64 



TEMPERATURE. 



Ill figure 2o the temperature amplitudes given in tables 23 and 24 (pages 59 and 60) Lave 

 been plotted against the solar energy amplitudes given above, the points for the different 



Fig. 23. ArapIituJe of temperature plottefl against amplitude of solar radiation. 



stations being indicated by various signs, and the months to which they refer by entering 

 against each the initial letters of the months. For the sake of comparison the months for the 

 Fram have been marked both by their own initials and by the initial letters of the correspond- 

 ing months in the south. 



The first striking feature of this diagram is the relatively small variations from a linear 

 relationship shown by most of the points. Neglecting for the moment the point for the 

 Barrier, twenty-four of the twenty-seven points lie \vithin the two dotted lines. This leads to 

 the conclusion that except for small differences the relationship between temperature amplitude 

 and energy amplitude is linear. 



The three points which do not lie within the dotted lines are the most important for 

 our present discussion. It will be noticed that they are the spring months of the Fram type, 

 being the Pram's observations for April and the Gauss's observations for September and 

 October. Both Mohn and Meinardus in their discussion of the temperature amplitude tried 

 to explain the low values of the amplitude in the summer, we now see that it was the high 

 amplitude in the spring which needs explanation. 



The abnormal temperature amplitude on the Barrier is clearly shown by the position 

 of the point for the Barrier near the top of the diagram, and it is the position of this 

 point which supplies the solution of the whole problem. 



We have already shown that the difference in temperature amplitude over the Barrier 

 and McMurdo Sound is due to the sea ice preventing low nocturnal radiation temperatures. 

 This action of the sea ice could not exist if there were a thick layer of loose snow upon it. 

 Now both the Fram and Gauss were surrounded by sea ice, it is therefore important to 

 examine the varying character of the snow covering of the ice during the year. Unfortunately 

 I cannot find in Mohn's discussion of the Fram's observations anv remarks on the character 



