6 RECORD AND DISCUSSION OF TEMPERATURES. 



compared instruments, the exact amount cannot be ascertained. November 23d, 

 1853, on the floe outside, the temperature was 10 lower, and again, December 20th, 

 1853, it was 6i lower than inside. On the 19th of January, 1854, the spirit 

 standard on the floe inside indicated 50, and at the astronomical observatory 

 the temperature was 58. 



To the local difference, in winter, of the temperature of the air incumbent on 

 land, and on ice-floes resting on a sea with a temperature not far from its freezing- 

 point, I have akeady alluded. During the first winter, the temperatures were 

 observed on the floe, but, during the second, on board the brig ; the mean differ 

 ence, for the five coldest months in the two years, amounts to 1.5, and for the 

 absolute minima it is but 0.9 the first being the colder in either case. This 

 result, together with the statement (p. 405, Vol. II. of the Narrative) that local 

 radiations were guarded against as far as possible, leaves no doubt that the recorded 

 temperatures during the coldest months of the first season are not sensibly affected 

 by any local radiation ; at the same time, it must be admitted that, in winter, the 

 ice-covered sea is, nevertheless, a source of heat which, propagated through this 

 cover, is expended by radiation into the colder atmosphere. 



Occasional omissions in the hourly record have been supplied by interpolation ; 

 these values are always indicated by being inclosed between brackets. The process 

 of interpolation will be found illustrated by an example at the end of the record. 



