SY R. J. TILr.YART). 625 



But we may not argue from this that the AgrioidiKv are de- 

 scended from the Protoneurino,, since such a supposition is nega- 

 tived on all other morphological grounds. Tlie wing-venation 

 of the Prototieuriiup. is reduced far beyond the Agrionine stage. 

 It is quite possible tliat the larva? of some archaic Agrionine 

 genus may still retain gills of constricted form. We can only 

 say that such a form must liave existed once within the Agrio- 

 niiKE, even if no longer existing to-day. The phylogenetic series 

 of gill-types (Constricted 8accus, Constricted Lamella, Nc^late 

 Lamella, iSubnodate Lamella, Denodate Lamella) is strictly 

 correct. But it does not follow that the series of generic types 

 which to-day exhibit these gill-types (e.g., jS'eosticta, Isosticta, 

 Caliagrioii^ Ischmira A giocnemis) is a true phylogenetic 

 sequence. The sequence Neosticta, Isosticta must be admitted. 

 Then there is a break: for, as far as we know, the larva* of the 



a b c. 



Text-ti^.-tU. 

 Pjvolution of the Aj^iiunid form of the Vertical Lciinelltir T^qie direct from 

 the Constricted Saccoid Type; a, Saccoid Type; //, <:, iiiteriuediate 

 staj^es; rl, Agiioiiid form of Vertical Lamellar Type. 



Protonenr'uKH have reached no higher stage than that shown by 

 Isosticta. There is also a break on the side of the A g rioiunce: 

 for, as far as we know, no larva of this subfamily possessing 

 vertical lamellar gills shows any older form than the Nodate of 

 (Jaliagrion. Thus the constricted stages still present in the 

 laivaj of Protonenrince have been passed and lofil by the more 

 highly specialised larvae of the Agrionitice: so that it is pure 

 chance that Caliagrioii picks up the sequence of forms at the 

 stage next beyond that to which Isosticta has attained There 



