Origin of Metamorphosis 357 



to complete its growth in the outer world, as do the 

 young of the marine animals just mentioned. Doubt- 

 less the larva represents in some respects an older 

 stage in evolution than the perfect insect. But it is 

 a well-developed animal which cannot, only because 

 it has yet no wings, be compared for a moment with 

 the free-swimming gastrula of a starfish or the veliger 

 of a water-snail. It has therefore been suggested, on 

 the other hand, that the insect-larva represents the 

 primitive adult insect, and that the transformations 

 through which it passes are entirely post-embryonic, 

 comparable for example to the growth of antlers in 

 the stag (71). 



This latter view receives strong confirmation from 

 the striking fact that no insect of any order is hatched 

 in the winged state. The way in which wings are 

 acquired varies greatly, but they are never fully de- 

 veloped until late in life. It seems certain, therefore, 

 that the growth of wings after hatching has always 

 been exhibited by the class, and that we may safely 

 take the development of the Cockroach as a type of 

 what occurred among the first winged insects. In 

 the second chapter evidence was brought forward to 

 show that the campodeiform larva is more ancient than 

 the eruciform ; and the few instances, such as the 

 Meloidar and Stylopidse, in which the campodeiform 

 precedes the eruciform grub in the life-history of 

 the same insect (see p. lio), clearly shows that the 

 latter form was gradually adopted with changing 

 conditions of larval life. As the higher insects ex- 

 hibited greater and greater divergence of habits and 

 surroundings in the winged and wingless stages of 

 their life-history, larva and imago became more and 

 more divergent in structure, and the resting pupal 

 stage with its profound internal changes became an 

 absolute necessity as a connecting link between the 



