MEMOIR XXIV.] THE ELECTRO-MOTIVE POWER OF HEAT. 333 



viation of the needle fell now to 165, being a deficit of 

 11. In this experiment care was taken that no heat 

 should be transmitted along the wire to the connecting 

 cups. 



The same was repeated with an iron wire of the same 

 length and under the same circumstances. The current 

 at first being 90, as soon as the spiral was made red-hot 

 it fell to 61, being a deficit, therefore, of nearly one third 

 the whole amount. 



To the increased resistance to conduction, occasioned 

 by an increased temperature, we are to impute the slight 

 rise of tension observed in thermo-electric currents. The 

 quantities are of the same order. 



We have next to show " that the quantity of electric- 

 ity evolved at any given temperature is independent of 

 the amount of heated surface, a mere point being just 

 as efficacious as an indefinitely extended surface." 



The quantities of electricity evolved by hydro-electric 

 pairs increase with the surfaces, but it is not so in ther- 

 mo-electric arrangements. A pair of disks of copper and 

 iron, two inches in diameter, were soldered together; 

 they had continuous straps projecting from them, which 

 served to connect them with the galvanometer cups. At 

 the boiling-point of water they gave 62; on being cut 

 down to half an inch in diameter, they still gave 62. 

 On the disk being entirely removed, and the copper 

 made to touch the iron by a mere point, its extremity 

 being roughly sharpened, the deflection was still 62. 



By means of a common deflecting multiplier, I ob- 

 tained the following results: 1st. A copper wire being 

 placed in a bath of mercury, the temperature of which 

 was 240 Fahr., I dipped into it a second copper wire, 

 the temperature of which was about 60 Fahr. ; the gal- 

 vanometer needles moved through 15. 



