464 



The results are shown in tabular form in Table 1. Columns A and B give the mean of the 2-hourly 

 observations of the dry and wet bulbs. Column C gives the mean of the warmest day, and column D of 

 the coldest day in each month. 



TABLE I. 



* Below the freezing point of mercury. 

 MEAN for each Month. 



The low temperature of the summer and the uniform temperature from April to September are 

 noteworthy. No great differences of the corresponding months of each year are shown, excepting in July, 

 which had -21'0 in 1903, and -8'5 in 1902, but even this amount of difference (12'5) may be surpassed 

 by two consecutive Januarys in England. Two years is too short a period to give a reliable average, but 

 the consistency of the figures throughout leads to the conclusion that the mean of the whole period (- T3) 

 is not far from the right value. 



It is curious that in each case December should have the only days showing temperatures above the 

 freezing point and be the warmest month. It would be more natural to find the highest temperature in 

 January or February, and possibly a longer series of observations might put January first. But the 

 figures all show that solar radiation has the most important influence on the temperature. Indeed, the 

 shape of the annual temperature curve is very striking; the rise from September to midsummer, and the 

 fall back to April, is what might be expected, but the uniform temperature of the six months from April 

 to September is quite exceptional. 



In the latitude in which the ship was ice-bound for two years, namely 77 51', the sun would be almost 



