BY K. J. TILLYAICD. 381 



cylindrical (rnther flat in the type, which is not fully matured), 

 shining nietallic-blackisli from 6 to 10; 5 showing green reflec- 

 tions; 1-4 brilliant metallic-green; 1, brownish on sides; 2, a pair 

 of small basal dorsal yellow spots, a narrow transverse yellow 

 band interrupted for 1 mm. on dorsum; 3, a fine yellow transverse 

 line close up to the suture; supplementary carinse on 3-6 black; 

 4-6 with no mar-kings; 7 with basal half of dorsum yellow; 8-10 

 with a touch of yellow low down on each side. Vulvar lamina 

 with two leaf-like folds or lobes about 1 mm. long, rounded. 

 Appendages 1mm., black, conical, downy, separated by the 

 triangular end of 10, under which projects a longish irregular 

 tubercle. 



llab. — Cape York, N. Queensland. A unique $, taken by Mr. 

 H. Elgner, on November 22nd, 1909, and now in my collection. 



It is now necessary to consider further the position of the 

 new Australian genera (described by nie in this, and the preced- 

 ing paper) in the subfamily Cordulihifc. These are Ausfrophya, 

 Austrocordulia, Pseudocordulia, Hesjjerocordulia, and Lathrocor- 

 diifia. This is a diflicult task, mainly because of the unsatis- 

 factory state of t)ur knowledge of the frue affinities of the 

 members of the group Cordulina* to which all these genera 

 belong. It must be borne in mind that, in a group of this kind, 

 containing the more highly specialised and advanced members of 

 the subfamih', ctmvergence of forms is very likely to occur. This 

 is particularlv the case as regards wing-venation, in forms where 

 the tendency has been towards reduction rather than amplifica- 

 tion. In the struggle for existence, it is not always the imago 

 that undergoes the most complete modification; in the Odonata, 

 at any rate, there are groups in which the main line of specialisa- 

 tion has been confined to the larvfe, the imagines remaining of a 

 faii'lv generalised tj'pe — and yet there is no reason to believe 

 tliat such forms are foredoomed to failure. The Gomphiiue 

 are a good illustration of this point. Now in the Curduliincf, 

 one can distinguish various tendencies at work, resulting in the 



* In my " Monograph of the genua Syvtkemis" I separated the Cordu- 

 liiniv into Syntfiei7iina, Macromina and Cordulina, leaving the latter group 

 for further revisiim and siihdivisicjii if necessary. 



