479 



the general circulation of the atmosphere ('Proc. Roy. Soc.,' vol. 74, p. 20), based largely upon the differ!in-<- 

 between the distribution of pressure at the Earth's surface and the distribution at the 4000 metres level as 

 computed by M. TEISSERENC DE BORT, came to the conclusion that a high-pressure area exists over the 

 Antarctic, and he pointed out that "the pole of extreme cold, which in the southern hemisphere must be 

 nearly coincident with the geographical polo, will have associated with it a component distribution for 

 easterly (wind) circulation similar to the low-temperature pole of north-east Siberia in the northern 

 hemisphere," a conclusion which seems to be completely justified by the mean pressure results shown in 

 Table I., and also by the wind distribution recorded at the " Discovery's " Winter Quarters, as well as at 

 Cape Adare and at the position at which the " Gauss " wintered in the ice. Dr. SHAW proceeds to say, 

 " the intensity of the cold in the south polar regions is undoubted, and the existence of the distribution for 

 an easterly circulation round a high-pressure centre due to the weight of the lower air follows directly 

 therefrom." 



THE SEASONAL VARIATION OF PRESSURE. 



The observations for a single year, or even two years, are insufficient to yield true mean values for the 

 individual months, and that this is so will be seen at once when the monthly means of pressure for the 

 first year of the " Discovery's " stay at her Winter Quarters are compared with the means for the 

 corresponding months of the second year; the mean for June, 1902, for example, is a quarter of an inch 

 lower than the mean for June, 1903, and the means for the two months of July differ as greatly, but in 

 the opposite way, the earlier value being in this case the higher of the two. 



TABLE II. Mean Monthly Pressure at the "Discovery's" Winter Quarters for the Years 1902-4. 



* 20 days' observations only. 



We must be careful, therefore, not to attach too much weight to the monthly means as such, but by 

 grouping the months into seasons we double the number of observations available for the spring and 

 autumn means and quadruple those for the summer and winter seasons, and with this augmented number 

 of observations we may fairly expect the casual fluctuations of pressure to cancel each other sufficiently to 

 yield a pretty close approximation to the true seasonal mean. Dealt with in this way, we get the mean 

 values of pressure for the four seasons shown in the following table, in which is also given for comparison 

 the corresponding means from Cape Adare, the " Belgica," and from the " Gauss." 



TABLE III. Seasonal Means of Pressure at the "Discovery's" Winter Quarters for the Years 1902-4; 

 and at Cape Adare, the Winter Quarters of the " Gauss " ; and the " Belgica," drifting in the Ice 

 Pack. Corrected to Latitude 45. 



