198 PROCEEDINGS OF THE AMERICAN ACADEMY 



NOTE. —Added June 19, 1893. 



Wiedemann's Annalen, No. 6, 1893, contains an interesting and 

 important article by Professor Kundt on the " Hall Phenomenon in 

 Iron, Cobalt, and Nickel." Taking exceedingly thin, transparent 

 films of these metals, and subjecting them to the action of a magnetic 

 field varying from a moderate intensity up to a value between 21,000 

 and 22,000 c. g. s., he has, with each metal, found the Hall effect and 

 the rotation of the plane of polarization of light to maintain a constant 

 ratio to each other throughout the whole range of magnetization. 

 Then, depending upon the work of Du Bois (Wied. Ann. 31, 1887) for 

 proof that the I'otation of the plane of jjolarization of light in these 

 metals is proportional to the magnetization (3) as distinguished from 

 the magnetic induction (58), he concludes that in any given plate 

 of iron, cobalt, or nickel the Hall effect is, other things being equal, 

 proportional to the magnetization. 



Professor Kundt credits me with having been the first to show the 

 probability of a close connection between magnetization proper and 

 the Hall effect in the case of nickel, but states that, in the case of iron 

 and cobalt, " investigations hitherto have not gone beyond the limits 

 within which the magnetization remains proportional to the magnetiz- 

 ing force." In this statement I think he does scant justice to some 

 experiments of mine published in the American Journal of Science 

 for August and October, 1888. These experiments, although they 

 did not go far enough to put the matter absolutely beyond question, I 

 felt to warrant me in making and publishing the following inference ; 

 " When a piece of iron, cobalt, or nickel is made to approach the state 

 of ' magnetic saturation,' the transverse current [Hall effect] obtained 

 from it increases somewhat less rapidly than the magnetic induction 

 through the metal, but experiments with very highly magnetized iron 

 and nickel indicate that this transverse current tends toward a fixed 

 limit rather than a maximum followed by a decline." 



The validity of Professor Kundt's conclusion that the Hall effect is 

 strictly proportional to the magnetization, depends upon the validity of 

 the law which Du Bois, as already stated, announced concerning the 

 relation between magnetization and rotation of the plane of polar- 

 ization of light. Now as this law was obtained by assuming for the 

 thin transparent films of " galvanoplastic" nickel and cobalt used by 

 Du Bois, the same magnetic coefficients found by Rowland in his 

 famous rings of cast nickel and cobalt, and as, moreover, with this 



