406 NEW AND RARE AUSTRALIAN AGRIONID^, 



To take an Australian example : — DiphUhia is more closely 

 allied to Podopteryx and Aryiolesf.es than it is to the great mass 

 of Calopterygidce. Yet because it has five antenodals, and 

 Aryiolestes only two, they are put into separate families ! 



The problem is admittedly a most difficult one, and the only 

 solution lies in the detailed study of the earlier types of the 

 Zygoptera, and especially of the rarer forms which may possess 

 the key to the puzzle. Though the Australian species are nearly 

 all classed as Agrion'id(e, yet the work I have so far been able to 

 do, shows tlie presence of at least one of the z groups of the 

 hypothesis outlined above Much more remains to be done, and 

 the detailed study of life-histories is reserved for other papers. 

 But, while following, for the present, de Selys' classification as 

 regards the Agrionidte, and his subdivision of the family into 

 " legions," I think the following points may be briefly stated 

 here, as indicating the lines suggested so far by the study of the 

 early stages of the Australian species : — 



i. The original Zygopterid stock has its nearest representatives 

 to-day in those species which possess an unspecialised larval 

 gizzard (sixteen folds, eight minor and eight major, with a large 

 number of similar teeth). This form appears to exist still in 

 most Caloptterygidre, possibly not in all, but it is found also in 

 Argiolestes and the Australian species of the legion Protoneura. 

 Biphlebia, Aryiolestes, and the legion Protoneura s^ve all stages on 

 an asthenogenetic line of descent, though the first two, as they 

 now exist, do not probably correspond entirely with the ancestors 

 of Protoneura most like them; in other words, a slight but 

 advantageous line of specialisation has given these forms equi- 

 librium, without forcing them to the extreme reduction-stage of 

 Protoneura. 



ii. Synlestes and Lestes possess a highly specialised larval gizzard 

 of a peculiar form. In spite of some obvious venational difl:er- 

 ences, Synlestes is essentially Lestine, and has no close relation- 

 ship with Argiolestes, its resemblance to that genus being due 

 purely to convergence through asthenogenetic reduction on similar 

 lines. This group is archaicly csenogenetic from the main 



