46 E. C. ANDREWS 



must override obstacles, even as the back v^aters of a high flood must. 

 Corrasion is here reduced to a minimum, for the simple reason that 

 these areas are not situated along graded channels. 



The overriding of drumlins by ice-masses is what must be expected 

 at any stage succeeding to the high -floods which produced them. 

 The drumhns are analogous to the great masses of debris piled up 

 along the banks or along the channels of a stream during high flood. 

 With the recession of the flood the load is dropped, and the babbling 

 brooklets of subsequent dry phases must either trickle through or 

 override the debris. But as for water, so for ice, the drumlins owe 

 their existence to former floods. 



(P. 27) "Could not a very deep glacier, having great pressure 

 on its bed, along with a steep gradient, giving high velocity, rapidly 

 abrade its bed?" Fairchild says "No!" But why not? It is only 

 along these old preglacial lines, whether belonging to the newer 

 canyon stage or to the older valleys of the flexed upland remnants 

 converging into the canyons, that we claim such erosive power. We 

 do not even need steep channel grades. A sharp convergence of 

 two canyon glaciers into one very little larger in cross-section, especially 

 if the walls are strong and deep, will furnish the velocity needed to 

 give the increased efflciency of erosion needed. 



But, as before remarked, over the bulk of the area — i. e., in inter- 

 stream channel areas — we do not expect much cutting work to be done. 



(P. 28) Fairchild states that basin excavation is less probable 

 than valley-widening. We maintain that in New Zealand sounds 

 the evident aim of the glaciers is to widen and flatten their floors. 

 At certain points of convergence, however, especially when the high 

 and strong wafls are confined, the thrusts find partial expression dur- 

 ing high floods in basin excavation the while undermining operations 

 are going on. (See ante, also diagrams and photographs.) We are 

 thoroughly acquainted in New Zealand with the tendency to widen 

 valleys by ice. 



(P. 31) These points have all been answered in other parts of the 

 report on the ice-flood hypothesis. Similarly the peculiar basins of 

 the Sierras, Cascades, Alaska, Norway, and New Zealand have been 

 shown to result from ice-floods along channels of profound preglacial 

 canyons, especially just below points of canyon convergence. 



