DY H. J. T1LLYARD. 



701 



Coniopterygidae, which may well have been descended from 

 tonus which originally possessed it. The loss of M is also to 

 be noted in most families of the Order; but this vein has been 

 found by me in the fossil Archepsy chops, as well as in recent 

 Psychopsidae (Section iii.). 



Another archaic existing family is the Sisyridae (Text-fig. 

 110), in which clear evidences of reduction from the Prohemero- 

 biid type may still be found; though the specialisation of the 

 larva for feeding upon fresh-water sponge, and the heavy re- 



2 * 1A 



Cuj Cu lb Cu M M 



M 3 M 2 



lext-Fig.110. 

 Forewing of Sisyra brunnea Banks, (fam. Sisyridae) ,-(x 17). Letter- 

 ing as on p. 535. 



duction in size, and consequently in density of venation, under- 

 gone by the imago, have considerably simplified the general 

 scheme. The pectination of R2+3 is as clearly evident as is 

 the retention of the original dichotomy of K 4+5 . Further, the 

 condition of M, when compared with that of Cu, shows the 

 strong probability of the former having been originally four- 

 branched, though most families of the Planipennia have this 

 vein only two-branched. It will be clear that, quite apart from 

 the smaller terminal twigs of normal Planipennian type, the 

 series of forks marked x, y } ,:, and cuf' are certainly homotypic, 

 and must therefore either all represent additions, or all be part 

 of the original archaic venational scheme. Now the condition 

 of Cu, apart from the terminal twigs, is quite primitive, and 

 shows the three-branched condition quite clearly, as in the Mega- 

 loptera, Trichoptera and Lepidoptera. The point marked cuf' 

 is clearly the secondary cubital fork, at which Ciij divides into 



