104 Description of the Steam Pyrometer. 
serve the quantity of steam it produces in cooling down to 212°. 
The actual temperature of the liquid being known by observation, 
and the quantity of steam by weight, every other quantity of vapor 
given by different temperatures of the same standard mass, will be 
produced by a proportionate quantity of heat. It will be seen that 
this method of proceeding takes no account of differences in specific 
heat at different temperatures. It comes at once to a simple expres~ 
sion of the heating power of a body measured by a single effect of 
the heating principle, that of conferring the elastic form on water, 
already raised to the boiling pomt. 
It will readily be conceived that the question of specific heats, of 
expansion, and contraction, and of course the variable rates of ex- 
pansion at different temperatures might be wholly disregarded, if we 
had an invariable standard by which to measure the portions of heat, 
that may at any time be present in a given portion of matter. ‘The 
latent heat of vapor supplies this standard. The following are some 
of the different results which have been obtained by those who have 
made experiments on this subject. 
Latent heat in vapor. Determined by 
950° =e . Watt. 
945 - - Southern. 
1000 - - - Lavoisier and Laplace. 
1040.8 - - Rumford. 
955.8 - - - Despretz. 
above 1000 - - Thomson. 
1000 - - - Ure, (corrected result.) 
mean 954 
I have in the preceding calculations assumed the latent heat, at 
990°. Should the results of Dr. Ure, which appear to have been 
made in a manner as unexceptionable as any yet published, be con- 
firmed and established by other philosophers, the facility of making 
calculations such as I have above presented, will be increased and the 
usefulness of the principle in pyrometry more fully established. 
Nore.—The experiment on melted iron, on page 101, is offered 
chiefly as an illustration. 'The apparatus then at hand did not ad- 
mit of all the exactness which the case allows ;—still the result is 
believed to be nearly correct. 
