68 



Recherches sur les Poissons Fossiles. 



branched out from these singular Sauroid fishes, which par- | 

 ticipate at the same time with fishes and reptiles, and that ; 

 the mixed character of this class was lost on the appearance 1 

 of the numerous reptiles which succeeded them, as we see 

 in the Ichthyosaurians and Plesiosaurians, participating in 

 their osteology with the character of the Whales in the class 

 of Mammalia, while the great terrestrial Saurians approxi- 

 mate to the Pachydermata, which seem to have been much 

 more slowly developed. 



This observation also agrees with those ideas of the philo- 

 sophy of nature, in which a regular organic developement 

 is represented in all created beings, constantly varying with 

 the different conditions of existence presented from time to 

 time on the surface of the globe, according to the changes 

 to which that surface itself is subject.* 



From such facts as these, may be seen throughout the 

 geological series of formations, two great divisions which 

 terminate at the gres vert. The first, the more ancient, afford 

 few traces of the Ganoides and Placoids. The second, 

 an approach towards existing beings, presents forms and 

 organizations much more diversified ; particularly the 

 Ctenoid s and the Cycloids, and a small number of species in 

 the two preceding orders which insensibly disappear, and of 

 which the living analogues are considerably modified. We 

 do not find the fishes of the first great period to present 

 amongst themselves any difference corresponding with that 

 which we now observe between the fishes of fresh water 

 lakes and rivers, and the fishes of the seas ; such distinc- 

 tions of fresh water and salt water strata extend no lower 

 than the oolitic series, so that the waters of remote times, 

 were probably circumscribed by less solid basins than at pre- 



* The views of tbe author are here not perhaps sufficiently worked 

 out, and the philosophy to which M. Agassiz refers in the preceding pa- 

 ragraph, although apparently derived from Lamarck, is still far from 

 satisfactory. 



