120 ME. H. BOLTON ON A GIANT DRAGON-FLY [June I914, 



or other of which groups this fossil wing is likely to belong. Any 

 principal veins that are lacking must necessarily have occupied 

 the middle area from which the wing-structure is missing. The 

 anterior wing-fragment consists, evidently, of two veins, the costa 

 and sub-costa, the hinder fragment showing clearly the presence of 

 the cubital and anal veins. As the anterior and posterior wing- 

 margins are present, no loss of veins can have occurred, except in 

 the middle of the wing, and it therefore follows that the absent 

 veins are the radius and media. The proximal anterior wing-margin 

 consists, not of a free alar expansion, as is usual in the Protodonata, 

 but of a coriaceous and tuberculate mass which is well marked ofE 

 from the costa along its whole length. For some time I held the 

 view that this coriaceous bar represented the costa, and that there- 

 fore the anterior wing-margin was supported by a united costa and 

 sub-costa, and the next, vein formed the radius. This view I do 

 not now consider tenable, but regard the coriaceous bar as an 

 unusual and highly-modified alar area. That other veins existed 

 between the anterior and the posterior pairs is proved by traces of 

 vein-fragments in the middle area, but so broken up and frag- 

 mentary as to be almost unrecognizable. 



The traces consist of three short portions of a single vein lying 

 in the middle line. These vein-traces are too small, and too far 

 back, to have formed any part of the stem of the radius, but might 

 well have been portions of secondary branching, or even of the 

 media. Assuming that the radius and the media occupied the 

 middle position to which I have assigned them, and that the obscure 

 remnants of a vein are portions of one of these veins, the wing 

 resolves itself into a more intelligible structure. At the same time, 

 its likeness to forms belonging to the Protodonata becomes very 

 evident. This fact is well borne out by a direct comparison of the 

 wing-fragment with the corresponding portion of Meganeura 

 inonyi Brongn. Other Protodonate features also become clearly 

 evident : namely, the parallelism of the principal veins, the co- 

 existence of similar cross- veins, and the distinctive character of the 

 anal areas in the two wings. So closely does the proximal part 

 of the wing of M. monyi agree with the Radstock fragment, that 

 there can be little doubt of some close degree of relationship. It 

 is, therefore, on the assumption that it is a Meganeurid wing 

 that I proceed to describe the general character of the Radstock 

 specimen. 



Description. 



The extreme proximal portion of the anterior margin of the wing 

 is swollen out into an elongated tubercular mass, which thins out 

 distally, dying away imperceptibly along the anterior margin. 



It can be traced quite easily over a length of 20 mm. ; near the 

 articulation of the wing it is covered with numerous low, smooth- 

 topped tubercles irregularly arranged. The extreme free margin 

 shows up in one place as a flat, straight, knife-like edge, crossed 

 by a series of fine striations directed antero-posteriorly. This 



