HALL AND CAMPBELL. — MAGNETIC EFFECTS IN SOFT IRON. 



G45 



Previous work^° has given us 1 \'M\') as the vahie of p at 0°C., with a 

 teinperature-coelHcient ().()().") 1<), and a Thouison-effect formula, rutuiing 

 from 32° C. to 182^' C, from which we get 



s = — (107 T -\-2T') 419 X lo- 



co 



T' being the absohite temperature. 



/S' of the last column is the estimated " thermo-electric height " of 

 our iron relative to lead. It will be discussed at considerable length 

 in connection with the formula of Voigt. 



Testing the Moreau formula with the data given, we find 



Plate 2 



Plate 1 



Temp. 



20° C. 



40° 



60° 



80° 



l lOO° 



45° 



+ 686 X 10" 

 763 



828 

 884 

 932 



+ 686 



+ 958 X 10-*^ 



934 



908 



883 



858 

 + 944 



(A-B) H- A. 



- 0.40 — ^ 



- 0.22 + 



- 0.10 — 



- 0.00 



+ 008 -J 



- 0.38 - 



(II.) 



Not all of the temperature-coefficients are sufficiently well known to 

 justify us in pushing the comparison through a greater range of tem- 

 perature, and in fact the values here set down for A and B are subject 

 to an error of several per cent, probably. But when all reasonable 

 allowance for error has been made, it seems unlikely that the Moreau 

 formula can hold in this iron through any considerable range of temper- 

 ature. The difference between the result from Plate 1 at 45° and that 

 for Plate 2 at or near the same temperature is less significant as throw- 

 ing a doubt upon the accuracy of our work than it would be if the two 

 plates had been cut in the same direction from the original iron bar. 

 Plate 2 is far more important, not merely because it has been studied 

 under better conditions, but because the main current in it w^as in the 

 direction of the fibres of the iron, the direction along which the meas- 

 urements for p and for s have been made. 



It appears that the Moreau formula holds in our iron for one tem- 

 perature, which in Plate 2 is near 80° C. Below this temperature the 

 static Thomson equipotential lines appear to be rotated through a 

 somewhat greater angle, by a given strength of field, than the equipo- 

 tential lines of an ordinary electric current ; but above this temperature 

 the opposite appears to hold. 



Comparison of gT'e -r- p in Table II with ^Tn in Table I shows an 



*" These Proceedings, 42, March, 1907, 



