268 THE PANORPOID COMPLEX (Introduction), 



relationships of the Order Mecoptera to those other Orders which 

 stand in closest relationship to it. In accepting Handlirsch's 

 division of the Holometabola into four Sub-classes, it is to be 

 understood that I do so only provisionally, for the convenience 

 of sifting and allocating the evidence that this arrangement 

 naturally offers. And, as it is obvious that the main weight of 

 the argument must centre around those Orders which are clearly 

 most closely associated with the Mecoptera, so it will be clear 

 that the Orders comprising the Panorpoidea must be most fully 

 reviewed. There is, however, a great deal of evidence to be ob- 

 tained from a study of the Neuropteroidea; and this Sub-class 

 can by no means be denied full consideration in the argument. 

 With respect to the Coleoptera and the Hymenoptera, the evi- 

 dence to be obtained from them at the present time is compara- 

 tively scanty, and has little weight compared with that derived 

 from the other Orders, though it may help to throw some light 

 upon the larger problem of the origin of the Holometabola as a 

 whole. These ideas I have tried to convey in the sub-title 

 selected for this paper. 



If, then, we accept provisionally, for clarity of argument, the 

 two Sub classes Neuropteroidea and Panorpoidea, in the sense 

 that Handlirsch defined them (but with some modiHcations in 

 the nomenclature of the Orders composing them), we may then 

 proceed to define the " Panorpoid Complex" as that assemblage 

 of Orders ichose ancestral characters can be shown to possess close 

 ajfiniti/ with the characters preserved to a yreat measure itt the 

 Order Mecojdera, without in any way binding ourselves to the 

 inclusion of any particular Order in the Complex. The Complex 

 itself is an elastic assemblage of Orders; and any particular 

 Order may be included in it, or removed from it, according as 

 the weight of evidence may determine. 



Throughout this paper, I shall use the name Mecoptera for 

 the Panorpatte of Handlirsch, and Trichoptera for his Phrygan- 

 oidea. The Rapliidioidea I hold to be not sufficiently distinct 

 from the Megaloptera to deserve ordinal rank, since van der 

 Weele's work (7) indicates the strong probability of their origin 

 from the more ancient Sialoid stem. Hence I shall merge them 



