THEORETICAL DISCUSSION. 265 



medium clouds for the Gauss Station for there is no doubt that if observations of the motion 

 of Erebus smoke and of the winds on the plateau are to be used in one part of the area 

 the medium and not the cirrus clouds must be used in another. This apphes also to the 

 air motion over the west Antarctic where Meinardus has also used the direction of motion 

 of the cirrus clouds, but in this area there is little difJerence in the direction of the alto- 

 cumulus clouds and the cirrus clouds* so that no radical mistake has been made in this region. 



Further the prevailing wind on the plateau near the magnetic pole is shown by Meinardus 

 as W.S.W. Now this direction was observed by David only when he was under the influence 

 of winds blowing down the glacier valley through which he ascended. As soon as he reached 

 the true top of the plateau the winds observed and the sa.strugi all pointed to the prevaiUng 

 winds on the plateau itself being from the south or south-east, a conclusion confirmed by 

 Mawson's Expsdition. 



Finally, although th? isobars over the plateau near the South Pole have been drawn out 

 by Meinardus to almost a point to make them roughly parallel to the wind there shown, 

 the result is very unconvincing for the wind is blowing straight out of the area of low pres- 

 sure. This would b3 even more striking if the tnie wind direction over the plateau as now 

 known had been eiatered; for this direction is parallel to the 150th east meridian and not to 

 the 180th meridian as shown on the diagram. 



If the reader will enter on figure 77 these corrected wind directions namely, on the 

 South Polar Plateau wind parallel to the 150 E. meridian, on the Magnetic Pole Plateau 

 S.S.E., over the Gauss Station N. 74° E., he will see at once that it is quite impossible to 

 reconcile them with the pressure distribution shown by Meinardus. 



If then we may accept figiu-es 81 and 82 as correctly representing the pressure distri- 

 bution at sea-level and at 3,000 metres respectively we reach the following conclusions : — 



(a) The pressure distribution over the surface of the plateau and also of that part of 

 the Antarctic at sea-level is anticyclonic. 



(6) At 3,000 metres over the part of the Antarctic at sea-level the pressiire distribution 

 is cyclonic. 



Thus Hobbs appears to be right in his main contention that the siuface of the land 

 both high and low is subject to anticyclonic conditions, but we are still faced with Meinar- 

 dus's main contention, that under such conditions evaporation will exceed precipitation and the 

 Antarctic should be denuded of its permanent snow-covering. 



This problem we must now consider ; it may be stated as follows : 



Owing to the anticyclonic conditions which predominate over the whole surface of the 

 Antarctic air flows towards the Antarctic in the upper atmosphere, then descends to the 

 surface and there flows outwards, at a much lower level than it entered. If the air were 

 saturated at the moment it entered the Antarctic it would be warmed up dynamically as it 

 descends and so when it reaches the surface it will be far from saturated. Even allowing 

 for a large amount of radiation, air under these conditions could not deposit appreciable 

 moisture on the surface and as Meinardus points out, the conditions when precipitation is 

 known to take place to the greatest extent, i.e., during cloudy weather and high winds, are 

 exactly those when cooling by radiation is least effective. We have therefore to explain how 

 air can enter the Antarctic in the upper atmosphere and leave it in the lower atmos- 

 phere and yet deposit moisture in the process. 



* According to the Argentine observations at South Orkneys alto-cumulus and alto-stratus S. 66° W., cirrus 

 S. 77 W., loc. cit., page 124. 



34 



