166 



Olio of the most useful theorems in tliermodyiiamies is ;i sort of supph^- 

 ment to the Second Law. and is (hie to Xornst. It states that the entropy 

 of a substance at the absolute zero of temperature is zero. This theorem is amply 

 justified by experiments on speeific heats, thermo-neutral ehemieal reactions, 

 and so forth. Boltzmann has shown that if the encrp:y be subdivided into a 

 large number of ecjual parts a cjuantity can be calculated, by means of the 

 theory of probability, which is proportional to the entropy as deduced with 

 the aid of Nernst's theorem, the proportionality factor depending ujjon the 

 magnitude of the elomentar}- amounts of energy. Some value of this amount 

 should give the value of the entropy exactly. This value, according to Planck, 

 would be the "elementar-quantum." Xernst's theorem maj' then be con- 

 sidered as another ground for belief in the basic truth of Planck's theory.'" 

 Additional evidence in favor of the theory is to be found in the phenomena 

 of fluorescence, the photo-electric effect, and others. 



We shall now examine Planck's assumptions and attempt to interpret 

 them physically. We know that Hertzian waves only differ in wave-length 

 from the luminous waves emitted by an incandescent bod}-. What is then 

 more natural than to assume that the atoms of bodies contain tiny Hertzian 

 resonators, or oscillators? We say "atoms" because the spectral lines of an 

 element appear in the spectra of its compounds. The "perfectly reflecting 

 walls" may be nothing more than a useful mathematical fiction, or may rep- 

 resent true physical bountlaries corresponding to the cell walls of Larmor's 

 elementary receptacles of energy. The assumption of a mechanical model is 

 of vastly less interest than those implying at least an atomic structure of 

 energy. If there are atoms of energy do t hey preserve their identity? Are 

 they invariable? Planck assumes a discontinuous emission but a more or 

 less continuous absorption of energy. May we not ask the <iucstiiin: "\vv 

 these disconl inuit ies due to the oscillator or to the energy itself"? If the 

 energy exists in discrete (piant it ies, why is it not absorbed as well as emitted 

 in multipl(>s of the "element ar ([Uanlum"? If absorbed in discrete amounts 

 are these identical with those emitted? .\j)parently not, for in the plieiKiniena 

 of fhujrescence the lludresceni light is ne.-iriy always of lower' fre(iuency than 

 (lie light which causes the (luorescence. or, speaking in terms of Planck's 

 hypotiiesis, the emergent (iu;mta are smaller than the incident (iuaiit;i, for 

 (he so-called iilnii/s af cnirtiii ;ire Larger in piiipiirl ion to the fre(|iicncy of tiie 

 light, so that "atoms" coi-rcsixinding to lilue light ;ire larger than those cor- 

 responding to red liiiht. if, then, we are to account for lluoresceni'c by me.ans 

 "Prejss. .\k;i(l. Wi.s. li.-lin, .Si:/,. Mcr., 4. pp. iVy\\\ lllll. 



