Miscellanies. 403 



foot, and the two little ke3's fell off. Nevertheless the two bars were 

 still magnetic, for not only did they preserve their adherence at h and 

 c but I could again suspend the little keys at a and d the poles alone 

 being changed. 



In repeating this experiment a moment after, only changing the 

 position of the magnets, that is, putting the pole S towards a and the 

 pole N towards d, the same phenomena were repeated, but in a con- 

 trary order. 



I had before discovered that two keys of middling size, subjected to 

 the influence of a strong magnet, and which, after that remain sus- 

 pended to each other, will preserve their adherence an indefinite 

 length of time, when removed from the magnet, provided it be done 

 with caution. It is evident then, that they remain impregnated 

 while in contact, for at the moment of their separation, the force 

 disappears. 



I afterwards tried the arrangements fig. 2, placing against the mid- 

 dle of the soft iron bar another similar one c d, but only about six 

 inches long, that the needle h might not be too immediately con- 

 trolled by the magnet ef. The three needles in this position, shew- 

 ed that the extremities a, b and c, had the same magnetism as the 

 nearest pole of the magnet. In withdrawing it the needles g and i 

 did not vary, but h at length, was reversed although less rapidly than 

 in the former experiment. 



But the reaction or reversion of the poles was manifested in the 

 most evident manner by disposing the parts as represented in fig. 3, 

 that is taking a single long bar, two short bars and two compound 

 magnets. In withdrawing the magnets simultaneously, not only are 

 the needles I and m reversed immediately, but the reaction must be 

 so rapid as to keep up a continued magnetism in the bars, for the two 

 little keys suspended at a and b still hold on, not having, I had al- 

 most said, time to fall off, although the needles clearly indicate that 

 the poles of the bar a b were reversed. In using but one magnet, 

 (i 1c for instance) the two needles / and m had their north end to- 

 wards d and /, and when the magnet was removed, tlie needle m 

 was reversed, the other remaining stationary. 



Must not this reaction be analogous to the secondary piles of 

 Ritter, which assume, when contact with the Voltaic pile is broken, 

 poles opposite to the latter, or to the little needles which have their 

 magnetism reversed, when a Leyden bottle is discharged through a 

 wire situated near them ? 



