68 Hart — Action of Light on Magnetism. 



direction of maximum susceptibility : it indicated, probably 

 that the light effect was local, and confined to the illuminated 

 surface. The effect varied directly as the intensity of the 

 illumination ; it was strongest for red light and independent of 

 direction of vibration of incident light ; and blackening the 

 bar made the action much slower. 



I first repeated and verified Bidwell's experiments. I pro- 

 cured eight pieces of the softest iron obtainable, of the follow- 

 ing dimensions : each was 20 cm long ; four were cylindrical 

 rods, with diameters 13 mm , 9-5 mm , 6'5 mm , 3'5 mm respectively : they 

 were designated rods A, B, C and I) ; four were of rectan- 

 gular cross-section, 12x3'5 mm , 12x2-5 mm , 12xl'5 mm , 12x0'75 mm 

 respectively : they were designated, rods E, F, G- and H. 

 They were treated in the same manner as in Bidwell's experi- 

 ments ; they were carefully annealed, being heated in iron 

 filings almost to the point of fusion and then cooled in ashes for 

 about 25 hours. They were first strongly magnetized and 

 then carefully and totally demagnetized. As a result of numer- 

 ous trials and careful manipulation, I was able to get distinct 

 effects from a gas jet at a distance of 6 inches. The follow- 

 ing is a characteristic set of experiments : the light was per- 

 pendicular to the axis of the bar- and only 3 cm of the north 

 end of each rod was exposed. The deflection represents 

 increase of magnetism as shown with magnetometer. 





Area exposed 



Cross-section 





Rod. 



to light. 



of rod. 



Deflection. 



A 



g. 15 sq cm 



l-33sq cm 



22-5 



B 



4-48 



0-71 



15-0 



C 



3-06 



0-33 



7-1 



D 



1-65 



o-io 



1-5 



E 



3-60 



0-42 



10-0 



F 



3-60 



0.30 



7-5 



G 



3-60 



0-18 



4-5 



H 



3-60 



0-09 



1-5 



The effect was proportional to the area exposed, but not 

 directly. It depended also on the intensity of the illumina- 

 tion. The action of the light does not appear to be here con- 

 fined to the illuminated surface, as the effect is approximately 

 the same, no matter which 3 cm of length of the bar is illu- 

 minated, and hence distance of illuminated surface from mag- 

 netometer has little or no effect. This is contrary to Bidwell's 

 results and is probably due to the very regular annealing. The 

 heating effect was also present. Light of long wave-length 

 had greater effect. I did not attempt Bidwell's experiments 

 in regards to polarized light and its action on the magnetism 

 of the iron, since no positive results were obtained by him. 



