Vlll 
APPENDIX. 
The extremes of temperature of each day were noted by a self-registering 
thermometer suspended amongst them ; the Table shews the highest and 
lowest degrees, as well as the mean temperature of each week ; the latter 
being the mean of the daily extremes divided by fourteen. 
The occasional stoppage of some of the chronometers, and the irregularity 
of others shew, that notwithstanding the precautions which were adopted^ 
the cold which was experienced was greater than they had been prepared to 
meet. 
No. 25 was the soonest affected, stopping, (on two occasions) at about 
15° Fahrenheit; the compensation of this watch was also faulty, as its rate 
varied exceedingly with changes in the temperature. The stoppages were 
conjectured to be owing to the congealment of the oil. No. 369 stopped at 10°, 
and 404 at 6° of the thermometer scale. It is also presumed, that No. 369 
was nearly stopped in the third week of November, when at the temperature 
of 1 5°, it lost in one day 47 seconds, and in the next above a minute, effects 
too considerable in a usually very steady going watch, to have been caused 
by defect in compensation. 
The chronometers of Messrs. Parkinson and Frodsham appear to have 
been far better prepared for the peculiar service on which they were employed, 
than any other of the box chronometers ; no instance occurred of any one of 
them being stopped by the cold. The rate of 228 experienced, indeed, con- 
siderable alteration in the three weeks of severest exposure in February 
and March ; 254 and 259 were affected, but in a much less degree, whilst the 
compensation of 253 could scarcely have been better. 
Mr. Arnold’s pocket chronometers, 2109 and 523, underwent severer triats 
in the course of the winter, from natural cold, than it is probable chronome- 
ters were ever subjected to before, having been used in lunar observations fre- 
quently for three and four hours together, at temperatures from — 20° to — 40°. 
and even so low as —45°. There was certainly a limit of exposure (in degree 
and continuance) beyond which neither of these watches would go, as 2109 
stopped on two occasions, and 523 once, as noted in the Table ; (it is likely 
also that 523 was about to stop on the 29th of November, when on being 
taken into the cabin, it was found to have lost 43 seconds on a chronometer 
with which it had been compared but a few minutes before.) But this limit 
was at a very low temperature continued for a considerable time, and when 
the cold was not sufficient to cause their stoppage, its effect on their usual 
rate was very small; this was especially the case with 2109, in which 
