FISHES OF LAKE SUPERIOR. 257 



herrings ; and again, the generic differences which occur among 

 the trouts, the graylings and white-fishes, and distinguish them from 

 ti-ue salmon, are far greater than that which exists between the chubs, 

 gudgeons, barbels or carps ; and the specific distinctions which may 

 be noticed in these different genera are again of an unecjual value. 

 So that we arrive at once to this important conchision, that 

 natural groups in the animal kingdom show naturally differences of 

 unequal value, and that all attempts on the part of naturalists to 

 equalize the divisions which they acknowledge in their researches, 

 must, as a matter of course, result in failure; and I have not the 

 slightest doubt that our classifications have not been more imjiroved, 

 and that we have made less extensive progress in the knowledge 

 of the true relationship between the various groups of the animal king- 

 dom, for the very reason that we have too often aimed at an arrange- 

 ment which the most familiar facts in nature plainly contradict. 

 Instead of this desired uniformity, we sometimes observe a numer- 

 ous group of closely allied species corresponding to another group 

 with few, but more distinct and more widely different species, and 

 even isolated types, the relation of which seems to branch in 

 all directions, without ever coming very close ta any other group. 

 Now, unless our classifications admit, as a natural limit, this diver- 

 sity, it will be impossible ever to form a system which will answer 

 to the natural afiinities really existing in nature. As I have said on 

 another occasion,* classification should be a picture from nature, and 

 not an artificial frame of our own invention, into which natural objects 

 ai'C more or less conveniently brought together. 



Another important point of view, of which naturalists should never 

 lose sight, is the relation which exists between animals now foimd 

 alive on various parts of the surface of our globe, and those known 

 to us only from fossil remains discovered in strata of a different geo- 

 logical age. 



The Lepidosteus, however isolated in the present creation, had 

 once many and very diversified representatives all over the globe. 

 Fossils of the same family of which the gar-pike is the type, have been 

 found all over Europe in the oldest fossiliferous beds, in the'^strata of 

 the age of the coal ; in the new red sandstone ; in the oolitic deposits, 



* See Principles of Zoologrj , by L. Agassiz and A. A. Gould, Vol. II. 



