188 CHARLES ROBERTSON. 



MALES. • 



Enclosure strongly rugose ; antenna long, joint 3 shorter than 5, rarely as long as 

 4 ; face without yellowish marks Tracliandreiia. 



Enclosure triangular, defined by an impressed line or by absence of hair, not so 

 strongly rugose ; joint 3 longer than 4 and as long as, or longer than, 5, 

 except in 0. personata 1. 



1. Face with yellowish marks 4. 



Face without yellowish marks 2. 



2. Glossa filiform, pubescent ; labial palpi long, joints subequal ; clypeus produced ; 



cheek bioad behind summit of eye; joint 3^4-6 lomeliSNa. 



Glossa and labial palpi as usual 3. 



3. Joint 3 longer than 4-f-5, cheek with a rounded angle Pfilaildreiia. 



Joint 3 not longer than 4-|-5 Andrena. 



4. Vein III5 absent; process small, triangular; angles of ventral segment 6 



reflexed ... .ParandrciiM. 



Vein III5 present 5. 



5. Glossa filiform, pubescent; labial palpi long, joints subequal. ■ ■ lomelissa. 

 Glossa and palpi as usual Opandrena or Pteraiidreiia. 



Trachandrena is a distinct group, whose forms may be separated 

 by a peculiar set of characters which are of little use in separating 

 the species of the other genera. 



Ptilandrena is an offshoot from Andrena, as here limited, the 

 loosely plumose scopiB being adapted for holding the large pollen 

 grains of Claytonia virginica and Geranium maculatwn, of which 

 the species are oligotropic visitors. 



The species in which the males have yellowish face- marks fall 

 into two groups, Opandrena and Pterandrena., the latter having 

 densely plumose scop?e for carrying the fine pollen of Compositse, of 

 which the specie^ are oligotropic visitors. P. lauracea, however, 

 has not been observed collecting pollen ; but I expect to find it get- 

 ting pollen from some of the Compositse. 



In Tomeiissa. the face- marks of the male and the tibial scopfe of 

 the female indicate a recent origin from Opandrena. The great 

 elongation of the glossa is correlated with the deep seated nectar of 

 Viola, of which the single species is an oligotropic visitor. 



Parandrena is developed from Opandrena by obliteration of vein 

 III5. I have two specimens having this nervure j)resent in one wing. 

 As a rule, specimens with this nervure wanting should be run 

 through the tables, just as if it were present, and should not be re- 

 ferred to Parandrena, unless there is quite a series of specimens 

 otherwise resembling the type. As I have stated before, it is the 

 most unstable element in the venation of bees. I have found it 



