70 Hart — Action of Light on Magnetism. 



removed. In my own repetition of Bidwell's experiment, the 

 magnetometer image did not always return to the same spot 

 after the light had been flashed upon the iron and removed. 

 There was in general a partial return, the law or cause of which 

 I was unable to discover. I then attempted a similar experi- 

 ment obviously suggested by the previous one, namely, that of 

 increasing by light the magnetic strength of a bar which was 

 in a steady field. I here also had great difficulty on account 

 of external vibrations and electrical disturbances. I steadily 

 increased the current up to different strengths and then flashed 

 the light on the exposed portion of the bar and noted increase 

 with the magnetometer. The first series here given is when 

 the field is of about 10-15 C. G. S. units and the second, of 

 about 350 C. G. S. units, 3 cm was exposed and the light was a 

 gas jet at a distance of 6 inches. 



Rod. 

 A 

 B 

 C 

 D 

 E 

 F 

 G 

 H 



It is a noteworthy fact that it is only the larger bars from 

 which no deflection was obtained. The conditions were pre- 

 cisely the same for the series and therefore for the larger bars 

 there is either an increased sensitiveness to vibrations coming 

 from without, or the action is dependent on the mass in some 

 manner not understood. It is probably only a small portion 

 of the iron which is so extremely sensitive as to be influenced 

 by light. A slight but sudden increase of the field up to the 

 maximum strength used in this experiment caused a rise in the 

 magnetism of one bar amounting to 125 divisions, while a 

 slight and sudden fall in the strength would also cause a 

 slight increase, when the sensitive state was obtained. 



All these results can be effectively accounted for by the 

 assumption of a regular closed magnetic circuit (in iron bars 

 carefully annealed, magnetized, and immediately demagnetized 

 as above described) in which the lines of force return through 

 a thin layer at the surface. The circuit is wholly or partially 

 closed in every case except that of complete saturation, and 

 the thickness of this surface layer depends on the magnetic 

 state of the bar. In Bidwell's experiments, we may suppose 

 the larger portion of the lines of force to return through the 

 surface layer, which is here of considerable thickness. Here 



Area exposed. 



Deflection — I. 



Deflection- 



-IT. 



6-15 S<J cm 



none 



none 





4-48 



none 



none 





3-06 



42 



none 





1-65 



6 



10 





3-60 



none 



none 





3-60 



36 



none 





3-60 



23 



15 





3-60 



V 



14 





