( fifi7 ) 



the volocilv of a.üilalioii incrc-isinu- w ilh ilu' hMiiporaliirc. Il is jo l)e 

 assiimeil tlial, iii (liis "lieal-inolion", evcrv oloch-oii (ravels alon<*- a 

 sti-aiühl line, iiiilil il sirikos against a |>arlicl(^ of llio iiiolal ; llic path 

 will llioi-ofbrc' he au ii'i'og'ulai' zigzag-lijio and, so long- as llicre is 

 no canst' (li-i\ing- llic cloclrons in a dcfinilc dii-cclion. an (dcincnl 

 of snrfaco will l)e traversed l»v (>(|ual nunihcrs of clocli-ims, ti'a\ellin<'- 

 to o|>|)osito sides. 'I'iiings \\ ill l»e dilferenl if tlie metal is (\\|)()so(l 

 to an electric force. 'J'lie motion of the electrons will still he an 

 irregular agitation; yet, motions in a deliuite direction will |»redo- 

 miiiatc, and this will show itself iji oni- ohserxations as an "electric 

 current." 



Now \\-e may infer from the relation hetween ahsorptioji and 

 emission that is re(|nired hy Kii{('iiMOi'i''s law, that the mechanism by 

 which llit^ emission of a hody is pi-oduced is the same as that to 

 Avhich it owes its ahsorhing power. It is thei-efore natural to e.xpect 

 that, if we conline ourselves to the case of great Avave-loigths, we 

 shall he ahle lo ex|ilain the emission of a metal \)y means of the 

 heat-motion of its tree electrons, Avithout recurring to the liypothesis 

 of ^-vibrators" of some kind, producing waves of detlnite periods. 



In the following pages this idea has heen worked out. After having 

 calculated the emissive |)Ower ^ve shall tijid that its ratio to the 

 ahsorhing power does jiot depend on the value of those (piantities 

 hy which one metal ditfers from anothei-. Accoi-ding to the law of 

 iviRCHHOFi', the result may he considei'ed as rejiresenting the I'atio 

 het-ween the emission and the ahsorj)tion for an arbitrarily chosen 

 hody, or as the emissive power of a pei-fectly black substance: it 

 w ill he found lo contain a certain constant (pianlit\ , w hose pli\ sical 

 meaning will aj>[)eai' from the theoi-y. 



§ 2. The ratio of which I have Just spoken is ijitimately comiected 

 with another impoi'tajit ]>hysical (pianlily, \ i/.. the density of the energv 

 of radiation in a space enclosed liy pei-fectly hla(d< walls, which ai"e 

 kept at a uniform absolute iem])erature 7'. If the electi'omagnelic 

 motions of which Ilie aetluM- in such a sp;u'e is the seat, arcMlecom- 

 posed into rays travelling in all directions, and each of which has 

 a deliinte w a\e-length, the energy pei- unit \'olume, in so fai' as 

 it belongs lo i-ays w ith w a\"e-lenglhs between ?. and / -\- </)., ma\' 

 be represeided by 



/•' (A 7') JL 

 /'' being a function whi(di many physicists have tried to d(>tenniiie. 

 Hoi/r/MANN and Wikn have shown by thermodynanncal reasoning 

 that the abo\e expression may be written 



