246 



Prof. J. A. Ewing. 



[Mar. 11, 



III. " Effects of Stress and Magnetisation on the Thermo- 

 electric Quality of Iron." By Professor J. A. Ewing, B.Sc, 

 University College, Dundee. Communicated by Sir 

 William Thomson, F.R.S. Eeceived February 24, 1886. 



(Abstract.) 



This paper comprises a revised version of one submitted to the 

 Royal Society in 1881, under the title " Effects of Stress on the 

 Thermoelectric Quality of Metals, Part I,"* along with much new 

 matter. It deals principally with the cyclic changes of thermoelectric 

 quality which an iron wire undergoes when exposed to cyclic variations 

 of stress (described in the abstract of the former paper), and with 

 the relations of these changes of thermoelectric quality to the changes 

 of magnetism which also occur as an effect of stress. Stress was 

 applied by exposing the wire to longitudinal pull by means of loads. 

 The changes both of thermoelectric quality and of magnetism exhibit 

 that tendency to lag behind the changes of stress to which in a 

 previous paper j the author gave the name hysteresis, and the effects 

 are sufficiently similar in regard to the two qualities to suggest that 

 the changes of thermoelectric quality occur as secondary effects of 

 changes of magnetism. To examine whether this is the case, simul- 

 taneous measurements of the magnetic and thermoelectric effects of 

 stress in an iron wire were made, and also independent observations of 

 the thermoelectric effects of magnetisation, without change of stress. 

 A comparison of these made it clear that stress causes change in 

 thermoelectric quality of iron directly, and not as a secondary effect 

 of magnetisation. If the wire be completely demagnetised to begin 

 with, and kept clear of all magnetisation during the application and 

 removal of stress, the presence of hysteresis is not less marked than 

 before. Experiments are given to show how the thermoelectric effects 

 of stress are modified by the existence of more or less magnetisation 

 in the wire ; and conversely, how the thermoelectric effects of mag- 

 netism are modified by the existence of more or less constant stress. 

 The influence of vibration in destroying the effects of hysteresis is 

 investigated, and also the result of exposing the wire to the process of 

 demagnetising by repeated rapid reversals of a continuously dimi- 

 nishing magnetising force, and it is shown that this process acts in 

 the same way as vibration in destroying the effects of hysteresis. 

 Residual effects of hysteresis are studied, as, for example, the diffe- 

 rence which presents itself when a wire is magnetised after having 



* Published in abstract in "Proc. Key. Soc," No. 214, 1881. 

 f " Proc. Roy. Soc," No. 216, 1881, p. 22. 



