Pliylogeny and Evolution of the Lejndoptera. 573 



sorts, such as in Hyhocampa and Attacus, and so fortli, 

 but all adhere to the special obtect formula of the fifth 

 and sixth abdominal segments only beinp; free. 



I conclude that this structure of the pupa is that which 

 affords far beyond any other, at once, a solid basis and 

 an extensible ladder, by which the imago can attain a 

 puptd fulcrum from which to reach and rupture the 

 cocoon and force itself out. 



'J'he questions of animal mechanics involved are, 

 however, I must admit, too complex and difficult for me 

 to analyse them successfully. 



A remarkable confirmation of this hypothesis is, how- 

 ever, presented by the butterflies. 



It seems very probable that when the Skippers had but 

 recently obtained obtect rank, they were not only still 

 capable of further evolution in the direction of consolida- 

 tion of punal segments, but had, if it is legitimate so to 

 express it, a distinct impulse towards further consolida- 

 tion. They would, however, no doubt ordinarily have 

 adhered to the usual obtect formula, but fur one circum- 

 stance, viz., they succeeded in doing without any sort of 

 cocoon. 



The mechanical problem of how to escape from the 

 cocoon by the way of least resistance, no longer 

 dominated them ; and accordingly we find in each group 

 of butterflies, that certain families have lost the power 

 of movement in one, two, or all three of the incisions 

 that are movable in the Ohtcctx. That the families with 

 least movement in each division happen to be the highest, 

 may b3 the result of what I have cnlled the impulse to 

 such progress, existing in the Skippers. Much more 

 probably the loss of movement occurred in each instance 

 for some special reason, probably in pursuance of pro- 

 tective devices that so dominate the evolution of 

 butterfly pup^, and as there was no going back, it 

 results inevitably that in each line of development, the 

 most solid pupa must be the most recent. 



In my first paper pi-esented to the Society in 1893, I 

 believe I gave the impression that I believed that all the 

 Lepidoptera might be arranged by their pupa3 in one line ; 

 I had not this idea myself, but I had not formed any 

 alternative view with sufficient definiteness to enable me 

 to advance it. 



I do not know that I can go much further yet, but I 



