1919] Crampton — Terga and Wing-Bases of Insects 63 



seem to be quite closely related to the Neuroptera. The line of 

 development of the Hymenoptera therefore probably arose near 

 the base of the neuropteroid stem, though the Hymenoptera are in 

 some respects intermediate between the Neuroptera and the 

 Psocidse. The Hymenoptera have many characters suggestive of 

 affinities with the Mecoptera, although the Mecopteron shown in 

 Fig. 3 is not so well suited as Panorpodes and other primitive 

 Mecoptera for demonstrating this relationship. Since the mecop- 

 teron line of development originated at a point quite far down on 

 the main neuropteron stem, it is merely to be supposed that the 

 Mecoptera will show evidences of a rather close relationship to the 

 Hymenoptera, which also occupy a position far down this main 

 stem, and similarly, since the Trichoptera branched off from this 

 main stem very near the point of origin of the Mecoptera, both 

 Trichoptera and Mecoptera show indications of affinities with the 

 Hymenoptera as well as with the Neuroptera; but the resemblances 

 between the Hymenoptera, Mecoptera and Trichoptera are the 

 most patent in the larval stages of these insects. Since the Hymen- 

 optera resemble Mecoptera in some respects, and since the Diptera 

 also resemble Mecoptera in many respects, it is to be expected that 

 there will be certain points of resemblance between the Hymenop- 

 tera and Diptera also; but I do not consider that the Hymenoptera 

 and Diptera are as closely related as MacGillivray and other 

 students of the wing venation have been led to suppose, from their 

 studies of this one set of structures alone. That the Siphonaptera 

 (the fleas) were descended from Diptera-like ancestors is admitted 

 by practically all recent investigators; but since these forms are 

 wingless, it has not seemed advisable to include a study of their 

 terga in a paper dealing largely with the wing-bases. 



The evidence of relationships among the insects here discussed, 

 as indicated by a study of the terga and wing bases, is in full 

 accord with the evidence from other structures as well, and the 

 conclusions here reached may be briefly summarized as follows. 

 The Neuroptera are as primitive as any representatives of the 

 superorder (the Panneuroptera) , and probably have departed as 

 little as any from the condition typical of the forms ancestral to 

 the group as a whole. The Hymenoptera are also very primitive, 

 and occupy a position far down on the main neuropteroid stem. 

 The Mecoptera and Trichoptera arose from neuropteroid ancestors, 



