4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I23 



the last immature instar of hemimetabolous insects. It must be noted, 

 however, that a pupa has essentially the form of the adult, and ex- 

 ternally is a preliminary adult stage with the internal organs not yet 

 completed. If the developing imago of the dragonfly within the larval 

 culticle shed the larval cuticle at the beginning of its metamorphosis 

 instead of at the end, it would come out as an incomplete adult, and 

 would thus be equivalent to a pupa, though it might complete its de- 

 velopment without another moult. According to Whedon (1929) the 

 dissolution of the larval abdominal muscles is not completed until 

 several days after ecdysis. 



The reason for the great degree of change made by the dragonfly 

 at the moult to the adult is the disparity between the larval strvicture 

 and the structure of the imago. The larva is specifically adapted in 

 almost every detail of its organization to life in the water ; the adult, 

 on the other hand, is equally adapted to the very different life it leads 

 in the air. In other words, through many millions of years the larva 

 and the imago have been evolving along quite different lines each for 

 its own purpose. The laws of heredity have been suspended, giving 

 the young insect freedom to follow its own evolutionary course, so 

 long as it finally reverts to the parental form. The embryo, therefore, 

 develops directly into the larva, and the hormonal condition within 

 the larva does not allow the forces of adult heredity to assert them- 

 selves until the larva has performed its part in the life of the species. 



Tillyard (1917) has suggested that the larvae of the Carboniferous 

 Protodonata probably lived in mud and rotting vegetation around the 

 shores of some stagnant lake, and here "began that series of adaptive 

 changes which finally led them to adopt a purely aquatic life." We 

 must assume, however, that the young stages of insects with meta- 

 morphosis at some early time in their evolution resembled the adults 

 of their species in the same degree that the young of modern insects 

 without metamorphosis, such as a grasshopper or a cockroach, resem- 

 ble their parents. The adult Protodonata were already perfect dragon- 

 flies in their way, and it is therefore highly improbable that at this 

 stage of odonate evolution the young could have had any likeness to 

 the specialized adults. The adult Protodonata themselves must have 

 had a long line of evolution extending back through a million years 

 or so before they reached the status of Carboniferous dragonflies. 

 The larvae, then, began their divergence from the adult line of evolu- 

 tion when the adults were insects of a much more generalized type of 

 structure than that of the Protodonata. In the body of a modern 

 dragonfly larva the only part that in any way resembles the corre- 

 sponding part of the adult is the thorax ; the specialized adult type 



