440 ME. r. W. HAEMER ON THE [^^Ug. IQOI, 



changes, moreover, are not uncommon; between Eeb. 27th and 

 March 7th 1892, for example, the temperature of the coast of 

 Labrador, in lat. 56° north, rose 61° Eahr., owing to the shifting 

 of the wind from north-west to south-east. 



The climate of the Northern Hemisphere could not have been 

 wholly cold during any part of the Pleistocene Epoch. Even when 

 the Glacial conditions reached their maximum, regions of com- 

 parative warmth must always have been the necessary complement 

 of those lying in the grasp of a perennial winter. On the other 

 hand, it need not have been everywhere warm during the Inter- 

 glacial episodes ; if mild seasons or mild periods were caused by 

 southerly winds in one region, colder conditions would have been 

 produced in others at the same time by winds from the north. 



The existence of extensive areas of high pressure, upon which air 

 from the higher regions of the atmosphere is constantly descending, 

 flowing outward from them at a lower level, involves also the 

 existence, in regions more or less contiguous, of areas of low 

 pressure of corresponding importance ^ ; reservoirs, in fact, into 

 which the air from the anticyclones may be poured, and from which 

 it may ascend, as by an aerial chimney-shaft, in order to complete 

 the vertical atmospheric circulation. If therefore high-pressure 

 centres existed over the ice-sheets, they would have been necessarily 

 accompanied by low pressure over the adjoining oceans, areas at all 

 seasons, during the Pleistocene Epoch, of comparative warmth and 

 atmospheric humidity. 



Statistical maps, such as figs. 4 & 5 (pp. 414 & 415), give an im- 

 perfect idea of the meteorological changes going on from day to day. 

 When in winter, for example, the Eurasiatic anticyclone advances 

 westward, and that of IS'orth America retreats also to the west, a low- 

 pressure system often exists in Davis Strait, as in fig. 13 (p. 442) ; 

 when the first-named anticyclone is situated farther north-west, 

 extending from Scandinavia to Greenland, the cyclone lies more to 

 the south' (figs. 12 & 14, pp. 439 & 443). 



Sometimes an anticyclone stretches from Greenland southward ; 

 at others that of the Azores sends out a tongue towards the north 

 (figs. 15 & 16, pp. 446 & 447). In that case, two cyclonic centres 

 are formed, one resting against the American, the other against the 

 European shores. The anticyclone of the Azores seems to be an 

 important factor : generally its centre lies to the east, as in fig. 5 

 (p. 415), but at times it moves towards the American coast. When 

 the Atlantic is strongly cyclonic, as in fig. ] 1 (p. 437), the anti- 

 cyclone seems to be driven southward, but it never disappears 

 entirely ; and it must always have been in existence during the 



^ These may be either large and shallow, or smaller with steep gradients, the 

 air ascending more rapidly through the latter. 



'^ The statistical chart for January (fig. 4, p. 414) shows the North Atlantic 

 cyclone lying well to the north. As a fact, however, high-pressure conditions 

 frequently exist during winter to the north of it : one or other of the conti- 

 nental anticyclones, and sometimes both of them at once, overspreading Green- 

 land (figs. 14, 15, & 16, pp. 443, 446, & 447). 



