OF ARTS AND SCIENCES. 197 



appears that an increase of magnetic induction is accompanied by a 

 lowering of the temperature of maximum Hall effect, but the number 

 of observed points from which curve B is plotted is too small to 

 warrant a conclusion upon this particular. 



Summary. 



The Hall effect in copper and in phosphor-bronze is affected but 

 little, if at all, by a rise of temperature from 20° to 3G0° C. 



The Hall effect was about one half as great in the phosphor-bronze 

 examined as in the copper, under like conditions. 



The temperature coefficient of the electrical resistance of the phos- 

 phor-bronze appeared to be about 0.00045, 



The Hall effect was observed in battery carbon. 



In nickel there is on the whole, in spite of certain marked differ- 

 ences, a strong resemblance between the temperature changes of mag- 

 netic permeability and the temperature changes of that " rotative 

 power " upon which the Hall effect depends. It is doubtful whether 

 an equally strong resemblance as to temperature changes could be 

 made out between either of these properties and any other property of 

 nickel. An equally broad statement cannot as yet be made concern- 

 ing iron or steel, but the evidence obtained, so far as it goes, is on the 

 whole in favor of such a proposition. 



NOTE. — Added May 19, 1893. 



Wiedemann's Annalen, No. 3, 1893, contains an " Explanation of the 

 Hall Phenomenon," by E. Lommel, which seems to me not well con- 

 sidered. Without reproducing Lommel's argument, which is easily 

 accessible, I will make the following criticisms. 



1 . The " explanation " attributes to the supposed Amperian currents 

 at the boundary of a magnetized body an entirely new power, in assum- 

 ing them to increase the strength of an ordinary electric current within 

 one half of the body and to decrease it in the other half. 



2. If the. Amperian currents in the boundary had the supposed 

 effect upon the distribution of the current within the body, it is very 

 doubtful whether this would produce the Hall phenomenon. 



3. Lommel's conclusion is that the Hall effect has one direction in 

 magnetic metals, and the opposite direction in diamagnetic metals. In 

 fact, the direction in iron is opposite to that in nickel. 



E. H. H. 



