EXPERIMENTS upon MAGNETISM. 327 



Perhaps this method of making experiments, by fub- 

 flances floating in water, and thus I'ubjedted in all their 

 motions to our e::amination, may lead to more important 

 difcoveries. The rates in which the magnetic attradlive 

 force decreafes, at different diftances, may, 1 think, be 

 coUefted from noting the velocity with which the floating 

 bodies move, at different diftances from the poles, or the 

 fpaces, which they pafs over in equal times. Nothing 

 obftiudts an accurate folution of this problem, but the 

 difliculty of obtaining meafurernents of fufliciently fmali 

 intervals of time. If the experiment were made upon a 

 large fcale, the difficulty might be removed. The remark- 

 able acceleration which is obfervable, when the filings 

 come near to their attracting point, feems to fatisfy the 

 eye, that the attraction increafes in a greater ratio, than 

 according to any law yet afligned. 



One magnet adts upon another, at a confiderable dif- 

 taflce, either by repulfion or attraction. Will not thefe 

 experiments lead to a rational conjecture, that in every 

 inftance, the aCtion is communicated by intervening mag- 

 netic fubftances. It aCts through atmofpheric air. But 

 this air may, from its conflituent principles, and it is 

 faid, does contain iron. The fmall particles floating in 

 the atmofphere, may be adled upon, like thofe floating 

 upon water. The tenuity of the particles will only render 

 the aCtion more fenfible. Each may become a magnet, 

 and thus by the aCtion of all the intervening affeCted par- 

 ticles, the action of one magnet may be communicated to 

 the poles of another diftant magnet. 



I have made an experiment, in order to afcertain, whe- 

 ther a magnet could exert its power in a Torricellian vacu- 

 um. A fmall quantity of filings was poured into a glafs 

 tube of fufficient length ; it was then filled with mercury, 

 and inverted in the fame fluid ; the filings floated on the 

 furface of the mercury in the upper end. The refult was, 



that 



