252 FOSSIL INSECT WING, OKDER PARAMKCOPTEKA, 



is due to the crowding out of this vein, during the rapid pro- 

 liferation of the branches of Rs that took place during the first 

 formation of the Prohemerobiid type of wing-venation. 



The family of Planipennia whose venation most closely re- 

 sembles that of Behnontia is certainly the Sisyridce, already 

 recognised as one of the most archaic types within the Order. 

 The pupal tracheation of this family has not yet been studied. 

 Until this can be done, we cannot say definitely that the Plani- 

 pennia are descended from the Paramecoptera; but enough will 

 have already been said to show that there is quite a fair pro- 

 bability of this being so. 



With respect to the Megaloptera, the forewings of this Order 

 either have Cu, actually fused with M for a short distance, or 

 the two are connected by what is apparently a cross-vein in the 

 position of the areulus. Both conditions can easily be derived 

 from the more archaic state seen in Behnontia, by means of a 

 gradual shortening of the areulus, leading to its final elimination, 

 and the consequent fusion mentioned. We should be careful to 

 remember that, within the single family Rhyacophilidce of the 

 Order Trichoptera, all stages may be found from the primitive 

 arrangement, closely resembling that of Belmonlia, through forms 

 in which pa becomes a short, transverse vein, to others in which 

 Oil! is actually fused with M. Within this family, the areulus 

 in the genera Glossosoma, Mystrophora, Agapetns, and Catagapetus 

 resembles that of Behnontia even more closely than it does in 

 Rhyacophila dorsalis (Text-fig.4). If, then, all stages from the 

 original type to the most complete reduction can be seen within 

 the limits of a single archaic family, how much more easily may 

 the same series of evolutionary stages be accomplished during 

 the passage from one Order to another? 



Thus, though we cannot prove certainly that the Planipennia 

 and Megaloptera are descended from the Paramecoptera, it will 

 be seen that there is quite a fair probability of this being the 

 case. If, when the Belmont Beds have been further worked, 

 no wings of the types associated definitely with these two Orders 

 can be found, then it will become more than ever probable that 

 they did not evolve until later on in geological time; and the 



