38 THE CANAttUN ENTOMOLOGIST. 



erect a separate family for it. If tliis tliird section is rejected, there is no 

 ground for the retention of Psithyrid.ne. The family Stelididasof Schmiede- 

 knecht is in the same case. 



The presence of a distinct malar space is a common thing in the 

 Apygidialia, occurring in all of the principal groups. It is rare in the 

 Pygidialia. 



The Pygidialia form a more recent, continuous series. I would 

 separate the Halictidaj from the Andrenidse on account of their structural 

 differences, their different flight, and the fact that they have produced 

 iheir own inquilines. The structural characters of Paranomia, etc., seem 

 iro justify their separation as a family. Macropis is separated in the same 

 \va\'. I do not think it is closely related either to Panurgidaj or to Melitta. 

 Halictoides is referred to Dufoureid.e. This family differs from Panurgidse by 

 the cell IIIn-2 being pointed on costa ; the mandibles bidentate ; labrum 

 without bn^al spac^' or process ; the sco]ia femorilegid, the females collect- 

 ing loose ]iollen ; the face without coloured marks and without foveas. 

 Both families show considerable variation in the structure of the mouth- 

 l)art3. Indeed, Rhophites, in Dufoureidie, has the labial palpi more highly 

 s|)eciaiizt;d than in any other bee I have seen, joints 1-3 being flattened and 

 4 being simple and lateral. Protandrena I would refer to Protandreninje, a 

 sub-family of Panurgidie. Panurgus is one of the exceptions among the 

 Andrenoidea in having crurilegid scop^e and collecting loose pollen. The 

 scopa is conseciuently less localized than in the local Panurgidfe, all of 

 which mix the pollen with honey. 



In Melectidio I would include a number of genera referred by 

 Ashmead to Stelididi^j — Ammobatoides, Biastes, Pasites, Neopasites. In 

 Ammobatoides pundatus the female does not show a distinct pygidial 

 area, but the male shows a distinct pygidial process. The postscutel in 

 Ammobatoides and Biastes differs from that of local species in being more 

 protuberant and surpassing the scutel. 



The Euceridoe and Eraphoridae are separated in families which seem 

 sufirciently distinct from Anthophoridoe. 



Finally, there remains a possibility that the Pygidialia and Apygidialia 

 had an independent origin from the pygidial and apygidial Sphecoidea. 

 In that case, the Antho[)hila would not form a natural group. 



In the description of die venation the nomenclature of Comstock 

 and Needham, Am. Nat. 32:414, 423, has been foUov/ed, except that 

 III, IV and V are used for R, M and Cu ; V^ for M^ ; V^ for Cuj ; 



