692 THE PANORPOID COMPLEX, Ui.j 



The Sialoidea are universally admitted to be more archaic than 

 the Raphidioidea. It is also generally agreed that the faindy 

 Cerydalidae (Text-figs. 105, 10(i) is more archaic than the family 

 Sialidae (Text-fig. 104), these being the only two families repre- 

 sented within the Suborder. It is in the comparison of the 

 venational types found in these two families that we meet with 

 our principal difficulties in the construction of the Archetype 

 of the Order. • 



The outstanding feature of the venation of the two Neurop- 

 teroid Orders, Megaloptera and Planipennia, as contrasted with 

 that of the Panorpoidea, is the generally strongly developed ten- 

 dency towards the pectination of Rs. This character is beauti- 

 fully illustrated in almost all Planipennia, but is by no means 

 so evident in the older Megaloptera. We have here to study 

 this peculiarity, and to decide what the archetypic condition was, 

 from which the pectinate type of Rs was originally evolved. 



A little consideration will show us that the tendency towards 

 pectination in Rs probably did not, at first, affect R 4+5 ,* but 

 only R2+3. For, in all the Sialidae, R4 +5 remains still dieh- 

 otomously branched, as shown in Text-figs. 47a, 104; while, in 

 the Corydalidae, where the pectinate form of branching of Rs is 

 generally much more in evidence, there are many examples of 

 the retention of the original dichotomic forking of R' 4+5 > as, for 

 instance, in Protohermes (Text-fig. 106) . Further, in those 

 genera in which pectination has become more complete, through 

 the loss of the original fork of R4 +5 (as in Archicliauliodes, 

 Text-fig. 105, Chauliodes, Corydalus and allies), specimens are 

 not infrequently found with this fork retained in one or more 

 wings. Thus we must conclude that, in the Archetype of this 

 Order, pectination of Rs, if present, had only affected R 2+3 . 



Our next difficulty lies in the presence of some forms in which 

 no pectination of Rs is visible at all, and from which we might 

 be led to the conclusion that the Archetype of the Order had a 

 dichotomic Rs. I refer in particular to the Australian genera 

 Austrosialis and Stenosialis, which are reduced archaic offshoots 

 from the family Sialidae. In Text-fig. 47, I compared the vena- 

 tion of Austrosialis with that of the Trichopteron Bhyacophila, 



*See, however, footnote on p. 552, and Text-fig. 39. The vein taken to 

 be R4+5 may possibly have been an originally branched R.5. 



