426 On the supposed Helical Path of the Electric Current. 



longitudinal magnetization, and the part (if any) due to a 

 solenoidal component of the current. The same remark applies 

 with equal force to any observation of the induction outside 

 the magnetized conductor. 



But if we use a hollow iron or steel rod in place of a solid 

 wire, and test the magnetic force inside it. we may at once 

 determine whether any part of the observed effect is due to a 

 circulation of the current round the rod. Let the rod be 

 long compared with its transverse dimensions, and let it be 

 uniformly magnetized. In the first instance, the magnetic 

 force at the middle of its length, either inside or outside, due 

 to its longitudinal magnetization, is 

 31 



where M is its magnetic moment and a its half length. When 

 a current is passed along the rod there will be (by Prof. 

 Hughes's experiment) a diminution of its moment. Let this 

 diminution of moment be AM. Then outside the rod, at the 

 centre, the magnetic force will be changed by the amount 

 AM 

 a? 



But inside, if any part of the effect is due to a solenoidal 

 arrangement of the current, an electromagnetic force will be 

 produced equal to 



— 47tC/j, 



where C is the current, and n the number of turns in this 

 solenoid per unit of length. The actual change of force at an 

 internal point midway between the ends will therefore be 



AM . n 



a? 



The question whether the second term of this expression has 

 any sensible existence, is to be answered by observing the 

 change of force inside when the current is made, and at the 

 same time observing the change of moment of the rod, or by 

 making simultaneous or successive measurements of the force 

 outside and inside. 



Various ways of carrying out this experiment will suggest 

 themselves. The two following plans, which I hope soon to 

 put to the test, appear practicable ; the second, in particular, 

 seems capable of a high degree of delicacy. 



I. Suspend inside the hollow rod ;t magnet of known 

 moment, and small enough to leave some clearance for swing- 

 ing. Time the oscillations of this magnet both before and 



