io6 



pressed back to the centre of the wing (Fig. 7, M 4 ), and now 

 lies in front of the free portion of the vein J/ 4 instead of be- 

 hind it.* 



Let us now consider the courses of the branches of the 

 radius. Here again we can gain help from a study of dip- 

 terous wings. Observe in the Bombyliid (Pantarbes) the 

 coalescence of the tips of veins 7? 5 and M^. In the Hymen- 

 optera a similar coalescence of veins R^ and M^ has occurred ; 

 but it has proceeded much farther, so that the free portion 

 of vein R^ in Pamphilius (Fig. 6, R^ is remote from the end. 

 of the wing and has the appearance of a cross-vein. 



In the Hymenoptera vein R,^ has been followed in its 

 migration along vein M 1 by vein J? 4 , which has now reached 

 a stage in Pamphilius that is quite similar to that reached by 

 vein R^ in Pantarbes. But like vein R b it has the appearance 

 of a cross-vein. 



From this it will be seen that the vein marked M t in Fig- 

 ure 6 is really compound, as it includes the tips of veins 

 R^ and Jt t . 



In PainpJiilius vein R l is curved away from the costal 

 margin of the wing to make room for a stigma (Fig. 7, S), 

 and vein R^ ends in the costal margin a short distance before 

 the apex of the wing (Fig. 6). Vein R^ has been lost in this 

 genus, but is well preserved in certain closely allied forms.f 

 and is, therefore, represented in the figure. 



While the tips of the branches of the radial sector have 

 migrated away from the apex of the wing, the bases of 

 these branches coalesce in the opposite direction ; from these 



* At the time that the figures in Comstock's ^lanual were prepared it was 

 believed that the media was typically three-branched. For that reason the vein 

 which we now regard as vein /l/ 4 was believed to be a cross-vein. The inter- 

 pretation given above accords better with what we have since learned to be the 

 typical form of the media. 



t See page 103, footnote. 



