Transparent Bodies by the action of Magnetism. 503 



gression, whose ratio would have been very little different from the 

 mean of the two preceding determinations, i. e. of 0-790. Now 

 we find in the memoir of M. Bertin* five series of experiments 

 relative to the decrease in the rotations observed at different 

 distances from a single branch of the electro-magnet. The two 

 first are relative to Faraday's heavy glass, and may be repre- 

 sented by two decreasing geometric progressions whose ratios 

 are respectively 0-78329 and 0-78330 for 10 millims. of increase 

 of distance; the third is relative to a flint prepared by M. Mat- 

 thiessen, and is represented by a decreasing geometrical progres- 

 sion whose ratio is 0-78233; the fourth is relative to the bisul- 

 phide of carbon, and is represented by a decreasing geometric 

 progression whose ratio is 0-78329; lastly, the fifth relates to 

 the heavy glass of Faraday, and is represented by a decreasing 

 geometrical progression whose ratio is 0-78329. The mean of 

 these five geometrical progressions is 0-78318, and differs very 

 little from 0-790. Thus the experiments of M. Bertin agree 

 entirely with the law which they seem to contradict. 



In order to show well the influence which the form of the 

 terminal surfaces of the electro-magnet exercises upon the law 

 of the decrease of magnetic action, I repeated the same series of 

 experiments by screwing one of my large armatures to the extre- 

 mity of the branch of the electro-magnet which I employed. I 

 obtained a law of decrease much more slow than in the prece- 

 ding case, which may again be represented by a decreasing 

 geometrical progression. On the contrary, when I replaced the 

 large armature by a cone of soft iron 24 millims. high, with a 

 diameter at the base of 45 millims., the decrease was much more 

 rapid, and entirely different from that which would have indicated 

 a decreasing geometrical progression. The following Tables con- 

 tain the results of these experiments : — 



Ratio of each magnetic 

 action to the preceding. 



096 

 0-92 

 0-91 

 0-90 



Mean . . . 0-925 



* Annates de Chimic et de Physique, .'3 ser. vol. xxiii. pp. 22, 23 



and 27- 



