THE HEIGHT OF THE ANTARCTIC CONTINENT. 297 



these two months has a definite value. Finally it is assumed that the pressure changes on 

 the surface of a high land at this height will be the same as those fomid for the free air. 



It is this latter assumption which it is necessary to investigate. 



Why should the pressure over the high continent bear any relationship to the pressure 

 in the free atmosphere over the low-lying area ? There is no air under the continent to 

 expand and contract with changes of the season and so to regulate the pressure on the 

 surface. Probably the idea which underlies the assumption is that the pressure at the ed"e 

 of the high land must ba in equilibrium with that in the surrounding free air, and that this 

 equilibrium decides the pressure over the whole of the adjacent continent. 



But we know quite well that the pressure conditions within any area cannot be deter- 

 mined from a knowledge of the pressiure conditions around the boundary of that area. The 

 interior pressure conditions can ba changed indefinitely without altering the boundary pressure 

 conditions provided that the air is free to move. For example, the 760 mm. isobar of an 

 anticyclone may occupy at one time exactly the same position as was occupied at another 

 time by the 760 mm. isobar of a cyclone. In this case the boundary pressure conditions 

 are identical but the interior pressure conditions are entirely different. 



Thus the pressure at the boundary of the high land might always be in equilibrium with 

 the pressiure of the adjacent free atmosphere without giving any indication of the pressure 

 over the interior of the high land. We must therefore conclude that the average pressure 

 on the siurface of the high land may baar no relationship to the average pressure existing 

 at corresponding heights in the free atmosphere. 



It may be contended, however, that the present problem is not concerned with the actual 

 pressure, but with the change in pressure between January and July, and if the pressure 

 in the free air is 11 mm. higher in January than in July it is probable that the mean 

 pressure over the high land will be higher by the same amount. 



If this were the case it would mean that the pressure over every element of the land 

 surface had changed by the same amount as the boundary, the consequence of which would 

 be that the pressure gradients would remain the same over the high land in January as in 

 July.* Now it is almost inconceivable that the pressure gradients, whatever they may be, 

 are the same over the Antarctic high land in the winter as in the summer. Let the pressure 

 distribution be cyclonic as supposed by Meinardus or anticyclonic as is more generally supposed, 

 the system is bound to be more intense in the wmter than in the summer, which means 

 that the gradients are steeper in winter than summer and the change in mean pressure greater 

 or less, according to which system is chosen, than the change of pressure at the boundary. 



The physical principles underlying Meinardus's theory thus indicate that the changes in 

 pressure over the high land from January to July may bear no close relationship to the 

 changes in pressure at corresponding heights in the surrounding free atmosphere. Hence cal- 

 culations based on the pressure changes in the free atmosphere may give an entirely wrong 

 value for the average height of the Antarctic Continent. 



We will, however, for the sake of a thorough investigation of this important and interesting 

 problem, now consider the meteorological data used by Meinardus to see the accuracy of his 

 calculations irrespective of the above conclusions. 



The pressure data. — ^The actual sea-level pressure is of only secondary importance in the 

 calculation, the result depending chiefly on the difference in pressure between January and 



* It 13 of course possible to conoeivo an entirely now pressure distribution which would give the correct 

 change in the moan pressure, but this would bo a matter of pure chance and therefore cannot be taken into 

 axount. 



38 



