198 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



this test, many a piece of steel which has been hardened with care 

 proves to be far from homogeneous. 



Occasionally great differences of resistivity may be found in a mag- 

 netized steel rod which yields a fairly uniform iron-filing diagram. 



The curve OKPE, of Figure 2 shows the induction flux {B) at differ- 

 ent points of the axis of a rod of Crescent Polished Drill Rod 29 cm. 

 long and 0.5 cm. diameter just after it had been magnetized to satura- 

 tion in a solenoid. Curve OGQE, shows the same quantity after the 

 rod had been exposed to steam for some time. AB is the common 



base of these curves. The distri- 

 f bution is in each case nearly 



.' '•...•*'*• uniform, and the iron-filing curve 



E ' . seems entirely so, but the resistiv- 



ity of the metal is far from uni- 

 form, as the dotted diagram ESCD 

 shows. This was obtained by 

 measuring the resistances of a 

 k large number of very short lengths 



of the rod and determining from 

 * ^ * the results values for the resistance 



per centimeter at about thirty 

 points on the axis. Of course a 

 small portion at each end could 

 not be treated in this way, and 

 the fact is indicated by the open 

 dots. One end of this bar was 

 in the soft state in which this ex- 

 cellent steel comes in the market ; 

 the other end had been heated red 

 hot and chilled, so that its resistivity was quite double that of the soft 

 end. This magnet was not so strong as a hardened magnet of this 

 steel should be, but was otherwise normal enough. 



Sometimes the iron-filing diagram belonging to a bar magnet seems 

 very irregular when the distribution of magnetism in the metal is not 

 very abnormal. Figure A shows a filing diagram belonging to a piece of 

 Crescent steel of the same dimensions as that just described, while 

 Figure 3 shows the values of B at different points in the axis. The 

 " centre of gravity " of the magnetism is in this case not far distant 

 from the middle of the bar. This same bar was remagnetized by rubbing 

 a point near its centre upon one pole of a large motor, and then gave 

 a filing diagram represented by Figure B. Here there are real conse- 

 quent poles, and the distribution of the induction flux in the bar is 



Figure 5. 



