THE ANTARCTIC CLIMATE 521 



ceniber, January, and February as summer, we may take it that the 

 mean winter temperature is — 16.8° 0. (1.8° F.), and the mean for sum- 

 mer — 1.5° C. ('29.3° F.). 



From his observations, Mr Arctowski concludes that between the sev- 

 entieth and seventy-first parallels of the southern hemisphere and amid 

 the ice of the Antarctic ocean— first, the mean temperature is lower than 

 that of the northern coast of Spitsbergen (Mossel, bay, 1872-'73, — 8.9° 

 C. (16.0° F. ) ) ; second, the minimum temperature is quite as low as the 

 minima observed on the east side of Greenland (Sabine island and 

 Scoresby sound) ; and, third, the mean temperature of the three summer 

 months is lower than the corresponding mean in the ice of the Arctic 

 ocean. If we consider that a considerable fraction of the seventieth par- 

 allel of south latitude is land, we can suppose that it may have a mean 

 temperature as low as the 70° N., and include a pole of cold with lower 

 temperature as the Asiatic or North American poles of cold. 



During the drift in the pack-ice hourly observations were made with a 

 marine barometer and with an aneroid. While Mr Arctowski has not 

 yet been able to apply exact corrections to these observations, the uncor- 

 rected values are near enough for present purposes. The lowest pressure 

 observed during the winter was 711.74 mm. (28.022 inches), and the 

 highest 772.14 (30.400 inches). The mean value of the monthly variations 

 of the barometer, amounting to 34.30 mm. (1.350 inches), shows very 

 clearly that the cyclonic belt extends beyond the polar circle. The 

 three months of almost continuous daylight (November, December, and 

 January) are characterized by a very small variation of pressure — only 

 23.95 mm. (.943 inch). The three corresponding months of winter have 

 also a mean less than those for the intermediate or equinoctial months. 

 The differences between the annual and monthly means show that Feb- 

 ruary, March, and April form a negative group, in which pressure is 

 relatively low; the three months of polar night form another group of 

 maximum barometric pressure; then follow August, September, and 

 October — months of decreasing pressure — a group which, although not 

 exactly negative, forms adistinct secondary minimum ; and, lastly, three 

 months of polar day forming a secondary maximum of pressure. The 

 existence of a direct, simple relation between the barometric pressure 

 and the progress of the sun is at once obvious. 



The winds blew from northerly and southerly points with almost 

 equal frequency, but easterly winds predominated over westerly. The 

 sky was usually overcast, most frequently with a thick layer of stratus 

 which formed a uniform gray covering and often persisted for days or 

 even weeks together, with only short breaks. The number of days dur- 

 ing which the air did not remain saturated -i. e., on which the hygrom- 

 eter indicated a humidity of less than 90 per cent— was October, 12; 

 November, 18; December, 22; January, 15, and February, 11. If ice 

 deposits from fog and similar precipitation are included, snowfall is re- 

 corded on 257 days and rain on 14 days of the year. Speaking generally, 

 it may he said that the weather was extremely cloudy; that fogs were 

 frequent; that snow fell on many days, and that the air was saturated 

 nearly the whole time. 



