Observations to determine the amount , &c. 207 
ing circle, and then to have computed the true altitude ; 
whence the actual refraction might have been deduced. 
The difficulties, however, attending the use of the repeating 
circle, during the winter of the polar regions, have already 
been alluded to on several occasions, in the accounts of the 
two preceding voyages of discovery. The most material of 
these consist in the extreme contraction of the spirit in the 
long level, when filled in the usual way ; the instantaneous 
freezing of the breath or other vapour on the glasses, oblig- 
ing the observer to hold his breath during each observation ; 
and the pain, amounting to the sensation, and producing the 
effects of burning consequent on touching intensely cold 
metal with the naked hand. The first of these was obviated, 
on the present occasion, by inserting a larger quantity of 
spirit than usual, so as to keep both ends of the bubble in 
sight, even during the most intense cold : this latter circum- 
stance, however, afforded the opportunity of remarking an 
increased sluggishness in the level at very low temperatures, 
arising possibly from a certain degree of thickening in the 
spirit, which required the instrument to stand unmoved for 
at least two minutes after the contact had been made, in order 
to insure an accurate reading. It is unnecessary to point out, 
how unfavourable to minute accuracy this circumstance must 
prove, in observing an object having quick motion, either in 
altitude or in azimuth. A set of zenith distances, consisting 
of only eight observations, cannot, indeed, under such cir- 
cumstances, be satisfactorily obtained in less than thirty-five 
or forty minutes. If to the difficulties already mentioned be 
added the annoyance sometimes experienced by the extinc- 
tion of the lamp for illuminating the wires during an obser- 
