182 DEVELOPMENT OF THE WING-VENATION OF ODONATA, 



the fine ends of the tracheae penetrate for a very considerable 

 distance. The same thing may be seen in the forewing of Ilemi- 

 cordulia {Yig.W A.) . What is of greater interest, however, is the 

 fact that, as in Stage 4, Aj does not help to form the distal border 

 of the loop at all, but is a weak trachea curving in half-way 

 between Ac and Cug to form the distal border of the first of the 

 two large cells forming the loop. Ag, on the other hand, is a 

 very strong, thick trachea, descending from Ac and forming the 

 basal side and part of the lower border of the loop. Cuob is 

 visible as a fairly strong trachea descending from Cuo into the 

 space below the imaginal wing-border. From these considera- 



■R+n; 



Iwb 



Fig. 13. 



Fig. 14. 



Y\^.\^.—Cordiilephya pygmcea Selys, i. Tracheation of full-grown 

 uymphal hiiidwing. (Lily Vale, N.S. W.; Feb., 1914). Notice the 

 narrowing of the imaginal wing-border. 



Y\g.\^.~Gordulephya pygmcea Selys, i . Basal portion of imaginal vena- 

 tion of hindwing. (Lily Vale; March, 1907). 



tions, it is quite clear that the loop possessed by the unreduced 

 wing ancestral to Cordulephya was definitely a cuhito-anal loop, 

 formed between Ao and Cu.,. Such a loop really distinguishes 

 the Eucorduliini and Lihelluliuce from all other Odonata. Hence 

 we may fairly place Cordulephya as a specialised side-branch of 

 the Eucorduliini^ which must have branched off from the main 

 stem at some period between Stage 3 and Stage 4. This agrees 

 exactly with the argument alread}^ advanced by me,"**' as a result 

 of a general study of its larva, that Cordulephya is a member of 

 the Eucorduliini, and has its nearest existing ally in the unre- 

 duced genus, Hesperocordulia, from Western Australia. 

 * " On the genus Cordulephya,''' These Proceedings, 1911, xxxvi., p.388. 



