" Skin "-effect in Electrical Oscillators. 



11 



A section of the magnetometer is shown at M in fig. 8. It 

 was of the plainest construction, consisting simply of a mirror 

 10 mm in diameter with two small steel magnets fastened on the 

 back by paraffine, and suspended by a silk fiber about 12 cms 

 long. Behind it was a block B, in which a small glass tube, 

 closed at one end, was wedged. The inner diameter of this tube 

 was approximately equal to the outer diameter of the detector 



Jro, 



3K35 



tF 



m 



tube, so that the latter moved snugly into the former, and, 

 when pushed in as far as it would go, was in a definite position 

 which could easily be recovered. 



The block B was fastened on the upper face of a board N 

 by a screw a around which it could be given a slow motion 

 by a screw S, thus easily making the detector approach or 

 recede from the magnetometer by a small amount. The tele- 

 scope and scale A were about 120 cms from the magnetometer. 



The position of the magnetometer was adjusted so that 

 when the detector, fully saturated, was pushed into the tube 

 T, the magnetometer deflection was 40 cms , the scale being num- 

 bered so that this read zero. After partial demagnetization 

 through the electrical oscillations, it was replaced in its posi- 

 tion behind the magnetometer and the alteration in reading 

 observed directly. 



It was found that the detector could be removed and 

 replaced with no observable difference in the reading ; and 

 successive magnetizations to saturation produced the same 

 deflection. 



Rutherford* showed that the demagnetization produced by 

 rapid oscillations is confined to the outer layers of the magnet. 

 The diffusion of the magnetic force from the dielectric into a 

 magnetizable substance is, indeed, similar to the passage of 

 electrical action into a cond actor. Hence for rapidly alter- 

 nating currents the magnetic action is confined to the outer 

 layers, the more rapid the alternation the thinner the layer 

 affected. Hence the reason for using fine wire and insulating 

 the pieces from each other. By so doing more surface is pre- 



* Loc. cit. • 



