706 PROCEEDINGS OF THE AMERICAN ACADEMY. 



mean loss, per degree, of the magnetic moment, when the magnet was 

 heated from about 10° C. to 100° C, expressed in terms of the moment 

 at the lower temperature, it appeared that in the case of rods 16 cm. long 

 and 0.9o cm. in diameter these mean losses were 



0.00042 for seasoned, chilled, cast-iron magnets 



0.00046 for seasoned magnets made of Crescent Steel Drill 



Rod 

 0.00046 for seasoned magnets made of Jessop's Round Black 



Tool Steel 

 while they were 



0.00070 for seasoned magnets made of Jessop's Square Tool 



Steel 16 cm. long and of cross-section nearly that of 



the round rods. 



In the case of shorter rods the difference was still more in favor of the 

 castings because, I suppose, they were more uniformly hardened in the 

 interior than the steel could be. 



If an iron casting which has been hardened and boiled is magnetized 

 in a solenoid either to saturation or to a degree which falls much short 

 of this, it is practically impossible to decrease- the moment by even so 

 little as a tenth of one per cent, by striking the magnet on end with a 

 wooden mallet or with a stone. I have tested many such magnets by 

 dropping eacli two or three hundred times upon, a stone slab, or by giv- 

 ing it hundreds of sharp blows with wooden clubs : the magnets get a 

 little warm during this harsh treatment, but when their temperatures 

 again fall to the original point the moments, which may have fallen a 

 small fraction of one per cent, regain wholly, so far as my observations 

 go, their old values. Some few of the specimens of special magnet steel 

 that I have examined are nearly equal to the castings in this respect. 



Prolonged boiling has, however, always reduced the moment of the 

 cast-iron magnets- very sensibly, and this loss may be as much as 20 

 per cent when the magnetizing field has been an extremely strong one 

 and the residual moment is very high ; if the casting has not been mag- 

 netized to saturation, the loss of moment by boiling is much less. If, 

 after a cast-iron magnet has been seasoned, its temperature be suddenly 

 raised from room temperature to 100° C, and then as suddenly lowered, 

 the magnet may not wholly recover its original strength until after the 

 lapse of several hours ; if, however, the upper limit be only 50° C, 

 there seems to be no sensible lag in the attainment of the whole of the 

 orirrinal moment after the testing. 



