Neuroptera and Coleoptera 359 



most ancient of the metabolic orders. In some families 

 of Beetles the campodeiform larva is to be found 

 almost in its primitive state ; in others a markedly 

 eruciform type has been developed. It is of interest 

 to notice that, as a rule, those Coleoptera with the 

 most highly modified larvae are distinguished in the 

 perfect state by a reduction in the primitive number 

 (five) of tarsal segments. Among the Neuroptera 

 the eruciform larva has only been assumed in a single 

 family (Panorpidae), but the campodeiform larva has 

 undergone more speciaHsation than among the Beetles ; 

 the cercopods have disappeared and the jaws have 

 become adapted for suction by the grooving of the 

 mandibles. In the perfect state the Neuroptera have 

 retained many ancient characters, the wings of the two 

 pairs being alike, and the abdominal segments but 

 little modified. The fore-wings of the Coleoptera, 

 on the other hand, have undergone greater changes 

 than those of any other insects, by becoming trans- 

 formed into dense chitinised elytra. With this excep- 

 tion, however, the Beetles are not a highly specialised 

 order ; the biting mouth is retained, the number of 

 abdominal segments is only slightly reduced, and the 

 prothorax remains distinct as in the lower insects. 

 It seems, therefore, that we are justified in regarding 

 the Neuroptera and Coleoptera as oif-shoots from a 

 common stock. And if we attempt to reconstruct the 

 presumed ancestor by uniting die most primitive 

 characters in the two orders, by imagining, for instance, 

 an insect with the body-form of the oil-beetle's active 

 larva, provided with four similar net-veined wings, 

 we realise the probabifity that the Neuroptera and 

 Coleoptera of to-day are the descendants of the 

 Platvpteroid insects of the Primary epoch. 



We are compelled to go back some distance in the 

 Primary period for the ancestors of the oldest meta- 



