200 FOSSIL INSECT WING, ORDER PARAMECOPTERA, 



of Ciij itself. The true Cu„ arises, as always, close to the base of 

 the wing, but is either a very weakly formed vein (as in Text- 

 fig. 6) or entirely lost 



Thus we see that, although the Diptera may well be directly 

 derived from the Paramecoptera, as regards their arculus- 

 formation and most other points in their venation, yet we have 

 to explain how it is that they have Ctij unbranched, as in the 

 Mecoptera. Either they originally had this vein distally forked, 

 which does not seem likely; or they are descended, not from the 

 Paramecoptera, but from one of the Orders in which Cu, is 

 simple, viz., the Mecoptera or the Paratrichoptera. 



We are hound, then, to conclude that the descent of the 

 Diptera from the Paramecoptera cannot be proved with certainty. 

 That there is a close affinity of some kind between the two Orders 

 appears certain: but the probability is that it is not a directly 

 ancestral one. 



(v.) Affinities with the Megaloptera and Planipennia. 



A remarkable feature in the wing of Belmontia is the position 

 of the fork of R4+5, quite close up to the first forking of Rs into 

 R2+3 and R 4+ 5 respectively. Now Comstock (1, pp.95, 147) has 

 shown very convincingly how the pectinate type of branching 

 found in the Megaloptera and Planipennia can be simply derived 

 from the older dichotomic type, by the recession of the origin of 

 R 4 from its original position on R4+5 across to R2+3 The con- 

 dition shown in Belmontia is intermediate between what are 

 usually regarded as the typical dichotomic and pectinate types 

 of branching; i.e., though definitely dichotomic, the origin of E 4 

 is so close to R2+3 that a very little change would produce the 

 pectinate type. Thus we see that, on this character, Belmontia 

 could well be the ancestor of both the Megaloptera and the 

 Planipennia. 



But, in order to establish this suggestion on a much firmer 

 basis, we must ask for evidence of the presence of the arculus in 

 the two Orders in question. For, if they do not possess this 

 distinctive venational structure, they cannot certainly be de- 

 scended from the Paramecoptera. Now, in a previous paper(7), 



