E. II Sellards — Types of Permian Insects. 253 



ham emphasized the fact that there exists in the wing of most 

 adult dragon-flies a persistently oblique cross vein lying usually 

 at or just beyond the separation of M M ; also a similar oblique 

 vein above, uniting the radius and M x . From the nymph wing 

 it was found that the trachea of the radial sector crosses M, 2 , 

 and that the oblique apparently cross veins are in fact parts of 

 the sector. An examination of this region on the Permian 

 genus Twpus reveals the presence of such an oblique vein 

 arising from M 2 at a point just beyond the separation of M. 1 

 and M 2 ; also an equally well-marked oblique vein, the sub- 

 nodus, connecting the radius and media. Bearing in mind 

 the complete agreement of all other veins of the radio-medial 

 area with those of the same area in modern Odonata, there can 



Figure 4. — Enlarged detail of the 

 region of the subnodus of Tupus per- 

 mianus, taken from the right front 

 "wing. Sn, subnodus ; 0, oblique 

 vein; Br, bridge; E, radius: Mi, 

 first main branch of media ; M 2 , sec- 

 ond main branch of media ; Es, ra- 

 dial sector ; M 3 , third, and M 4 , fourth 

 main branches of media. Enlarged 

 4 diam.: drawn with camera lucida. 



Figure 5 — Same area from the 

 left hind wing of Tupus permianus. 

 Enlargement and lettering as in fig- 

 ure 4. In this wing there is a slight 

 depression (not shown in the figure) 

 running from the subnodus toward 

 the oblique vein. 



scarcely remain a doubt as to the meaning of these two oblique 

 cross veins. They clearly indicate a crossing of the radial sec- 

 tor. The area of the wing in the region of the crossing of the 

 sector is shown enlarged in the accompanying camera lucida 

 sketches. The fundamental significance of the crossing of the 

 sector as a bond or relationship between Paleozoic and modern 

 forms is too obvious to require further emphasis.* 



It will be observed that in this Permian genus there is appar- 

 ently a loss of that part of the sector lying between M x and M 2 . 



* Handlirsch's statement, Eevision American Paleozoic Insects, p. 689, 

 with reference to the Protodonata, that the "intersection of the longitudi- 

 nal veins" is "still entirely wanting," can not, I feel sure, be maintained 

 in view of the evidence here presented. The agreement of the veins of the 

 entire radio-median area is altogether too complete to admit of explanation 

 on any grounds other than that of strict homology. 



