44 SURFACE GEOLOGY, 
The most important feature of this kame, if we compare it with others 
in New Hampshire, is, that along its entire extent it constitutes a single 
continuous ridge, which runs by a very direct course nearly in the middle 
of the valley, having no outlying spurs, branches, parallel ridges, or scat- 
tered hillocks of the same material associated with it. The kames in the 
Merrimack valley and in eastern New Hampshire also average much 
coarser, and more frequently contain angular boulders, while in some 
places they show a gradual transition from sand and water-worn gravel 
to unmodified moraines. 
This remarkable ridge shows the course of the glacial river by which 
the floods from the melting ice, laden with gravel, sand, and clay, found 
their way between ice-walls to the open valley below. All the material 
which was thus brought down was probably gathered from the melting 
surface of the ice-sheet; and the pebbles were rounded in being carried 
along by its streams. Near the mouth of the channel in which these 
waters flowed, a portion of their gravel and sand was deposited with the 
alternation of summer and winter. Elsewhere, kames may have been 
formed by rivers beneath the ice-sheet; and when many boulders are 
contained in them, or found on their surface, they seem to be most read- 
ily explained by supposing them dropped from a melting roof of ice. It 
is at least plain, that if any kames have been formed under the ice, they 
must contain many boulders derived from this source. In nearly all the 
kames of New Hampshire it seems more probable that the angular mate- 
rials and large boulders, which we find associated with these water-worn 
deposits, were brought by the same currents, frequently in floating masses 
of ice. Their infrequency here puts it beyond doubt that the kame of 
Connecticut valley was formed in an open ice-channel. It is probable 
that this did not extend at one time over the whole distance where we 
find the kame, but that it was gradually formed as the melting advanced 
northward, which was at so slow a pace that for a long time walls of ice 
enclosed the deposits of the glacial river. After these walls melted, the 
gravel and sand remained in a long, high ridge, which became nearly 
covered by the subsequent slow deposition of the high alluvial plain. 
When the river entered upon the work of excavating its present chan- 
nel in the alluvium, the kame was a barrier which confined erosion to the 
area on one of its sides and protected its opposite side; so that this ridge 
