72 Nie lS, OWLIN AZUL (OMF GIR OLOGY, 
therefore no proof that the deposit containing them was not 
made by a stream. 
The transportation of coarse materials by streams may be 
facilitated by the co-operation of ice. When the rigors of win- 
ter are past, and the snows upon the hills and mountains are 
being rapidly melted, the swollen streams break up the ice which 
has imprisoned them, and hurry the broken masses forward with 
their increasing floods. As these masses of ice are torn away 
from the banks of the streams, they sometimes carry with them 
very considerable quantities of earth, sand, gravel and even 
bowlders, to which the ice had frozen. These materials are 
deposited at some point further down the stream, or in the lake 
or sea to which the river flows. 
But however we may conceive the work of rivers to be 
abetted, the material which they transport could never be 
deposited on land surfaces outside the limits of their valleys. 
It might be distributed widely by the waves and currents of the 
seas and lakes into which the rivers carried it, but it could never 
be deposited on the tops of hills or mountains. Furthermore, 
the deposits made by streams, except in the case of alluvial 
cones, have surfaces which are nearly plane. They are never 
marked by such hills and depressions as often characterize the 
drift. The difference is not one of degree, but one of kind. How- 
ever vigorously we may suppose streams to have acted, there- 
fore, and however we may suppose their proper work to have 
been modified by co-operating agencies, they fail to account for 
the distribution of the drift or for its topography; and not only 
do they fail to account for the distribution and topography of 
the drift, but its distribution and topography are such as to con- 
stitute positive and conclusive evidence against the reference of 
the larger part of the drift to rivers acting under any circum- 
stances whatever. Since the distribution of the drift disproves 
all hypotheses which ascribe it to rivers, the argument need not 
be pursued further. But if we were to examine in detail each 
of the characteristics of the drift with reference to its produc- 
tion by streams, we should find that their combined testimony is 
