238 Mr Anderson's Correction for the Effects of Humidity 
The next example I shall quote, is one in which General 
Roy’s method errs about 13° in defect ; its object being to de- 
termine the height of the Peak of Snowdon, above Caernarvon 
Quay. The difference of elevation was ascertained, by a care- 
ful trigonometrical measurement, to be 3555.4 feet. 
At the time when the observations were made, which was at 
8 h 27 m A> Mt? ^ is stated, there was “ a fog above? The ba- 
rometer and thermometer, at the two stations, then stood as sub- 
joined : 
Bar. 
Att. Therm. 
Det. Therm. 
Caernarvon Quay, 
29.984 
56| 
55J 
Snowdon Peak, 
26.271 
42 1 
43 
As a fog is mentioned to have existed above f at the time 
the observations were taken, the point of deposition may be as- 
sumed to have been 5° below the temperature of the air, at the 
lower station, and 2° below it, at the upper. According to this 
supposition, which cannot be far from the truth, we should have 
f=. 382, and f- =.274. The several coefficients determined, as 
in the preceding examples, and multiplied together, give 3560.8 
for the height in feet, which differs only 4f feet from the true 
height. 
The only example given, by General Roy, which seems to 
set the corrections for humidity at defiance, is the one which was 
furnished by his observations, to determine the height of Moel 
Eilio above Caernarvon Quay. Though the weather, at the 
time the barometrical observations were taken, must, from what 
is stated, have been exceedingly damp, the result, by the com- 
mon formula, contrary to what generally happens, greatly ex- 
ceeds the geometrical height ; nor is it possible to reconcile them, 
by any supposition that can be made, respecting the hygrome- 
tric condition of the atmosphere, were it even supposed (what 
cannot be admitted) that the air was in a state of absolute dry- 
ness, over both stations. Whether this anomalous result is to be 
ascribed to some inaccuracy in the geometrical measurements of 
General Roy, or to an error in his barometrical observations, I 
am the less disposed to hazard an opinion, as the trigonometri- 
cal data, which he employed to determine the height, do not af- 
ford sufficient checks to detect errors, if they existed. Two an- 
gles of elevation were, indeed, measured at different stations ; 
