184 DEVELOPMENT OF THE WING-VENATION OF ODOyATA, 



fig.5) SO as to meet Cuo just at its point of greatest curvature. 

 Thus A. and Cu, between them, form two sides of a large trian- 

 gular space, the suhtrimigle (st). One of the set of cross-veins 

 developed transversely^ between Cuj.o and Aj now becomes 

 specialised by slantwise lengthening, so as to run from the first 

 bend of Cu^.^ towards the bend of Aj, and thus develops into a 

 firm support for both bends, as the inner side of the suhtrimigle. 

 In Austrocordulia, owing to the small amount of transverse 

 elongation of the triangle, st is rather small, and consists of one 

 strongly-formed triangular cell. But in Synthemis, owing to 

 further transverse elongation of the triangle, st has increased 

 very much in area. Hence is developed the system of support- 

 ing cross-veins, which divides the subtriangle into three cells, 

 and which persists throughout the rest of the phylogenetic series 

 with only slight variations. 



Stage 4 {Eucorduliini and Libellulince) shows very little ad- 

 vance from Synthemis. The growth of Ai is stronger, the bend 

 much sharper (approaching a right angle), the subtriangle larger 

 and more strongly formed, but essentially of the same structure 

 (Fig.llA). It tends, however, to become yb?*r-5io?eo? through a 

 second bending of A^, and at this same point A^ gives off a fairly 

 strong branch, which runs at first considerably basad, then bends 

 completely round, and finally runs distad to meet Aj itself after 

 its "break-back" from Cu.j. Thus there is formed a small and 

 complete " anal loop," represented in the imaginal venation by 

 two strong supporting cells lying under st. 



In Plate xi., fig.L^, the very interesting forewing of Cordule- 

 phya nymph is figured. It shows several peculiarities. Ac runs 

 very slantingly, parallel to the basal side of the triangle. The 

 bridge-vein is extremely well developed. The postnodal area 

 of the wing is excessively shortened, and the tip of the wing is 

 formed irregularly, as if a small part of the margin had been 

 broken off. Tins last peculiarity is to be seen also in Hemicor- 

 dulia, and probably in other Eucorduliini, while the shorten- 

 ing of the postnodal area is also noticeable in this group. These 

 are two further links in the chain of evidence which binds Cor- 

 dulephya to the Eucorduliini. 



