4 AKT. 13.-K. HO\DA, S. SHIMIZU AND S. KTJSAKABE : 



cylinder was stretched hy u Aveak spring- in the usual way. The 

 deflection of tlie mirror attached to tiie rotating cylinder was observed 

 by means of a scale and telescope. The details of the apparatus will 

 be easily understood from the above figures. 



The dimensions of each jiart of our arrangement were as follows : — 



Length of the coi I = oO.O cm, 



Its internal diameter = o.O ,, 



1-M = .'ÎTO.? „ 



Outer radins of the pulley \ = 8.93 ,, 



Inner „ „ „ „ J = 7.15 „ 



Diameter of the rotating \ = O.IGO ,, 



cylinders ) = 0.280 „ 



Diameter of the copper wire = 0.008 ,, 



Scale-distance = 2o0.8 ,, 



If a ferromagnetic rod is twisted tlirougli a small angles, the 

 thin vertical wire is pulled down through a. siuali distance Iif, R 

 being the outer radius of the ])ulley. This causes an elongation of 

 the weak spring attached to the copper wire, and consequently the 

 rotating cylinder, whose radius is r, is turned through an angle R^/r 

 Hence the angle of torsion is magnified in the ratio R : r ; in the 

 actual calculation, we must take into account the thickness of the 

 thin wire. The ratio was in our case 106 : 1 ; with this arrang-e- 

 ment, we were able to measure a change of angle amounting to only 

 1.92" X 10~^ per cm of the ferromagnetic rod. 



8. The measurement was conducted in the following order. 

 The specimen to be tested was fixed in the axial line of the magnetiz- 

 ing coil so as to lie nearly in a uniform field. If the steel pivot on 

 one end of the bar carrying the specimen was left free, and a 

 magnetizing current passed through the coil, a deflection of the mirror 

 was observed, though there was no twisting couj^le. The deflection 



