concerning Volt a' s Contact Force. 4b'9 



then the B aggregation there will be a force urging the 

 corpuscles to flow from A to B. 



In a gas the negative corpuscles are by far the most 

 mobile ; indeed the positive corpuscles seem unable to move 

 except with their atoms electrolytically. It may be that 

 even inside a metal the negative corpuscles are likewise by 

 far the most mobile, being readily handed on from one atom 

 to the next, this process constituting metallic conduction; 

 while the positive corpuscles remain attached to their atoms, 

 which except in an electrolyte or a substance with some 

 electrolytic properties are not subject to locomotion at all. 

 Assuming this mobility of corpuscles of one sign solely, they 

 are urged along a gradient of temperature in a metal for two 

 reasons : one a mechanical force equivalent to a dp/dx acting 

 on an element of volume, or \jn of this per corpuscle ; the 

 other an electrical force dY/dx acting on its electric charge e. 



When there is no current these forces must balance. 80 



idp__ ^y 



n dx dx ~~ ' 

 or 



dp — nedY. 



Now proceeding as usual without compunction to utilise 

 the gas- analogy 



p = nmRT, 

 we have 



nedY = mRd(nT) 

 or 



dY = m/e . RTdlog(nT). 



This is the E.M.F. that acts along the gradient of tem- 

 perature, giving rise to a reversible evolution of heat and 

 representing the coefficient of the Thomson effect, a, called 

 the specific heat of electricity in a metal, such that the work 

 done in transferring a charge q up a difference of temperature 

 dT\sqdY = qadT, or 



.= ^RT^log(nT). 

 Hence for a circuit of two metals a and b 



<r a —(Tb= ReT -Trfi log -p. 



di ° nl 



And by ordinary thermodynamics this is related to the 

 Peltier effect at the junction and to the resultant E.M.F. in 



