196 PROCEEDINGS OP THE AMERICAN ACADEMY 



are directly available, for the reason that none of them are curves of 

 constant magnetic induction. Curves H, I, J, and K are derived 

 from curves given by J. Hopkinson, Royal Society Proceedings, June, 

 1888, p. 317, to show the effect of temperature changes on the magnetic 

 properties of " an impure nickel." Hopkinson's own curves, like those 

 given by Perkins, merely furnish data for these curves of constant 

 magnetic induction, being either curves of constant temperature or 

 curves of constant magnetizing force. From the history and descrip- 

 tion of the nickel specimens it seems likely that the nickel examined 

 for the Hall effect resembled the nickel examined by Perkins more 

 than it resembled that used by Hopkinson. The magnetic induction 

 per square centimeter in the various cases is shown by the following 

 table. 



An examination of curves D to K shows certain general resemblances. 

 Most of these curves grow steeper with rising temperature, until a 

 point lying between 200° and 250° is reached, then continue to ascend 

 with diminishing steepness, and, attaining a maximum height before 

 300° is reached, afterward plunge suddenly downward. A similar 

 statement holds true of the Hall effect curves B and C, as we have 

 seen, but the turning points occur at considerably lower temperatures. 



Some of the permeability curves in Figure 2 seem likely to cross 

 the base line at temperatures between 300° and 400°, but experiment 

 gives no warrant for such crossing, and there is much evidence in favor 

 of continuing them as asymptotes to the base line. Curves B and C 

 are probably asymptotic to the base line, although further experiments 

 upon this point would not be superlluous. 



In the group of curves D to G, as in the group H to K, it is evident 

 that an increase of magnetic induction is accompanied by a lowering 

 of the temperature of maximum permeability. In curves B and C it 



