yO SMITHSONIAN MISCELLANEOUS COLLECTIONS [VOL. 5° 



verse section, is circular or punctiform at the base (pi. x, figs. 31, 

 1 ) , and shaped like a narrow slit farther up. It appears to terminate 

 some distance below the apex. I have never been able to see an ori- 

 fice at the end of a hook. This hook-cavity is, as a rule, in open con- 

 nection with the cavity of the costal vein, but this communication 

 may be interrupted by soft matrix or chitinous layers secondarily 

 formed. Sometimes the walls of the costal vein itself are thickened 

 to such an extent that its lumen is obliterated and the vein itself con- 

 verted into a solid rod, from which it is to be inferred that these 

 cavities generally are not of any importance. The costal vein is 

 sometimes considerably broadened in the frenal region, sometimes 

 throughout the whole hook region (pi. vn, fig. 6) ; sometimes it is 

 dilated at the point of insertion of each hook (pi. vii, fig. 4). In most 

 cases, however, it shows no distinct differentiation in the hook region 

 except a slight incurving at the base of the hooks. The distal hooks 

 are inserted in the upper surface of its wall, either in the middle of 

 its broad face or nearer to the wing margin (pi. x, figs. 27-29). 

 It is interesting that the Tenthredinids and Anthophilids, which are 

 systematically very far removed from each other, also show the 

 greatest differences in the mode of attachment of the hooks, while 

 in the families systematically intermediate transitional forms of 

 hook attachment are seen. The difference in the mode of insertion 

 of the hooks appear to be correlated to the differences in the for- 

 mation of the costal vein, and not solely dependent upon differences 

 of habit. For the investigation of these conditions in the Ten- 

 thredinids I had at my disposal, besides Tenthrcdo mesomelcena 

 L., abundant and excellent specimens of Sir ex gigas L., an ideal 

 material for this work on account of its large size. In both 

 species I found the hooks quite similarly inserted. I will therefore 

 describe them together. Sire.v gigas is, like the other Tenthredin- 

 ids, remarkable for the thinness of the wing-lamellae and the spa- 

 cious cavities in its wing veins. Its hooks are very much flattened, 

 distinctly ribbon-shaped. The attachment of the hooks involves con- 

 siderable modifications of the wall of the costal vein at the points of 

 insertion. The hooks are inserted in circular openings of the vein 

 wall, the margins of which are raised up to form elevated rings, 

 both outside and in. The outer elevation is very slight (pi. x, figs. 

 27, 28, 29, R), while the inner one attains considerable dimen- 

 sions. These elevated rings are not cylindrical, but form semi- 

 globular pans (pi. x, figs. 27, 28, 29, P). In sections through the 

 wings of fully developed specimens caught on the wing the cavity of 

 the pan is not in connection with the lumen of the vein. In young, 



