400 STUDIES iX AUStRALlAS' MECOPTERA, ii., 



ti'aclieEe all come off from a single stem. This must be regarded 

 as a higher specialisation than that found in the Mecoptera, or 

 in any Holometabolous Ordei'. For, in all of these, it would 

 appear that tu-o is the minimum number of wing-tracheae yet 

 reached in the course of evolution, i.e., one from each of the main 

 tracheal groups. 



It is interesting to compare the courses of these two trachea^ 

 in the four Holometabolous Orders which show reduction. Of 

 the four, the only one which retains the media intact and separate 

 from the base onwards is the Mecoptera. In this Order, the 

 trachea belonging to the cubito-anal group passes into the media, 

 which is as far forward as it could possibly get. This shows 

 clearly that, in the ancestral form of the Mecoptera, the median 

 trachea had already become attached to the cubito-anal group, as 

 it has in most Orders of Insects. In the Hymenoptera, the media 

 is suppressed in the imaginal venation, and only a bare vestige 

 of its trachea is to be seen in a very ancient family, the Siricidce. 

 The principal trachea of the cubito-anal group passes along the 

 cubitus, and may or may not be accompanied by a separate anal 

 trachea behnv it. In the highest forms, this latter becomes fused 

 with the cubital trachea; so that the condition of two trachea? 

 only, in the wing-rudiment, is there fully reached. In the 

 Trichoptera and Diptera, in which the media is fused basally 

 with either the radius or the cubitus for a greater or less dis- 

 tance, the trachea of the cubito-anal group passes likewise along 

 the cubitus. 



From a study of this character only, then, it is clear that the 

 Mecoptera present a more archaic stage than that seen in the 

 Trichoptera and Diptera; and hence there is no evidence here 

 against the presumption that both these Orders are derived from 

 the older Mecoptera, as Handlirsch supposed. The Hymenoptera, 

 on the other hand, cannot be derived from the Mecoptei-a, since 

 they show a more archaic stage in the Siricida' and some other 

 families (where more than two trachea3 are still present), even 

 though they stand far in advance of the Mecoptera in having the 

 media eliminated from their venational scheme. 



