1892.] 
495 
[Davis. 
plain on their further free course towards the sea, abundantly 
warrants our belief in the occurrence of similar streams during 
the decay of the New England ice-sheet. Bursting forth now 
here, now there, along the ice margin, a subglacial stream would 
bring a large supply of crystalline rocks gathered from the high- 
lands and a smaller supply of bedded rocks from the lowlands and 
spread them all out in front of the ice. If the discharge were on 
an open country, the decrease of velocity of the stream below 
the point of escape would require the deposition of a large part of 
its load, and thus it would become a building stream, like those 
that characterize the ice front of Greenland and Alaska. If it 
emerged from the ice into a body of standing water, the detritus 
would at first accumulate close to the point of escape, growing 
higher and higher till the water surface was reached, and thus 
forcing the stream to mount over the obstruction that it had 
formed ; as the delta then increased forward, the channel of the 
subglacial stream would be worn and melted to a larger size, 
and with increase of size and continually varying current 
and load the channel would become more or less clogged. The 
channel would become enlarged chiefly where the stronger 
current flowed, and thus a cumulative process would determine 
the formation of a tunnel or subglacial tube for the escape of the 
rushing stream. Most of the detritus brought by the stream 
would be carried forward, but the coarser parts would be left to 
fill the growing tunnel. Here we find reason for the natural 
selection of finer material for the frontal delta, and of coarser 
gravel and stone for the esker ridge, the two commingling at the 
head of the delta where both forms are joined. Bearing in mind 
that a large area of back country may now and then have 
delivered its drainage by a subglacial course, or an inverted 
siphon, it need not excite surprise that stones of considerable 
size up to a foot or more in diameter should be carried across 
from the lowlands and up to the delta level, there to be spurted 
forth with the lighter gravel and sands. The smaller of the 
stones thus delivered on the delta top might be dragged over the 
surface beds towards the front, but as a rule they would remain 
near the head where the coarser materials are now found. This 
is well shown in the cut on the circuit railroad. 
