PHENOMENA OF REGRESSION AND CONVERGENCE 221 



to Douville, there is in this no question of direct 

 descent from the Triassic Ceratitae, but a regression 

 which has brought back to the adult state the 

 Ceratite phase of certain early Ammonoids. Even 

 the genus Neolobites of the Cenomanian stage con- 

 stitutes a final term of this regression, in which the 

 saddles and the lobes remain whole and go back to 

 the Goniatite phase. In the same way, the effacing 

 of the ornamentation observed in the last whorl in 

 most of the genera of the Ammonites, or else the 

 uncoiling of the spiral, so frequent in various families 

 of Cretacean Ammonites, have often been con- 

 sidered as regressive characteristics. Modern palae- 

 ontologists are more inclined to see in this, as does 

 Hyatt, an indication of a senile or gerontic phase 

 which reproduces, nearly at least, the infantile or 

 nepionic phases by which the evolution of each 

 branch begins. We shall return to these facts when 

 we study the causes of the extinction of species. 



The natural tendency presented by certain 

 groups of fossil animals to modify in the same 

 direction and by a kind of parallel course, the 

 characteristics of the individual or of the group 

 taken collectively, enables us to understand how 

 in certain cases, appearances of connection, of resem- 

 blance or even of more or less complete confusion, 

 between species belonging to genera whose starting 

 points have been totally distinct, may be produced. 

 These curious facts, carefully studied by modern 

 palaeontologists, have received the name of phe- 

 nomena of convergence. 



Theoretically an absolutely perfect convergence 

 going so far as the confusion of generic characters 



