PALEOLITHIC RACES 



337 



beneath, and has actually deepened its valley. Thus the terrace 

 was cut out during a period of erosion which followed upon a 

 period of deposition (b, fig. 2). 



The second terrace involves a similar succession of events ; 

 it points to a return to the earlier conditions, when the river, 

 powerless to erode, spread out a second sheet of shotter over the 

 newly excavated valley floor (c, fig. 2) ; then came renewed activity, 

 and the second terrace was carved out. The same is true of 

 the third and fourth terraces, and thus we have repeated, time 

 after time, an alternation of periods of deposition and periods 

 of erosion, of overburdening and acceleration. Such are the 

 immediate inferences from the facts. 



We must now take a step further, and attempt to account 

 for this alternation of processes. 



The interdigitation of the terrace with its moraine shows 

 that ihe terrace, or rather the sheet of shotter from which it was 



Fig. 2. 



carved out, was deposited during an interval when the glacier 

 was comparatively stationary, i.e. during the interval in which 

 it built up its terminal moraine. But when a glacier is stationary 

 the amount of water discharged from it is comparatively small, 

 the annual discharge is indeed precisely equal to the annual 

 snowfall by which the glacier is replenished : when the glacier 

 is advancing the discharge is even less. Under these circum- 

 stances the resulting river was scarcely larger than the cor- 

 responding river which now represents it, and its power to 

 erode was at a minimum. 



If now we are to endow this river with greater volume and 

 velocity we must assume that the glacier commenced a retreat, 

 or in other words that more ice was melted away from it than 

 was made good by the annual snowfall ; and this retreat must 

 have continued for no inconsiderable period — it must have lasted 

 at least as long as was necessary for the sweeping away of the 

 previously deposited shotter and the deepening of the valley. 



Thus, if this reasoning be valid, we are led to greatly enlarge 



22 



