478 TROMOPTERA 



most easily distinguished, through the long preefurca and the very 

 distinct divergence of the radial and cubital veins whereby the radial 

 vein appears to originate rectangularly from the cubital vein quite close 

 to (or only shortly before) the discal cross-vein ; very few genera remain 

 doubtfully placed in this subfamily, though in Henica Macq. { — LagocMlus 

 Loew) and in some other genera the small cross-vein is placed {Lomatia-\\kQ) 

 towards the end of the discal cell, and in Mulio (and its close allies) the 

 prsefurca is rather long and the divergence of the veins considerably 

 before the discal cross-vein. The Systropincc, though the last subfamily 

 to be suggested, are also well defined through the remarkable elongation 

 and bareness of all parts of the structure, and through the somewhat 

 hollowed-out back of the head. It is not easy to decide the nearest 

 relationship of the Systropincr, but it is curious that they seem to be 

 allied to the most strongly bristled genus in the whole family, viz., 

 Toxopliora, because of the similarity of the peculiar structure of the 

 scutellum and the basal joint of the antennae; if this surmise be correct, 

 it is not so difhcult to trace the relationship through the ToxopJiorince 

 to the Lomatince, though a great difference of opinion may exist as to 

 the boundary line. Tlie Toxophorince in the sense adopted by me do 

 not form a very coherent group, especially as regards Edimus (including 

 Thevenefimyia) and Tomomyza ; Schiner included Amictus (including 

 Thlijjsomyza) and Gyllcnia in his Lomatince, but I have laid more stress 

 upon the remarkable " humpiness " of the Toxcphorince and have therefore 

 included those genera which also show many points of divergence from 

 the Lomatinm in their more bristly nature, longer proboscis, and cylindrical 

 abdomen. The remarkable dorsal development of the prothorax in 

 Toxopliora and Lepidoplioixi may distinguish those two genera from the 

 others included by me in the subfamily, but they seem to be allied to 

 Edimus and Thcvenetimyia which have the most Lomatia-\\kQ figure. 

 Tomomyza is of still more doubtful location and may well belong to the 

 Lomati'iia', but has a more cylindrical abdomen and an upright discal 

 cross-vein placed near the middle of the discal cell, and also seems to be 

 allied to CyUenia. The Lomatinoi (restricted by the omission of the genera 

 just mentioned) become a rather symmetrical subfamily and are distin- 

 guished from the Boiiibylina^ by a more oblong flatter abdomen and many 

 minor details which are mentioned under the characters of the subfamily. 

 The Bombylince are also not quite homogeneous, and when the genera which 

 have a simple cubital vein are better known it is very probable that they 

 will form another subfamily, the Platypygina:, distinguished further by 

 the almost total absence of pubescence and by the humped Cyrtidous-like 

 figure. 



I have been able to examine specimens of the following genera : 



BoMBYLiN^ : Atiastoechus, Bfnnhylius, Vischistus, Geron, Heterostyhim, Lordotus^, 

 Fhthiria, Platypygvs, Ploas, RhaJtdopselaphus, Syaioechus, Usia, and the 

 rather aberrant Corsomyza and Lastoj»-osoj>a (head veiy transverse), Seri- 

 cosoma and Pantarbes (antennae widely separated, and abdomen conical). 



