136 Professor Forbes on the Befr action 



tiplier their attempts at its solution *. But, like previous expe- 

 rimenters, they employed a metallic mirror to concentrate the 

 rays of the moon, which, acting in the usual manner of dispers- 

 ing the heat of the thermometer, produced so great a cooling 

 effect, as completely to neutralize any positive results. 



10. It occurred to me, however, from the consideration of 

 M. Melloni's very decisive experiments as to the permeability 

 of screens of different kinds to heat from various sources, that the 

 moon's heat must, in very great proportion at least, radiate 

 through glass. And this on several grounds ; as, 1 . because the 

 sun's heat, of which this may be considered as an integral part, 

 does so with scarcely any loss ; 2. because heat, accompanied by 

 light, always does so, and generally in proportion to the brilliancy 

 and refrangibility of that light ; and, 3. because the lunar rays ha- 

 ving passed through the whole thickness of the atmosphere must, 

 according to the experiments of De la Roche, fully confirmed 

 by Melloni, have parted with the greater part of that species of 

 heat most easily stopped, and hence arrive at the earth in a 

 state comparatively capable of passing through glass and si- 

 milar substances. If this opinion be correct (nor can I entertain 

 any doubt upon it), and if we substitute a lens for a mirror to con- 

 centrate the lunar rays, we shall profit by all, or nearly all, of their 

 heating effect, whilst such a lens, instead of promoting the radia- 

 tion of the heat of the thermometer to the sky, will entirely stop 

 it (because heat of this description does not pass sensibly through 

 the thinnest glass), and thus its disturbing influence will be en- 

 tirely prevented. 



11. I employed, therefore, a polyzonal lens made by Soleil 

 of Paris, in my custody, to concentrate the moon's light. The 

 diameter of the lens is 30 inches; its focal distance about 41 

 inches, whence we may compute the size of the lunar image to 

 be a circle 0.38 inch in diameter. Comparing this with the di- 



* Annales de Chimie et de Physique. December 1831. 



5 



