646 BIILI, SYSTEM TECIIXIC.IL JOUkX.il. 



obtaiii (L — k'l). to Ix- falk'd ecj), as tlu- artiial i'iK'rg\' fxi)i-n(li'il in 

 l)ultiiiv; the electron across the boiindar\- of the metal.* 



In tlu- [process which I ha\e just described, the iiipul of heat con- 

 sists (il llu- following; Icrnis: XL which ji"^'^ '" extract the A' electrons 



uch Is hlierat(.-(l 



from the metal in the Inst chamber, — X\ L + -.'<tl')\\ 



^ a 1 ' 



when .V electrons condense into the metal in the second chamber. 

 and —iVirfZ' which is absorbed 1)>' the electrons in tra\ellini; throuuli 

 the wire. The output of work is .V^7" during the expansion of ihe first 

 chamber, — NkT — NkdT during the contraction of the second cham- 

 ber. The input of entropy is XL T during the evaporation in the 



fu'st chamber. 



X\L 7'+ ",,,, )(/7' durinu the condt'iis.ition in the 

 ^ dl ' 



second ch.unlH'r. and ( —.V.? 7")f/7' dnrini^ the flow of electrons through 



the wire. 



We now cnm|)Kii' tlu- cycle b\' changing llu' pressure and tem- 



peratiu'e of the gr.unme-molecule of electron-gas in the fu'st chamber 



from p, T to p-\-dp. T-\-dT, after which it becomes equi\-alent with 



the gramme-molecule in the second chamber at the beginning of 



the ]>roccss. Calculated in the usual way — isothermal contraction 



at 7' from p to p+dp, isobaric exjiansion at p+dp from 7' to T+dT 



—we lind: input of heat, [^^ Xkd'L - Xk'r\d{ln p) dl'\d'l'\ output of 

 icork. XkdT~XkT\d{ln p) dT\d'J'; input of entropy. [llXk T)dT- 



xk\d(in p) dr\dr. 



The two processes together constiluli' a cumplete rt'\fr>ible c\-cle. 

 We therefore e<|uate the sum of the inputs of enlroi)\ to zero, and 

 obt.iin : 



<'« d<t> , o d(ln p) ,, 



and ec|uate the difference of the inputs of he.ii .nid the oiiiputs of 

 work to zero, which gi\es: 



-V*--v+o*=fJ (17) 



'This definiti()n sugKosts a llurni.il jiuiIumI oI mrasiiriiiK /-, wliiili li.is several 

 limes iH'en piil into practiie. The experiments are <lirtiriill ami ihe data imisl 

 l)e lorrec ted fur many iiUliiences, hut the Ix-st results indicate that tL-kT) is ap- 

 proximately equal ti. (•/. (if (,S|. Ihe d.il.i of Darisson and (■ernier indicate a slight 

 difTerencc. wliicli m.i\ lie .in inipiirl.ini lest of snj,'Ki'sled theorii-s of cdndiiilinn. 



