BY R. J. TILLYARD. 609 



narrower pterostigma {pt) of the hindwing, in the absence of 

 the di.stal forking of R 3a in the forewing, and in the fact that 

 M lb and M.y do not fuse to form a Y-vein in the forewing, as 

 they do in the hind. The wonderful preservation of the arche- 

 dietyon, and of the sockets of the macrotrichia in this fossil 

 lia^ already been emphasised (5, 25) ; but I think it is neces- 

 sary here to draw attention again to the fact that the distribu- 

 tion of the macrotrichia upon the veins has been made use of 

 in determining the courses of the main veins. It would not 

 have been possible to determine the correct course of M 4 with- 

 out these structures, since the basal piece of that vein, though 

 much more strongly formed than any cross-vein present, is bent 

 at right angles to M 3) and would undoubtedly have been con- 

 sidered a cross-vein, if it were not for the presence of large 

 macrotrichia upon it (Text-fig. 57) . Also, the inter-median 

 cross-vein (im), closing the median cell (mc) distally, is bent 

 obliquely, and might easily have been mistaken for a piece of a 

 main vein, were it not for its very weak formation, and the 

 absence of macrotrichia upon it. 



It is most unfortunate that we do not know what the basal 

 parts of the wings of this fossil were like. With our present 

 knowledge of Belmontia to guide us, it would seem to be ex- 

 tremely probable that the cubito-median Y-vein was present, 

 fully formed. In this respect my original restoration would 

 need to be corrected (5, PI. ix.) . 



In a comparison of this fossil with Belmontia, it should be 

 noted that, besides being much the larger insect, Archipanorpa 

 has the larger number of branches of Rs, viz. ten in the fore- 

 wihg, as against seven in Belmontia. Further, it is quite clear 

 that Archipanorpa is undergoing stenogenesis, seeing that M 4 

 and Cu la are already partially fused in both wings, and M lb 

 with Al 2a ' in the hind, not to mention the very obvious crowding 

 of the distal branches together. All the branches of Rs in this 

 fossil are clearly original archaic dichotomies, as they are also 

 in Belmontia; there can be no question of the addition of extra 

 brandies from the tip backwards, in either form. Thus we have 

 to accept Archipanorpa as less reduced and more archaic than 

 Belmontia, as regards the condition of its radial sector, in spite 

 of the fact that it is an Upper Triassic form, while Belmontia is 

 Upper Permian. 



The most crushed together of the branches of Rs in Archi- 



