46 COLORATION OF LARV-E. 



belly, too, may bear black marks, but only in such cases 

 where there are glands, which the larva can exsert 

 at will, and when it has the habit of throwing the 

 abdomen over the head (as does Croesus) for the double 

 purpose of exposing the glands, and whipping away 

 ichneumons. That the larvae can drive away these 

 insects by means of the abdomen, I have noticed more 

 than once with Croesus septentrionalis. 



Many greenish-coloured larvae give out odours and 

 secretions, but they differ in habits from those just 

 described. They are small larvae with flat bodies ; 

 they feed on the upper side of the leaf, eating only 

 the cuticle, so that in this way it becomes white. Now, 

 as these larvae are gregarious, and are not only covered 

 with secretions, but can also give out bad smells, they 

 are enabled to surround themselves with a fetid atmos- 

 phere, which makes their presence as effectually known 

 as if they had bodies of bright contrasting colours. 



A priori we might expect that species which are very 

 closely related and similarly marked as imagines would 

 also resemble each other in larvae. But no conclusion 

 could be more astray from the actual state of the facts. 

 There are, indeed, some genera and groups in particular 

 genera in which the larvae and imagines are coloured 

 and marked alike in the embryonic and developed 

 states, such as, for instance, with Dineura (so far as 

 we know), but others which closely resemble each 

 other in the imago form are utterly dissimilar in the 

 early one. A striking example of this is found in 

 Croesus. The larvae of the three British species have 

 the same forms and the same habits, but as regards 

 coloration they are utterly distinct. This difference in 

 coloration is, I think, readily explainable by the larvae 

 of C. septentrionalis and C. latipes being more active 

 and more offensive, as is shown by the bad odours they 

 give out. C. varus, on the contrary, is not quite so 

 active, and does not use the ventral glands so effec- 

 tively, but to make up for this it is of the same green 

 as the alder with only a few slight black lines along the 



