THE PALEONTOLOGICAL DATA 93 



Devonian insect and consequently one of the oldest insects known, the 

 subcosta as figured by Scudder is clearly of this type. In the Plecoptera, 

 which is a very generalized order of insects, probably more generalized than 

 any other of the recent orders of winged insects, except perhaps the 

 Orthoptera, the subcosta is of this type (Fig. 82 , a) . In very young nymphs 

 of Odonatathe subcostal trachea is two-branched (Fig. 227), andunbranched 

 in the wing of a mature nymph (Fig. 228). In the Lepidoptera, the sub- 

 costa is two-branched in Hepialus; and MacGillivray has since shown that 

 the subcostal trachea is two-branched. 



Passing to the orders in which the tracheation is reduced, and conse- 

 quently of no aid in the solution of this problem, the subcosta is two- 

 branched in many of the Trichoptera, in the more generalized Hymenop- 

 tera, and in Protoplasa, the most generalized of the recent Diptera. It is 



Sg Sc 



' \ 



Fig. 82. — The three types of the subcosta; a, subcosta of 



Nemoura; b, subcosta of Rhyphus; c, subcosta 



of Corydalus. 



unnecessary to go farther to show that we had good reason for regarding 

 the two-branched type of the subcosta as primitive. 



Let us now turn to an examination of the paleontological evidence bear- 

 ing on the primitive type of the subcosta. At first sight this evidence does 

 not support the conclusion drawn from the study of living fonns; for in the 

 Palseodictyoptera, which in most respects are the most generalized of 

 fossil insects, the subcosta was not branched (Fig. 77). 



I do not feel, however, that this evidence is conclusive. The most 

 generalized of the Pateodictyoptera are the Dictyoneuridc-E, of which 

 Stenodictya lobata (Fig. 77) is the best preserved form known. While the 

 members of this family have retained the prothoracic and abdominal 

 wing-rudiments, and have not yet attained definite cross-veins in the wings, 

 both primitive characteristics, the mesothoracic and metathoracic wings 

 have evidently been specialized in certain ways that will be discussed later. 

 And it is quite possible that along with these specializations there has been 

 a loss of one branch of the subcosta. One must keep constantly in mind 



