DESCRIPTIONS OF AUSTRALIAN iESCHNINJ!. 9 



opinion, to intensify the exceedingly close relationship between these two 

 genera as compared with others. Conversely, therefore, I am quite unable 

 to subscribe to Oockerell's latest proposition * of setting up the Anacini as 

 a distinct tribe co-ordinate in value with the yEsclmini. No doubt the 

 venation of Anax is peculiar, and shows some striking differences from that 

 of other ^schnines ; but this change is, in my view, of very recent date, 

 and in no way comparable to the two diverging tendencies that have been 

 operating for ages on the two diverging groups — the ^schnini and the 

 Brachytrini. We have in Anackesclina the connecting-link between ^schna 

 and Anax, just as we have in Procordulia the connecting-link between 

 Somatochlora and Hemicordulia. Nobody would propose to erect a tribe 

 Hemicordulini ; yet to me the development of Anax from ^scJina offers an 

 exact parallel to that of Hemicordulia from Somatochlora, except that the 

 former has been more vigorously csenogenetic, and so shows a greater degree 

 of differentiation. The same causes, also, have operated in both cases — viz. 

 the invasion, by an offshoot of a temperate group, of the warmer tropical or 

 subtropical region, and the attempt to accommodate itself, both in habits of 

 breeding (in still water) and in manner of flight, to the domain of the 

 Libellidince. 



On the other hand, when one considers the y^schiini inter se, it becomes 

 evident that Anax has so far outrun its nearest relatives in csenogenetic 

 development that some recognition of the differences between them is 

 necessary. It seems, therefore, to me advisable to retain the three separate 

 series in this tribe, as suggested by Walker — viz., the ^Eschna series, the 

 Anax series (including only the genus Anax, into which I propose to merge 

 Hemianax for reasons given below), and the Gynacantha series. This 

 division is strongly supported by the study of the larvse, so far as I have 

 been able to carry it out. The development of the rectal papillae in Anax- 

 nymphs is very remarkable ; and even though it be the direct outcome of the 

 invasion of still water by this genus (the papillae giving an undoubtedly 

 superior respiration- activity to these larvse), yet the fact ought not to be 

 overlooked in classification. In the case of Gynacantha, I have made what 

 to me seems a most unexpected and extraordinary discovery, that the nymph 

 of G. rosenhergi possesses a splendid set of strong setce on the lateral lobes of its 

 labium. If this be a characteristic of the larvse of the group, it will indeed 

 mark them out as very distinct from all other ^schnid larvse. Again, I can 

 only regard this as another csenogenetic character assumed directly as the 

 outcome of the invasion of still water by these nymphs, directly comparable 

 but in no way related to the similar development in the labium of Libellulidse 

 and of those Agrionidse that breed in still water. It would be most inter- 

 esting to see whether the rectal gills of Gynacantha have developed papillae 



♦ Proc. U.S. Nat. Mus. vol. xlv. (1913) pp. 577-583. 



