BY R. J. TILLYARD. 



883 



I 



due to tlie fact that trachea Rs in these forms is itself branched. 

 Now, unfortunately, we do not know the ontogeny of the larval 

 wing in these subfamilies. The suggestion I would make is 

 that, in these archaic forms, we have trachea lis and trachea 



1 



Text-figJJ. — Su>;gested phylojfeny of the Od(jiiate wing in the region of 

 radius, media, and branches, a, archaic ancestor of Zj'goptera and 

 Anisoptera; h^-d^, phylogenj' of Zygoptera; /(j, reduction of Rs; 

 t'l, h)ss of T\s — M eyapodagrio7iine stage; d^, capture of Ms ^er 

 ■sa/tiDu by M„ = Lesfid stage (original course of A-ein Ms preserved as 

 the bridge Br in the iniaginal venation); ''o-e„, phj'logeny of An- 

 isoptera; l>2, Rs crosses Mj (cf. Text-fig. 2/;); c„, Rs crosses Mj and 

 M.J ; d„, Rs captures Ms per m//nm (cf. Text-fig. 5); e„ , Ms sup- 

 pressed (the original course of vein Ms preserved as the bridge Br 

 in iniaginal venation). 



Ms both present, and that, in the last larval instar, the latter 

 trachea becomes hitched on to Rs, just as M.. became hitched on 

 to Ms in the Neostiota aberi'ation. In that case, the whole 



