THE NEW ZEALAND SOUND BASINS 41 



But even in periods of maximum ice-flood many broads or diver- 

 gences in the channels (canyons) would form spots where aggradation 

 or partial glacial stagnation would occur. The deep fiord basins, 

 with their associated lower precipices, steep, even upper slopes, 

 aligned chff bases, and "hanging" valleys, mark points of maximum 

 ice-scour, and occur in exactly the positions where we would expect 

 to find them; but there are other spots at which we would expect to 

 have evidence of much less efficiency of ice-corrasion. This is even 

 so. At these points spurs are not observed shriveled up to their 

 containing walls, nor are rock basins found. These points correspond 

 with the smoother waters of a stream-flood. In the canyons them- 

 selves these evidences of weakened erosive power may occur only 

 at some considerable height above the channel bases. 



Again, as the ice-flood in places welled over the canyon sides after 

 the manner of a river overflowing its banks, so we could expect very 

 little comparative scouring to occur at these points. There will be 

 no natural drainage lines to follow at this level, and the ice-overflow 

 will here expend most of its energy in opposing gravity as it rises 

 above the upland irregularities. Aggradation with pronounced 

 stagnation results from this. Old drainage lines may, of course, be 

 found, and here, especially on declines,^ ice-corrasion may be con- 

 siderable; or the ice-cap may open onto the steep sea-front and set 

 up incipient valley formation here. "Hanging" or even graded 

 valleys may result thus, if the ice-flood be of considerable duration 

 in time. 



Consider now the effect of diminishing the ice-supply. Decrease 

 of volume is marked by loss of speed, and this by a wonderfully 

 decreased capacity for work. Long before ice-drought conditions 

 shall have set in active, corrasion will be suspended, and aggradation 

 will be the order- of the day. For during the period of flood the 

 channel bases have been graded so as to be adjusted^ to the heavy 

 burdens passing over them. Now an ordinary stream^ is wholly 



1 The writer would insist on the idea being kept in mind that a glacier should 

 corrade more on a decline than on an incline. 



2 A stream-flood may occupy several days only, while the glacial period (summa. 

 tion of ice-floods) probably lasted very many thousands of years. 



3 We should say this law applied to viscous substances generally. 



