22 



imaginal stage, which is again markedly different from all the stages 

 preceding it. 



Originally it was supposed that the caterpillar had within it " the 

 germ of the future butterfly;" that it had only to throw off a certain 

 number of larval skins to disclose the pupa, and the pupal skin to 

 liberate the imago, the latter being supposed to be within the 

 various skins all the time. On this theory that the larva hid, as 

 it were, the imago, the term " larva " ( = a mask) was a very appli- 

 cable one to the young of the holometabolous insects. This theory 

 of the early entomologists, too, was largely founded on fact, for 

 Swammerdam observed that in the full-fed larva of Cidex he could 

 point out all the limbs of the future nymph concealed beneath the 

 skin; and he has also noticed that beneath the skin of the larvae 

 of bees just before pupation took place he could see the antennae, 

 mouth-parts, wings, and legs of the adults. Our views, however, on 

 the nature of metamorphosis have undergone a radical change since 

 Weismann discovered the imaginal germs within the larva, and since he 

 pointed out the fundamental processes concerned in the change from 

 larva to pupa, and pupa to imago. 



Larvae themselves have been divided into two classes : i \) The cam- 

 podeoid form, resembling the active nymphs of the heterometa- 

 bolous insects. (2) The cruciform larvae, including those of saw-flies, 

 Lepidoptera, Hymenoptera, Coleoptera, etc. 



The campodeoid is considered on many grounds to be the older 

 form, and it is urged that the larvae of Melo'c and of Sitaris prove 

 the cruciform type of larva to be derived from the campodeoid, 

 since these beetle larvae when young are of the latter, and when fully 

 grown of the former, type, the campodeoid form being lost as soon 

 as the larvae begin their parasitic mode of life. The cruciform larvae 

 appear to be due to the degeneration of the true legs, consequent 

 upon living a cofnparatively stationary life, the prolegs having been 

 developed later, when the competition for existence no longer enabled 

 the species to compete successfully without a certain amount of com- 

 paratively rapid movement. The most specialised type of the cruci- 

 form larvce may be seen in certain dipterous and hymenopterous 

 larvae which have lost the power of movement from place to place, 

 and either bathe themselves in their food, or take no active part in 

 obtaining it. 



With regard to the origin of the eruciform larva, Packard considers 

 that the first steps were taken in the Neuroptera (as restricted by 

 Brauer), in which, though the larvae are campodeoid, there is a true 

 resting pupal stage. He states that in Mantispa there are two larval 

 stages which give the key as to the mode in which the change was 

 brought about. The larva of this insect has a true campodeoid 

 form when newly hatched, with long four-jointed legs, which would 

 enable it to move about freely after its prey. It begins, however, at 

 once to live a sedentary life in the egg-sac of a spider, and, before its 

 first moult, it loses the use of its legs, while the antennae are partly 



