CROSBY: ORIGIN OF ESKERS. 
409 
or wholly ceases. We have already noted the absence of tributary 
streams, and if such existed, their beds also, at least in their lower 
courses, would necessarily be aggraded with the bed of the main 
river. On either side of the tunnels the ice rests heavily upon the 
ground, and the marginal lakes of valley glaciers as well as crevasses 
filled with water show that the basal contact must usually be so tight 
as to permit a movement of water only by seepage, which could, at 
the best, effect the removal of only the finest or clayey part of the 
drift. The absence of an adequate available supply of material is, 
to my mind, the most serious of all the objections to the subglacial 
theory of eskers. Anyone who considers the great extent and depth 
of many of our delta plains, and the vast volumes of material required 
to form them and their tributary eskers, will not doubt that, if 
derived from the subglacial drift, the latter should show extensive 
erosion over the areas to the northward. But we look in vain for 
evidence of such erosion, although its record should be very distinct 
in the cases where many millions of cubic yards of sand, gravel, and 
bowlders have been delivered through a single narrow channel, not 
to mention the still greater volumes of clay and quartz flour which 
we know must once have been incorporated with the coarser detritus. 
If the subglacial hypothesis be true, the region to the north of some 
of our extensive sand plains ought to have been, in large part, swept 
bare of till or ground moraine; but we do not find it so, and it 
could not be so, as long as the ice sheet rested upon and protected it. 
Davis concluded that the Newtonville and Auburndale eskers 
must be products of subglacial streams acting on subglacial drift, 
because they contain fragments of slate and conglomerate which 
must have been derived from the ledges within two to four miles to 
the northward, and he thinks it improbable that englacial drift could, 
in so short distance, rise to the level of superglacial streams. The 
proportion of material from nearby sources is, however, very small, ■ 
certainly not more than ten and possibly not more than five per cent, 
and the elevation of this small fraction of the drift a hundred feet or 
so in, say, three miles does not impress me as offering any special 
difficulty. Still farther within the Boston Basin, as in the vicinity of 
Newton Upper Falls and West Roxbury, at distances of from four 
to eight miles from the granitic northern border of the basin, the 
proportion of drift of local origin in the eskers and sand plains is 
still very small, the granitic rocks from the northern highlands 
