longitude of winter harbour. 
Ixxxv 
Of the altitudes in the preceding Table, those on the 5th and 12th of 
October were observed with a repeating circle, and all the others with a 
sextant and artificial horizon, being corrected for index error. 
As it may be conceived that the circle would have been a preferable 
instrument for such purposes, during the time that the ships were stationary, 
it is necessary to explain the reason of its not having been employed. 
In temperatures below Zero, two inconveniences were found to attend 
its use ; first, the bubble of the level became enlarged by the contraction 
of the spirit beyond the extent of the ivory scale designed to register the 
distance of its ends from the Zero of the scale ; thus when the level was 
nearly horizontal, both ends of the bubble were concealed by the brass 
frame work in which the level is enclosed ; secondly, the several joints 
and pivots of ^the instrument gradually stififened, and after a short exposure 
were effectually set fast. 
As it was thought that the latter inconvenience might be in part occasioned 
by the thickening of the oil, the circle was taken to pieces so far as could 
be ventured, and thoroughly cleansed ; but although certainly improved, 
it was not so much so as to become useful, since after six minutes^ exposure to 
a cold of —20°, the instrument could scarcely be moved in azimuth by force 
applied to the pillar, and no other cause remained, as was believed, to which 
to attribute the obstruction, than the unequal contraction of the two metals, 
brass and steel, which are used in its construction. 
It is probable that could this difficulty have been overcome, means might 
have been found of preventing the enlargement of the bubble of the level ; 
which indeed was partly accomplished by the conveyance of heated air 
through a tin tube, filled beneath the level, (the terrestrial telescope being- 
removed), having a chimney to carry the air clear of the instrument. This 
contrivance, though it embarrassed, yet did not impede the use of the circle, 
while it delayed the too great contraction of the spirit for above a quarter 
of an hour at a very low temperature ; it might also have been improved, 
could any method have been devised of preserving freedom of motion in the 
joints and pivots. 
So long as the apparent time could be deduced by the sun, it was just 
possible to obtain a single set of zenith distances before the inconvenience 
amounted to an absolute impediment ; the temperature not being lower than 
— 6°, and the circle having been thoroughly warmed by the cabin fire before its 
exposure ; but when the sun no longer rose to a sufficient height above the 
