558 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



On the Effect of bending German- Silver Wire. 



As we had used, in the research on the Thomson effect in iron, 

 spirals of gerraan-silver wire wound on a cylinder 1 era. in diameter, 

 assuming that the bending produced no important change in the 

 thermo-electric performance of the wire, investigation of the effect 

 of this particular curvature seemed now especially appropriate, even 

 if somewhat belated. 



Four pieces, each 150 cm. long, of just such annealed german-silver 

 wire as had been used in making our thermo-electric couples, were taken, 

 and each was mounted on a glass tube about 1cm. in diameter and 

 16 cm. long in the manner indicated by Figure 4. C and C are copper 

 wires lashed to the cork K^ and soldered to the german-silver wire, 

 which, beginning at A, enters the tube through a hole in the side and 

 runs straight to the farther end, where it emerges under the cork K 2 



Figure 4. 



to follow a spiral course to B. The distance between turns of the 

 spiral is about 0.3 cm. 



Each tube thus prepared was made to connect two metal vessels con- 

 taining a thick oil, each end extending through a tight-fitting cork, K 3 

 or K 4 , about 5 cm. into the oil. One body of oil was kept at or very 

 near the temperature of the room; the other was heated to 15 or 20 

 degrees above this temperature. 



The two junctions with the copper being near together in one body of 

 oil were very nearly at the same temperature. The arrangement really 

 gave, then, a thermo-electric couple consisting of straight german-silver 

 wire combined with spirally wound german-silver wire in one continuous 

 piece, the junctions of this couple being in the two vessels of oil with 

 a difference of temperature of some 15 or 20 degrees. 



With this condition of things, tests of the four couples, which we 

 may call S 1? S 2 , S 3 , and S 4 , respectively, were made on two days, 

 October 6 and 7. On each day S x and S 2 gave deflections in one direc- 

 tion and S 3 and S 4 gave deflections in the opposite direction. Dividing 



