Magnetomotive Force. 215 



each side, so that the centering has to be very true for the 

 cradle to swing free. The circle is read through the glass 

 sides of the case. 



The magnet employed is made from a cylindrical bar of 

 cast tool-steel. It consists of eighteen pieces, fitted and 

 hardened as above described. They were then fastened 

 together in two lengths of 9 pieces each with wooden splints, 

 and placed between the terminals of an electro-magnetic 

 magnetizer constructed for the purpose. When screwed up 

 firmly between the poles of the magnetizer a current was 

 transmitted through the coils. In this condition it was tapped 

 with a hammer for some time. When removed, each com- 

 pound bar retained a considerable permanent magnetism. 



The chief difficulty was to observe with sufficient accuracy 

 the small deflection produced by the separated pieces. Small 

 differences in this small quantity produce large differences in 

 the calculated resistance (or permeability) of the steel. The 

 measures now given probably reach to an accuracy of about 

 a tenth of a degree. Up to this point I have endeavoured 

 to take count of all the errors of the apparatus. The final 

 measures are : — 



Deflection due to 18 pieces joined up = 13 o, 



„ „ separated = l o, 05*. 



Dimensions of magnet: — centim. 



Whole length = 28*50 



Length of each piece = 1*58 

 Diameter = 1-97 = 211 



Let r be the resistance of one of the steel pieces expressed 

 in centimetres of a similar air-cylinder, 



aR the resistance of one end ; 



then 18r+2aR 1-05 AO 



_______ = __ = . 08 nearly 



(assuming the forces proportional to the deflections). 



Then r= '053 x a centim. 



and length of piece = 1/58 centim. 



,. „ length of air _'053x„ 



steel of same resistance 1*58 



= -034x«. 



* After the pieces had remained separated for some days in the bifilar, 

 I noticed that the reading- had changed in the direction of increased 

 moment. A set of readings gave 13°-0 and l 0- 9. After standing for 

 some days joined up again, 13° _ and l°-8. These latter values corre- 

 spond to p = 15 nearly if a = -Q, to /-i— =29 nearly if a= :*3. There appears 

 to have been a sort of spontaneous rearrangement of the magnetism of 

 the little pieces in the direction of less resistance, probably with diminu- 

 tion of a. The spontaneous change was an increase of moment, not a 

 diminution. The point requires further examination. 



