iSgq-] 



PSYCHE. 



373 



characters, it probably represents some 

 other species in another genus. 



I quite agree with all good system- 

 atists that color alone is not of generic 

 value; but Mr. Cockerell, with others, 

 must surely admit that it is of great 

 value and assistance in recognizing 

 certain genera when used with salient 

 structural characters. The genus Phil- 

 o.xanthus was not based upon color alone 

 as Mr. Cockerell intimates, and as he 

 may readily see if he will again consult 

 my description. 



Mr. Cockerell seems loath to see 

 Perdita, upon which he has done so 

 much work, split up into subgenera, 

 although inspired evidently by my 

 beginning he does not hesitate to pro- 

 pose the generic name " Perditella " 

 for what he calls " a really good sub- 

 genus " for his Perdita larreae^ marcialis 

 and larrearum. As so frequently 

 occurs in Mr. Cockerell's work, as for 

 instance in his subgeneric divisions of 

 the Coccidae, he has not given a single 

 essential character for the recognition 

 of this genus. The only character 

 given, namely, " second submarginal 

 cell small, triangular" will not hold 

 good for the three species mentioned, 

 while the so-called large stigma is 

 essentially the same as found in Perdita 

 and Cockerellia, although it appears to 

 Mr. Cockerell's eyes larger on account 

 of the smallness of the marginal cell. 



Perditella however, may possibly be 

 retained for /(j/7r«, marcialis, and luteola. 

 It is singular however, that all of the 

 specimens of these species that I have 



yet seen are males, and, I suspect, they 

 may really represent the opposite sex 

 of Philoxanthus. In the latter event the 

 characters made use of in my table will 

 then be of sexual and not generic 

 value. 



Mr. Cockerell sums up his remarks 

 by saying : " I do not say that Perdita 

 should not be divided into two or more 

 genera; probably it will ultimately 

 have to be split into half a dozen but it 

 will be necessary to proceed with 

 caution." 



This is just about the number of 

 genera I have recognized. 



"Proceed with caution," is good 

 advice, and it is a pity Mr. Cockerell 

 has not followed it himself before erect- 

 ing some of his recent new bee genera 

 since most of them will be found to be 

 synonyms. I find no less than four 

 in the following group: Hemihalictus 

 Ckll. equals Dufourea Lepeletier, Hes- 

 peraspis Ckll. equals Rhophitoides 

 Schenck, Callandrena Ckll. equals 

 Biareolina Dufour, while Pseudopanur- 

 gus Ckll. equals Panurginus Nylander. 



In order to show how Perdita and 

 allied genera are separated, I have 

 concluded to publish my MS. generic 

 table of the family Panurgidae, which 

 is as follows : — 



Table of Genera, 



Marginal cell at apex more or less remote 

 from the costa or obliquely, broadly trun- 

 cate. ....... 5 



Marginal cell towards apex acuminate, attain- 

 ing the costa. 



