516 Dr. Herschel's Experiments on the solar, and 



*' than visible rays, and that its report therefore ought not to be 

 " depended upon/' Now, although it does not appear from the 

 148th experiment that such a supposition can have much foun- 

 dation, yet, since those experiments were not made with a view 

 to ascertain the proportion of heat contained in each part of the 

 prismatic spectrum, we cannot lay so much stress upon them 

 as the accuracy which is required in this case renders necessary. 

 Let it therefore, contrary to our 100th experiment, be admitted, 

 in order to explain the phenomenon of the red glass, which 

 stops so much light and so little heat, that all the heat which 

 it intercepts consists entirely of the rays which are visible, and 

 that every one of the invisible rays of heat is transmitted. 

 Then will 999,8 intercepted rays of light be equal to 606" rays 

 of heat; and the remaining 394 will be the number of rays 

 we are now to place to the account of the invisible heat which is 

 transmitted. 



Having thus also got rid of this difficulty, we are next to 

 examine how other facts, collected in the same table, will agree 

 with our new concession. A violet-coloured glass, for instance, 

 stops 955 rays of light ; these, at the rate of 999,8, or say 1000, 

 to 606, must occasion a deficiency of 5J9 rays of heat. But, 

 by our table, this glass stops only 489 of them ; and there will 

 thus be 90 rays of heat left unaccounted for. To enhance 

 the difficulty, this glass, by our 164th experiment, stops also ± 

 of the supposed 394 invisible rays, which will amount to an 

 additional sum of 98. And our 111th experiment shews, that 

 actually a great number of these rays, that otherwise cannot be 

 accounted for, come from the store of heat, the rays of which 

 are of the refrangibility of red light. 



A dark-blue glass stops 801 rays of light ; these, if light and 



