MAGNETISATION ON THE THERMOELECTRIC QUALITY OF IRON. 363 
This is in agreement with observations published as long ago as 1856 by Sir W. 
Thomson, who has laid the foundation of the subject in Part III. of his great series of 
papers on the “ Electrodynamic Qualities of Metals.” (Phil. Trans., vol. 146 (1856), 
p. 709, or Reprint of Papers, vol. II.) 
§ 6. If, however, the application of load be continued up to and beyond the limit of 
elasticity of the metal, the negative effect passes a maximum, becomes much reduced, 
and generally even changes to positive before the wire breaks.'" 
A characteristic example of this action is given in fig. 1, Plate 21, which shows the 
thermoelectric effect of loading an annealed iron wire up to the breaking point. The 
initial electromotive force (which has a small negative value) is due to accidental defect 
of homogeneity in the wire. The maximum negative E.M.F. is reached at or near the 
limit of elasticity (12 kilos.), and the subsequent drawing out is associated with a 
rapid change towards positive. After the wire was broken the load was removed, and 
the broken ends joined. A strong jjositive deflection then showed that the part 
which had suffered longitudinal strain, and was now freed from stress, w 7 as thermo- 
electrically positive to the other portions. This result has also been previously 
pointed out by Sir W. Thomson! as well as by Magnus. J 
§ 7. Another experiment is given in fig. 2, where a a shows the effect of the first 
loading of another piece of annealed iron wire of larger section than the last. In this 
case the loading was stopped at 35 kilos., before the wire broke, and the load was 
gradually removed. The thermoelectric changes which took place during the removal 
are shown by the curve b b. During removal of load the E.M.F. again changed to 
negative, passed through a negative maximum, changed to positive, and finally, when 
the load was entirely removed, exhibited the positive value which Magnus and 
Thomson have remarked. 
The load was then gradually reapplied, and this resulted in the complex changes of 
E.M.F. shown by the curve c c —a curve by no means coincident with that found 
during the removal of the load. The loading was continued past 35 kilos. At 36 
kilos, a sharp change of gradient occurred as the wire (hardened by the previous 
stretching) began again to draw. At 38 kilos, the wire broke. 
§ 8. Fig. 3 shows the results of another experiment of the same kind, in which another 
piece of the same iron wire was stretched (after annealing), first by the application 
and removal of 35 kilos., then of 40, and finally broken, in the third loading, by a 
stress of 43 kilos. The thermoelectric currents were observed for every kilogramme 
of load, during application and removal. In the first loading (a a) while the wire 
was still soft the currents were irregular, but at 30 kilos., when the wire began to 
* Of. a paper by Mr. G. W. von Tunzeuiann on “ The Production of Thermoelectric Currents in 
Wires subjected to Mechanical Strain,” Phil. Mag., ser. 5, vol. 5 (1878), p. 339, where this result has 
been previously stated. In other respects the present writer’s observations do not altogether agree with 
those of Mr. Yon Tdnzelmann, 
f Loc> cit ., § 109. \ Poggbnd. Annal., vol. 83 (1851), p. 469. 
3 A 2 
