93 



the obtaining of mechanical effect from heat, bjr means of a perfectly 

 reversible arrangement, depends in a definite manner on the trans- 

 mission of a certain quantity of heat from one body to another at a 

 lower temperature. There is a degree of uncertainty in the present 

 application of this principle, on account of the conduction of heat 

 that must necessarily go on from the hotter to the colder parts of 

 the circuit ; an agency which is not reversed when the direction of 

 the current is changed. As it cannot be shewn that the thermal 

 effect of this agency is infinitely small, compared with that of the 

 electric current, unless 7 be so large that the term B y^, expressing 

 the thermal effect of another irreversible agency, cannot be neglected, 

 the conditions required for the application of Carnot and Clausius's 

 principle, according to the demonstrations of it which have been 

 already given, are not completely fulfilled : the author therefore con- 

 siders that at present this part of the theory requires experimental 

 verification. 



1. A first application of the theory is to the case of antimony and 

 bismuth ; and it is shewn that the fact discovered by Seebeck is, 

 according to equation (c), a consequence of the more recent discovery 

 of Peltier referred to above, — a partial verification of the only 

 doubtful part of the theory being thus afforded. 



2. If y denote the quantity of heat evolved, [or— © 7 the quantity 

 absorbed] at the surface of separation of two metals in a compound 

 circuit, by the passage of a current of electricity of strength y across 

 it, when the temperature t is kept constant ; and if <p denote the 

 electromotive force produced in the same circuit by keeping the two 

 junctions at temperatures t and f', which differ from one another by 

 an infinitely small amount, the magnitude of this force is given by 

 the equation 



p=0/A («' — <) id) 



and its direction is such, that a current produced by it would cause 

 the absorption of heat at the hotter junction, and the evolution of 

 heat at the colder. A complete experimental verification of this 

 conclusion would fully establish the theory. 



3. If a current of electricity, passing from hot to cold, or from cold 

 to hot, in the same metal produced the same thermal effects ; that 

 is, if no term of 1 a^ depended upon variation of temperature from 

 point to point of the same metal ; we should have, by equation (a) 



