proceeded so far that both branches cross the anal furrow 

 and end in the anal vein remote from the margin of the wing. 

 (See Fig. 6.) 



It should be noted that vein O/ 2 is rarely preserved in this 

 order, even in the more generalized forms. We have found 

 it only in the genus Pamphilius. In Macroxyela* the posi- 

 tion of the fork of the cubitus is indicated by a bend in this 

 vein. 



If the branches of the media be now examined, it will be 

 seen that vein J/ t (Fig. 6) extends longitudinally near the 

 centre of the distal part of the wing, its primitive course 

 being modified slightly if at all. Vein Af t follows a course 

 similar to the course of this vein in the Bombyliid ; so also 

 does the medial cross-vein (Fig. 6, m). A comparison of 

 the position of cells J/,, ist M^ and 2d J/ 2 in the Bombyliid 

 and in the typical hymenopterous wing (Fig. 7) is very in- 

 structive. 



Returning to Pamphilius (Fig. 6), we see that vein M 3 

 coalesces with the first anal vein, crossing the anal furrow 

 near the margin of the wing. It is evident that the forces 

 that are causing the branches of the cubitus to migrate along 

 the first anal vein and towards the base of the wing are ex- 

 erting a similar influence on this vein. It is also evident 

 that vein M t and Cu^ coalesce at the tip, and that the migra- 

 tion of the united tips of these veins (marked On in the figure) 

 towards the base of the wing has so modified the course of 

 that part of vein J/ 4 which is still free that this part of this 

 vein extends towards the base of the wing. This change is 

 very similar to the change in the course of vein O/ a in the 

 Empidid. 



A curious result of this change in the direction of the 

 course of vein J/ 4 is that the cell J/ 4 has been closed and 



* Comsttfck, loc. ctt., Fig. 735. 



