Fig. 1. 





12 







1 





o 



R\ 





/s 



s 









-ITT - 



it (> 



-1 



Twist on the Thermo-electric Qualities of Iron. 227 



gathered from observations with small loads ; but only a few 

 of such were made — made simply for the sake of sounding, 

 so to speak. 



In fig. 1 is given a series of observations I obtained early 

 in the course of my investigation. It 

 seems to establish the existence of the 

 phenomonon of thermo-electric hyste- 

 resis with respect to twist (in my own 

 ease, combined with a certain amount 

 of longitudinal tension). 



Starting from the axis of zero twist, 

 we have an ascending trend and then a 

 maximum point — a feature which also 

 exists in soft iron once stretched per- 

 manently and subject to the influence of increasing longi- 

 tudinal tension. This is evidently a result of the interaction 

 between the hysteresis and the actual and immediate effect 

 of twist. As the amount of twist is increased, the latter soon 

 overpowers the former, causing the curve to descend. On 

 being untwisted the wire seems to respond, yet with the 

 hysteresis in action. The return branch falls below the direct, 

 continues to ascend even beyond the axis of zero twi^t, and 

 not far from it attains again a maximum point, which is 

 almost symmetrically situated with respect to the axis against 

 the first maximum. Further untwisting produces the same 

 state of things as on twisting. In short, the whole appear- 

 ance of the curve is very similar to that of the well-known 

 curve of magnetization and twist. 



Jt is hardly necessary to mention after what has been said 

 above, that as several sources of disturbance were but imper- 

 fectly eliminated in my experiments, some curves I obtained 

 were not so good as the one just cited ; for example, the maxi- 

 mum points were both found on the 

 positive or both on the negative side of 

 the axis of zero twist. It sometimes 

 happened that one of the cycles gone 

 through in succession was shifted 

 relatively to the other (fig. 2). This 

 may be partly explained by the fact of 

 my having merely taken scale-readings. 



It is to be regretted that I did not 

 try many samples of soft iron to see 

 whether the hysteresis phenomenon 

 does exist in them all. But as the 

 samples I tried were taken at random, it seems exceedingly 

 improbable that the phenomenon is confined to them alone. 

 {b) Such, 1 believed, might be taken as the normal type for 



Fijr. 2. 



