ctiri-il column is subject to some uncertainty, arising from the level of the mer- 

 cury not being perpendicular to the direction of the tube. In the wheel ba- 

 rometer there are several sources of error, which, though so small in amount 

 as not to injure it for domestic or popular use, yet are such as to render it alto- 

 aether unfit for scientific inquiry. 



A contrivance called a vernier, for no'ing extremely small changes, is usu- 

 ally applied to the vertical baromete and supplies the place of an enlarged 

 scale. It consists of a small pr 5 ^aied plate, which is moveable by a screw 

 or otherwise, and which slid" on the divided scale of the barometer. By 

 means of this subsidiary .oaie, we are enabled to estimate magnitudes on the 

 principal scale amoun'iug to very small fractions of its smallest divisions. 



The principle o f oie vernier is easily explained. Let B A, fig. G, represent 



B C 



