880 Thermodynamics of Radiation. 



less concordant on account of the feebleness of the radiation 

 to be measured, which is about 200 times less for rocksalt 

 than for quartz at 1000° 0. 



Fie:. 2. 



+5 







Walker ^^v. 



\ 







/ y 



CALLENDflR 











c ^5 s 





PlANCH 





\\ 







R/tYLE/OH^-^ 







250° 500 750° 1000" 



TeMPfR/ITUfff. CeMTIGRADE 



Difference of Rubens' Observations on Quartz Reststrahlen 

 from Planck's formula, compared with other formulae. 



The agreement of the proposed formula (8) with direct 

 experiments on radiation is seen to be satisfactory. As 

 indicated in the previous paper, the agreement with atomic 

 theory as regards (1) the number of atoms N = 6"12 X 10 23 in 

 a gramme atom, and (2) the atomic unit of energy per unit 

 frequency B&/N = 6*34 x 10 _2 >, is equally satisfactory, accord- 

 ing to the estimates of these quantities obtained from other 

 sources. The variation of specific heat at low temperatures 

 can also be represented by the thermodynamical formula 

 with fewer arbitrary hypotheses than by Planck's. These, 

 however, are questions involving many speculative elements, 

 and are of little weight compared with the thermodynamical 

 argument on which the formula is founded. The Doppler 

 effect must occur in the isothermal emission of an energy- 

 stream, and has not been considered in this connexion. That 

 it should lead directly to Carnot's expression T(dj^/dT) v for 

 the latent heat per unit volume, is too striking a confirmation 

 of the principles of the classical thermodynamics to be 

 disregarded. According to my view, it affords an additional 

 relation, which suffices, in conjunction with Wien's law, to 

 fix the distribution in full radiation. 



