16 



fifty species of fish represented in these collections, it is easy for us to 

 tell, by a simple examination of our numerical data, which of these fifty 

 species have come out in our nets in each other's company the most fre- 

 quently, and in what situation or habitat each of these most frequent 

 associates has been found most abundant. In so far as frequency of 

 habitat occurrences and frequency of associate occurrences coincide, we 

 have evidently a local association distinguished, together with its char- 

 acteristic and accustomed habitat. 



I tested the utility of these simple ideas during the summer vaca- 

 tion of 1905, by an application of them to our Illinois collections of the 

 so-called darters species of the subfamily Etheostominse in a way to 

 prove, what, indeed, we already knew as a matter of common observation, 

 that, taken as a whole, these darters are an associate group, and that 

 their characteristic habitat is what also we already knew it to be the 

 rocky rapids of small streams. The essential correctness of the method 

 was thus verified ; and I was also able to distinguish six species of darters 

 peculiarly typical of the group, to be regarded as especially characteristic 

 of it because they were found more than two and a fourth times as fre- 

 quently associated with each other as they were with the seven remain- 

 ing species; and likewise because they were about two and a half times 

 as frequently associated with each other as were the seven remaining 

 species among themselves. 



It was easy to show, on the other hand, by similar methods, that 

 the sunfishes (see Fig. 23 and 24), although as much alike to a general 

 observation as the darters, are not a homogeneous ecological group, but 

 that they are so variously related to different habitats to different fea- 

 tures of their environment that several of the sunfish species are much 

 more frequent associates of fishes widely different from themselves than 

 they are of each other; that the various sunfish species often belong, in 

 fact, to different zoological associations. Indeed, it was found by re- 

 peated use of this method of analysis that it was a rather common thing 

 for closely related species of fishes near neighbors in the taxonomic 

 system to be in some sense averse to each other's company to avoid 

 each other, seemingly, and to find their closest and most familiar asso- 

 ciates in fishes far removed from them in taxonomic relationship. 



I have supposed this to be an expression of the disadvantages of 

 close competition between closely similar species, and of the advantage, 

 consequently, of such differentiations in habit and such separations in 

 ecological preference as would carry these natural competitors into non- 



