BY R. J. TILLYARD. 207 



decreases in diameter, and finally leaves the wing-base on its 

 costal side as a fine thread-like trachea. It then runs downward 

 a-nd forward to join the second j^edal trachea(ip^) somewhat below 

 the point where the visceral trunk enters it. 



In all Odonata, the distance from the origin of the alar trunk 

 to the point at which it gives off the anal wing-trachea (A) is 

 very short; while the distance from the point at which the costal 

 wing-trachea (C) is given off, to the point where the alar trunk 

 ■ends on the pedal trachea, is much longer, especially in Agrionidce. 

 The diameter of each alar trunk is very much smaller at its end- 

 ing on the pedal trachea than it is at its origin from the dorsal 

 trunk. 



It seems, therefore, very evident that the oxygen-supply of the 

 developing wing in Odonata is derived from the anal end of the 

 alar trunk, being drawn from the main dorsal trunk, which 

 receives its supply from the branchial basket in Anisoptera, or 

 from the caudal gills in Zygoptera. This fact is, in my opinion, 

 the cause of all the chief peculiarities of the Odonate wing, as 

 will be shown in Section B of this part. 



The diagram in Fig. 20 shows a lateral view of the tracheation 

 of the thorax in a larva of Austrolestes leda Selys, directly after 

 the last larval ecdysis. The drawing was made under the camera- 

 lucida by means of strong, artificial, transmitted light, in which 

 all the tracheae appeared almost black on a pale yellowish-green 

 ground. Making allowances for the greater obliquity of the 

 thorax in the Zygoptera, the results of my dissections of ^Eschnid 

 larvae agree very fairly closely with this diagram. I thought it 

 advisable to publish the diagram of Austrolestes rather than of 

 uEschna, because the former can be much more easily studied, and 

 the complete result verified; while, in the latter, only repeated 

 dissections enable one to piece the whole scheme together, and 

 thereby invite the unwitting introduction of error. 



Section B. — General theory of the unique development of the 

 Odonate wing -venation. 



We are now in a position to enunciate a theory concerning the 

 unique peculiarities of the wing-venation of Odonata, in the 

 following words. 



