198 DEVELOPMENT OF THE WING-VENATIOX OF ODONATA, 



base, where it finally coalesces with A' from the base to Ac. 

 Beyond Ac, it takes a slanting direction from Ac towards Cuob, 

 which is now only about half its original length; all the other 

 subcubital cross-veins becoming similarly shortened. A similar, 

 but less intense narrowing also takes place along the costal 

 border. As a result, the base of the wing becomes quickly 

 narrowed to less than half its former width, and, in this shape, 

 it is withdrawn from the wing-case at metamorphosis. The ex- 

 pansion Avhich follows mainly affects the " blade " of the wing, 

 the petiolate stalk undergoing a comparatively small lengthening. 



A further point of interest in Zyyopterid wings is the develop- 

 ment of Cu. In all Zygopterid wings, there is very little curving 

 of this trachea. In some cases (as in Caliagrio7i), Cug contiimes 

 the line of Cui.2, while, in others, it is about as much curved as 

 Cuj, but, of course, in an opposite direction (see Synlestes, Plate 

 xii., fig. 5). In every case, Cus is less strongly developed than 

 Cui, the latter being a very strong and almost straight trachea. 

 In the Protoneurioii, Cu^ undergoes a gradual reduction, until, in 

 the most asthenogenetic forms (Plate xii., figs. 9-1 2), it is seen to 

 be only a very short trachea terminating somewhere about the 

 level of the origin of M. from Mj..^. Its length varies very much 

 in the different genera. In many genera, Cuo is completely sup- 

 pressed, and Cu then appears as a short, unforked trachea, 

 slightly bent downwards under the arculus. In such forms, it is 

 not surprising to observe the complete disappearance of trachea 

 A, except perhaps for a minute basal portion not easily observed: 

 since the asthenogenetic tendency must affect A before it can 

 touch Cu. 



These highly specialised Proto7ieurini are, without doubt, from 

 the point of view of the reduction of their tracheal system, the 

 most advanced members of that very prolific and successful 

 phylogenetic line, which probably includes, in the form of numer- 

 ous more or less successful side-branches, the whole mass of the 

 CalopterygidfP.^ and also all the AgrionidcK possessing regular 

 quadrilaterals. 



Let us now pass to the question of the origion of Rs in Zyyop- 

 tera. In all except one of the forms so far examined, no connec- 



