Lifting Power of the Electro Magnet. 



515 



taining what increase of c would produce L — L'" in these three instances. 

 For 115 it is G-32 ; for 116, 6-29; but for 117 it is only 3-66. As, however, 

 the last is not to be depended on, because it goes far beyond the largest c, on 

 which the interpolation is founded, the close agreement of the other two seems 

 to confirm the supposition. It must, however, be remarked that even the 

 highest of these ratios will not entirely explain the excess of No. 113 over 

 No. 112. It deserves to be noticed that the separation in these experiments 

 always took place between the keeper and the discs, never between the discs 

 and the magnet. 



It was not within the scope which I proposed to myself to try the power 

 of this magnet at sensible distances, but in a few instances I did so, interposing 

 at the keeper plates of brass 0''12 thick. 



Table XXVI. 



The great diminution caused by these small intervals is remarkable, espe- 

 cially where the helices are close to the keeper, and part of the cylinders is 

 uncovered; the force then being only the 29th part of its ordinary amount. 

 In No. 122, where the long cylinders are entirely covered with spires, it is 

 much less considerable, the force being nearly one-fourth ; and even the removal 

 of the slide in the following one does not alter this fact, which is in accordance 

 with the results of Table xv. It follows from tliis that in electro-magnetic 

 machines, the magnet should be of considerable length, and covered uniformly 

 with the spires. At the double distance 0'-24 the law of decrease is nearly 

 the inverse square of the distance ; the deviation from it being due to the dis- 

 tribution of the magnetism on the polar surfaces ; but this is evidently not the 

 law which prevails at the distances of normal contact. 



