456 ON BURNING GLASSES AND M1RROUS. [MEMOIR XXX. 



have been founded is to attack it at its origin, and at- 

 tempt to produce an impression on the shining bod} 7 . 

 These methods are essentially distinct. To borrow an 

 illustration from acoustics, it is one thing to try to 

 modify a sound on its passage through the air, and a 

 very different thing to exert an influence on the sound- 

 ing body. 



When Bancalari's experiment on the influence of the 

 poles of a powerful magnet on a flame was first publish- 

 ed, I repeated it at once, expecting that the oscillations 

 of the shining particles were constrained to take place in 

 one plane by the magnetism, and that the light emitted 

 would be polarized. The result, however, did not seem 

 to prove this. 



A similar experiment was then made with the electric 

 spark from the prime conductor of a machine. It was 

 compelled to cross between the poles of a powerful elec- 

 tro-magnet. But when the magnetism was on it did not 



O O 



seem that the light was polarized. 



De la Rive has shown that the voltaic arc between 

 charcoal points is greatly disturbed when it passes be- 

 tween the poles of a powerful electro-magnet. In the 

 hope that this would produce the expected disturbance, 

 I examined an arc formed between points of copper, plat- 

 inum, and gas carbon; but though the sounds emitted 

 were strong, resembling the sudden tearing of a piece of 

 cloth, I could not perceive that the light was polarized. 



In like manner the induction spark from a contact- 

 breaker and the phosphorescent light from fluor-spar were 

 tried without success. I still think, however, that with 

 better means than those thus employed the experiment 

 would succeed. 



At the commencement of this Memoir it was stated 

 that we should consider, 1st, the manner in which a ray 



