283 



easier transmitted by salt scratched by sand-paper in two directions 

 at right angles, than luminous heat. Thusi, a plate of salt which, 

 when well polished, transmits 92 per cent, of heat derived from a 

 lamp, and sifted by a glass plate, and also 92 per cent, of heat 

 wholly unaccompanied by light, transmitted, when roughened, only 

 17 per cent, of the former and 45 per cent, of the latter. 



" A thin plate of mica, when similarly scratched with emery-paper, 

 so as merely to depolish it, transmitted much more nearly the same 

 per-centage of heat from different sources than when bright ; shew- 

 ing, that the loss of polish affects the transmission of the more re- 

 frangible rays much more sensibly than that of the others. 



" Yet this effect is not attributable to a variation in the ratio of 

 the reflection of heat of different kinds at the surfaces of the plate. 

 For, in thejirst place, I have proved, and already communicated 

 the fact to the Royal Society (see Proceedings for April 1839), 

 that reflection takes place at a polished surface, with almost, if not 

 exactly, the same intensity for all kinds of heat ; and, secondly, I 

 have found, by direct experiment, that, at least for the higher angles 

 of incidence, reflection is most copious from rough surfaces for heat 

 of low temperature, or the same kind which is most freely trans- 

 mitted, proving incontestibly that the stifling action of rough sur- 

 faces is the true cause of the inequalitv. 



" That there is a real modification of the heat in passing through 

 a roughened surface, as well as through laminated mica and the 

 smoky film, appears from direct experiments which I have made on 

 the heat sifted by these different media ; which, when transmitted 

 by any one of these, is found in a fitter state to pass through each 

 of the others ; and this modification is found to be more perceptible, 

 as the character of the heat is more removed from that which these 

 media transmit most readily, that is, as the temperature of the 

 source is higher. Thus, heat derived from a lamp, has 36 per cent, 

 transmitted by a certain smoked plate of rock-salt. But if the 

 heat transmitted by the smoked salt has previously been sifted or 

 analyzed by transmission through another plate of smoked salt, 

 through laminated mica, and through roughened salt, the per- 

 centage is raised from 36 to 44 in the two former cases, and to 

 40^ in the latter, proving incontestibly the specific action of these 

 transmissions in arresting the more refrangible rays. 



" I next considered, that, as a moderate number of scratches ap- 

 peared to produce this modification, it might be practicable to ob- 

 tain the effect by transmitting heat simply through fine wire-gauze. 



