COilRECT CHAMBER BAROMETER. 83 



the opposite column must be supplied by the branch D E F ; 

 let it then descend to the station H h. Now from the con- 

 struction of the weather glass here described, O o is the 

 only base from which we can measure the length of the mer- 

 cury contained at any time in C C accurately by the scale; 

 consequently the part of the column situate in H h and O o 

 will be neglected in the observation ; and the height of the 

 barometer will appear to be less than it really is by the space 

 H O; that is, the top of the column will not ascend so far 

 as' it ought to do above the middle poinl of the scale. On 

 the contrary, when the mercury descends in the leg B C it 

 will rise in the leg D E F, in consequence of a quantity o£ 

 this fluid coming into it from the opposite branch. As oft 

 then as the weather glass denotes a height less than 29*5 

 inches, the surface of the mercury in the leg D E F will 

 ris6" above O o, the true base of the scale; consequently 

 when the column comes to be measured by the index, it will 

 appear longer than it really is by the height of the mercury's 

 surface jn the leg D E F above the base O o ; that is, the 

 upper Extremity of the column in the branch B C will fall 

 below the point 29*5, but not so far as it ought to do. Thus 

 it has been made evident, that the motion of the mercury 

 in the shorter leg diminishes the range of the barometer 

 when it comes to be estimated by the scale; which circum- 

 stance pointis out the necessity of a correction, and perhaps 

 the easiest method of doing it consists in bringing the sur- 

 face of the mercury in the leg D E F to a coincidence with 

 the base O o, before setting the index attached to the scale. 

 My apparatus for this purpose is nothing more than an ivory The method of 

 piston K of a cylindrical figure, about ]| inch in height ^^'■"^^""g ^lie 

 and I an inch in diameter. Its lower extremity is a little ther dass by % 

 convex and immersed in the mercury; it is also easily moved piston, 

 vertically by means of the handle or stem G L, which passes 

 through a cap covering the mouth of the tube E F, and 

 having a hole, 1, in its centre, wide enough to receive the 

 handle. iTh is part of the piston is also attached to the frame 

 A a by ia loop or socket of brass, L, in which it would move 

 lertically with perfect freedom, were it not for the pressure 

 or-ii bent spring which is situate between L and 1, and acts 

 ujthn the handle G 1, with a force, that keeps the piston iu 



G 2 any ^ 



