560 PROCEEDINGS OF THE AMERICAN ACADEMY. 



mark of the meter-rod. The result was usually a small deflection of 

 the galvanometer needle, which deflection remained fairly constant until 

 the ice was withdrawn, when it soon disappeared. The ice was then 

 placed opposite the 26 cm. mark and the deflection was noted. In this 

 way progress was made, hy 1 cm. stages, until the 75 cm. mark was 

 reached. Then a return set of observations was made, beginning at the 

 75 cm. mark and ending at the 25 cm. mark, with 1 cm. stages. 



For any one point the return observation usually showed a deflection 

 of the same sign and the same order of magnitude as the deflection 

 shown by the forward observation. The mean of the two deflections was 

 taken for each point examined, and the mean deflections, or quantities 

 proportional to them, were plotted on co-ordinate paper, distances along 

 the wire being measured on a horizontal line and the deflections being 

 represented by vertical distances above or below this horizontal line. 



M r« so *» M 



Figure 5. 



FlGUBE 6. 



Through the points thus placed a curve was drawn free hand, which 

 indicates roughly the variations of thermo-electric condition revealed by 

 progressive observations along the wire. Cj in Figure 7 is the figure 

 thus obtained with an unannealed specimen, which we will call No. 1, of 

 wire from the same piece which has furnished the german-silver of our 

 couples. The maximum height of this curve, which occurs over the 

 12 cm. point, corresponding to the 37 cm. point of the wire as shown in 

 Figure 5, indicates an electromotive force of 16 X 10~ 8 volt, approxi- 

 mately, as a result of touching the wire with ice, when the room 

 temperature was about 18° C. 



The curve in question crosses the median line, of no electromotive 

 force, 22 times in the 50 cm. of wire investigated. There was no 

 obvious reason for these alternations of thermo-electric effect, or indeed 

 for any thermo-electric effect. The wire, in the part tested, was free 

 from any serious looking kinks or other imperfections. 



