ZooL.— Vol.. III.] KELLOGG— NET-WINGED MIDGES. 20I 



It seems unnecessary in the face of the descriptions and 

 figures of the new species here recorded, and of the con- 

 trasting characters as indicated in the foregoing analytical 

 tables, to argue for the soundness of this suggested revision, 

 based on venation; or rather for the unsoundness of the 

 old classification based on continuity and bisection of the 

 eyes. It may indeed be necessary to revise again the 

 generic classification when another half dozen species are 

 found, but this further revision cannot possibly be based on 

 eye characters; it will probably be based on venational 

 characters. It is not a long step from a slightly free radius2 

 (=anterior branch of second vein) as shown by Bibiocephala 

 doanei and B . elegantiihis to one wholly fused with radiusi 

 or radiuss (with the first or second vein), as is the case in 

 Blefharocera and Pktlorus. Indeed so obviously is this 

 difference simply a matter of degree of serial modification 

 or specialization that it is a question whether the species of 

 Philorus and Bibioce^hala should not be looked on as 

 members of a single genus, distinguished from Blefharocera 

 by the possession of a medio-cubital cross-vein (cross-vein 

 connecting veins four and five). 



The series formed by Bibiocefhala grandts, B . comstocki, 

 B. elegantuhis, B. doanei, and Philorus bilobata, with regard 

 to the addition of the radial vein as shown in text-fig. i, is 

 very suggestive. It is certainly a pretty demonstration of 

 what that formerly mysterious little triangle, which was 

 found and puzzled over in Bibioce^hala doanei (see the 

 original description of doanei as a Liponetira in Psyche, 

 Vol. IX, pp. 39-41) really is. 



This serial modification (by increasing fusion of one 

 radial branch with another) has been elsewhere prettily 

 shown by Comstock (Manual, pp. 442-443) after Winnertz, 

 in the case of certain Mycetophilid species, in which a series 

 almost exactly similar to the one just shown above is 

 described, except that the fusing vein is not radiuso but is a 

 vein formed by the already complete fusion of radius2 and 

 radiuss. 



