CAUSES OF THE GLACIAL TERIOD 



575 



would conform exactly with Peiick's conclusion. The temperature would 

 fall because of the draining away of the warmth of the lower atmosphere. 

 This in itself would tend to decrease precipitation somewhat. A slight 

 increase in storminess would be needed in order to balance this. Thus if 

 the Alps had somewhat more storminess than now, but were surrounded 



Figure 20, — The Distribution of Loess. Its Relation to Quateniari/ Glaciation and to 

 oresent Deserts and subarid Regions. (After De Martonne) 



(Keproducea from Fig. 310, p. 664, in De Martonne"s Traite de Geograpliie Pliysique) 



by an area .of decrease, we should get the conditions which seem actually 

 to have existed. 



, Th0 distribution of , loess. — A more striking case than that of the Alps 

 is found in the distribution of loess. This is illustrated in figure 20, 

 which is taken from De Martonne's Traite de Geographic Physique. As 

 the map plainly shows, there are two chief types of loess. One consists 

 of deposits on the leeward side of modern deserts. In places such as 



