METEOROLOGY 



developed towards the north-west. These radiants were 

 produced by the apparent convergence, due to perspective, 

 of long belts of cirro-stratus clouds; they could be 

 observed swinging round for several hours from north- 

 west to true north, and even east of north. When this 

 was the case a blizzard was certain to be impending. Un- 

 fortunately, during a blizzard, the air was generally so 

 thick with snow that we were unable to see the top of 

 Erebus. At the end of a blizzard the air current over 

 Erebus became suddenly reversed, the steam-cloud 

 swinging round from the south to the north. After a 

 time, following on the conclusion of a blizzard, a high- 

 level current was seen to be floating the cirrus clouds from 

 the south-east towards the north-west, and the steam of 

 Erebus would stream out towards the north-west. We 

 could not account for this high-level south-easterly cur- 

 rent. It looked like a reversal of the usual upper wind, 

 and it appears to be a fact new to meteorological science. 



As regards the sequence of events during a blizzard, 

 they would seem to be as follows: First, there would be 

 gentle northerly winds at Cape Royds for perhaps one 

 or two days; temperatures would be low, and it would 

 appear as though the air flowing south was coming to 

 supply the void which otherwise would be caused by the 

 contraction of the atmosphere near the pole. Then would 

 follow two or three days of absolute calm, the temperature 

 meanwhile continually falling. We may suggest hypo- 

 thetically this meant that a great mass of air near the 

 pole was constantly becoming heavier and denser as the 

 result of the increasing cold. Sooner or later, perhaps 

 a week after the northerly wind ceased to blow, this 

 heavy mass of cold air would seem to force a passage 

 for itself equatorwards. It would commence rushing 

 out as a south-easterly blizzard, and as soon as this 

 rapid current was started, and even before it had 



409 



