THE WINGS OF NEUROPTERA 



151 



tus (Fig. 143) that I collected in Florida. In this wing the base of vein 

 R2+3+4 has been switched along a cross-vein to vein Ri, so that there 

 appear to be two radial sectors. 



It was the existence of this abnormal wing that suggested the probability 

 that in wings like those of the MyiTneleonidce, where there arc many cross- 

 veins, the suppression of the dichotomy had been attained by the switching 

 of the base of vein R4 to vein R.2+3 along a cross- vein; and it was thought 

 that an examination of a larger series of wings would reveal the existence of 

 examples of .the successive stages of this switching. 



Ver\' many examples were found of wings in which the base of vein R4 

 appears to be at a halfway station in its journey along a cross-vein from 



r ca 



Fig. 143. — Wings of A'europly)ix appendiciilatus. 



vein R5 to vein R2+.3. This condition is very common in the Ascalaphidae 

 (Fig. 142). 



It is only rccenth" that I have been able to find examples of the succes- 

 sive stages of the switching of the base of vein R4 from vein R5 to vein 

 R2+3. These were found in the Osmylidaj, a family in which the dichot- 

 omy of the radial sector is normally suppressed. 



Figure 144 represents the wings of an atavistic individual of Osmylus 

 hyalinahis, in which the ])rimiti\-c dichotomy of the radial sector has been 

 retained; the homologies of the wing-veins are indicated in Figure 145. 



That such cases may be common in the Osmylidae is indicated by the 

 fact that in the small collection of osmylids to which I now have access two 

 individuals present this condition. In one Osmylus hyalinatus the radial 

 sector is dichotomously branched in both fore wings and in one hind wing. 

 In an Osmylus tessellatus the radial sector is dichotomously branched in the 

 fore wings; but in the hind wings the dichotomy has been suppressed. 



