644 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. xxii. 



tures. The radial cross-vein is intermediate in its location between 

 Slrex and Tremex^ bein^ almost interstitial with the free part of R., 

 which is fully preserved; the transverse part of the base of media 

 has begun to disappear and the radio-medial cross-vein to function for 

 it; the apices of the veins forming the appendiculate cell of the 

 front wing have faded out, so that there is found exactly the same 

 condition in both wings that exists in Trem.ex ; the free part of the 

 second anal vein is preserved and the anal area of the wing is large. 

 All these characters ally this genus more closely with a form like Sirex 

 (iJhlcornls than with Tremex^ near which it has been placed. 



The Siricidse are a well-circumscribed group, defined by the condi- 

 tion of the area between costa and Sc+R+M, by the relation of media 

 and the medio-cubital cross- vein, by the position of the radio-medial 

 cross-vein, and by the prominent appendiculate cell at the apex of the 

 front wing. 



Megalodontidx. — This family (tig. 92) contains four genera and 

 about twentv-tive species, and is peculiar to Europe, Asia, and north- 

 ern Africa. It represents a line of specialization very similar to that 

 found in the Siricida?. This is shown by the condition of the base of 

 media and the medio-cubital cross-vein. The cells R, 1st Rj, and 2d 

 Ri + Rg are arranged in a row. The base of the radial sector is trans- 

 verse, as in the Lydid*, but that portion of the sector between the 

 anterior end of the radio-medial cross-vein and the posterior end of 

 the radial cross-vein is not so strongly curved, so that the cell 1st R^ is 

 almost a parallelogram. The Megalodontidte differ from the Siricidfe 

 in that the cells R and 1st R^ are subequal in size, the cell M^ is propor- 

 tionately much smaller, the cell 2d R1+R2 is of about the same length, 

 but is much broader, while the apex of the cell is like that of the Xye- 

 lidte and Lydida?, and is not appendiculate. The anal cells are like those 

 of the typical wing. The area between costa and Sc-j^R+M is broad, 

 and all trace of the subcosta is wanting. The vein forming the apex of 

 the cell M is a straight one; each of the components of this vein is not 

 separately bowed on the basal side with an emargination between them, 

 as is so characteristic of this vein in the Siricida?, while it is more 

 strongl}^ inclined than in this latter family. The hind wings are just 

 like those of the Lydidaj, except that the subcosta is wanting. The 

 generalized condition of the apex of the cell Ri+21 obliquely truncated 

 and not pointed, is notable. 



The forms considered here as constituting the family Megalodontidfe 

 are invariably classified by systematists as a subfamily of the Lydidte. 

 That they do not have any affinities with this family is shown b}^ the 

 difference in the position of the medio-cubital cj"oss-vein, by the dif- 

 ference in the position of the cells R and 1st Rj, by the entire loss of 

 the subcosta in both wings, by the loss of the free part of Cua, by the 

 straightening of the cubital vein, so that the curve so characteristic of 



