65 
1913-14.] Inclination of Curves of Thermoelectric Force. 
ing the mounting of a fresh and freshly annealed wire) was the accompanying 
curve, the figures for which appear in Table I., obtained, in which only one 
of the thirty plotted points lies appreciably wide. 
Table I.— Table showing Loads carried by Strained Wire (Mean Diameter 
0-682 mm.) and Corresponding E.M.F.’s, reduced to Mean Temperature- 
Range of 75° C. June 4, 1913. 
Load (kg.) . 
E.M.F. (volt x 10 6 ) 
0 
0-600 
0-5 
2-101 
1 
3-459 
1-5 
4-732 
2 
5-932 
2-5 
7T20 
3 
8-261 
3-5 
9-318 
4 
10-300 
4-5 
11-289 
5 
12T41 
5 5 
12-920 
6 
13-718 
6-5 
14-471 
7 
15-092 
7-5 
15-649 
8 
16-101 
8-5 
16-299 
9 
16-833 
Q-F, 
17-080 
10 
17*211 
10-5 
17330 
11 
17-363 
11-5 
17-302 
12 
17-178 
12-5 
16-970 
13 
16-661 
13-5 
16-272 
14 
15-800 
14-5 
15-260 
It may be remarked, with reference to this curve and the table of values, 
that the loads given as abscissae do not include the weight (175 gm.) of 
the weight-pan. This accounts largely, though not entirely, for the fact 
that “zero” load on the curve shows an E.M.F. of 0-60 xlO -6 volt; the 
unavoidable manipulation of the annealed wire in mounting it (and 
consequent slight strain) gave rise to an observed E.M.F. at (true) zero 
load of about 0*066 X 10 -6 volt, which accounts for the remainder of 
the discrepancy. During the experiment — which lasted some eight con- 
secutive hours — it was, of course, impossible to secure that the temperatures 
inside the steam jacket and cold-water jacket which surrounded the 
two junctions of the couple should remain constant. Besides the unavoid- 
able small variations due to the replenishment of the boiler and other 
incidental parts of the experiment, there comes into play a very con- 
siderable thermoelastic cooling effect* at the junctions as the load is 
increased. Thus, though these junction-temperatures, which were read 
between every two galvanometer readings, proved fairly steady in the 
neighbourhood of 97° C. and 21° C. respectively, varying at most by about 
half a degree, for greater accuracy all the observations of E.M.F. used in 
plotting the curve were reduced by interpolation to a mean temperature- 
difference of 75° C. The reduction was easily performed from a series of 
temperature-E.M.F. curves for different constant loads, prepared in connec- 
tion with the original experiments. 
Examination of the Curve . — The curve, when drawn to a large scale 
* Loc. cit ., pp. 36-43. 
VOL. XXXIV. 
5 
