ANCESTR Y AND CLA SSIFICA TION. 233 



than some butterfly eggs; no defect nor unful- 

 filled purpose can be conceived in any of the 

 organs of the butterfly. So too in the changes 

 from caterpillar to butterfly, what could more com- 

 pletely accomplish its end than the chrysalis ; it 

 is a perfect mummy. Only in the caterpillar is 

 there any opportunity for additional change. 

 The first step toward the formation of a new 

 stage I believe may be seen in the differences in 

 the appendages of immature and of full-grown 

 caterpillars, described on an earlier page. I 

 have not yet ascertained any common feature by 

 which the peculiarities of the earliest stage of 

 caterpillars may be contrasted with those of later 

 stages, unless it be the comparative length of the 

 appendages. But certainly no caterpillar has 

 been examined which does not show between 

 these stages differences in form, in the relation of 

 parts, or in the disposition and character of the 

 appendages, greater than occur between caterpil- 

 lars of different tribes. While if we go outside 

 the families of butterflies we shall find here and 

 there conspicuous examples of this tendency to 

 hypermetamorphosis, which the mere mention 

 of the young larva of Meloe and Sitaris will recall 

 to the mind of any entomologist. All these I 

 conceive to be examples of the present working of 

 the same law, which from simple transformations 

 like those of our grasshoppers and plant bugs 



