ELASTIC FORCES OF AQUEOUS VAPOUR. 563 
by refraction, in consequence of the unevenness of the blown 
tube of glass. It was for the purpose of discovering these de- 
viations, and correcting them when necessary, that the two baro- 
meters were adapted to a long vertical metallic scale divided 
into centimetres. The distances between these divisions were 
read off with the cathetometer, before the tube of glass was in- 
terposed ; the same was done when the tube was placed round 
the barometers and filled with water, and it was observed 
whether the distances remained the same. In those cases where 
the distances were but slightly different, it was easy to keep ac- 
count of the displacement, and to make the correction for the 
observed height. The tube forming part of my apparatus pre- 
sented very great irregularities of this kind, so that I judged it 
prudent not to use it: the corrections which must have been 
made for the heights observed were frequently considerable, and 
gave rise on that account to a still greater uncertainty. 
I replaced the tube of glass by one of iron plate of the same 
diameter, cylindrical at the bottom, and terminated at the top 
by a square case, two opposite sides of which were formed of 
flat glass, chosen for the purpose. The divisions traced upon 
the scale then retained the same distances when viewed directly, 
or through the water of the tube. 
For raising the temperature, a furnace is placed under the iron 
vessel, and from time to time the water agitated in the tube ; when 
the temperature approaches that point at which the observation 
is to be made, the greater part of the charcoal is removed, and the 
water is continually agitated in the tube, until the thermometer 
placed at the height of the barometric chambers ceases to rise ; 
a considerable time elapses before this takes place, and the dif- 
ference of height in the two columns is measured at the moment 
of its maximum. In order to make a second experiment, the 
glowing charcoal is again placed in the furnace and the ascent 
of the thermometer again determined; when this is effected the 
charcoal is removed, and the operation repeated as before. 
Thus two determinations at very nearly the same temperature 
are obtained. 
This method gives very accurate results for temperatures that 
_ are the same or a little higher than the surrounding tempera- 
ture ; but it ceases to do so at temperatures a little more ele- 
vated. The water then separates so easily into layers having 
different temperatures, that a uniform temperature can only be 
VOL. IV.— PART XVI. 2s 
a“ 
