100 WIND. 



frequency of high winds was the same during September in both years, yet September, 1911, 

 had 25-6 per cent, of calms and September, 1912, only 33 per cent. 



There is obviously a third factor at work, and this is the cold layer of air which forms 

 near the surface during the winter months. Wlien this layer is present the wind moves over 

 its upper surface, while the layer itself remains stationary. It is only when the upper wind 

 becomes great that the cold layer is removed ; thus during such periods the vnnd conditions 

 are a succession of calms or high winds. The presence of the cold layer is shown by the 

 low temperature and we find that months with a high proportion of calms have low tem- 

 peratures. Thus March, 1911, and July, 1911, had practically the same number of high winds, 

 but the former had no calms while the latter had 17' 1 per cent. But the mean temperature 

 of July, 1911, was -21-1°F., while that of March was +7-2T. 



The winter conditions in the two years 1911 and 1912 are very instructive. The whole 

 of McMurdo Sound was firmly frozen over during the winter of 1911, but it was open during 

 the greater part of the winter of 1912. Thus while the cold layer could form with ease 

 over the thick ice in 1911, it could not form over the open water in 1912. It is therefore 

 not surprising to fijid that from May to September, 1911, with a mean temperature of — 16-5°F. 

 the percentage of calms was 22-4, while during the same period in 1912, with a mean 

 temperatuie of only — 6-5°F., it was as low as 5' 5. 



The contrast between the conditions during these two winters is a good example ot how 

 the temperature and wind interact on one another. If there are not many storms, the ice 

 can form over the sea and a stagnant layer of cold air develop above the ice. Over the 

 cold layer moderate winds pass \vithout disturbing it, thus giving a calm near the ground 

 and so reducing still further the mean wind velocity. A few high winds remove this layer 

 and also the ice and in consequence the gradient wind extends down to the ground and so 

 increases the mean wind velocity and also the mean temperature. 



Thus during the winter the wind conditions are in a state of unstable equilibrium. High 

 winds make the conditions favourable to more winds, while light winds produce conditions 

 favourable to calms. 



Daily Variation of Wind Velocity. 



Curves of the daily variation of wind velocity in each season are shown in figm-e 32. 



These curves are based on all the data available, namely : November, December, .January 

 two years ; February, March, October three years ; April, May, June, July, August, September 

 four years. The data will be found in the volume of tables. 



As data for the first two years are only given for two hourly intervals it has been 

 possible only to combine two hourly values for the second two years. The mean velocity 

 during the two hours midnight to 2 a.m. has been plotted against 1 a.m. and so on. 



Except in the winter months the daily variation of wind velocity has a maximum in 

 the early afternoon and a minimum soon after midnight. This is the usual form of the daily 

 wind variation and is explained by the convexion currents which are set up in the day time. 

 The upper air usually moves with a greater velocity than the lower air and the convexion 

 currents act as a connecting link and convey momentum from the upper to the lower air. 

 Thus when the convexion currents are most active in the early afternooii the lower air moves 

 faster than during the night when, owing to the absence of convexion currents, the upper 

 air moves over the lower air without tending to drag it along. 



Evidence of appreciable convexion currents over McMurdo Sound even when the Sound 

 is frozen over and the ice covered with snow was found from the vertical temperature 



