570 REGNAULT ON THE 
For making observations at higher temperatures, the glass is 
fitted into its frame E F G H, the vessel is filled with very clear 
water, and the operation conducted in precisely the same man- 
ner as in the first series, page 565. In order to estimate the 
prismatic deviations of the rays of light produced by the inter- 
position of the water and the glass, care must be taken, as in the 
first series of experiments at page 564, to make a mark upon 
the tube 40, and to divide the tube ah into centimetres. By 
this method the results recorded in Table II., Series A, B, C, 
were obtained. 
It is of importance that in all the experiments the barometer 
should be perfectly correct, and that the apparatus be so ar- 
ranged as to allow of comparison being easily made between the 
barometer 40 and a standard barometer, at any moment during 
the experiments. For this purpose, the board on which the 
barometric tubes are fixed has attached to it a piece of metal r 
in which moves a vertical screw v, terminated at both extremities 
by rounded points. This screw is turned until its lower extre- 
mity is exactly at the level of the mercury in the cup. The glass 
is not placed in its frame, and the distance from the level of the 
mercury in the barometer o 4 to the upper point of the screw is 
measured with the cathetometer. To obtain the exact height of 
the barometric column, it is only necessary to add to this distance 
the length of the screw between its two points, which can easily 
be done with the same cathetometer, by raising the screw to a 
convenient position. 
The delicacy of the barometer can likewise be tested in another 
manner, which is equally susceptible of great exactitude. By 
pouring a larger quantity of mercury into the cup U, or by 
lessening the quantity of that liquid, the capacity of the baro- 
metric chamber can be made to vary considerably ; if the baro- 
meter is perfectly free from air its height ought to be the same, 
whatever may be the capacity of the barometric vacuum; it is 
however no longer the same when the barometer contains the 
least trace of air*. 
The same apparatus is very convenient for determining the 
elastic forces of aqueous vapour at low temperatures; only in 
this case, I do not use the sheet-iron vessel V V! V", but in its 
* This last mode of verification has long since been shown by M. Arago, 
who made use of the same principle in constructing barometers that a traveller 
could fill upon the spot without haying occasion to boil the mercury. 
