Ill] THE WINGS 53 



cases a brace was not developed. In many Gomphinae (D) further 

 strengthening was gained by convexity of the posterior border. 

 In the Libellulidae, the gradual adoption of a soaring or " skimming " 

 habit of flight, instead of the original method of " darting," lessened 

 the strain on the pterostigma, which finally tended to become both 

 smaller and unbraced in the highest forms (E, F). 



In the Zygoptera, a still earlier and more complete departure 

 from the original method of flight quickly relieved the stress on 

 this part of the wing. Hence it is only in the oldest Calopterygidae 

 (Epallaginae, Thorinae, G) and in the Lestidae (H) that the archaic 

 regular elongated pterostigma is found. In the Calopteryginae 

 (N, o) this organ is either completely absent, or represented by 

 a "false" pterostigma of a whitish texture, traversed by fine veins, 

 and not enclosed on either side by strong cross-veins. In the 

 Pseudostigmatinae also, the pterostigma is either absent or false. 

 It exhibits a peculiar abnormality in Mecistogaster (M), where it 

 forms a kind of costal hump. In the Agrionidae in general there 

 is a decrease in length in the pterostigma from the older to the 

 more advanced forms. The highly developed Agrioninae have it 

 lozenge or trapeze-shaped. In a few genera, such as Anomal- 

 agrion (L), it becomes abnormal, and may sink below the level 

 of the costa. In these advanced forms, the general weakness of 

 the "sculling" flight renders the pterostigma almost unnecessary, 

 as regards its original purpose of strengthening the wing. In 

 Ischnura (j, K) a secondary sexual difference is developed, the 

 pterostigma of the fore- wing in the male becoming particoloured, 

 while that of the female remains unicolourous. A similar device 

 occurs in the Libelluline genus Hemistigma, but in both sexes. 



The Nodus (fig. 19). 



The phylogeny of the formation of the nodus in the Dragonfly 

 wing has been until quite recently purely a matter for speculation, 

 the palaeontological evidence available being capable of more than 

 one interpretation. On the one hand, it seemed clear that the 

 shortening of the subcosta, to end up at a point somew r here near 

 the middle of the costal border, must have been the essential con- 

 dition precedent to nodus-formation. But, on the other hand, 



