CLASSIFICATION OF THE BEES. 



51 



B, — contaius but a single family : 



Fam. VIII. Megachilidse.— Megachile, Chalicodoma, Lithurgus, Trachusa, Osmia. 

 Heriades, Trypetes, Chelostoma, Anthidium. 

 C, — contains two families : 



Fam. IX. Sphecodidre. —Sphecodes. 



Fam. X. Prosopidse. — Prosopis. 



Section III. 

 This section he has divided into two groups : 



A.— Inquilines or Commensales, i.e.. parasitic bees living in the nests of social 

 species. 

 Fam. XI. Psithyridas.— Psithyrus (-= Apathus). 

 B. — True parasitic bees. 



Fam. XII. Melectidse.— Meleda, Crocisa, Nomada, Epeolns, Epeoloides, Biastes, 



Pasites, MelUtoxena, Ammobates, Ammobatoides . 

 Fam. XIII. Stelididfe.— Stelis, Ccelioxys, Dioxys. 



This arrangement of Dr. Schmiedeknecht's, and his separation of 

 the old Apida? into many well defined and readily recognizable 

 families, I consider a decided improvement over all schemes of 

 arrangement proposed by those who have preceded him, and will, I 

 feel sure, in time meet with the hearty support and approval of all 

 students of the bees. 



Dr. Henry Friese, in his work, which is substantially a continua- 

 tion of Schmiedeknecht's, follows closely the latter' s arrangement, 

 except that he treats his families as subfamilies, and has recognized 

 14 subfamilies in place of 13 families. This additional subfamily 

 is made by separating from Schmiedeknecht's family Stelididse, the 

 genus Ccelioxys and others, which he calls the Coelioxinre. 



Dr. Friese reverses the order of arrangement, proposed by 

 Schmiedeknecht, and begins with the Prosopinse or less specialized 

 bees, and ends with the Apinse, or the most highly specialized. 



It is so nearly identical with Schmiedeknecht's arrangement as 

 not to require a repetition here. 



In the following pages it will be seen that I have in the main 

 recognized most of the families, as proposed by Schmiedeknecht, but 

 instead of thirteen I have fourteen, not entirely agreeing with either 

 Schmiedeknecht nor Friese. How this is done may be readily dis- 

 cerned by an examination of my table of families, which is to follow. 

 Three or four of Schmiedeknecht's families I consider unnatural, 

 or at least not equivalent to family rank, or to the rank of his other 

 families, and two are suppressed. His family Sphecodidre I have 

 merged as a subfamily with the Andrenidse, since it agrees in every 



TRANS. AM. ENT. SOC. XXVI. MAY - 1899 ' 



