498 M. A. Berlin on Circular Magnetic Polarization. 



In the following table is shown the comparison of the re- 

 sults calculated in this manner with those furnished by expe- 

 riment. 



This comparison is the last verification which can be made 

 of our formula. We might, it is true, vary d, and then, as 

 a?=0, we have 



z = A{l+r^-% 

 or 



A A , 



s—A= —r^. 

 r" 



We conclude that the quantity z— A ought to decrease in 

 geometrical progression as the distance of the poles increases 

 in arithmetical progression. But experiment no longer con- 

 firms this conclusion, which is owing to the coefficient A no 

 longer being constant, but varying with the distance of the 

 poles ; for these, by reacting one upon the other, change the 

 intensity of their magnetism. If this reaction did not take 

 place, the action of the two poles in contact with the flint-glass 



