HALL. — THERMO-ELECTRIC HETEROGENEITY IN ALLOYS. 545 



four such couples had been used for some time, with repeated heatings 

 of the junctions to a temperature of 218° in air (sometimes contaminated 

 by vapors of various kinds) the difference appeared to be greater, 

 amounting in one case to slightly more than 2 per cent. These couples 

 had been subject to many vicissitudes of handling; but no satisfactory 

 explanation of the increase in their range of sensibility can at present be 

 given ; it is possible that the first estimate of this range was too small 

 and the last too large. 



4. The german-silver wires used in these couples were about 1.5 m. 

 long. When one junction of such a couple was immersed to a depth of 

 50 cm. in air heated to the neighborhood of 180° C. and the other to a 

 like depth in air at a temperature near 140° C, the intermediate part 

 of the german-silver wire being exposed to the air of the room with no 

 other shield than the wall of a glass tube within which it was enclosed, 

 the mean electromotive force per degree of the couple sometimes changed 

 as much as 5 per cent when the temperature interval was reversed, the 

 junction which had been in the 180° place being put into the 140° 

 place and vice versa. All attempts to explain such discrepancies as due 

 to local electromotive forces in parts lying outside the german-silver 

 failed. Attempts to account for them by means of local peculiarities 

 in the german-silver wire led to the experiments described in the next 

 three sections of this summary. 



5. Four couples, each made from a single piece of annealed german- 

 silver by leaving one part of it straight while the other part was formed 

 into an elongated spiral wound on a cylinder about 1 cm. in diameter, 

 were tested. The mean electromotive force of the four appeared to be 

 about 0.07 per cent of the electromotive force of a couple made of 

 german silver and copper, the spiral wire being thermo-electrically a 

 little farther away from copper than the straight wire is. No great con- 

 fidence should be placed in the numerical accuracy of this result. 



6. Four couples, each made from a single piece * of annealed german- 

 silver by leaving one part of it unchanged while the other part was 

 stretched permanently 10 per cent, were tested. As a mean result it 

 was found that such couples have an electromotive force about 0.6 per 

 cent as great as that of a copper and german-silver couple, the stretched 

 german-silver being thermo-electrically farther removed from copper 

 than the unstretched wire is. 



1 These four pieces were from the same spool of wire (all called 18 per cent 

 nickel) as the others, but not from the same strand. 

 vol. xli. — 36 



