6 PROCEEDINGS OF THE NATIONAL MUSEUM voi. 68 



5. Third submarginal cell of fore wing normally higher than long, the vein 



closing it outwardly interstitial with the cross vein from hind margin of 



stigma, or even proximad of it gaudii group, p. 13. 



Third submarginal cell of fore wing subquadrate or longer than high, the 

 vein closing it outwardly much beyond the cross vein from hind side of 

 stigma 6. 



6. Head very conspicuously narrowed posteriorly, the sides converging from di- 



rectly behind eyes when seen from above ampla group, p. 14. 



Head normal, sides when seen from above generally parallel or subparallel 

 for a short distance behind eyes 7. 



7. Space between antennal socket and eye not over one-half as wide as one 



socket; posterior ocelli separated from eye by a distance not greater than 

 the transverse diameter of either ocellus; anterior extremities of gular 



carina visible from side except in nigripes inaequalis group, p. 15. 



Space between antennal socket and eye distinctly more tlian half as wide &s 

 socket 8. 



8. Posterior ocelli separated from eye by a space at least 1.5 as wide as one 



ocellus micrommata group, p. 17. 



Posterior ocelli separated from eye by a space less than 1.5 as wide as one 

 ocellus 9, 



9. Gular carina not elevated and deflected outwardly at or near its anterior 



extremities, never visible when head is viewed directly from the sides; 

 vein emanating from apical costal margin of stigma extending along costa 

 as far as or farther than length of the cross vein emanating from hind side 



of stigma nitida group, p. 20. 



Gular carina with its anterior extremities visible when head is seen from the 

 side, or if these elevations are not visible, or indistinctly so, the vein 

 emanating from apical costal margin of stigma bends down at, or very close 

 to, its base and does not run along costa as far as stated in last section. 



atrata group, p. 22. 



Bradley separates two species, eriopis Bradle}* and arenivaga 

 Bradley, from all the others on the structure of the basal segment of 

 the abdomen. In these two this segment has no basal neck and there 

 is no dorsal, and scarcely any lateral, constriction between it and the 

 second segment, I have seen no sjiecimens that possess tliis character. 

 Bradley figures the hypopygium of arenivaga and, tliough the details 

 are not very clear, it evidently is very similar to that of atrata Blake, 

 as is also the venation of the fore wing, wliich he figures, too. 



Bradley included 30 species of Brachycistis in the restricted sense 

 in his ])aper. Of t]iese 1 have seen 22 and some of the others are 

 included in my keys on the strength of characters given by Bradley. 

 A few are omitted as it is impossible for me to include tliem owing to 

 the descriptions lacking characters which are essential to my placing 

 them in my paper. The omitted species include perpunctata Cockerell 

 petiolata Fox, ioachinensis Bradley, nuda Fox, suhquadrata Fox, and 

 the two already referred to. I believe that the absence of the carina 

 on the u])per margin of occiput will readily distinguish the first- 

 named sj)ecies from any other yet described, and nuda Fox appears, 

 to me to be a synonym of atrata, whicli Bradley evidently suspected 

 from his remark that it was "doubtfully distinct from atrata." 



