HALL AND CAMPBELL. — ^LA.GNETIC EFFECTS IN SOFT IRON. 641 



With Plate 2 we found 



1910. hTk. Arm*. ^. 



al 



Dec. 28 + r>4it X 10"' 29.8° 8.8 (n-s) 5600 



•' 29 580 " 31 8° 8.0 (s-n) f^^rjO 



Means + 568 " 30.6° 8.5 5600 



Dec. 30 + 614 X 10-* 59.2° 7.9 (s-n) 5470 



" " 649 " 61.1° 7^ (n-s) 5550 



Means + 632 " 60.2° 7.8 5500 



The temperature-coefficient of j^Tn in Plate 2 is, according to these 

 figures, 



(632-568) -^ 568 (60.2-30.6) = 0.0038. 



The + sign of f,Tn means, according to our definition, that the 

 isothermal lines are by the magnetic field rotated in the same direction 

 in which the magnetizing current fiows. 



The difference of temperature actually produced between the arms 

 by this action was about 0.0054°. 



The magnet current was usually put on about 45 seconds before the 

 final galvanometer reading showing the Leduc effect was taken. Pro- 

 fessor Campbell, who did most of the experimental work in this research, 

 is under the impression that the Leduc effect did not attain its maxi- 

 mum value as soon as the full field-strength was reached, but that it 

 continued, according to the testimony of the galvanometer, to grow for 

 a time after that moment. We cannot speak positively, as yet, on this 

 interesting particular. 



Relations of eTe, eTn, etc. : The Moreau Formula and 

 ITS Suggestions ; the Voigt Formula. 



Attempts have been made by various writers to correlate some or all 

 of the various effects which are dealt with in this paper and to connect 

 them with some of the more generally familiar electrical and thermal 

 properties of metals. Naturally the electron theory has been used of 

 late in such efforts ; but we shall consider first a suggestion or propo- 

 sition made by Moreau,^ about ten years since, in which no mention 

 of the electron theory was made or intended. Indeed, Moreau says, 

 "As, moreover, the variations of A^[g7^/,] are explained by those of 

 c[e7^e], since the Hall effect is the primordial phenomenon, one is justi- 

 fied in supposing that the molecular state of the plate has a consider- 



f C. R., 130 (1900), pp. 122-124, 412-414, 562-565. 



VOL. XLVI. — 41 



