342 Dr. Ure’s new experimental researches 
“ be attended with no little difficulty. It would be better t&> 
6 ‘ have several thermometers suspended at different heights, 
“ in the body of the water, and to take the arithmetical mean 
M of their indications. Or otherwise, which would probably 
s< be more exact, we might employ a thermometer having a 
“ cylindrical bulb, equal in length to the column of vapour. 
“ It would then be necessary that the column of water should 
“ rise sufficiently above this vapour to allow the thermo- 
“ meter bulb to be equally immersed, or we must make on 
“ its indications the small correction mentioned p. 59, in 
“ order to reduce the temperature of the cylinder of mercury, 
“ to the temperature of the reservoir. The employment of 
“ such a thermometer may appear at first sight sufficiently 
difficult, since it seems that the length of the cylindrical 
“ reservoir must be very considerable, if the elastic force of 
the vapour be great.* ” 
He then proceeds to show how this difficulty may be ob>* 
viated (as indeed it had previously been by Mr. Dalton), by 
taking barometer tubes successively shortened, as the force 
of the steam is augmented by heat. He proposes to use four, 
between the freezing and boiling points of water, each being 
two decimeters, or nearly 8 inches long, and the thermometer 
bulb having also that length. The plan which I imagined, as 
it completely obviates the source of errors arising from the 
large and variable space occupied by the vapour, supersedes 
the necessity of employing M. Biot’s singular remedy. It 
likewise avoids other complications, introduced by the heat- 
ing and consequent elongation of the mercurial column itself 
* Traite de Physique. Tome I. p. 268. 
