concerning Volta's Contact Force. 361 



it, without the need for any local performance of work or any 

 corresponding local E.M.F."* 



I should be glad to know whether any objectors to what I 

 may call the Maxwellian view — the view of the precise local 

 correspondence between E.M.F. and reversible heat-effects — 

 do seriously base their opposition on this ground of a real 

 physical specific heat of electricity. The only other alter- 

 native that I can see (short of Helmholtz's theory, to be 

 mentioned later) is to say that the metallic junction is 

 really a chemical junction too, that the energy does not 

 take the form of heat because it takes that of chemical 

 combination, and to point to a weighable quantity of alloy 

 or other compound produced at that place by the passage of 

 a current. Whether or not any chemical compounds are 

 really formed at a given junction, then becomes a simple 

 question of fact. 



Since a voltaic circuit never has less than three junctions, and since 

 we are trying to apply to it as much of the thermoelectric laws as it will 

 stand, it may be well to write down these laws for the case of a metallic 

 circuit with any number of metals, and therefore with a thermoelectric 

 diagram having discontinuities at the several junctions, — everything dis- 

 continuous but temperature. Each metal, referred to the standard, has 

 a certain thermoelectric power at each temperature, and accordingly the 

 resultant E.M.F. of the circuit is 



E = cyclejp^. 



We may also express the E as a sum of the E.M.F.'s at the several 

 localities, the junctions on the one hand, and the temperature-gradients 

 in the metals on the other ; so that 



E = s(n) -f 2(e). 



Further, as regards the n function of each metal relatively to the 

 standard 



JJqu = P a T everywhere ; 

 and also 



©« = J <r a dt ; 



so, maintaining this reference to the standard, it is still true that 

 E = 2jn^ZlogT. 



But Tlab, though still equal to P a &T, is no longer equal to TdE/dT, 

 unless by E is meant E a fi, that is the E.M.F. of a non-existent small 

 portion of circuit formed solely of the two metals A and B. 



The fact is, that at a junction there are now three values of P and of n, 

 viz. Uoa, n &, n a b ; and it is only when we constantly maintain the 

 reference to the standard metal o that we can say E=2 ( TLdlogT; other- 

 wise there are discontinuities not allowed for. We cannot say that 



f "?p + f 8 ^+ f 4 5^ + . . ,, + pHp 



* Cf. Phil. Mag. March 1886, pp. 271-276. 

 Phil Mag. S. 5. Vol. 49. JNo. 299. April 1900. 2 C 



