566 [May 7, 



From (II.) we learn that any number of metals with their ends at 

 the same temperature may be introduced without effect, or 



[a, 6] t + [b, c-] t =[a, c] t ...... (IV.) 



This equation will always be true if 



[,y],-H.-M* .... (V.) 



whence we may write (III.) 



or, in other words, the electromotive force of a couple may be consi- 

 dered as the difference of the electromotive force of two metals, each 

 of which is found by subtracting its tension at the higher tempera- 

 ture from that of the lower one. 



Everything therefore depends on a knowledge of the value of what 

 may be called the electric tension of each metal at the various tem- 

 peratures. This for every metal is a function of temperature, and 

 may be called, in the language of the paper, a function of the nature 

 (or name) of the metal and the temperature. 



(The nature of the metal may be altered otherwise than chemi- 

 cally.) 



If the temperature of the metal vary in any way throughout its 

 length, then if it be homogeneous, the electromotive force will depend 

 only on the temperatures of its extremities. 



In a circuit of one metal, the author considers that at the junction 

 of the ends there may be a real discontinuity of temperature while 

 there is a continuity of electric current. He regards the explanation 

 of the effect by the stratum of air between the unequally heated ends 

 to be unsatisfactory. Mercury, as is known, will not produce 

 thermo- currents in this way. The author considers that the texture, 

 &c., as well as the chemical nature of the substance, influences the 

 value of the thermo-electric function. He also shows the possi- 

 bility of the thermo-electric inversions first discovered by Professor 

 Gumming. 



