ANNUAL VARIATION. 89 



The coaclitions are entirely diltereiit in the south. Practically the whole region within 

 the Antarctic Circle is snow-covered land, and completely surrounding this continent is an open 

 ocean, the temperature of which in high latitudes even in midsimimer does not rise above 

 the freezing point. On the return of the sun, the air over the Barrier warms up in propor- 

 tion to the insolation received. At mid.summer the temperature is still 10°F. below the freez- 

 ing point. Compared with the north, there is little inflow of heat from lower latitudes because 

 the prevailing wind is almost entirely from the south, also the few winds which do blow 

 from the north come from a cold ocean instead of from a warm land surface. Thus local 

 insolation is practically the only source of heat even at midsummer, hence when it decreases 

 after the solstice, the temperature decreases with it. 



The low summer temperatures in the Antarctic have been a source of difficulty to meteoro- 

 logists. Mr. Dines, in discussing the results of the Discovery Expedition, wrote * : — • 



' The low mean temperature of the summer is also strange, and very difficult to 



explain The insolation in the Antarctic in summer is greater 



than the insolation on any other part of the earth at any other time, this being 

 the date at which the sun is nearest to the earth, and if temperature depended 

 only on solar radiation we should have the highest terrestrial temperature occurring 

 in December in the neighbourhood of the South Pole.f A large mass of ice pre- 

 vents the temperature from rising above the freezing point, because the air, being 

 nearly pervious to the radiation, takes its temperature chiefly from the surface with 

 which it is in contact, and if that surface be ice or snow, it cannot be above 

 32° F. But ice, except in its power of evaporation, is in no way more efficacious 

 than any other rock in checking a rise in temperature up to its own melting point. 

 Why then does not the mean summer temperature at least reach the freezing 

 point as indeed it does in the north polar regions, where the insolation is less 

 intense ? ' 



The answer to the (piestion is now quite clear. Of the solar energy which falls within 

 the Antarctic Circle, such a large proportion is lost by direct reflection from the snow that 

 the remainder is not sufficient to raise the temperature of the aiik to the freezing point before 

 the solstice is reached, and the energy commences to decrease. The Arctic Circle on the other 

 hand is during the summer surrounded by snow-free continents, the temperature of which 

 rises well above the freezing point, and warm air is carried from them across the whole 

 polar region. Insolation, as Mr. Dines points out, is less in the north than the south, but 

 insolation plus warm winds is much greater in the north than the south ; hence the tlifferenco 

 in the respective maximum temperatures. 



During the period of decreasing insolation, the temperature over the Barrier falls slightly 

 less rapidly than it rose giving a small lag behind the insolation, so that the mean temperature 

 in April, when the insolation ends, is 4°F. higher than in August, when the sun appeared 

 after the winter. This, however, is to be expected, for tho snow sui-face and the air above 

 it have some slight heat capacity. The lag, however, is very much greater in tho north. 

 At the end of the summer, the ice in the north is thinner than at the end of the winter, 

 ther<; is some open water, and the surrounding land surfaces are still free from snow, and, 

 therefore, have a relatively high temperature. All these effects supply heat, and keep the 

 temperature high, thus when the sun sets for the last time in October, the temperature in 

 the north is 24-9°F. higher than when it rose in February. 



* National Antarctic Expedition, 1901-1904, Meteorology, Part I, page 465. 



t In tliia sentence Mr. Dines neglects the atmosphere. If the atmospliere is consijercd, the region of maximum 

 insolation is 35^ S., while tlie South I^lc receives less heat than the Equator. See Plate B 15 of Angot's work. 



12 



