IQOS] 



(lassification of the Corduliiuae 



277 



longitudinally, is established; there is nothing like a bisector in 

 the Macromian line. 



The anal loop, although just beginning to l)e recognized and 

 used in the systematic study of the Odonata, has undoubtedly 

 played a role of first importance in the evolution of the wings of 

 the Corduliinae. Within the restricted group just defined the 

 tip of the anal loop has developed in two divergent ways: 



Fig. 8. Tj^pical structures of Corduline wings: (a) the wings of Tetrago- 

 neuria cynosura Say; (b) the anal loop of Gomphomacromia paradoxa; (C) the 

 same of Syncordulia gracilis; (d) base of hind wing of Macromia. Veins: (c), 

 costa; (Sc) subcosta; (R) raditis; (M) media; (Cu) cubitus; (A) anal; (Rs) radial 

 sector; other branches numbered from front to rear. Other parts, (ar) arculus; 

 (n) nodus; (st) stigma; (s) supertriangle; (t) triangle; (u) subtriangle; (m) 

 membranule; (x) anal triangle of the male; (1) anal loop; (y) radial supplement; 

 (z) median supplement. 



(i) It has extended itself outward parallel to the vein Cui in a 

 bilaterally symmetrical point, and the tendency of the triangle 

 to be concurrently retracted to the level of the arculus has been 

 least where the elongating loop has most closely paralleled vein 

 Cui. (Idomacromia, etc.) This tendency may be said to char- 

 acterize in a general way Williamson's groups III and IV, with 

 Neoph^^a added to group IV. The Triangle has reached the level 



