[ loa ] 



panied with a fmall table of corredions for converting the obferved 

 differences of altitude into the true to be applied at reading off 

 the obfervations. 



But as fome perfons may chufc to eonftrud thefe barometers 

 who could not rely on their fkill in making thefe calculations, 

 they may find the above corredion mechanically, as follows: 

 before you put the inftrument together, let the tube exceed the 

 required length, perhaps three or four inches : break off three 

 inches, and referve them till the barometer is finifhed ; apply it 

 in its cafe to the fcale, and carefully note the height it ftands at. 



For greater fecurity, let this obfervation be made three br 

 four times, then take it out of the cafe or mounting, and opening 

 the cylinder, without lofmg any of the mercury, fill your referved 

 tube of three inches, and pouring the contents into the cylinder, 

 replace all in the mounting. Now obferve the height the 

 mercury ftands at ; the difference of this and the former height 

 gives precifely the effed on the level of the mercury in the 

 ciftern, occafioned by the addition of three inches of the contents 

 of the tube ; and this quantity will ferve as an argument to con- 

 ftrud a fcale of corredion for the inftrumenVwhich cannot err. 



It is to be obferved, that when you have thus attained what 

 may be called the error of the fcale, you (hould withdraw the 

 mercury poured in, to leave the more room in the ciftern, if 

 wanted. 



Example 



