Vol. 70.] EROM THE RADSTOCK COAL MEASURES. 121 



tubercular structure I regard as the equivalent of the free alar 

 expansion seen in front of the costal vein in JHeganeura monyi. 

 It forms a rigid bar, enormously strengthening the anterior wing- 

 margin, and rising proximally into a prominent shoulder. The 

 costa is sharply marked off from the hinder margin of the tuber- 

 culated alar mass, and passes directly outwards as a broad strap- 

 shaped vein, broadest at its point of origin, and very slowly 

 diminishing in width as it passes outwards towards the tip of the 

 wing. It must have formed the greater part of the frontal margin 

 of the wing. At its articular origin the costa seems united to 

 the next vein (sub-costa) by a common root, but examination shows 

 that this apparent union is caused by a slight backward movement 

 of the proximal end of the costa over that of the sub-costa, and that 

 the two may have been quite free one from the other. Some 

 distance bevond the termination of the alar mass the free marsrin of 

 the costa gives origin at regular intervals to a series of low spines, 

 projecting freely from the front edge of the wing, and gradually 

 inclining outwards towards the wing -apex. A few scattered 

 tubercles and a faint ridge occur along the middle line of the vein 

 and eventually die out, leaving the distal portion of the costa 

 smooth — except for the marginal spines. The hinder edge of the 

 costa is straight, and interrupted along its course by a series of 

 parallel branches, which pass directly backwards and join the 

 sub-costa. Of these lateral branches eleven are still whole and 

 clearly marked, while portions of nine others can be distinguished. 

 It would, therefore, appear that the two veins were joined along 

 their whole length by a parallel system of straight cross-branches 

 with no intervening network. Several of these cross-branches 

 seem to be jointed at their anterior third. 



The sub-costa arises in close contact with the costa, diverging 

 rapidly until the two are about 56 mm. apart. Once the maximum 

 divergence has been attained, the sub-costa remains fairly parallel to 

 the costal margin, although there are indications that the two come 

 together some distance bevond the ed°;e of the winer-fraerment. 



This is mainly brought about by the long flattened convexity of 

 the costal margin. The costa and sub-costa are nowhere more than 

 5 to 6 mm. apart, in the main but 4 mm., the interval narrowing 

 down to 3 mm. at the broken distal edge of the wing. While the 

 sub-costa possesses the flat strap-shaped character of the costa, it is, 

 on the whole, a more delicate and slightly narrower vein. A few 

 tubercles are disposed in an irregular median line along its length. 



Nowhere along the hinder margin of the sub-costa can be traced 

 any portion of the general integument of the wing ; the only indi- 

 cation of a backward continuation is seen in the ragged edge of the 

 proximal portion of the vein, and in traces of a few branch veins 

 passing directly backwards. 



Radius and media. — I have already given reasons for supposing 

 that the intermediate venation of the wing is missing, and that in 

 all probability this venation consisted of the radius and the media 

 with all their branchings. 



k2 



