NO. 1438. WINGS OF THE TENTHREDISOIDEA—MAcaJLLIVRAY. (545 



this vein in the Lydidje is wanting, by the lack of coalescence between 

 cubitus and Sc+R+M, which extends for ahuost one-half the length 

 of the vein Sc+R+M in the Lj^didtv, and l\v the diti'erence in the 

 shape of the cell M^. The facts just given show conclusively that the 

 Megalodontidw are more closely allied to the Siricid^e than to any 

 other family, while there are an abundance of characters for retaining 

 them as a distinct family. 



Cephidse. — A family of about a dozen genera and moderately numer- 

 ous in species of intercontinental distribution. It is not closely related 

 to any of the families described hitherto. The front wings (tigs. %'?>- 

 96) are long and narrow, the hind wings have been remarkably reduced 

 in the anal region, and the insects are very rapid fliers. The area 

 between costa and Sc+R+M has been eliminated in both wings. 

 Media separates from radius distant from the stigma, in about the same 

 region as in the Xyelidte. The cell R is long, subequal in length to 

 the cell Rj, and is in line with the cells 1st Rj and 2d Rj + R.^, as in the 

 Siricidse and Megalodontida?, but ditfers from these families in that 

 the apex of the cell R is almost opposite the apex of the cell 1st Rj. 

 The radio-medial cross-vein is transverse. The portion of the radial 

 sector between this cross-vein and the stigma is longitudinal, while 

 the portion of the sector between the radio-medial cross-vein and the 

 radial cross-vein is subtransverse, and might readily be mistaken for 

 a part of the radio-medial cross-vein if it were not for the angle at its 

 anterior end. The radial cross-vein is oblique and appears to be the 

 base of R.,. The cell 2d Rj + R., is very long, as in the Siricidte, but is 

 pointed at tip. The anterior end of the medio-cubital cross-vein is 

 joined to media in certain species distinctly beyond its origin^ while in 

 others it arises from the angle between R and M. The cell M^ is as long 

 as or longer than the cell R. The cubitus coalesces with Sc+R+M 

 for only a very short distance at base. The cells 1st A and 2d A are sub- 

 equal in width, and the free part of the second anal vein is situated just 

 beyond the middle of cubitus and is perpendicular to the first and third 

 anal veins, while the contraction of the third anal vein is wanting. 

 The hind wings are just as distinctive as the front wings. With the 

 exception of the subcostal area, costa being coalesced with R+M, the 

 number of veins is the same as in the Xiphydriida?. The most dis- 

 tinctive character is in the arrangement of the cells. Beginning at 

 the base of the wing (tig. 95) the cells 1st A, Mj+Cu+Cui, M,+lst 

 M,,, R+Rj+R^, and Rg are arranged in an oblique row from the base 

 to the apex of the wing. 



Practically all systematists are agreed in considering the Cephidse 

 as a distinct group worthy of family rank. So far as the wings are 

 concerned, they are the most distinctive of any group of the Tenth re- 

 dinoidea, and are only indirectly related to any of the other families. 

 They are generalized, so far as the origin of media is concerned, but 



