300 THE HEIGHT OF THE BARRIER AND THE (SOUTH POLAR PLATEAU. 



forgotten, however, that the mean obtained from coast stations in all probabihty gives a value 

 too low by several millimeters. Table 154 shows that the mean of all the observations taken 

 in the Antarctic is a difierence for January-July of 3- 00 mm. or 3- 87 mm. according as 

 actual or smoothed values are considered. From this I think we are justified in saying that 

 the average pressure difference at sea-level over the Antarctic from January-July is as 

 likely to be +3 mm. as it is to ba 0, which is the value accepted by Meinardus. 



If wa apply the value +3 mn. instead of mm. in Mainardus's formula, leaving all 

 other assumptions unchanged, we find that the average height of the land within the Antarctic 

 Circle works out to be 966 metres, which is very different from Meinardus's value of 1,350± 

 150 metres. 



The tempsmture data.— It is as nscsssary to know the average temperature in January 

 and July over th3 whole Antarctic as it is to know the pressure, and again it is the differ- 

 ence, much more than the actual values, which affects the result. 



After a careful discussion of all the data available in 1909, Meinardus concluded that the 

 average temperature at sea-level over the whole Antarctic in January is — 2-7°C. and in 

 July -25-2''C., giving a difference of 22-5°C. 



Again Meinardus neglected the fact that all the observations used by him were obtained 

 at coast stations. Who would think of determining the average range of temperature over any 

 other continsnt from a fiw scattered observations at coast observatories ? Here our discussion 

 of the Barrier temparatures comes to our help. In table 4-5, page 83, it is shown that the 

 difference in temparature between January and July is 21°C. at Cape Evans and 29 "C. 

 on the Barrier. This indicates that the temperature difference January- July may be 8 C. 

 more at a land station than at a coast station in the same latitude only a few miles distant. 



It therefore does not seem unreasonable to conclude that, considering how much of the 

 Antarctic is far removed from the ssa, Meinardus's value of 22-5"'C. for the difierence in 

 temperature between January and July obtained entirely from coast stations is much too 

 small. 



If Meinardus's value for the July temperature is lowered by only 5°C., leaving all other 

 of his assumptions unchanged, the average height of the surface within the Antarctic Circle 

 is reduced to 1,050 metres. 



The vsrtical temperature gradient.— Meina,rdus assumes that the vertical temperature gradient 

 is -•5°C. per 100 metres both in January and July. Now the balloon observations made 

 in McMurdo Sound and by Barkow in the Weddel Sea have proved that during the winter 

 there are large temperature inversions in the lower atmosphere. There are not sufficient balloon 

 observations to give an average value of the temperature gradient during July, but there 

 cannot be the slightest doubt that it is much less in July than in January. In fact it is 

 not improbable that over inland areas where the temperature is very low and where the 

 wind is much less than near the coast there is a nearly permanent temperature inversion 

 up to heights of at least 2,000 metres. It therefore appears to me that the average vertical 

 temperature gradient over the whole Antarctic in July is as likely to be as ■5°G. 

 per 100 metres. 



Putting this value in Meinardus's formula and leaving all other values unchanged the 

 average height of the surface within the Antarctic Cii'cle is raised to 1,610 metres. 



It is of interest to see what is the effect on the calculation of height if all three of 

 the above new values are introduced into the formula. In the following table the new and 



