274 R. S. TARR RECENT ADVANCE OF GLACIERS IN ALASKA 



area broadens seaward, starting with a width of 3 or 4 miles at the 

 mouth of the mountain valley and expanding to a width of over 10 miles 

 near the sea. 



In the mountain valley the crevassing has produced a jagged, pin- 

 nacled ice surface (plate 20, figure 1), but out in the expanded portion 

 of the Malaspina piedmont area the crevasses form great rents, with many 

 table-top areas between — remnants of the former level-topped ice-plateau 

 Owing to the elevation, the snow had not completely disappeared from 

 this section in July ; so that, excepting along the glacier margin, melting 

 had not yet greatly modified the area between the crevasses. This fact 

 clearly points to the conclusion that the crevassing is a result of move- 

 ment in the season of 1906 ; for if the broken surface had been exposed 

 longer to the air, melting would necessarily have produced much greater 

 effect in rounding off the table-topped areas between the crevasses. 



That the advance of Marvine glacier was in progress during the sum- 

 mer of 1906 was abundantly proved. The margin of the glacier was 

 being pushed forward into the form of a jagged ice-cliff resembling that 

 of Atrevida glacier (plate 31). During this forward push the ice had 

 been broken into great blocks, and the morainic soil, which varied in 

 depth from 2 to 15 feet, had been greatly disturbed. The broken ice- 

 blocks, stained by a veneer of debris washed down over them, reminded 

 one of some of the frost-riven granite cliffs of New England, and at a 

 distance looked far more like rock than ice. Ice-blocks were seen to 

 tumble from the glacier margin as we passed along it, and the morainic 

 soil was constantly falling from the cliff and being incorporated in the 

 new talus slopes and alluvial fans. 



By exposing to the air so much ice that had previously been deeply 

 blanketed beneath moraine and forest, melting was greatly increased, in- 

 numerable new streams were developed (plate 23), and the volume of 

 Kwik river greatly augmented. That this process was then actually in 

 progress was clearly proved by the changes .which occurred during the 

 month that elapsed between our journey to Blossom island and our 

 return. In this interval some of the streams had grown noticeably in 

 volume and headed farther back in the crevassed ice, in some cases having 

 developed ice-tunnels; the talus slopes and alluvial fans had grown in 

 area, and in places the ice-margin had become distinctly altered in form. 



By the forward thrust and breaking of the ice, the forest growth on it 

 was being destroyed (plate 22). Many of the trees were inclined at 

 various angles ; others had fallen down the ice-front or into the crevasses ; 

 and large numbers were seen or heard to fall during our passage along 



