Rods of Iron, Nickel, and other Metals. 



387 



Ail experiments of a similar character with wires with a 

 permanent torsion of 2,3,4, &c. showed that for a small 

 magnetizing force the deflexion rose almost proportional to 

 it, and that after a certain point further torsion was not pro- 

 duced by increasing the magnetizing force, and that ultimately 

 with a greater magnetizing force the rod began very slightly 

 to untwist. 



8. The next experiment was made to discover any relation- 

 ship which might exist between the initial torsion of the rod 

 and that added to it by the force of the magnetic field when the 

 magnetic force was constant. The following results were ob- 

 tained. Up to some value of initial torsion between 10 and 

 15 the deflexion produced by magnetization is a maximum, 

 and then it decreases as the torsion rises. 



Initial Torsion. 



Scale-divisions. 



Initial Torsion. 



Scale-divisions. 



1 

 2 

 3 

 4 



3-00 

 500 

 6-50 

 7-25 



5 



10 

 15 



800 



10-00 



8 00 



By initial torsion I mean that the rod received a permanent 

 torsional set of 1 , 2 , 3, &c. complete revolutions. 



9. I noticed in the case of the last two values that after 

 each magnetization the rod began to slowly untwist when 

 the current was taken off. After each experiment the reading 

 fell below the figure it started at. As an instance, the initial 

 position was at 60 divisions ; when the current was on, 69 

 was reached, and when the current was taken off the reading- 

 was 59. 



Subsequent experiments showed that when a rod having 

 permanent torsional set was used in induction experiments 

 with a ballistic galvanometer, the induction followed in value 

 the slow change of torsion of the rod. In some cases I noted 

 the slow change from last February to the present time (July 

 30) . Each induction experiment is modified by its predecessor 

 having apparently changed the constitution of the rod. 



10. In the next experiment this step by step unwinding of 

 torsion after magnetization is shown in the case of a rod of 

 nickel. The rod was placed in exactly the same position as 

 the iron rod of the former experiments. The behaviour of 

 nickel appears to be the same as that of iron, but in nickel the 

 increase of torsion with magnetization is not so great. The 

 twelve vertical lines, in the appended diagram, represent pro- 

 longations of the scale. In each case No. 1 indicates the position 



2 D2 



