316 RADIANT HEAT. 



to the direct influence of a luminous hot body, and comparing it with 

 the ratio similarly observed when a glass screen was interposed. 



The screen acquiring, and therefore radiating, heat from the first 

 moment of the experiment, will affect the thermometers in a ratio (as 

 before observed) differing little from equality; and these equal quan- 

 tities added to the terms of the ratio of the direct effects of the 

 luminous body, will, of course, diminish the inequality of that ratio. 

 This cause of error may not have operated to any great degree, but 

 its tendency is obviously to a diminution of the ratio. 



Notwithstanding this, the observed result in all cases with a lamp, 

 or with iron raised to a "bright red heat, was, that the ratio of the 

 effect on the black to that on the white thermometer was increased by 

 the interposition of ihejkcreen. 



A summary of the results of two sets of experiments, (conducted 

 with some slight variation,) and in the second of which the tempera- 

 ture acquired by the screen was carefully noted, is as follows : 



Rise of thermometer (centig.) in 1 mia. 

 Glass screen. No screen. 



( (1 ) 1° 2 

 Iron bright hot < L'( ft "1 



Argand lamp • • 



White Black. White. Black. 



(1.) 1°.25 2°.75 7°.0 8°.75 



... 1 .25 2 .95 3 .75 



(1.) .6 2 .0 1 .8 3 A 



(2.) 1 .3 2 .35 - : 2 .35 3 .2 



These numbers are the means of several repetitions. 



The necessary conclusion from ||iis difference in the ratio of the 

 direct and screened effects is, that the portion of heat which has the 

 property of permeating the screen, has also the property of affecting 

 the two surfaces in a ratio different from that in which the part inter- 

 cepted acts upon them. 



As in researches of this kind great numerical precision is unattain- 

 able, I was especially, at every step of the inquiry, anxious to devise 

 as many variations of the experiment as possible; these all tended to 

 confirm the results just given. 



Thus I used a large differential thermometer having its bulbs dif- 

 ferently coated, and exposed each of them in turn to the luminous source 

 of heat, the other being completely screened, and invariably found 

 the ratio of the effects on the black and white bulbs considerably 

 greater when affected only by the transmissible part of the heat than 

 when exposed to the whole. As before, the part added on the remo- 

 val of the screen was of a nature tending to add to the terms of the 

 former ratio quantities in a ratio much nearer equality, viz: that which 

 the effects of simple radiant heat would give when acting respectively 

 on the two bulbs. 



Other variations of the fundamental experiment were as follows: 



A differential thermometer having one bulb black was exposed to 

 the radiation from luminous hot bodies, first with and then without 

 the interposition of a glass screen, the same position being preserved. 



If the screen had no influence, it is evident that in whatever pro- 



