BY R. J. TILLYARD. 415 



the Caddis-flies, Alder-Hies, Snake-Hies, and Lacewings; in other 

 words, his single Order comprised the insects which we now dis- 

 tribute between the three Orders Trichoptera, Megaloptera, and 

 Planipennia. His Order was divided into four groups, Phry- 

 ganeina, Myrmeleonina, Hemerobiina, and Corydalina. The 

 latter was again divided into four families, according to the 

 following key (7, p. cci.): — 

 Ocelli S :— 



Body clothed with hair Itkonesidce Newm. 



Bod3 T naked : — 



Antenna? simple Gorydalidce Newm. 



Antennae pectinated Ghaiiliodesidn i Newm. 



Ocelli SicUicke Leach. 



Thus it will be seen that the Ithonidce were associated with 

 the Megaloptera, instead of with the Planipennia, though it is 

 now generally agreed that they belong to the latter Order. 



Inadequate as we now know Newman's classification to have 

 been, yet there is no denying the fact that the Ithonidce are, 

 amongst all Planipennia, the most closely related to the Megalo- 

 ptera. In my previous paper, I gave reasons why it appeared 

 very likely that the larva of It hone might be aquatic (9, p. "28 1 ). If 

 such had proved to be the case, and if, as might further have been 

 expected, the larval type of the Ithonidce had shown close affinity 

 with that of the Megaloptera, and its pupa had proved to lie free 

 in the earth, without forming any cocoon, then the separation of 

 the two Orders Megaloptera and Planipennia could no longer 

 have been maintained. There are, indeed, many present-day 

 entomologists who refuse to recognise these two Orders, on the 

 ground that the imagines cannot be separated by any characters 

 of sufficient importance. Such authors, of whom we may take 

 Comstockd) as an example, treat these two groups as Suborders 

 within a single Order Neuroptera. Their position is quite 

 logical. The Planipennia, however, are so distinct from all other 

 Neuropteroid insects in possessing their unique suctorial larva, 

 in which the mandibles and maxillae are developed together to 

 form a pair of sucking-jaws, that it is, from the point of view of 

 the Phylogenist, unwise to merge them with any other group; and 



