Hart — Action of Light on Magnetism. 69 



I then extended these experiments of Bidwell in an attempt 

 to learn more in regard to the nature of this action of light on 

 magnetism. Every cause of molecular disturbance favors the 

 magnetization of a bar subjected to a magnetizing force, and 

 also favors its demagnetization after it has been withdrawn 

 from the field. Vibration has a specially marked effect upon 

 iron. Ewing* has shown that if a bar of this metal be kept 

 from the slightest vibration one can obtain residual magnetiza- 

 tions much greater than those shown in steel bars, but the 

 least vibration causes the acquired magnetism to vanish almost 

 completely. If the action of the light is of the same nature 

 as that of a mechanical shock on a magnet and not dependent 

 on the special structure of the magnetic field in Bid well's 

 experiment, we ought to obtain effects from light in experi- 

 ments like those of Ewing, if they are sufficiently delicate. I 

 have been able to get very good results from a series of experi- 

 ments with this object in view. The bars were magnetized in 

 a magnetic field of about 400 C. G. S. units and the current 

 was steadily and rapidly shut off by means of a continuous 

 current rheostat. The slightest vibration caused a sudden fall 

 in the residual magnetism. Extreme care was necessary. Only 

 a small portion of the bar was exposed to the light, as it was 

 found that the magnetizing coil could not be removed without 

 causing the sudden fall in the magnetic strength of the bar. 

 Electric disturbances due to the surface-road and electric light 

 wires had the same effect, and it was necessary to work during 

 the quiet portions of the day. The following is the average 

 result of a number of experiments. Only 3 cm of the bar was 

 exposed, that is, there was the same surface exposed as in pre- 

 vious experiments, and the light was that of an ordinary gas 

 jet at a distance of six inches ; other conditions were the same. 

 The direction of the deflection indicated in every case a 

 decrease of magnetism. 



Rod. 



Area exposed to 



light. 



Deflection. 



A 



6-15 S( i ' 



:m 



44-0 



B 



4-48 





26-1 



C 



3-06 





12-5 



D 



1-65 





6-4 



E 



3-60 





13-1 



F 



3-60 





8-4 



G 



3-60 





4-3 



H 



3-60 





0-9 



These results differ from those of Bidwell in the fact that 

 the deflections here are permanent, even after light has been 



*Proc. Roy. Soc, 1890, p. 342. 



