THE BAROMETER. 285 



and rise likewise, both producing an aeriform fluid in the chamber above the 

 surface of the mercury, which presses upon that surface with an elastic force 

 ind produces a corresponding diminution in the height of the column of quick- 

 silver, sustained by the atmosphere as already explained. This imperfection 

 may be avoided by previously heating the tube. The particles of air which 

 adhere to its inner surface being thus expanded by heat, will fly off by their 

 elastic force, and the particles of moisture will be converted into vapor, and 

 likewise disengaged from the surface. 



All the effects now explained may be produced by filling the tube with mer- 

 cury in the first instance and then boiling the liquid in it, which may be easily 

 accomplished. The heat will not only expel all liquid and gaseous impurities 

 from the mercury itself, but also will disengage them from the inner surface of 

 the tube. These precautions being taken, the column of mercury sustained in 

 the tube will indicate by its weight the true amount of the atmospheric pressure. 

 But in order to be able to compare the result of any one barometer with any 

 other, it is necessary that the weights of equal bulks of the liquid mercury 

 used in both cases should be the same ; and for this purpose we must be as- 

 sured that the mercury used is pure, and not combined with other substances. 



We have just seen how all substances in the liquid or gaseous form may 

 be extracted from it. Impurities may still, however, be suspended in it in a 

 solid form. 



To remove these it is only necessary to enclose the mercury in a small bag 

 of chamois leather : upon pressing this bag the quicksilver will pass freely 

 through its pores, and any minute solid impurities which may be contained in 

 the mercury will remain in the bag. Pure and homogeneous mercury being 

 thus obtained, we have advanced another step toward the certainty that the in- 

 dications of different barometers may correspond ; but there is still one other 

 cause of discordancy to be attended to. Suppose a barometer to be used in 

 Paris, and another in London, at a time when the pressure of the atmosphere 

 in both places is the same, but the temperature of the air at Paris is higher 

 than the temperature of London. The mercury in the one barometer will have 

 a higher temperature than the mercury in the other. Now it is well known 

 that when mercury or any other body is heated, its dimensions increase. In 

 other words, bulk for bulk, it becomes slighter. Consequently, if two columns 

 be equal in weight, that which has the higher temperature will have the greater 

 altitude. Hence it appears, that under the circumstances supposed, at a time 

 when the atmospheric pressure is the same in London as at Paris, the barom- 

 eter at the latter place will be higher than at the former. To guard against 

 this source of error, it is necessary, in making barometric observations, to note 

 at the same time the contemporaneous indications of the thermometer. Tables 

 are computed, showing the changes in the height of the mercury correspond- 

 ing to given differences of temperature. It is evident that in comparing the ( 

 results of the same barometer observed at different times, it is equally neces- 

 sary to note the difference of temperature, and to allow for its effects. This, 

 however, is a refinement of accuracy which is not attended to, except in ob- 

 servations made for philosophical purposes. 



One of the difficulties attending barometri ; observations arises from the very 

 minute changes produced in the height of the column by slight variations in 

 the atmospheric pressure. The whole play of the upper surface of the column, 

 in the most extreme cases, does not exceed three or four inches in a given 

 place ; and mercury being a very heavy fluid, a variation in the pressure of the 

 atmosphere, of sensible amount, may produce scarcely any perceptible change 

 in the" height of the column. One of the most obvious remedies, at first viiw, 

 would seem to be the use of a fluid lighter than mercury. In the same psopor- . 



-s^X^-Xy-* 



