DAILY VARIATION— SEPTEMBER TO MARCH. 



63 



amplitude of solar energy between the spring and summer to account for the decrease of 

 temperature amplitude. 



As the temperature of an ice surface cannot rise above the freezing point, it is obvious 

 that the temperature amplitude will be decreased as soon as the mean temperature closely 

 approaches that point. This is in all probability the reason why the warmest months of the 

 McMurdo type of daily temperature variation have a smaller amplitude than those on either 

 side. Even when the mean temperature of the month is low, as at Framheim and McMurdo 

 Sound, there are a few days in the month when the maximum temperature might have gone 

 higher if the melting of the ice had not prevented the complete warming up of the surface. 

 It must be remembered that the temperature ampUtude of the surface itself is many degrees 

 greater than that of the air, and therefore the surface might reach and strive to pass the 

 freezing point on many days with air temperature well below that point. 



But this explanation does not account for the rapid decrease in amplitude which takes 

 place in the Fram type a long time before the mean temperature gets near to the freezing 

 point. Thus in the case of the Fram the amplitude falls from 5-9°F. in April to half its 

 value in the following month, yet in the latter month the mean temperature was practically 

 twenty degrees below the freezing point. The same was the case, but to a lesser extent, 

 with the Gauss. In November the maximum air temperature only once reached the freezing 

 point and throughout the month the mean maximum temperatiue was 7'5°F. below the 

 freezing point, yet in spite of this tha temperaturb amplitude decreased from 9-3°F. in October 

 to 7-5°F. in November. 



Suggested Explanation of the Variation in the Amplitude During the Period September to 

 March. — There can be no doubt that the prime factor in determining the daily temperature ampli- 

 tude is the solar energy amplitude. The first step in the investigation should, therefore, be 

 to plot the observed temperature amplitude3 against the ca'culated energy amp'itudes to see 

 if there is any close relationship. This has been done for two stations of the Fram type and 

 two of the McMurdo type, namely, (a) Fram and Gauss : (6) McMurdo Sound and Snow Hill, 

 and the observations made on the Barrier have been included for reference. Mr. Normand 

 of the India Meteorological Department, has very kindly made the calculations of the solar 

 energy amplitude by the method described by Mohn on page 588 of his discussion of the 

 meteorological results of the Fram Expedition. This method takes into account both the 

 energy received directly from the sun and also that diffused from the illuminated atmosphere. 

 The following table contains the results- of the radiation calculations :— 



Table 27. 



Solar Radiation (gram. cols, per min. per sq. cm. of horizontal surface). 



