[ 240 ] 
one conftant temperature; and thus he became pof- 
feffed with a general notion, of the necefiity of 
reducing the obfervations of the barometer, to fome 
ilandard temperature, upon all occafions. 
But, for the particular purpofe of computing dif- 
ferences of elevation, inftead of attending to the 
correction of the obferved heights of the quick- 
silver at all, it will be a readier way, to make the 
correfponding correction immediately, upon the 
difference of the tabular logarithms. If any quan- 
tity be diminithed or increafed, in a given propor- 
tion, the logarithm of its proportion to any other 
given quantity (that is, the difference of the tabular 
logarithms) is diminithed or increafed by a given 
magnitude ; namely, by the logarithm of the given 
ratio ^ in which the variable quantity is altered. 
The logarithm of the ratio of x to a; -f- _' 6 __ a,*, or of 
96 12, to 9613, is 452 Therefore, if as before, 
exprefs the difference of temperature in degrees of 
bird’s Fahrenheit, ,, x 452 is the correction to be’ 
applied to the difference of the logarithms of the 
obferved heights of the quickfilver in the barometer. 
This correction is to be fubftraCted, if the tempe- 
rature, at the upper ftation, hath been the cooler of 
the two, and to be added, in the contrary cafe. I am 
indebted to the astronomer royal for the firft 
hint of this elegant method of applying the cor- 
rection in queftion. 
(;tz) The rea tier will here recolk-Ct, that I fpeak of the 
tabular logarithms in the language of M. de luc. The loga- 
rithm of the ratio of 9612 to 9613 is, in the more common 
{tile, 0.0000452. 
The 
