NATURAL GROUPING OF BREMID.-E. 395 



ject these latter to a critical examination in the light of ex- 

 tensive biological observations on some of our American species. 

 Before doing so, it seems desirable however to review briefly 

 several efforts at grouping which have been made on the basis of 

 (i) coloration and (2) structure. 



GROUPING ON THE BASIS OF COLORATION. 



The Austrian entomologist von Dalla Torre ('82) is apparently 

 the only one who attempted to subdivide the Bremidae on the 

 basis of coloration. However, since many species of the genus 

 Bremus (Bombus] which are not closely related show very striking 

 similarities in coloration, e.g., Bremus americanorum 3 and Bremus 

 terricola, this method of grouping is a very artificial one and 

 for this reason has not been favorably received by other workers. 

 In this connection the present writer would like to confess that 

 he himself, not long ago (cf. Plath '220, pp. 40-41), made a 

 similar mistake with respect to the genus Psithyrus. 



GROUPING ON THE BASIS OF STRUCTURE. 



There have been several attempts made to group the Bremidce 

 on the basis of structure. The first of these was by the Russian 

 General Radowszkowski ('84) who divided the genus Bremus 

 into eleven groups. This scheme of grouping was later some- 

 what modified by Franklin ('i2/'i3) in his "Bombidse of the New 

 World," in which the then-known, eighty-five American species 

 of the genus Bremus are divided into seven groups, a procedure 

 which has been followed by other American workers. This 

 method of grouping has been further extended by Franklin 

 ('i2/'i3) to the American species of the genus Psithyrus which he 

 divides into three groups. 



Another scheme of grouping which has much in common with 

 Radowszkowski's ('84) is that of Vogt ('n) who divides the genus 

 Bremus into nine subgenera, which Ball ('14) and Kriiger ('16 

 and '20) later increased to ten and thirteen respectively. The 

 last-named author, in addition to making a few modifications in 



3 Dr. Joseph Bequaert and the writer have recently studied De Geer's descrip- 

 tion and figure of Bremus pennsylvanicus, and, like Dr. T. H. Prison ('23), have 

 come to the conclusion that it is better to use the name americanorum until the 

 type specimen of De Geer is located. 



