434 ME. F. W. HAEMBR ON THE [Aug. I9OI 



freezing-point, a comparatively small amount of precipitation, none 

 of which is for the time being lost to the land, may cover the latter 

 with a thick mantle of snow. It is not necessary that snow should 

 fall constantly, or perhaps even very frequently, in order that per- 

 manent ice-fields should establish themselves. The prevalent winds 

 blow, for the most part, outward from Greenland, and the yearly 

 precipitation in the centre of that country is probably but small. As a 

 matter of fact, however, sufficient snow does fall there to have produced 

 a great thickness of ice, and the latter can only have been derived 

 from moist oceanic winds occasionally caused by some such an 

 alignment of the isobars as that shown, for example, in fig. 13 (p. 442). 

 "We need not suppose, therefore, that westerly winds were prevalent 

 in Scandinavia during the maximum glaciation of Europe, or 

 easterly winds in Labrador during that of North America ; but such 

 winds may have — indeed must have — occurred there from time to 

 time. The northern part of the Atlantic area would probably have 

 been at all seasons, during the Pleistocene Epoch, one of atmospheric 

 disturbance, as it is at present in winter : the storms being then, for 

 reasons before given, more violent, and the rainfall more copious. 

 It will not therefore be difficult to understand that, although the 

 prevalent direction of the winds must always have been more or less 

 outward from the ice-sheets, a considerable amount of moisture 

 may have reached these from the ocean. If snow falls faster than 

 it melts or evaporates, it may accumulate, time being given, to 

 almost any extent.^ The absence of warm winds in summer may 

 be perhaps more necessary for the growth of an ice-sheet than 

 great precipitation.^ 



It will perhaps be pointed out that in Dr. Buchan's maps (figs. 17 

 & 19, pp. 450 & 454) the alignment of the statistical areas of high 

 and low pressure, and of the isobars, in the Northern Hemisphere, 

 is shown to be generally from south-west to north-east, whereas in 

 my hypothetical restoration of the meteorological conditious of the 

 Glacial Period (figs. 18 & 20, pp. 452 & 456, and 21 & 22, pp. 458 

 & 460) some of them are represented as pointing from south-east to 

 north-west. Seeing that currents, either oceanic or atmospheric, 

 passing in the Northern Hemisphere from south to north must tend 

 to be deflected eastward, and those flowing from north to south 

 westward, it may be asked whether a prevalent south-east to north- 

 west arrangement of the isobaric lines in former ages is physically 

 probable, or even possible. A reference to the Daily Weather Charts 

 will show, however, that such conditions occur even now not infre- 

 quently. When the centre of the Icelandic cyclone lies to the east of 



^ My old friend and master, Searles V, Wood, Jr., in his paper on ' The 

 Cause of the Glacial Period' (Geol. Mag. 1883, p. 296), stated his reasons for 

 beheving that the stupendous mass of land-ice under which the glaciated 

 area of North-eastern America was buried may have accumulated under a 

 precipitation less than that which now takes place in the same region. 



^ There must have been more precipitation, however, over the southern parts 

 of the ice-sheets than there is now in Grreenlancl, as the sun's heat in summer, 

 and its melting-power, M'Ould have been greater in the former case than in 

 the latter. 



