Physical Constitution of the Sun. 353 



If in the present observations we may neglect the slight vari- 

 ations observed by Pouillet in the specific heat of platinum at 

 high temperatures, the observed variations of the quantities of 

 heat developed are proportional to the variations of temperature 

 of the wires. The relation between the emissive power of a 

 body and its temperature, for homogeneous rays, could then be 

 represented by an analogous function, and thereby the determi- 

 nation by experiment of KirchhofFs function in the way indi- 

 cated, at least in its dependence on the temperature, be facilitated. 



For practical objects, however, of the determination of the 

 temperature of opaque incandescent bodies in an optical way, 

 even a theoretical knowledge of that function would not be ne- 

 cessary, it would be sufficient to determine for several kinds of 

 rays, in as different parts of the spectrum as possible, the con- 

 stant in question, by which the light-emission as an exponential 

 function is connected with the temperature of the incandescent 

 body. Since this function must be the same for all opaque black 

 bodies, independent of their other qualities, only the photome- 

 tric comparison of the part in question of the spectrum with the 

 analogous place in the spectrum of another opaque body of 

 known temperature would be requisite, in order to deduce from 

 the observed ratio of intensity, with the aid of that function, 

 the unknown temperature of the second body. 



To give a clear idea of the course to be pursued, suppose the 

 following arrangement of the apparatus. In front of the slit of 

 a spectroscope two right-angled reflection- prisms are fixed with 

 the reflecting faces directed to opposite sides, so that the rays 

 from two sources of light, one on each side of the spectro- 

 scope, are spread out into two adjacent spectra. One of the 

 sources is a galvanically incandescent plate of platinum, for 

 which, in like manner as in the above-described experiments, the 

 constant of the function of the light-emission is empirically de- 

 termined for a defined class of rays. Some direct calorimetric 

 determinations of the temperature with different degrees of in- 

 candescence could without difficulty be effected under the given 

 conditions, for a first approximation, to form the basis of the 

 method. If now the other half of the slit be illuminated by an 

 opaque incandescent body of sufficient extent, so that the appa- 

 rent surface fills up the entire opening of the slit, the photome- 

 trically determined ratio of intensities of the two spectra at the 

 places in question will, with the aid of the function ascertained 

 for the platinum plate, immediately determine the temperature 

 of the second body. For the photometric variation of the quan- 

 tities of light the polarization principle would be suitable. For 

 this purpose the rays from each of the two sources to be com- 

 pared could, before they enter the reHection-prisms, be caused 



Phil. Mao. S. 4. Vol. 46. No. 307. Nov. 1873. 2 B 



