REPORT ON RADIANT HEAT. 
297 
plate but an assemblage of facts, or alleged facts, determined 
with more or less accuracy; few indeed with any great pre¬ 
cision, many resting upon very vague evidence, and in several 
instances the results of different observers exhibiting a wide 
discrepancy or even direct contradiction : whilst, with very few 
exceptions, any general laws can hardly be said to be established 
with that certainty which can substantiate their claim to be 
received as legitimate physical theories. 
In offering suggestions for the advance and improvement of 
this branch of science, the first and most essential point to 
which attention ought to be directed, is the improvement, or 
rather invention, of the means of obtaining accurate indications 
of radiant heat, down to its most minute and feeble effects. In 
reference to this point, good determinations are much wanted of 
the degree to which the expansion of the bulb influences the 
accuracy of air thermometers. The improvement of mercurial 
thermometers so as to produce an instrument of extreme sen¬ 
sibility to the minutest effects of heat, is an object the attain¬ 
ment of which would probably be more important than that 
of any other means for accomplishing the end in view. But 
other methods founded on good principles should be diligently 
sought for and tried ; for example, it might be matter of in¬ 
quiry whether we could render available to this purpose the 
incipient melting or softening of some substances by a very 
slight increase of heat, or the evaporation of volatile liquids. 
But it is more particularly desirable that the instrument of 
MM. Nobili and Melloni should be tried, and a precise exami¬ 
nation set on foot of its real accuracy and the causes of error 
to whose influence it may be liable. This is the more neces¬ 
sary from the very remarkable character of many of their 
results ; whilst the alleged sensibility of the instrument, as 
they describe it, is such as almost to exceed belief. 
When we shall have succeeded in obtaining that prime re¬ 
quisite, an unexceptionable measure of minute effects of radiant 
heat, we may then proceed with some hopes of success to ex¬ 
amine the points on which there at present prevails so wide a 
discrepancy between different experimenters. 
The polarization of heat is perhaps the question which of 
all others requires the most extreme sensibility in our thermo¬ 
meter, or rather tliermoscope, in order to its satisfactory deter¬ 
mination. It may be tried either directly with the simple heat 
from nonluminous hot bodies ; or with luminous sources, with 
and without a glass screen, comparing the total compound result 
with that due to the transmissible part or heating power of light 
alone, and thence deducing the part due to simple heat. The 
