136 



WIND. 



disappears. It is a simj)le matter to calculate the height at whii^h this occurs. On the mean 

 of the year the following value.s are probably coiTect : — 



Table 73. 



With these values the pressure over the Sea and Barrier are equal at -5,500 feet = 1,680 

 metres, which height is oon.siderably below the height of Erebus whicli is approximately 

 13,000 feet. At the top of Erebus the gradient from the south to the north is almost exactly 

 the same as that from the north to the south at sea level. The isobars will almost certainly 

 run nearly parallel to the edge of the Barrier and vAW therefore be almost due east and west. 

 Hence the gradient wind at the height of Erebus will be due west and this is the cause of 

 the great preponderance of wind from the west shown by the Erebus smoke. 



At first sight it would seem strange that at the height of the medium clouds the \nnd 

 blows so much more frequently from northerly directions than it does either below or above 

 as shown by the surface winds and Erebus smoke respectively. The gradient on the ground 

 is from south to north and the gradient is reversed well below the height of the medium 

 clouds where it becomes from north to south. In the free air above Erebus a nort.h to 

 south gradient produces west to east motion, but in the confined space below Erebus, where 

 the medium clouds form, the air cannot move to the east and therefore moves along the 

 gradient, i.e., from north to south. Thus mo.st of the westerly winds shown by the Erebus 

 smoke are represented by northerly windi at the height of the medium clouds. Thus northerly 

 motion is much more frequently shown by the medium clouds than by Erebus smoke. 



Air motion at the height of the cirrus clouds, above 16,000 feet. — ^Tbe cirrus clouds form 

 at a distinctly higher altitude than that normally reached by Erebus smoke and figure 47(1) 

 shows that the air motion revealed by them is essentially different. Reasons will be given 

 in the next section for believing that the winds in the south-east quadrant both at the height 

 of Erebus smoke and the cirrus clouds are connected with the southerly blizzards and if we 

 provisionally neglect these, the diagram for the cirrus cloud is almost the same as that for 

 the Erebus smoke rotated through a right angle. Thus, in so far as the motion is not affected 

 by the local blizzards Erebus smoke shows an almost exclusive motion from the west to 

 the east while the cirrus clouds show an ecpially exclusive motion from the nortli to the 

 south. Without the south-easterly %\-inds the mean motion at the height of the cirrus clouds 

 would be from somewhat to the west of north, Including these winds brings the mean motion 

 to N. 27° E. which is approximately N.N.E. 



Blizzards and upper air motion. — ^An examination of the four diagrams in the first coluirm 

 of figure 47 shows that at all heights in the atmosphere the frequency of the winds 

 from the south-east quadrant is greater than would be expected fifun the frequenc}' in the 

 south-we.'.t and north-east quadrants. 



The question at once arises as to the connexion between this relatively large frequency 

 in all layers of the atmosphere and the surface blizzards. To investigate this question the 

 upper currents were classified according to the simultaneous winds at the surface. Owing to 

 the small number of observations of the cirrus clouds it was not possible to classify the 

 cloud motion according to all wind diiections and velocities on the gi'ound ; it was therefore 



