H. A. Rowland — Studies on Magnetic Distrihuiic 



27 



curves for steel are much more 

 ould thus give greater values to 

 r in the formula, a result to be 

 expected. We also observe in 

 both figures the great change 

 in distribution due to the direc- 

 tion of magnetization. In the 

 case of the electro-magnet this 



little 



than 



change in scale ; but in the per- 

 manent magnet there is a real 

 change of form in the curve. It 

 seems probable that this change 

 of form would be done away 

 with bj using a suf3Scient mag- 

 netizing-power or magnetizing 

 Results fromjelectro-magnets:— -^^ apphcation of permanent 

 magnets ; for it is probable that 

 3 as originaUj. ~ the fall in the curve E is due 

 from Table XII magnetized to the magnctizing-force hav- 

 )site to Its original magnetism. • ■, „„«;„;„^r ^„ „u , 



npletelj 



thee 



Dg been sufficient to change 

 partially at the 



On comparing the distribution on electro-magnets with that 

 n permanent magnets, we perceive that the curve is steeper 

 g toward the end in electro-mag- 



nets than in permanent mag- 

 nets. At first I thought it 

 might be due to the direct ac- 

 tion of the helix, but on trial 

 found that the latter was almost 

 inappreciable. I do not at 

 present know the explanation 



ments on the distribution of 

 ,%abfe''S\ *° ^^ original magnetism on permanent mag- 

 Scaie four times that of fig. 5. ^^ts, and SO I shall only con- 



sider this subject briefly. I 

 have already given one or two results in Table XII. 



The following tables were taken from two exactly similar 

 Stub's steel rods not hardened, one of which was subsequently 

 used in the experiments of Table XIL They were 12-8 inches 

 long and -19 inch in diameter. 



The coincidence of these observations with the formula is 

 very remarkable, but still we see a little tendency in the end 

 observation to rise above the value given by the formula. 



