254 FOSSIL INSECT AVING, ORDER PARAMECOPTERA, 



Beds, we see the obvious differences of the more complete system 

 of cross-veins in the latter (these were, however, very weakly 

 chitinised), the more symmetrically arched and regularly 

 arranged dichotomies of the veins, the absence of the arculus (as 

 in almost all Mecoptera, owing to the partial fusion of Cuj with 

 the main stem of M), and the unbranched condition of the very 

 typically Mecopterous Cu 2 . M 2 _ 4 is six-branched in theforewing 

 of Permochorista, but only five-branched in Belmontia, as also in 

 the existing genus Tcmiochorista, which stands very close to 

 Permochorista in its general scheme of venation. On the other 

 hand, lis is only four-branched in Permochorista, while in 

 Belmontia it is seven-branched 



Thus it is quite clear, apart from the question of time, that 

 neither of the two Orders Mecoptera and Paramecoptera can be 

 derived from the other. For the Mecoptera are the older in 

 possessing the more regular dichotomies, the more regular system 

 of cross-veins, the more numerous veinlets in the costal space, 

 and the larger number of branches of M; whereas the latter have 

 the more archaic type of arculus, the forked Cu,, and the larger 

 number of branches of Rs. We must, therefore, postulate, as 

 common ancestor to both Orders, an even more archaic type of 

 Holometabolon, which possessed the archaic features of both. 



Turning next to the Protomecoptera, we must compare Bel- 

 montia with Archipanorpa (3, p. 188, Pis. viii.-ix.), the only known 

 genus of that Order, from the Upper Trias of Ipswich. The 

 system of cross-veins in this genus is intermediate in complete- 

 ness between that of the Mecoptera and of Belmontia, and the 

 same may be said to be true of the regularity and symmetry of 

 the dichotomous branchings of its veins. Archipanorpa is de- 

 finitely more archaic than Belmontia in possessing a well-devel- 

 oped costal vein, distinct from the costal border of the wing, and 

 in having a larger number of branches of Rs and M. It pos- 

 sesses a forked Cu, but the anterior branch Cu lB is partially fused 

 with M 4 ; in this, it is clearly more specialised than Belmontia. 

 The region of the arculus is not preserved in either wing of 

 Archipanorpa; and hence we cannot complete the comparison. 



Thus it is clear again that, apart from the question of time, 



