The Transformation of the Mexican Axolotl. 587 



of the flies. They could have been quite different 

 without the form of the imagines having been 

 thereby modified, since the stages of insect meta- 

 morphosis vary independently of each other in 

 accordance with the conditions of life to which 

 they are subjected, and exert scarcely any, or only 

 a very small form-determining influence upon each 

 other, as has been amply proved in the preceding 

 essay. In any case the power of these larvae (the 

 Cecidomyiaf) to propagate themselves asexually 

 was first acquired as a secondary character, as 

 appears from the fact that there exist numerous 

 species of the same genus which do not " nurse." 

 In the form which they now possess they could 

 never have played the part of the final stage of 

 the ontogeny, nor could they formerly have pos- 

 sessed the power of sexual reproduction/ 2 In 

 brief, we are here concerned with true larval re- 

 production, whilst in Triton we have reversion to 

 an older phyletic stage. 18 



11 See the first essay " On the Seasonal Dimorphism of 

 Butterflies," p. 82. 



" [Eng. ed. It has frequently been objected to me that the 

 existing Axolotl is not a form resulting from atavism, but a case 

 of " arrested growth." The expression "atavism " is certainly to 

 be here taken in a somewhat different sense than, for example, in 

 the case of the reversion of the existing Axolotl to the Ambly- 

 stoma form. Further on, I have myself insisted that in the first 

 case the phyletic stage in which the reversion occurred is still 

 completely preserved in the ontogeny of each individual, whilst 

 the Amblystoma stage has become lost in the ontogeny of the 

 Axolotl. If, therefore, we apply the term " atavism " only to 



