9 
predaceous fish it would need the swimming capacity and the raptatorial 
teeth of a pike-perch or a pike; to get effective access to the abundant 
stores of molluscan life, gastropod and bivalve, in our streams and 
lakes, it would need either a suctorial mouth or strong pharyngeal jaws 
with crushing teeth, or both of these, and thus it might become the 
equivalent of a sheepshead, a sunfish, or a sucker, as other conditions 
should determine. To continue as a plankton-eater, it would need the 
numerous, long, and slender gill-rakers of a paddle-fish or a lake herring, 
and with these, especially if it had a suctorial mouth and a very long in- 
testine, it might be able to sift and strain from the silt of the bottom 
the finer organic particles derived from the debris of aquatic vegetation 
and from the wash of the land. Other specialties of differentiation in 
structure or in habit might open to it less usual food resources, as with 
the darters and the top-minnows; and a mere deviation or degradation 
of taste might add the carrion of the stream to its menu. 
Evolution of food habits must thus have taken the course of struc- 
tural evolution—an advantageous specialization in various degrees and 
in various directions from a generalized, undifferentiated original. Some- 
times added specialties of advantageous equipment have brought in their 
train limitations or prohibitions in other directions, which have shut a 
species out from certain food resources in making others more available. 
The same set of gill-arches, for example, can not serve at once as a 
plankton-net and a shell-crusher; but, generally speaking, the structural 
differentiations mentioned have enlarged the resources of the fish in some 
directions without reducing them in others. Even so definitely predaceous 
a fish as the Great Lakes trout, which lives habitually on the abundant 
herring of the lakes, has been known to devour salt pork, ham bones, 
chicken bones, raw potatoes, corn cobs, rags, spoons, tin cans, silver 
dollars, and in single instances, a watch and chain, an open jack knife 
seven inches long, and a two-foot piece of tarred rope. 
From this it would appear that these structural differentiations 
have not necessarily followed upon differentiations of preference, fitting 
the fish to get more easily and abundantly the kind of food which it had 
already come to prefer; they seem to have arisen independently of any 
peculiarities of choice, and may, indeed, have forced the species, in a 
sense, into directions which it would not otherwise have been inclined 
to follow. 
