Woodworlh.l 21 S 
the draiiia.^e, and that tliey were mostli/ ohstru'icil diul rinsed lnj the trans- 
pnrtatidii and deposition of modified drift. The waninfj ice-fiekls were 
then deeply incised by brooks and rivers pouring over them in the descent 
to their border and to the adjacent land lately uncovered by the glacial 
retreat. Hydrographic basins of the ice-sheet probably extended 50 to 
200 miles or more from its margin, resembling those of a belt of country 
along a sea c(>ast; but the glacial rivers, and their large and small 
branches, had much steeper gradients than those of the present river 
systems on the land surface, and often or generally they flowed in deep 
ice-walled channels, more like canons than oixlinary river valleys. Much 
englacial drift, Avhich had become supei'glacial, was washed away by 
rains, rills, and small and large streams from the ice surface; and the 
osar gravel ridges are the coarsest sediments progressively deposited 
near the ice-front in such channels which were cut backward into the 
retreating edge of the ice by the superglacial streams. 
Mr, Upham^ states that "precisely the same explanation of tlie 
mode of formation of the osars was reached independently fifteen 
years ago by Dr. N. O. Holsf^ in Sweden and by the present 
writer [Mr. Upham] in New Hampshire. Four years earlier 
, . . nearly the same view had been first published by Prof. 
N. H. Winchell in Minnesota." 
Professor Shaler in 1884 argued for the deposition of eskers 
by subglacial streams, but he did not then suppose the deposits 
were made inside the ice. Later the subglacial origin of eskers 
is stated thus by Professor Shaler : — 
I am disposed to hold .... that the serpent kames have been formed 
in the following manner. The outflowing glacial streams excavated 
channels within the ice Avhich they kept free as long as the currents were 
strong enough to scour their channels. In the closing stages of the ice- 
sheet, while the front was no longer advancing or perhaps inclined to 
retreat, these arches were filled in by material borne by the diminished 
currents.^ 
I have italicized in Mr. Uphani's statement the portion which 
parallels with that made by Professor Shaler regarding the pro- 
cesses concerned in the deposition of drift at the close of the last 
ice-epoch. Mr. Upham describes the filling of subglacial stream 
passages with water-laid drift, but apparently ignores the deposits 
resulting from this process. -The theory advanced by Professor 
1 Amer. geol., v. 8, p. -380, 1891. 
2 Forhaiidlingar Stockholm geol. forening, bd. 3, p. 97-112, 1875. 
3 Geology of Cape Ann, Mass. Ninth Annual report U. S. geol. survey, p. 550. 
