260 M. G. Kirchhoff on the History of Spectrum Analysis 



above consideration proves that the proposition is possibly, but 

 not that it is necessarily true. 



The proposition under consideration is subsequently not car- 

 ried out so generally as it is enunciated in the above-cited words, 

 but only for a very special case. In connexion with these words 

 he says, " and I shall select for the proof a definite form 

 and description of enclosure — the conclusions arrived at ren- 

 dering it highly probable (if not rigidly demonstrated) that the 

 same adaptation will hold good for every enclosure however 

 irregular or varied." The case which he considers is that of a 

 body limited only by one plane with a black surface placed op- 

 posite and parallel to the plane. In the proof which Stewart 

 in this case gives for his assertion, there is, lastly, an error 

 which is rendered evident in the results at which the author 

 arrives. He concludes, " We have, therefore, two laws neces- 

 sary to the equilibrium of temperature : 1st, that the absorp- 

 tion of a particle is equal to its radiation, and that for every 

 description of heat ; 2nd, that the flow of heat from the interior 

 upon the surface of a substance of indefinite thickness is pro- 

 portional, ceteris paribus, to its index of refraction, and that 

 for every description of heat/' This second law (corresponding 

 to the equation which, as already mentioned, expresses a possible 

 property of the internal radiation of a body) is not correct ; the 

 magnitude therein considered is not proportional to the index of 

 refraction but to its square*. On a later occasion f Stewart 

 himself states (without, however, noticing his former contra- 

 dictory assertion), " Now if R denote the radiation of lamp- 

 black, and p, the index of refraction of an uncrystallized medium, 

 it may be shown that the internal radiation as thus defined is 

 equal to Up?" 



4. At the close of Prof. W. A. Miller's paper on coloured flames, 

 before alluded to, he states, "It may be interesting to remark, 

 in connexion with the speculations on the absorptive action of 

 the sun's atmosphere, that if solar light be transmitted through 

 a flame exhibiting well-marked black lines, the lines reappear 

 in the compound spectrum, provided the light of day be not too 

 intense compared with that of the coloured flame. This may be 

 seen in the red light of the nitrate of strontia, and less perfectly 

 in the green light of the chloride of copper. It would there- 

 fore appear that luminous atmospheres exist in which not only 

 certain rays are wanting, but which exercise a positive absorp- 

 tive influence on other lights." In his lecture " On Spectrum 



* See Kirchhoff, Untersuchungen, &c. 2te Ausgabe, Berlin, 1862, S. 37. 

 t Report of British Association for the Advancement of Science, 1861, 

 p. 107. 



