THE BAROMETER. 



283 



for the purposes to which it is intended to be applied, the surface F may be 

 regarded as fixed, and the whole change in the height of the column may be 

 taken to be represented by the change in the position of the level E. All or- 

 dinary barometers are constructed in this manner. But it is not difficult to ad- 

 just a scale upon a tube which will give with accuracy the actual variation in 

 the length of the column by means of the change in the level of the surface 

 E. Let us suppose that the cistern P D has a flat, horizontal bottom and per- 

 pendicular sides, and that the magnitude of the bottom bears a certain known 

 proportion to the bore of the tube. Suppose this proportion to be that of a 

 hundred to one. If the pressure of the atmosphere increase, so as to cause the 

 column of mercury sustained in the tube to be increased in height by one inch, 

 then as much mercury as fills one inch of the tube will be withdrawn from the 

 cistern ; but as the base of the cistern is one hundred times greater than the 

 bore of the tube, it is evident that this inch of mercury in the tube would only 

 cause a fall of the hundredth of an inch in depth of the mercury in the vessel. 

 Consequently it follows that the increased elevation of an inch in the column 

 produces a depression of a hundredth of an inch in the surface F. Thus it 

 appears that the increased length of the column E F, is produced by the sur- 

 face F, falling through the one hundredth of an inch, while the surface E rises 

 through ninety-nine hundredths parts of an inch. The same will be true 

 whatever change takes place in the height of the column. We may therefore 

 infer generally, that whatever variation may be produced in the surface E, the 

 consequent variation produced in the height of the column is greater by a 

 ninety-ninth part. 



If, then, the top be so graduated that a portion of it, the length of which is 

 one hundredth part less than an inch, be marked as an inch, and all other di- 

 visions and subdivisions marked according to the same proportion, then the 

 indications will be as accurate as if the surface F were fixed, the tube being 

 divided accurately into inches and parts of an inch. 



Fig. 3. 



The barometer is represented mounted and furnished with a scale, in fig. 3 

 The glass tube is surrounded by one of brass in which there is an aperture cut 



