[ 122 ] 



to this barometer a fcale be adapted in which a line ^l- of an 

 inch be made to reprefent one inch, when the mercury falls 

 through the length of this line its altitude will be really leffened 

 by one inch. And thus the divifions of this fcale will reprefent 

 the true variations and altitudes of the mercury in inches, 

 as corredly as the common fcale can do in any large fixed 

 barometer. This correded fcale is to be divided into aliquot 

 parts fimilar to thofe in the common fcale of inches ; and to its 

 divifions are to be annexed the fame figures or numbers that 

 are annexed to the correfponding divifions of the common 

 fcale, 



The eafieft, and, I believe, the mofi; accurate method of 

 forming a corred fcale for a portable barometer, is this : put 

 it up by a fixed barometer, whofe veffel, for the flagnant 

 mercury, is fo large that you may be fure the furface of the 

 mercury in it will not rife or fall perceptibly on its rifing or 

 falling in the tube ; fo that the common fcale, annexed to this 

 large barometer, will always point out the true variations and 

 altitudes of the mercury in the tube. Mark, at the fame time, 

 the points at which the mercury ftands in the tube of each 

 barometer. When you find that the mercury in the fixed ba- 

 rometer has varied through any given fpace, fuppofe one inch, 

 then take accurately the length of the fpace through which it 

 has varied at that time in the portable barometer ; this 

 will be the length of a line which is to reprefent one inch 

 in the corred fcale for that portable barometer. That this 

 obfervation may be accurate it Ihould be repeated often. In this 



way 



