228 NEW YORK STATE MUSEUM 



illustrate in their different species the tendency of AP and M- 

 to fuse together. Limnophila toxoneura [pi. i8, fig. 2] 

 shows a moderate fusion; L. brevifurca [pi. 14, fig. 6] 

 shows fusion almost to the wing margin ; and L. q u a d r a t a 

 [pi. 18, fig. 6] shows the fusion complete; and the other 

 species of Limnophila figured on plate 18 show various inter- 

 mediate conditions. Rhaphidolabis [pi. 19, fig. 2] shows these 

 veins almost fused, and the nearly allied Plectromyia [pi. 30, 

 fig. 4] shows them wholly fused. 



In the hinder fork, I have seen no evidence of any tendency 

 for M^ and M^ to fuse, but both tend to atrophy. Both, though 

 commonly developed in the more generalized neuropteroid in- 

 sects, have been found well developed among all the Diptera 

 only in the fossil Rhabdinobrochus above cited. One of the 

 hinder branches of media is quite persistent. In fact it is 

 nearly always present, but it may be found well developed or 

 weak or broken or absent in different species of the genus 

 Dicranomyia. It is usually fused basally with vein Cu^ for a 

 distance, but I have never found it fused beyond the level of 

 the median cross vein, and I do not believe that it ever dis- 

 appears by total fusion in the Tipulidae, although it does so 

 among certain of the higher Diptera. Dicranomyia 

 immodesta [pi. 27, fig. 3] shows it persistent while the 

 median cross vein has disappeared, D. c i n e r e a [pi. 27, fig. 4] 

 shows it interrupted, and attached to the end of the flexed 

 'median cross vein, the two together simulating very decep- 

 tively a persistence of vein M-. This appearance is wholly de- 

 ceptive, however, as any one may satisfy himself by a little care- 

 ful comparative study. In ? D. whartoni [pi. 27, fig. 5] it 

 has wholly disappeared, save for the slightest bit of a spur on 



^ Critical comparative study is sometimes necessary for determining where 

 vein R ends and vein M begins. In Paratropeza [pi. 21, fig. 4], for 

 example, there are five branches of veins reaching the wing margin between 

 the tips of Ri and Cui. Is it Rs or M that is three branched? How shall 

 this be determined.? Only by comparison of allied forms. It will thus be- 

 come sufficiently clear that Rs is three branched. The other interpretation 

 would be inconsistent; for (i) the first fork of the vein Rs is always m- 

 volved in the formation- of the cord ; vein R4+.-, is here set ofif posteriorly at 

 a right angle from the base of vein R^+a; and (2) in allied forms the cross- 

 vein M joins at its anterior end vein Mi-f-2, not vein M^; that is, it is 

 situated on the proximal side, not on the distal side of the upper median 

 fork. That fork is therefore not present in Paratropeza. This ap- 

 parent confusion is due to the elimination of the r-m cross vein by the long 

 basal fusion of veins Rj^b and Mi+a. 



