4o fcffett of Magnetifm on Time-piecet. 



j a' 23'' in its rate. It muft be obvious to every perfon, thai 

 even this difference, great as it was, would be increafed oj: 

 diminifhed as the wearer fhould happen to carry in his 

 waiftcoat pocket a key, a knife, or other article made of 

 iteel. This circiimftance, taken along with the amount ol 

 the variation occafioned by the polarity of the balance, was 

 fully fufficient to produce all the irregularity obfcrvcd in the 

 going of the watch. 



I then took away the ftcel balance, fubftitutcd one made 

 of crold, and, having brought the watch to time, obferved its 

 rate of goinsx, and found it as uniform as any watch of the 

 like conft.rutr.ion; for, though it was a duplex efcapement, 

 which is perhaps the belt yet invented, at lead: for common 

 purpofes, it had no compenfation for the expanfion and con- 

 traction occafioned by heat and cold, and therefore a perfect 

 performance was not expected. 



Steel balances being commonly in ufe, and, on that ac- 

 count, eafieft to be procured, and being on many accounts 

 preferable to any other, I was unwilling to abandon them 

 entirely; but refolvcd to take the precaution of always trying 

 them before I fhould apply them to ufe. The mode I adopted 

 was, to lay them upon a dice of cork fufficient to make them 

 float upon water, and I was in hopes that out of a eonfidcr- 

 able number I might be able to felect fufficient for my pur- 

 pofe ; but to my furprife, out of many dozens which I tried 

 in this manner, I could nol felecl one that had not polarity. 

 Some of them had it but in a weak degree, and not more than 

 one or two, out of the whole quantity, appeared to have it 

 fo flronp- as the one which gave birth to thefe experiments 

 and to the prcfent paper, which is perhaps more prolix than 

 could be wiflied : but the fubjett appeared to be not unin- 

 tereftiii?, and I hope the remarks I have offered will not be 

 altogether ufelcfs, as every thing that can tend to add to the 

 perfection of time-pieces, or to remove any caufc that operates 

 againft their perfection, is of fomc importance. 



My only motive in fending this for publication in the 



Philo- 



