242 STUDIES IN AUSTRALIAN NEUROPTERA, iii., 



(10) Differences in the form of the three intra-cubital cells 

 (icitj, ic?t.,, iciCg) in the forewing. 



(11) Differences in the anal region: in particular, the curvature 

 of 3A, and its relationship with the wing-boider. 



It is possible that a careful study of these points might reveal 

 a satisfactory method of subdividing the genus. No characters 

 which are correlated merely with size (e.g., greater or less num- 

 ber of branches from Rs) can legitimately be used in this con- 

 nection, since species of different sizes are often most closely 

 related. 



Phytogeny of the Chrysopidcp,. 



The results obtained in this paper appear to me to be suffi- 

 ciently striking and suggestive to enable us to understand the 

 phylogeny of the Chrysopidce along fairly broad lines, if not in 

 every minute detail. We are still unable to make the full com- 

 parisons between the scheme of venation in Chrysopa aiid that 

 of allied groups, because we have not yet obtained the scheme of 

 the pupal tracheation in those groups. For example, it would 

 be very instructive to compare theChrysopid venation with that 

 of the Apochjysidfe. But the pnpal tracheation of the latter is 

 not known, and indeed is never likely to be discovered, since no 

 larva of this family has yet been found. All we can do is to 

 compare the imaginal venations, using them with the greatest 

 care, and applying stringent tests to our conclusions. The same 

 holds good if we desire to compare the venation of Chrysopa 

 itself with that of other genera of Chrysopidce. The nearer 

 the relationship, the closer will be the comparison: but we have 

 to remember that, so far, no other genus except Chrysojm has 

 had its scheme of tracheation worked out, so that this final court 

 of appeal is denied to us. 



It seems safe to lay down the following rule to enable us to 

 distinguish at sight a true media from a pseudo-media, in what- 

 ever wing it might occur: — The apparent median vein, is a true 

 viedia if the branches of Rs nearest the base of the wiiiy do not 

 descend directly on to it, but run long iludu tally above and more 

 or less parallel to it. If, hotvever, one or more of these branches 



