110 Measurement of Heights hi/ ihe Barometer, [Aug; 



same temperature by means of Table 4; I then take the differ- 

 ence of the logarithms : this gives a basis for finding the proba- 

 ble temperature at the lower station according to the most ordi- 

 nary law of decrease. For this purpose it will sufiice to reduce 

 the difference to these decimals, and multiply it by the constant 

 number 122. The product expresses in degrees and decimals, 

 the quantity by which we must augment the temperature indi- 

 cated by the higher thermometer to get approximately that at 

 which the lower ought lo stand. By this means the correction 

 for the temperature of the air is applied. 



The factor 122 expresses the mean rate of decrease as deduced 

 from the collection of observations before-mentioned, and agrees 

 with the law of decrease adopted by some philosophers from 

 theory; but though this may inspire some confidence in the 

 observations upon which the value of that factor depends, yet it 

 inspires none for this mode of measuring heights. I find that 

 this method does not sufficiently correct the deviations of the 

 simple method ; and moreover the errors instead of compensat- 

 ing each other, are almost always on the same side, and tend 

 generally to diminish the real height. This tendency is remark- 

 able ; it has little to do with the decrease of temperature 

 assigned ; for we perceive it equally in the uncorrected method, 

 It rather concurs with other circumstances to make me suspect 

 what has been most generally considered the mean height of the 

 barometer at the level of the sea. If I can sufficiently confide in 

 my own observations, this mean is taken too low. If it be raised 

 to 30*089 inches, for the hour of noon, and temperature 12*6°, 

 there will be rather more equality between the cnances of error 

 in the different modes of calculation in which this mean is 

 employed. 



fiut whatever may be the corrections which we maj^ ^pply to 

 the approximate methods which have caused this digression, 

 they will be always very faulty methods. We shall never have 

 recourse to them but with distrust, and only in order to estimate 

 within about 30 feet, the elevation of a place when we have no 

 means of obtaining a more precise determination. The baro- 

 meter after all does no more towards the measurement of heights 

 without corresponding observations than the repeating circle 

 witliout an exact determination of distances. 



Mode of conducting the Observations. 



To understand the theory of barometrical measurements is by 

 no means difiicult : it is even still easier to learn to calculate the 

 observations well, but the most difficult thing is to perform them 

 well. Very skilful persons have often given extremely deficient 

 observations, as well from want of good instruments, as of good 

 methods; and always from having considered as more easy than 

 it really is, what appears to be a very simple physical experi-» 



