382 AUSTRALIAN CORDULUN^, 



formation of wliat ma}' be termed strong and weak groups of 

 imagines. For instance, the Macromina on the strong side, and 

 the Synthemina on the wealc side are two successful results 

 of one line of development. Thougli, in wing-venation, the 

 Macromina have progressed far beyond the Sytithemina, yet the 

 balance is probably maintained by the better larval development 

 of the latter (as I have already shewn, these larvae resist drought 

 and starvation to a degree hitherto unprecedented in known 

 forms). It has, therefore, occurred to me that, if we knew the 

 larvae of those forms which I have grouped together in the 

 Cordulina, we might find the same "strong" and "weak" 

 tendencies of the imagines (balanced by the opposite tendencies 

 in the larvte) clearly enough indicated to enable us to subdivide 

 the group into two sections co-ordinate witli the Macromina and 

 Synthemina. 



As I cannot claim a very intimate knowledge of some of the 

 genera that are not Australian, and as the life-histories of so many 

 species still remain quite unknown, I shall confine this discussion 

 to the Australian genera, togethei' uith a few of the better known 

 Old World forms when necessary. A cursory glance at the Aus- 

 tralian members of the Cordulina suggests, at once, a " strong " 

 and "weak" line of development. The strong group, typified 

 chiefly by robust head and thorax, long legs, more or less corduli- 

 form abdomen, and elongated anal loop, together with complete 

 recession oi the hind wing-triangle to the arculus, is clearly repre- 

 sented by the great genus *SW2«tot'A/o?-«( bipolar) and its offshoots, 

 Procordalia and HemicorduJia. The iveak group, typified by 

 smaller head and thorax, short legs, thin iiearly cylindrical abdo- 

 men, and shortened anal loop, together with incomplete recession 

 of the hindwing-triangle, is represented by Syncordulia and its 

 allies, amongst which I number Austrophya, Austrocordulia, 

 Pseudocordulia, and LathrocorduJia. 



Were these two gi-oups represented by a single type of lar\a, I 

 should be content to leave the Cordulina as a single group 

 coordinate with Macromina and Synthemina. But the discovery 

 of the remarkable larva X*(the actual species is still undeter- 



* "Oil some Experiments with Dragonfly Larvje." 'I'liese Proceedings, 

 1910, XXXV., p. 667, 



