44 
food, consisting mainly of insects, crustaceans, alge, and fragments of 
aquatic plants, is as different from that of the preceding species as is its 
distribution. The little heterodon, on the other hand—about two inches 
long—is essentially a lake and pond species, and its abundance, not only 
in the lake region of northeastern Illinois but also along the larger rivers, 
is explained by the fact that it is mainly in the lowlands of the river bot- 
toms that lakes and ponds are to be found in Illinois. It is in such sit- 
uations that it finds the bottoms of mud and sand which seem to attract 
it, its frequency ratio there being 71 as compared with 22 over rock and 
sand, and 7 over mud. The food of this species is consistent with this 
preference of location, being mainly Entomostraca and small larve of 
gnats. It is, indeed, essentially a plankton-eater, being of the size to 
make the plankton of its favorite resorts its most convenient and abundant 
food. 
I know well that these specific details are hardly fit for a general 
jiecture, but they are the materials of my generalizations, and I must ask 
you to indulge me to the extent of two more examples, chosen from an- 
other family of fishes—that most interesting division of the perches 
commonly known as the darters. These are the johnny darter (Fig. 
19; Map XC), and another species, Cottogaster shumardi (Fig. 20; Map 
LXXXVIII), which has no English name. They are particularly inter- 
esting, because the johnny darter, although very abundant all over the 
state, seems to avoid the larger streams, having a frequency there of 
only 3 as against 53 for creeks, while the Cottogaster, although compara- 
tively rare, occurs almost wholly in the larger rivers and in the bottom- 
land lakes in their immediate neighborhood. Whether there are differ- 
ences in food corresponding to their distribution we cannot tell, be- 
cause the food of the rarer species has not been studied. 
Many other instances might be given of the fact that fishes can be 
separated into groups according to their habitats as well as by differences 
in their food, but that the groups so formed are of very unequal scope. 
It is as if, in classifying fishes structurally, we should find that there 
were some families which combined the characteristics of nearly all the 
others; that other kinds present many such common characters, but a 
smaller number; and that only a few had differentiated so far from the 
common mass as to have fixed distinguishing characters of their own. 
An ecological classification, while quite possible, and indispensable also, 
ought not to be framed in imitation of the classifications of the taxonomist, 
but must have objects and methods of its own. 
