40 E. C. ANDREWS 



resistance to the mighty ice-fiood. The differential strength of the 

 converging glaciers in their upper and lower parts will be pronounced. 

 Undermining action will set in. The upper portions will be slowly 

 worn back to steep, even slopes, but the cliff bases will experience 

 such tremendous lateral thrusts that perpendicular and even over- 

 hanging slopes will be produced at points somewhat lower than half- 

 way down the valley sides. The disposition of the converging valleys 

 will indicate which side of the lower canyon is selected for special 

 attack. In exceptional cases^ bottom thrusts will also, near these 

 points of canyon convergence, excavate rock basins possibly several 

 thousands of feet in depth below baselevel. 



Figs. 6, 9, and lo illustrate these stages in wall and bottom 

 excavation in the narrow crystallines. 



With the undermining action by the heavy lateral thrusts of the 

 lower portions of the glacier, the lower ends of many weaker tribu- 

 tary valleys will be shorn off, and their mouths will be left hanging 

 along the aligned cliff bases. 



Figs. 5, 6, II, 12, and 14 illustrate these points well. The dia- 

 grams are taken from photos of Milford and the Arthur River. No 

 one, we venture to say, could see lower Milford Sound and fail to 

 see in it a heavy ice-flood working along lines of preglacial drainage. 



As might have been expected, the northern wall of Milford has 

 been selected for special attack. This is well seen just below the 

 junction of the Arthur and Cleddau canyons, but the point is still 

 further accentuated a little farther down, where the immense glacier 

 of Harrison Cove was picked up. The great corrasion of the eastern 

 wall of Harrison Cove is also noticeable as compared with that of 

 its western one. 



Beyond these points come the evidence of minor spur truncation 

 only, the contracted sound entrances, and the shallowing of fiord 

 waters. As we might have expected, the huge rock basins and asso- 

 ciated "hanging valleys," double cliff slopes, etc., are seen in places 

 like this where the ice-fall is pronounced. Beyond this plunge to 

 baselevel we should expect comparative stagnation.^ 



1 The depths attained will be a measure of the thickness of the ice-sheet. 



2 From considerations of lack of gathering ground for ice, the cessation of cor- 

 rasive power after fall to baselevel, etc. 



