388 DISTRIBUTION OF HEAT IN THE SPECTRUM. [MEMOIR XXVIII. 



od of investigation, on principles which I will now ex- 

 plain. 



For the sake of clearness, restricting our thoughts for 

 the moment to the more familiar case of the visible spec- 

 trum, if we desired to ascertain the true distribution of 

 heat, would not the proper method be to collect all the 

 more refrangible rays into one focal group, and all the 

 less refrangible into another focal group, and then meas- 

 ure the heat that each gave ? If the view currently re- 

 ceived be correct, would not nearly all the heat observed 

 be found in the latter of these foci, and little, if indeed 

 any, be found in the former ? But if all the various re- 

 gions of the spectrum possess equal heat-giving powers, 

 would not the heat in each of these foci be the same? 



Let us give greater precision to this idea. Using Ang- 

 strom's wave-lengths the length at the line A is 7604, 

 and that at H 2 3933, and these lines are not very far 

 from the less and more refrangible ends of the visible 

 spectrum respectively. The middle point of this spec- 

 trum is at 5768, which may therefore be called its optical 

 centre. This is a little beyond the sodium line D, which 

 is 5892. Now, if by suitable means we reunite all the 

 rays between 7604 and 5768 into one focus, and all the 

 rays between 5768 and 3933 into another focus, are we 

 not in a position to determine the true distribution of 

 the heat ? Should the heat at these two foci be sensibly 

 the same, must not the conclusion at present held be 

 abandoned ? 



If in these investigations the rays of the sun be used, 

 it is necessary to restrict the examination to the visible 

 spectrum, excluding the invisible red and invisible violet 

 radiations. On these the earth's atmosphere exerts not 

 only a very powerful but a very variable action, and, 

 what is still more, an action the result of which we can- 

 not see, so that we are literally working in the dark. 



