. 
I thought that this method would be still more applicable to 
the determination of the elastic force of mercurial vapour be- 
tween those limits of temperature at which it is extremely weak. 
I introduced into the globe of the apparatus, figs. 1 and 2, with 
the barometric tubes outside the vessel V V', about 300 grammes 
of very pure mercury, and produced a vacuum a great number 
of times, warming the mercury to dry the apparatus ; finally, I 
My objection applies to the case of an air-thermometer compared with a mer- 
curial thermometer (t. vi. p. 373). It appears to me that this method ought to 
be rejected for low temperatures, which are so easily obtained stationary by 
liquid baths. M. Magnus would certainly not have met with such great differ- 
ences in his experiments on the tension of aqueous vapour at low tempera- 
tures, if he had made use of liquid baths, 
