DISPERSION OF FRESH-WATER FISHES. 117 
is a considerable difference between the fatna of 
the Columbia and that of the Sacramento. The 
species which these two basins have in common 
are chiefly those which at times pass out into the 
sea. The rivers of Alaska contain but few species, 
barely a dozen in all, most of these being found 
also in Siberia and Kamtschatka. In the scanti- 
ness of its faunal list, the Yukon agrees with the 
Mackenzie River, and with Arctic rivers generally. 
There can be no doubt that the general ten- 
dency is for each species to extend its range more 
and more widely until all localities suitable for its 
growth are included. The various agencies of 
dispersal which have existed in the past are still 
in operation. There is apparently no limit to 
their action. It is probable that new “colonies” 
of one species or another may be planted each 
year in waters not heretofore inhabited by such 
species. But such colonies become permanent 
only where the conditions are so favorable that 
the species can hold its own in the struggle for 
food and subsistence. That various modifications 
in the habitat of certain species have been caused 
by human agencies is of course too well known to 
need discussion here. 
We may next consider the question of water- 
sheds, or barriers which separate one river basin 
from another. 
Of such barriers in the United States, the most 
important and most effective is unquestionably 
that of the main chain of the Rocky Mountains. 
This is due in part to its great height, still more 
to its great breadth, and most of all, perhaps, to 
