CLASSIFICATION 19 



from some mandibulate and winged form. The conversion of mandibu- 

 late into suctorial organs may be seen within the order Collembola, but it 

 is highly improbable that Hemiptera arose from forms like Collembola. 

 Hemiptera are exceptional among insects with a direct metamorphosis in 

 their highly developed type of suctorial mouth parts. 



Metamorphosis offers, upon the whole, the broadest criteria for the 

 separation of insects into primary groups. All the orders considered thus 

 far are characterized either by no metamorphosis or by a slight, or so- 

 called " direct," or "incomplete," transformation. The following orders, 

 on the contrary, are distinguished by an " indirect," or " complete," 

 metamorphosis, which appears in Neuroptera and attains its maximum 

 development in Diptera and Hymenoptera. 



With Neuroptera the eruciform type of larva appears, as a derivative 

 of the earlier thysanuriform type. The larva of Mantispa, as Packard 

 has shown, actually passes, during its individual development, from the 

 primary, thysanuriform stage to the secondary, eruciform condition. 



Mecoptera form an isolated order, though their caterpillar-like larvae, 

 with eleven or twelve pairs of legs, suggest affinities with Lepidoptera 

 and, more remotely, with the tenthredinid Hymenoptera. 



Trichoptera, while much like Mecoptera in structure and metamor- 

 phosis, are undoubtedly closely related to Lepidoptera; in view of the ex- 

 tensive and deep-seated resemblances between caddis flies and the most 

 generalized moths (Micropterygidae) there is little doubt that Trichoptera 

 and Lepidoptera originated from the same stock. 



The origin of the coherent group Coleoptera is by no means clear, al- 

 though thysanuriform larvae occur frequently in this order. Packard 

 suggests that both beetles and earwigs arose from some thysanuroid form 

 or that the primitive coleopterous larva sprang from some metabolous 

 neuropteroid form. In any linear arrangement of the orders the position 

 of Coleoptera is largely arbitrary, and here the order is intruded between 

 Lepidoptera and Diptera simply for want of a more satisfactory place. 



Lepidoptera, Trichoptera and Mecoptera are probably branches from 

 one stem. Lepidoptera, Diptera and Kymenoptera are regarded by 

 Packard as having had a common origin from metabolic Neuroptera. 



Among Diptera, such larvae as those of Culicidae are comparatively 

 primitive, according to Packard, and larvae of Muscidae are secondary, or 

 adaptive, forms. 



Siphonaptera used to be regarded as Diptera and are probably an off- 

 shoot from the dipteran stem. 



The most primitive hymenopterous larvae are those of the sawflies 



