250 Mr Anderson’s Corrections for the Effects of Humidity 
wliich it demands, increases both the chance of error, and tlie 
labour of measurement ; and hence this method of determining 
differences of elevation is never resorted to, unless in cases where 
the object to be attained is sufficiently important to afford an 
adequate remuneration for the time and trouble .which must be 
devoted to it. 
Barometrical measurements, if they could be relied upon, 
would frequently be greatly preferable, both to levelling, and 
the geometrical methods. It is to be feared, however, that, not- 
withstanding the numerous and refined corrections, by which the 
formulae for deducing heights by the barometer have been gra- 
dually modified and improved, these formulae will not always 
lead to the same result, in different states of the atmosphere. 
This is chiefly to be ascribed to the neglect of one very import- 
ant element in the calculation, namely, the humidity of the air, 
which has either been entirely disregarded, or introduced into 
the formulae, by such as have attempted to make a correction 
for its influence, in so vague and general a form, as to be inap- 
plicable to cases, in which there is a considerable deviation from 
the mean hygrometric state of the atmosphere. The new coef- 
ficients, which I propose to apply to the common formula, will, 
I am convinced, account for several anomalous results, which have 
hitherto received no satisfactory explanation, as well as remove 
all the sources of error which attach to barometrical measure- 
ments, when they are made in different states of the air, in re- 
spect of humidity. To point out, in- a more distinct manner, 
the nature of these coefficients, it may be proper to bring before 
those, who have not particularly attended to the subject, a view 
of the formula which is commonly employed, as it has been im- 
proved by the labours of De Luc, Sir G. Shuckburgh, and Ge- 
neral Roy. 
It may be remarked, then, at the outset, that, if the air pos- 
sessed perfect elasticity, and uniformity of temperature through- 
out its whole mass, it is demonstrable, that, if /3 and h repre- 
sented the heights of the mercurial columns in the barometer, 
at the upper and lower stations, respectively, the difference of 
elevation A, expressed in some assumed measure of length, would 
be expressible, by an equation of the form 
