190 ME. E. B. POULTON ON THE EXTEENAL 



or spines possessed by those that pupate in tubular galleries cut in plant-stems or formed 

 by rolled-up leaves. 



Conclusions as to the Nature of Lepidopterous Metamorphosis.— -Many writers have 

 pointed out that the form of metamorphosis which consists of three sharply separated 

 stages has been, in all probability, derived from a form in which many closely similar 

 stages Gradually led up to the final sexually mature form. The present metamorphosis 

 of Lepidoptera &c. has been derived from the more ancestral form, still witnessed in the 

 Orthoptera, by the omission of intervening stages, and also by the subsequent special- 

 ization of the final stage. In estimating the position of the lost stages it is most impor- 

 tant to gauge the morphological relation of the pupa to larva and imago. Directly we 

 attempt this comparison we find that, whatever morphological feature we adopt as a 

 criterion, the position of the pupa is immensely nearer to the imago than to the larva. 

 The o-reat morphological break is between larva and pupa, an interval so wide as to dwarf 

 the minor differences between pupa and imago. 



At the change of skin which separates the two former stages we suddenly pass from a 

 stage with simple eyes, without wings or external generative organs, into a stage with 

 compound eyes, wings, and well-marked external generative organs. We may therefore 

 safely conclude that many stages have been lost between larva aud pupa. 



At the time when these stages intervened the stage represented by the pupa was very 

 near to the final form, if not the final form itself. This conclusion follows from the 

 close morphological similarity of pupa to imago, and from the presence of distinct 

 external generative organs. 



The suppression of intervening stages has left the first or larval stage in an extremely 

 ancestral condition, so that the larva in Lepidoptera is far more ancient than the first 

 stage of those insects (Orthoptera), which still retain the more ancestral method of meta- 

 morphosis. These, therefore, have lost the early stages, while Lepidoptera &c. have 

 lost all the stages intervening between the earliest and a very late stage. 



It is probable that there are very few, if any, lost stages between pupa and imago, 

 but the differences between them are due to subsequent specialization in the latter. 

 Such specialization is frequently of quite recent date. 



It is most interesting to inquire for the possible reasons which determined the loss of 

 the intervening stages and the concentration of metamorphosis. It is quite clear that 

 the loss is associated with, and in fact rendered possible by, the quiescence of the pupal 

 staee, during which the tissues can be broken down (histolysis) and re-developed in the 

 form of the imago. Thus the great morphological interval between larva and imago can 

 be crossed without the need of intervening stages. 



One interesting result of pupal quiescence and of histolytic change is the conclusion 

 that there is no definite pupal stage, as far as the internal parts are concerned. The 

 external parts will be shown to possess a clear and obvious morphological meaning, but 

 a meaning which only becomes clear on the supposition that the internal parts possessed 

 an equally definite significance at some time in the past. The sculpture on the surface 

 of the pupa — its parts and their arrangement — point to a very definite stage; but beneath 

 the surface we find either a gradual transition from larval to imaginal organs or the 



