MULL. METEOROLOGY. BAROMETER. OO 



itself; as will easily appear by computing the effect of 

 such changes on the mercury according to the known 

 rules. 



The twofold action of air thus included is easily under- 

 stood. When the column approaches its greatest elevation 

 its motion is retarded by the increased elasticity of the 

 confined fluid, a balance of forces being at length pro- 

 duced. Hence imperfect barometers show the greatest 

 differences at their highest elevations; while even these 

 all vary, in proportion to the quantity of air confined, and 

 to the vacant space above the column. These inequalities 

 are also the most sensible in rapid fluctuations, since the 

 fall of the mercury will thus commence before the rise in 

 such barometers has been completed. The effect of in- 

 creased temperature is in some cases of the same nature, 

 by increasing the elasticity of the confined air. But it 

 also acts by producing a depression of the column at times 

 when the mercury would otherwise be at rest. 



From a knowledge of this circumstance a ready test 

 for a barometer presents itself; namely, that of heating 

 the tube above the column by the application of a hot 

 body, and noting if any depression follows. From this know- 

 ledge also may be deduced an useful practical rule ; that 

 of bringing all barometers used in measuring altitudes as 

 nearly as possible to the same temperature ; as the only 

 means of diminishing an error which to a certain extent 

 appears unavoidable, arid which in ill constructed instru- 

 ments may be considerable. 



It is now necessary to say that the limits of the errors 

 from this cause, were estimated by exposing the barometers 

 of imperfect constructions above mentioned to various tem- 

 peratures, at different elevations, and under all the variety 

 of change that occurred throughout one winter. Even 

 where the construction was so bad that the instrument 

 must have been unsaleable, and the changes of tempe- 

 rature were much greater than can ever occur in nature, 

 the differences were never such as to amount to the least 



