at Electrodes to various Electrical Phenomena. 149 



The results also show that the cold platinum had 71*10 per 

 cent, and the hot platinum 44*08 per cent, less amount of 

 " surface-resistance *" as a cathode than as an anode ; also 

 that the total amount of such " resistance " in the two 

 directions in the circuit, when one of the junctions was heated 

 to 94° C, was 195*75 per cent, greater in the direction of 

 the thermoelectric current than in the opposite one. 



The results obtained with these two liquids show that 

 the amount of " surface-resistance " at each junction of such 

 thermoelectric couples varies very greatly with temperature ; 

 and that the amount of such " resistance " to the passage of 

 a current in either direction was more largely decreased by 

 rise of temperature with the dilute nitric than with the sul- 

 phuric acid. I had previously found with thermoelectric 

 couples composed of metals and electrolytes, " that by rise of 

 temperature ' transfer-resistance ' was usually and considerably 

 reduced/' (Proc. Eoy. Soc, 1885, No. 236, p. 210.) 



The direction of thermoelectric current produced in each 

 of these two cases was the same as that indicated by the 

 relative amounts of " surface-resistance " in opposite directions 

 at the heated junction, but the reverse of that indicated by 

 the relative total amounts of such " resistance " in opposite 

 directions in the entire circuit. 



As in each of the two couples, the total amount of " surface- 

 resistance"" in the circuit, when one of the junctions was 

 heated, was very much greater in the direction of the thermo- 

 electric current produced than in the opposite one, I conclude 

 that the direction of that current was not chiefly determined 

 by the alteration of such " resistance " by rise of temperature. 

 The thermoelectric action and the u surface -resistance '* 

 therefore appear to be largely distinct phenomena ; and as 

 both liquid and metal in those couples were therm oelectro- 

 positive substances, the direction of the current was probably 

 caused chiefly by the tbermoelectro-positive potential of the 

 liquid increasing more rapidly than that of the metal by rise 

 of temperature. We may also infer that the strength of the 

 thermoelectric current in those couples was very largely 

 increased by the great diminution of " surface-resistance."" 



In a paper on " Some Relations of Heat to Yoltaic and 

 Thermoelectric Action of Metals in Electrolytes " (Proc, 

 Royal Soc. 1884 ; No. 233, pp. 251-290), I have shown by 

 numerous experiments that " when metals in electrolytes 

 were heated they were more frequently rendered electro- 

 positive than negative, in the proportion of about 2*8 to 1*0 ; 

 the proportion in strong solutions was about 3*27 to 1/ J 



Second. — With regard to voltaic couples. As heat affects 



