4 PREFACE. III. 



Those ichthyologists who may deny the validity of all the species of the genus 

 Cottus which I adopt, considering some of them as mere varieties, must still acknow- 

 ledge that these varieties are worthy of being known. 



There are two species of Cottus proper, the originals of which I have not seen, 

 viz.: C. cognatua, Rich., and my C. Fabricii (C. gobio, Fabr.). No specimens of 

 either of these species are preserved in any American collection. 



An intimate knowledge of the Cottoids will enable us more readily to appreciate 

 the specific characters of a great many other groups, whose geographical distribution 

 is less extensive, but in which the species are altogether as numerous and as 

 uniform. The species of a genus are generally uniform whenever they are nume- 

 rous : this is a necessary consequence. When the type possesses but a few cha- 

 racters, it is not liable to so many diversities amongst the species, and these cha- 

 racters, in being shared, appear as if exhausted, leading superficial observers to the 

 belief that species pass into each other by gradual steps. 



Not only among fishes, but in other classes also we meet with natural groups 

 composed of numerous species, which are allowed to vary beyond the limits of their 

 circle of activity, and for the study of which great skill must be displayed in order 

 to recognize the species in the midst of these variations. 



