290 GEOGRAPHICAL AND GEOLOGICAL DISTRIBUTION. 



of Italy present us with other instances of the permanent establish- 

 ment of marine types in fresh waters. Indeed, it is scarcely neces- 

 sary to seek for examples of this kind, seeing how very readily the 

 many marine fishes which periodically ascend the inland streams 

 during the spawning season accommodate themselves to the newly 

 imposed conditions. Of so little importance does a change of 

 medium appear to be in many cases that it is frequently very diffi- 

 cult, or impossible, to indicate whether a given group of fishes is 

 more properly of a marine or fresh-water type. The numerous in- 

 stances where certain species of a genus are of one habit, and other 

 species of the same genus of the opposite habit, render the deter- 

 mination of this question still more difficult. 



An accommodation similar to that which has been noticed in 

 the case of marine fishes also obtains with many of the more strictly 

 fresh- water forms ; i. e. , they descend without inconvenience into 

 the briny oceanic medium. This we see in the case of the trout, 

 the charr, in several species of Coregonus, and especially among 

 the toothed carps (cyprinodonts) and sticklebacks. How far these 

 may wander out to sea is not exactly known, but there is no special 

 reason for supposing that they might not proceed, at least in some 

 instances, to very considerable distances. A species of the cyprino- 

 dont genus Fundulus (F. nigrofasciatus) was obtained by the offi- 

 cers of the "Challenger" from the pelagic fauna of the Atlantic, 

 midway between St. Thomas and Teneriffe. It is manifest, there- 

 fore, that an arm of the sea is not an impassable barrier to certain 

 forms of fresh- water fishes ; and not impossibly some brackish- water 

 forms may occasionally find their way completely across the oceanic 

 expanse. The irregular distribution of certain types or species 

 thus receives a partial, or, at any rate, a possible solution. 



The total number of species of strictly fresh-water fishes recog- 

 nised by Gunther is nearly two thousand three hundred, of which 

 four are lung-fishes, thirty-two ganoids, twelve lampreys, and the 

 remainder teleosts or bony-fishes. Of the last nearly one-third are 

 comprised in the family of the carps (Cyprinidae) and somewhat 

 more than one-fourth in that of the cat-fishes (Siluridae). The 

 Characinids9 and Chromides, forms from tropical America and 

 Africa, are represented by somewhat more than two hundred and 

 fifty and one hundred species respectively, the salmonoids by one 

 hundred and thirty-five species, and the toothed carps (cyprino- 



