10 MONOGRAPH OF THE FRESH WATER III 



marine or fluviatile, shows a very intimate relationship, both amongst themselves 

 and the neighboring groups of Scorpamoids and Scisenoids. 



The typical genus of the marine division is that of Acanthocottus, numerous in 

 species, and distributed over the arctic and temperate regions of both hemispheres. 

 In the Pacific Ocean we have the genera Trachidermis and P<><l<iln-ux, composed 

 each of but a few species confined to the seas of China and Japan. The first is a 

 diminutive of Acanthocottus, from which it differs in having a prickly skin and 

 teeth on the palatine bones. The second is intermediate between Trachidermis 

 and Acanthocottus, without forming, however, any direct passage from one to the 

 other; it has the smooth skin of Acanthocottus and the palatine teeth of Trachi- 

 dermis. As to the general form and physiognomy, one would consider Podabrus as 

 the eccentric type of Acanthocottus : the two species which it embraces, differing 

 greatly from each other in that respect. 



The genera Hi-utifi-i/>f'rus and //tWA/f/Wn///* exhibit a tendency towards Scor- 

 paeuoids ; both are provided with teeth on the palatine bones, the skin of the former 

 being somewhat prickly, and that of the second, partly covered with scales. 



The typical genus of the fresh water division is that of Cottus, numerous in species, 

 distributed over the whole range of the temperate region of both hemispheres. On 

 the north-western shores of America there is the genus Cottopsis, the amplification of 

 Cottus, provided with a prickly skin and teeth on the palatines, like Trachidermis. 



Finally, the genus Triglopsis seems to recapitulate both Acanthocottus and 

 Cottus, and to foreshadow the group of Sciacnoids. 



Thus eight genera, five marine, and three from the fresh water, constitute 

 actually a natural group to which the name of COTTOIDS is well appropriated. 



2. GENEALOGY OF TIIE COTTOIDS. 



The history of the Cottoids prior to our epoch rests, as yet, only upon the dis- 

 covery of a few fossils in Europe. The family appeared for the first time on the 

 surface of the globe, sometime during the last period of the cretaceous epoch, where 

 it was represented by two genera, Jl< rt/>/(x-f/Jt>i/iix and C'affipteryjr, precursors of the 

 group of Tri'jli'l"; their affinities leave no doubts on this point. But these two 

 -enera die out with llie end of their epoch. 



To them succeeds the irenus Cottus in the tertiary epoch, where it was the only 

 one of the family, and, indeed, had but few known species. Two of these, Cottu* 

 /-/'//sand C. 2>/>.>/''<"' "-. Agass., come nearest to ('. </<>}>!<> and other fresh water 



cies of the -re mis now living, and a third species, Ci>/fii* <///<*, Afjass., is more 

 intimatflv related to Culliix .-<,//,/<'/'</<>, and /;/(/*//*, precursor of the mannc type of 

 the L'eiius. which w now call AatuUkoootttlt, 



It is not improbable that Cottoids may have existed in North America before 

 the pre-ent epoch. The study of the cretaceous and tertiary deposits situated 

 within the limits of the zone inhabited by living species will furnish us information 

 on this point. 



The jreoloiMi-al on t he A.-iat ir Cont im nt . ha\ e not yet furnished us 



