704 LIFE-HISTOKIES OF AUSTRALIAN ODONATA, NO. III., 



systematic classification. Hitherto placed with the Cordulegas- 

 terince, they must now be removed from that subfamily, and either 

 added to the true JSschnince, or placed next to the latter in a 

 small subfamily of their own. The argument in favour of this 

 change may be stated as follows : — 



(1) In studying the position of the Petalia-gvonp of genera, 

 from a knowledge of the imagines only, too much stress has been 

 laid upon one characteristic, viz,, the fact that their eyes onl}' 

 just touch, as is also the case with Cordidegafiter. Mainly on 

 this character, they have been included in the C orduUgasterince^ 

 this subfamily having been distinguished from the true ^Eschnhice 

 by the fact that, in the latter, the ej^es are closely contiguous for 

 a considerable distance. As a matter of fact, this character is 

 reall}^ of less importance than others which have been entirely 

 ignored by systematists in placing this genus. In the Libellidince 

 we can recall the case of two closely allied insects, viz., Croco- 

 themis erythi^cea Brulle, and Rhodothemis rufa Rambur, in which 

 the chief difference is that the eyes of the former are, like those 

 of nearly all Libellidince, closel}' contiguous; while in the latter, 

 the eyes barely touch, and, in some specimens which I possess, 

 are absolutely separated. Yet in all other respects these insects 

 are closely allied, and their divergence in this one characteristic 

 is only sufficient to place them in two diff*erent but closely united 

 genera. We ought, therefore, in the case of Petaliasmd allies, to 

 examine whether, in other respects, they show a remarkabl}'- 

 close similarity to Cordulegaster or not. That they do not, is 

 clear from the fact that the ovipositor of the females of Petalia 

 is quite unlike that of C or dideg aster, and is of the true jEschnine 

 form, and still suitable for the placing of ova in submerged 

 tissues, a power lost to C or duley aster, whose ovipositor is so 

 modified that the insect is unable to penetrate tissues, but can 

 only deposit its eggs in mud and shallow water by laboriously 

 supporting itself vertically with its wings, and dragging its 

 abdomen in the mud. Again, as regards wing venation, the 

 shortness of the wing-triangle, with its one cross-vein, is nearly 

 paralleled by Gomi^hceschna and Brachytron, admittedly ^^c^mne^ 



