206 Glacial Phenomena in the Yellowstone Park. [March, 
The great ranges to the east and south of the lake are not known to 
contain a single exposure of crystalline rock. That there are no 
such formations in the whole drainage of the Upper Yellowstone is 
established by the fact of the almost total absence of granite 
pebbles on the shores of the lake or in the bed of the river. The 
home of this wanderer must be sought elsewhere. To the north, 
beyond the valley of the third canon and the East fork, lies the 
granite highland previously mentioned. To the north-west, be- 
yond the valley of Gardiner river, at the southern end of the 
Gallatin mountains, is another exposure of granite at an elevation 
sufficient to have given origin to it. The distance in either case 
is upwards of twenty miles. From the great falls the river de- 
scends in a northerly direction until it strikes the base of the 
granite highland ; here it unites with the East fork and turns to 
the west along the south base of that highland, following the line 
of the great displacement until it passes the granite gateway of 
the second canon (see accompanying map). To reach its present 
position from the northern locality, the boulder must cross the 
course of the great valley of the East fork and the third canon 
and ascend the river, as it now exists, a distance of twenty miles, 
avoiding on its way, by a circuitous route, the intervening Wash- 
burn range and the opposing mass of Amethyst mountain. If 
from the Gallatin mountains, it must first have crossed the valley 
of the upper Gardiner river and afterwards a considerable spur of 
the Washburn mountains — a journey of twenty miles south-east. 
Notwithstanding the fact that this pathway would, with anything 
like the present topography, seem to present fewer obstacles to 
the advance of a glacier than that from the north, I cannot regard 
it as at all probable that this was its course. The mass of the 
Gallatin mountains is not great. Glaciers originating in its short 
abrupt valleys would have no great longitudinal extent, and would 
probably advance no farther than the deepest part of the valley 
that lies along their immediate base. 
The great ranges to the north are of sufficient extent to give 
birth to ice rivers of the grandest proportions. The present dis- 
tribution of the erratic fragments of granite tends to strengthen 
the impression that they had their origin in the north. If this 
be admitted, it becomes at once' clear that the erosion of the 
grand canon has been accomplished since the close of the glacial 
period, or at least that a second erosion has taken place if a canon 
did exist prior to the glacial epoch. 
