MOTION OF THE UPPER ATMOSPHERE. 



137 



decided to separate out the upper motions when the wnd at the surface was 20 niiles per 

 liour or more from the soutli, and compare these with the roniainins observations. The former 

 class will then represent blizzard \\-inds and a comparison with the remaining observations 

 will show if the blizzards d(i affect the motion of the upper atmosphere. 



The following table gives the numerical results and they have been plotted in the second 

 and third columns of figure 47. 



Table 74. 



Upper air motion, according lo /rimls at the surface. 

 Percentage frequency. 



Examining first the diagrams for medium clouds we see that with blizzard winds the 

 frequency from the S.S.E. and S.E. has grown at the expense of the winds from the N.N.W. 

 and N.W. Rememljoring that these clouds are contained in the channel over McMurdo Sound, 

 this is the result we should expect, if the pressure conditions causing the blizzards extended 

 so high as the medium clouds. It is interesting to notice that there is a not inconsiderable 

 number of winds at the lieiglit of the medium clouds which move against the high southerly 

 winds at the surface, wliicli only emphasises the fact pointed out several times previously 

 that large dift"crences of air motion exist separated by very small vertical distances. 



The diagrams for Erebus smoke are interesting, for the southerly wind diagram is very 

 similar to that for the remaining winds, except that the frequency of the south-east winds 

 is more than doubled. This seems to point to blizzard conditions being superposed on the 

 normal conditions, and not that the normal conditions undergo a change which results in 



la 



