BY R. J. TILLVARD. 



433 



since the fossil evidently lacks the essential discoidal cell, not to 

 mention a nodus, which is almost certainly absent also. We are 

 therefore driven back to consider the claims of the Protodonata. 

 The strongly built, probably coriaceous, border is already 

 known in the Carboniferous genus Meganeura. The new fossil 

 also resembles this genus in the manner in which M appears 

 basally as a very weak vein diverging gently from the strongly 

 built R. The Protodonata, like the Odonata, are notable for 

 the unbranched condition of Ks. Now, in Jleganeiira, there are, 

 in the forewing, f^'o radial sectors arising fairly close together, 

 and running closely parallel for a considerable distance. If, 

 owing to the narrowing of the wing, in the course of e\'olution, 



Text-tig. U). 

 Diagrams to show the structure of tiie radius and media in Protodonata: 

 a, suggested ancestral condition, from which there can be derived: — h, 

 the condition found in Ai'.roplana mirahilis, n.g. et sp., bj' fusion of 

 the two radial sectors (Rsj, Rso) near their bases; c, the condition 

 found in the forewing of Meijaneura monyl Brongn., with radial 

 sectors still separate, but anterior branch of media (M,) eliminated; 

 and d, the condition in the hindwing of the same species, witli the 

 ladial sectors eliminated but the media not reduced. 



these two sectors were to fuse together, we might expect to get 

 some such formation as we find in this new fossil, viz., a single 

 Rs with two origins. In the hindwing of Meyaneura, however, 

 we find that Rs is entirely absent. But there is a pecidiar 

 anterior branch of M, which takes its place, and which branches 

 dichotomously exactly in the same way that the corresponding- 

 branch of M in the hindwing of the new fossil does. All 



32 



