307 
trigonometrical measurements at Spitsbergen. 
example, it is very probable that the mercury may have been 
heated by the warmth of the person who carried it up the 
hill, which heat it would not part with so readily as the ther- 
mometer suspended by its side : I believe this to be a very 
frequent source of error in barometrical measurements, and 
that the insertion of the attached thermometer in the mer- 
cury itself in the cistern, is a great practical improvement. 
This error would have most effect, when the difference was 
greatest between the temperature of the air and that of 
the human body ; but it would render the computed height 
less than the correct, whereas Dr. Irving's measurement is 
already in excess. So also, as the barometer does not seem 
to have been furnished with a means of adjusting the scale 
to the level of the cistern, and as no correction appears to 
have been made for the descent of an inch of mercury from 
the tube into the cistern, its true height may have been, from 
this cause, actually less than the observed ; but this also would 
encrease the elevation of the hill. 
I had supposed it possible that the view of the summit, which 
was the station of the barometer, might have been inter- 
cepted from the low ground on which the base was measured, 
and from whence the angles of elevation were taken ; but 
this conjecture was not borne out on the spot. I must leave 
it, therefore, in its former uncertainty, though I trust with 
this difference, that the question is no longer of the same 
interest or importance as before. 
Having thus detailed the particulars of this comparative 
measurement, I may be allowed to notice, that I have had 
much disappointment in not having it in my power to try 
the experiment on a hill of greater elevation ; but those 
