INTRODUCTION 23 



not by gradual increase. The true reading of the matter, 

 therefore, is not that the ancestral condition is repeated, but 

 that the larval condition of the ancestor is retained, because 

 the larva is still hatched, and still lives in the same way ; 

 but the structure after metamorphosis is different, because 

 the fish has acquired different habits. On the theory of 

 natural selection, we must suppose that those individuals 

 have been selected whose eyes were most symmetrical in the 

 larval stage, and most asymmetrical in the adult condition. 

 But we have no evidence that among symmetrical fishes 

 individuals occasionally occur in which one eye moves up 

 towards the edge of the head during growth. Even if slight 

 variations of such a character were proved to exist it would 

 be difficult to believe that they would be great enough to 

 make any difference to the fate of the individuals possessing 

 them when the fish took to lying on the ground. The 

 theory of independent variation and selection as applied to 

 flat-fishes is unsupported by evidence, while the conclusion 

 that the metamorphosis of these fishes is the direct result of 

 the change of conditions is in harmony with all that we know 

 of the effect of physical conditions on individual organisms. 



In these two cases, that of the frog and that of the flat- 

 fish, the larval condition is either unmodified or less modified 

 from the ancestral condition than the adult. But in numer- 

 ous other cases the larva has been modified in adaptation to 

 new conditions, while the adult has remained nearly the same. 

 This is particularly conspicuous in many insects. I will not 

 discuss at length the question of the tracheal gills of aquatic 

 larvae, for although they are probably secondary adapta- 

 tions, there are some who regard them as representing 

 an ancestral series of organs, from some of which the wings 

 were derived. But if this be the case, the entire absence of 

 wings and tracheal gills in the terrestrial larvae shows that 

 the latter by no means recapitulate the ancestral history. 



