282 THE PANORPOID COMPLEX (Introduction), 



have been the case. Bearing in mind the complex nature of the 

 win'^-venation in the most archaic types of insects, we are bound 

 Xfo conclude that, at the beginning of the evolution of the wing, 

 there must have been a period of great enlargement, with a rapid 

 and abundant production of veins in all parts (following upon 

 the rapid increase in the tracheation of the developing organ). 

 After this had reached its maximum, there began a period of 

 arrangement and reduction, during which the very beautiful and 

 perfect wing-types at present existing were evolved. We may 

 term the process by which any part of a wing becomes broadened, 

 with consequent production of further venation in that area, 

 Platygenesis (Gr. TrAan's, broad); while the opposite process, in 

 which the wing-area in question becomes reduced or narrowed, 

 with consequent reduction or elimination of some of its existing 

 venation, may be termed Stenogenesis (Gr. (rTev6<;, narrow). The 

 development of the costal space in the Psj/chopsidce, of the 

 enlarcred anal area of the hindwing in Anisopterid Odonata, and 

 of the wide anal fan in the Orthoptera, Perlaria, and other 

 insects, are good examples of Platygenesis; while the process of 

 Stenogenesis can l)e followed out very fully in such an evolu- 

 tionary series as the Diptera Nemocera. 



Scheme of the Work. 



We have already stated that phylogenetic conclusions, to be 

 acceptable, must be based upon an examination of as many char- 

 acters as possible. In undertaking an analysis of a large number 

 of characters, two methods of procedure are possible. We may 

 either select each Order in turn and study its characteristics 

 fullv, with a view to the determination of its Archetype: or we 

 may select any given set of characters, as, for instance, those 

 offered by the Wing-venation, and study them as they are ex- 

 hibited throughout the whole of the Orders under discussion; 

 proceeding to deal with other sets of characters in subsequent 

 parts of the work. 



This second method must obviously be the one chosen, for 

 only by it can the necessary comparisons be made between the 

 same sets of characters in different Orders. Thus we have to 



