and Polarization of Heat. 139 
quantity, and half of that may be estimated as a sensible impres- 
sion, we may measure an effect of z}> of a centigrade degree, 
and perceive (by unassisted vision), an effect of 745. 
14, In the case of the moon’s rays, concentrated 3000 times, 
we have seen that it is improbable that even the last effect was 
produced. The whole sensitive extremity of the pile being larger 
than the moon’s image, was not brought into action; but if we 
compare their relative dimensions,* we shall still find that it is 
improbable that the direct light of the moon would raise a ther- 
mometer one three-hundred-thousandth part of a centigrade degree, 
at least in this climate. 
15. The value of the thermo-multiplier consists not so much in 
the minuteness of its indications, which may easily be equalled 
by employing large enough thermometers, but in the certainty 
and rapidity of its action. Air thermometers, such as I compared 
it with, though the size of the balls was inconsiderable, required 
so long a time to assume their temperature, that, when exposed 
simultaneously with the thermal pile to the source of heat, the 
latter had almost assumed its maximum effect before the others 
had sensibly moved ; and it is obvious that, in delicate experi- 
ments, where constancy in the producing cause is presumed, ra- 
pidity of execution is essential. In short, with an air thermo- 
meter (which requires from 10 to 15 minutes to give a single re- 
sult), the greater part of the experiments to be described would 
have been impossible from this cause alone, and the remainder 
would have been tedious beyond measure. It will therefore be 
conceived that were thermometers enlarged so as to give as mi- 
nute indications as the multiplier, they would be utterly unma- 
nageable. 
16. Of all the researches of M. Meixonr on radiant heat that 
* The moon’s image contained 0.114 square inches, whilst the area of the pile is 
about 0.40. Hence little more than a fourth of the pile was brought fully into ac- 
tion ; but any dispersed light (for which we have made allowance), would act on the 
neighbouring parts. 
