imensions do not often occur within the
drift, but are usually resting above it with their sharp angles and
rough surfaces unchanged, having travelled evidently upon the glacier
and not under it. But such large boulders, polished and scratched like
the smaller pebbles, are to be found everywhere imbedded in American
drift, while the angular fragments of rock resting above these
triturated masses are comparatively rare.[H] It is evident from this
that the ice overtopped the rocky inequalities of the land, and that the
detached fragments remaining beneath the icy covering underwent the same
action from friction and pressure to which the whole mass of drift was
subjected. The distribution of the few angular boulders scattered over
the country no doubt began when some of the higher portions of the land
had emerged from the mass of snow and ice; and they are most frequent in
New England, where the mountain-elevation is greatest.

The mineralogical character of the loose materials forming the American
drift leaves no doubt that the whole movement, with the exception of a
few local modifications easily accounted for by the lay of the land, was
from north to south, all the fragments not belonging to the localities
where they occur being readily traced to rocks _in situ_ to the north of
their present resting-places. The farther one journeys from their
origin, the more extraordinary does the presence of these boulders
become. It strikes one strangely to find even in New England fragments
of rock from the shores of Lake Superior; but it is still more
impressive to meet with masses of northern rock on the prairies of
Illinois or Iowa. One may follow these boulders to the fortieth degree
of latitude, beyond which they become more and more rare, while the
finer drift alone extends farther south.

It is not only, however, by tracking the boulders back to their origin
in the North that we ascertain the starting-point of the whole mass; we
have another kind of evidence to this effect, already alluded to in the
description of the _roches moutonnees_. Wherever the natural surface of
any hill, having a steep southern slope, is exposed, the marks are
always found to be very distinct on the northern side and entirely
wanting on the southern one, showing, that, as in the case of many of
the _roches moutonnees_ in Switzerland, the mass moved up the northern
slope, forcing its way against it, grinding and furrowing the northern
face of the hil