Part III.] OR POISONED BY VEGETABLE GROWTH ? 
19a 
ficial wash of rain, increasing as it descended, 
would many times in the year desolate the 
whole of the lower parts of the hills and moun¬ 
tains ; and the lower parts of the hill-sides 
would have less soil than the upper parts. The 
channels of torrents and rivers prevent this 
effect now; they are nature’s ditches and gut¬ 
ters : so that, in this light, rivers may be re¬ 
garded as a conservative, not a destructive 
power. But rivers are mere labourers, or acces¬ 
sories, in the affair. The wash of rain is the en¬ 
gineer which has laid down the gradients of 
this preventive surface-drainage over the entire 
area of the earth. The source of the valley is 
always much higher up than the source of the 
river; I mean, than the spring source of the 
river: for the snow source or glacier source, 
being both superficial sources, I consider the 
same as the rain source of the valley. The river 
has no power of making a valley above it; but 
a torrent of rain water has the power of scoop¬ 
ing a valley below it. Even on Salisbury plain , 
which is comparatively flat and covered with the 
closest greensward, these dry valleys, or rather 
continuations of valleys, above the heads of 
rivers exist: and that the cause which caused 
them still works may be argued from the valleys 
o 
