106 SCALE OF THE BAROMETEE. 



•* than the observations denote them. About ^V should 

 ** have been allowed upon them." Not to mention the dif- 

 The proper ficulty of obtaining the exact areas of the top of the column, 

 correctiofi sel< j^^^ of the bulb, the latter of which is continually varying 

 on account of its spherical form; it appears from Mr. Dal- 

 tons second observation, that philosophers do not always 

 make the necessary corrections, even when they have suf- 

 ficient data to do so. 

 barometer The barometer which I generally make use of has a scale 



scales for this ^^'^^^^ ^^ ^^^ bulb, as well as one at the top of the column, 

 purpose. These scales are both divided into equal portions of an inch : 



in making an observation, therefore, with this barometer, I 

 have nothing to do, but to take the height of the mercury 

 in the column, and of that also in the bulb*, and by sub- 

 ' tracting the latter from the former, the true altitude is im- 



mediately obtained. In the annexed 6gure, pi. IV, fig. 1, 

 1 have given a sketch of an improved construction of a ba- 

 rometer, upon the same principle. 

 The same ef- This barometer consists of a tube, bent into the form of a 

 fie scale. siphon, and hermetically sealed at the largest end, which 



must exceed 31 inches. The other end is open, and 4 or 

 5 inches will be a sufficient length for it. If both the legs of 

 the siphon be of equal size, it is evident, that, when the 

 mercury rises one inch in the largest leg, it will fall one 

 inch in the shortest; and vice versa. The scale is to be 31 

 inches in length, and graduated in the usual manner at 

 the top ; but it must be movable, so as to slide freely, 

 upwards and downwards, in a groove, which is to be set in 

 the frame of the instrument. When we wish to take ar^ 

 observation, we have only to fix the bottom of the scale 

 in the same horizontal line with the surface of the mercury 

 in the open tube ; and the height of the column may be in- 

 stantly noted, as with the common barometer; only in tliis 

 case, no correction will be necessary for the rising or falling 

 of the surface of the reservoir. An index should be affixed 

 to the lower extremity of the scale, to facilitate the adjust- 

 ment of it to the height of the mercury; and another 

 moyable index, with a vernier scale, may be added to the 



• 9oth the scales are graduated from the same point. 



top, 



