Experiments oh the tnutual Gravitaikn ef Sod'ifi, 4a|^ 



immediately found that it was not ftiff enough, as the attrafllon of the wefghts drew the balW 

 fo much afid'e as to make them touch the fides of the cafe. It was thought proper, how* 

 erer, to make fome experiments before a ftifFer wire was fubflituted in its place. 



In this trial the rods by which the leaden weights were fufpended were of iron, the pro- 

 bability of magnetlfm in which feemed to be of no confequence, as the arm contained no- 

 thing magnetical. Some trials however, for greater fecurity, were maJe with the rods 

 alone. Tlie efFe£t was fuch as might juRify a very flight fufpicion of magnetifm ; for which 

 xeafon the rods were changed for others of copper. "When the weights were hung on, the 

 •ttraction on the ball was very fenfible, as it carried the latter through fifteen divifions, or 

 if inch; and the regularity of the refults was fuch, that the extremes did not differ from 

 each other more than one-tenth part. It was found that the effe£l of the attra£tion feemed • 

 to incrcafe for half an hour or an hour after the motion of the weights ; an effeft which 

 might be fuppofed to arife from a want of clafticlty, either in the fufpended wire or fome- 

 thing it was faftened to, which might make it yield more to a given preflure, after a long 

 continuance of that preffure, than it did at firft. This fufpicion was put to the trial by 

 keeping the wire for two or three hours in a ftate of torfion amounting to fifteen divifions, 

 and then returning it by means of the handle K to its original fituation. It was not found 

 by two repetitions of the experiment that the wire had acquired any fet or permanent twift. 

 The arm was then fufpended by a differ wire, with which the apparatus, inftead of perform- 

 ing its vibration in about fifteen minutes, employed only feven. The motion of the arm 

 was in this cafe not quite fix divifions between the two oppofite pofitions of the weights,. or 

 the deviation from the natural ftation of the arm was half that quantity. 



A variation of the power of the weights upon the balls, in fome refpedt fimilar to that 

 obferved with the fmall wire, was alfo found to take place in the prefent experiments. It 

 happened that the cafe in which the arm was enclofed was placed nearly parallel to the 

 magnetic eaft and well ; and therefore, if there was any thing magnetic in the balls and 

 weights, the balls would acquire polarity from the earth — and the weights alfo, after re- 

 maining for fome time in either of the near pofitions, would acquire a polarity of the fame, 

 kind, and attraft the balls more ftrongly on that account; and on the contrary, when re- 

 moved to the oppofite near pofition, if the magnetifm continued permanent in the weights 

 for fufficient time, the magnetic power would operate in the way of repulfion. This was 

 in fad the kind of effe£l pointed out by the obfervations. The accurate habits which fo 

 eminently diftinguifh all the experimental proceffes of Mr. Cavendifli did not however per- 

 mit him to reft contented with the indu£lion to which thefe fafts feemed to point. He al- 

 tered his apparatus fo as to produce at pleafure a rotation in the weights upon their vertical 

 axes, without opening the room or difturbing their actual fituation. By this contrivance he 

 had it in his power to difpofe the imaginary magnetic poles in fuch pofitions as were beft 

 calculated to deteft their operation. The arm was not affected; and confequently it fol- 

 lows that the irregularity in queftion did not arife from magnetifm. This canclulion was 

 ftill more firmly eftablilhed by fubftituting two aflual magnetic bars in the place of the 

 weights ; for it was not obferved that any alteration was produced by reverfing the pofitioai 

 of their poles. 



The next circumftance that fuggefted itfelf was, that the effeft might be owing to a dif- 

 ference of temperature between the weights and the cafe i for it is evident, that if the 



' weights 



