102 BRITISH LEPIDOPTEEA. 



that the very perfect resemblance which larvae bear to their surroundings, 

 so perfect as often to render them practically invisible, as well as the 

 development of long tussocks of dense hair, that is so loose that it is shed 

 without injury to the larva almost as soon as touched, and also the for- 

 mation of huge spiny structures, have all been brought about in preference 

 to offensive structures, because the exceeding delicacy of structure 

 of the larva would prevent it taking any active part in any physical 

 attempt to combat the attacks of its enemies. Such offensive mea- 

 sures as are occasionally adopted, e.g., the syringe of Centra, are so 

 very exceptional, that they can almost be neglected from the considera- 

 tion of the general question. The development of mutant tubercles 

 and flagella are probably meant only to startle ichneumons that 

 injure and attack the larvae in quite another way, whilst protective 

 coloration, and, indeed, all forms of passive defence, alone are used 

 against their physically stronger foes. 



There can be, we think, little doubt that all the purely defensive 

 structures of insects hairs, flagella, glands, etc., have been developed 

 in response to the increasing attacks of enemies. That we know very 

 little about the subject is very evident, and workers have here an un- 

 limited field for observation. Almost every specialised larva responds 

 in some particular way to its environment, and here the field natura- 

 list has the whole domain to himself. Each peculiar structure has 

 its own particular use, and it is only by observing closely the habits of 

 the animal in nature, that the use can be learned. It requires great 

 skill, power of observation, and unbounded patience, but it is a work 

 that will repay the labourer with interest, for all the care he may 

 bestow upon it. 



CHAPTEE IX. 



CLASSIFICATION OF LEPIDOPTERA. 



PROBABLY nothing relating to the Lepidoptera has undergone such a 

 complete revolution during the last decade as our notions of the 

 classification of the order, and, to a great extent, our views are 

 governed by the observations of a few workers, of whom Chapman, 

 Comstock, Dyar and Packard are the chief. In a paper "On the 

 Classification of Lepidoptera " we ventured the opinion that " no 

 scheme based on a single set of characters belonging to only one stage 

 of an insect's existence could possibly be even approximately perfect. 

 It is possible to conceive that, especially in those orders in which the 

 methods of life differ so greatly in the various stages, and different 

 means of defence and protection are thus rendered necessary, an insect 

 may be very greatly modified in one particular stage, without any 

 corresponding modification in the other stages being at all necessary. 

 It may happen to be of advantage for the larva to be of a generalised 

 type, and for the imago to be much more specialised, or vice versa. If 



Tram. Ent. Soc. London, 1895, pp. 343 et seq. 



