656 PROCEEDINGS OF THE NATIONAL 3IUSEUM vol.xxiii. 



The Eprobosoidea differ iti so many important particulars from the 

 remaining families — such as the method of reproduction, manner of 

 living, much tougher integument of the body, structure of the 

 proboscis and of the antennae — as to justify their separation into a 

 group equivalent to all of the other Diptera. Between these two 

 divisions there are no intermediate forms. This is the position first 

 assigned them b}- Latreille, and in this he has been followed by 

 Meigen, Westwood, Walker, Haliday, and Bigot, while Loew and 

 Osten Sacken make them one of three primary groups. 



In the present arrangement the Tipulida^ are placed hrst in the list, 

 since they are evidently the lowest, most generalized of all the Dip- 

 tera; their comparatively large size, elongated form, weak organiza- 

 tion, numerous, many-branched veins, and long, many-jointed antennae 

 all tend to confirm this supposition. The Mj^cetophilidas arc placed 

 at the opposite end of the first superfamily for the reason that in sev- 

 eral forms the legs, and especially the antennae, are comparatively short 

 and robust, thus approaching the members of the second superfamily; 

 thus the genera PJatyura and IIrS2)erinKS approach very close to 

 Pleela, in the Bibionida?, which begins the second superfamily. The 

 genus Rliypliiis is closely related to Hhachlcerus^ in the Leptida?, for 

 which reason the Kh3'phida3 are placed at the end of the second super- 

 family, while the Leptida> begin the third. The latter, the Taban- 

 oidea, are bristleless flies, further distinguished from the two following 

 superfamilies by the greatly widened empodia; the genus Pangonia^ 

 in the Tabanida?, with its unusually large calypteres, frequently elon- 

 gated proboscis and reported habit of hovering over flowers, like a 

 humming-bird, naturally connects with the genus Eid(»icJni.s, in the 

 Acroceridie; and the relation of the latter to the Nemestrinidte is a 

 rather close one. The members of the following superfamily, the 

 Bombylioidea, are usually more or less bristly, and are essentially 

 flower-visiting flies among which the habit of hovering over flowers is 

 of rather frequent occurrence, v, hile the singular course of the veins 

 in the apical part of the wings of many serve still further to connect 

 them with the Nemestrinidie. The Asiloidea are usually provided v,ith 

 stout bristles and are almost without exception predaceous, the habit 

 of hovering over flowers being unknown. 



The family Lonchopterida3 is retained in the Orthorhapha, notwith- 

 standing the fact that de Meijere, from a recent study of the earh' 

 stages of LoncTioptera lutea^ while admitting that the family is in many 

 respects intermediate between the Orthorliapha and Cyclorhapha, 

 concludes that it has slightly more relationship with the latter than 

 with the former. In Lonclioptera ^ however, there are four posterior 

 cells in each wing, while the Cycloi'hapha never have more than three 

 of these cells; in the Orthorhapha Brach3'cera and in the Xemocera 

 with a discal cell the possession of more than three posterior cells is 



