1894.] NATURAL SCIENCES OF PHILADELPHIA. 301 



genera of this subfamil}'. The following table is baf^ed chiefly on 

 that in Cre^^son's " Synopsis" : — 



Anterior wings witli ttiree discoidal cells, therefore with two recurrent 

 nervures. , 



Abdomen with a tolerably long petiole . . . Pemphredon Latr. 

 Abdomen with the petiole not longer than the hind coxae. 

 Posterior tibiae spinose or subserrate ; labium emarginate at tip.'* 



DiouoNTUs Curt. 



Posterior tibiae (excepting the calearia) unarmed ; labrura 



pointed at tip Passalcecus Shuck. 



Anterior w ings \\ith two discoidal cells, therefore only one recurrent 



nervure 



Anterior wings with one subniarginal cell . A^imoplanus Giraud. 

 Anterior wings with two submarginal cells. 



Petiole short; recurrent nervure joining the first transverso- 



cubital nervure Spilomena Shuck. 



Petiole long; recurrent nervure received in the middle of the 

 llrst submarginal cell Stigmus Jur. 



Subfamily BEMBICIJMJE. 



Under this head I unite the Larrida?, Bembicidte, Nyssonid^e, 

 Philanthidie, and ^Nlellinidte. The characters of these supposed 

 families are not sufficient or constant enough to sustain them in 

 such a rank, and are valueless in some cases, even as characters of 

 minor importance. As in the case of the Bembicidte, it is easy 

 to take such types as Bembex, Monedula, forms with rostrate 

 clypeus, and separate them into a family, apparently distinct from 

 the Nyssonidie and Larrida;, if these genera are compared, say, 

 with Larra and Ni/sso^i; but certain genera will be encountered, 

 Xeolavra and Buthynostethus for instance, whose proper position will 

 remain undetermined. Neolarra combines both Larrid, Bembicid 

 and Nyssonid characters, yet it will fit in neither of the families de- 

 fined. Bothijnostethus inclines to both the Larridie and Nyssonid?e, 

 and seems to be a connecting link between them. Sflzus and Sphe- 

 cui.<, although placed in the Bembicidie, possesses the neuration and 

 non-rostrate labrum, characters which bind them to the Nyssonidie. 

 It must not be forgotten that the labrum of the Nyssonidre is promi- 

 nent, indeed in Gonjtcs very prominent. Thirteen tribes of this 

 subfamily seem to be indictated, which number will undoubtedly 

 have to be reduced in the future. 



^* I have not seen Poleinistus Sauss., described as occurring in Madagascar 

 and Mexico. It is related to Passahxcus and Diodoiitiis. 



