298 



THE HEIGHT OF THE BARRIER AND THE SOUTH POLAR PLATEAU. 



July. To realise this it is only necessary to remember that there would be no problem if 

 the pressure actually found by observation at sea-level were 11 mm. higher in January than 

 in July, for then the pressure over the whole globe would be equal in these two months 

 without the presence of the continent. The high continent is only necessary to account for 

 the fact that sea-level observations do not give this diSerence in pressure between January 

 and July. 



From his discussion of the pressure data Meinardus concluded that the average pressure 

 at sea-level over the whole Antarctic was the same in January as in July. 



Since Meinardus \vrote the following additional data have become available : two years' 

 observations in McMurdo Sound, and one year's observations at Framheim, Cape Adare, L'ile 

 Petermann and in ths Weddel Sea. All these new observations show a lower pressure in July 

 than in January. 



In the following table all the available data tcom Antarctic stations have been collected.* 

 The average difference of pressure is given for each station as determined from the actuals 

 for each month, and then again after the data have been smoothed in the usual way by 



applying the formula b (smoothed) = — j to the monthly values. 



With the large unperiodic changes of pressure from month to month which are such a 

 featiu-e of the Antarctic conditions there can be Uttle doubt that the smoothed values are 

 the better, for they eliminate to some extent large irregularities and therefore more nearly 

 approach what would ultimately ba obtained from a longer series of observations. 



In obtaining the means of the groups of stations in columns 5, 6, 8 and 9 the differences 

 have been weighted according to the number of months from which the individual means have 

 been derived. 



Table 154. 



Pressure differences Jamuiry-Jithj in mm. 



* For actual values used and references see Volume' III. 



t Five days' observations for December, thirteen for January and twenty-eight for February only are available. 



In obtaining the smoothed values these have been weighted according to the number of days, double weight being 



^,5 December -|-2xl3 January -\-2& February, 

 given to January thus ^^ ^ 



% See figure 1 and table 1 of Veroffen des K. Preuss. Met. Institutes, No. 265, Bd. IV, No. U 



