396 Prof. Draper on the Production of Light by Heat 



and further, that the color of the heat emitted depends upon the 

 temperature of the radiating source. It is one thing to investi- 

 gate the phenomena of the exchanges of heat-rays of the same 

 color, and another when the colors are different. A perfect the- 

 ory of the exchanges of heat must include the principle of ideal 

 coloration, and, of course, so too must a law of cooling applica- 

 ble to any temperature. 



There is another fact to some extent considered by Dulong and 

 Petit, but not of such weight in their investigations, where the 

 range of temperature was small, as in ours, where it rises as high 

 as nearly 3000° F. ; I mean the difference of specific heat of the 

 same body at different temperatures. At the high temperatures 

 considered in this memoir, there cannot be a doubt that the capa- 

 city of platinum for heat is far greater than that at a low point. 

 This therefore must control its rate of calorific emission, and prob- 

 ably that for light also. 



From these and similar considerations, we should be prepared 

 to discover that as the temperature of an incandescent solid rises, 

 the intensity of the light emitted increases very rapidly. 



foregoing reasoning. 



perimentai proofs 



The apparatus employed as the source of the light and meas- 

 ure of the temperature was the same as in the preceding experi- 

 ments, — a strip of platinum, brought to a known temperature by 

 the passage of a voltaic current of the proper force, and connect- 

 ed with an index which measured its expansion. . ~ 



The principle upon which I have determined the intensity o 

 the light is that first described by Bouguer, and recently introd ^ 

 ced by M. Masson. After many experiments I have been led to 

 conclude that this is the most accurate method known. 



Any one who will endeavor to determine the intensities of lign 

 by Rumford's method of contrasting shadows, or by that of equal y 

 illuminated surfaces, will find, when every precaution has bee^ 

 used, that the results of repeated experiments do not *J°. 

 There is moreover the great defect, that where the lights dltter 

 color it is impossible to obtain reliable measures, except by reso 

 ing to such contrivances as that described by me* y hts 



Bouguer's principle is far more exact: and where the ig 

 differ in color, that difference actually tends to make the ies^ 

 more perfect. As it is not generally known, I will indicate 

 nature of it briefly. f , j te 



Let there be placed at a certain distance from a screen ot ^ ^ 

 paper, a candle so arranged as to throw the shadow of a ru e,^ 

 other opake body, on the screen. If a second candle be p ^ 

 also in front of the paper and nearer than the former, there is 



* Phil. Mag., August, 1844 



