240 AMERICAN HYMENOPTERA. 



5. Bombus vagans, B. bolsteri and B cockerelli. 



6. " frigidus and B. couperi. 



7. " flavifrons, B. centralis and B.pleuralis. 



8. " sitkensis and B. atnbiguus. 



9. " nielanopygus, B. sylvicola and B. gelidus. 



10. " ephippiatus a.n6. B. pulcher. 



11. " huntii a-nd B. laticinctus. 



12. " separahis a.nd. B. mormonorum. 



13. " ru/ocinctus, B. mexicensis and 5. henshawi. 



If some other worker on this family should reduce the 

 thirty-one species here listed to thirteen, as indicated, the 

 writer could present no very substantial reasons why they 

 should not be so classified. 



Where the variation of a species has made it necessary 

 to describe several color gradations, the writer has, in most 

 cases, designated those gradations as ''' color variants'' and 

 numbered them consecutively instead of giving them names. 

 Most entomologists would probably have given these grada- 

 tions names and ranked them either as varieties or as sub- 

 species. As different gradations are often found together 

 in the same nest, being the offspring of the same queen, it 

 seems unreasonable to the writer to give them any particular 

 rank and the term "color variant" has been chosen as the 

 one which seems to the writer to most accurately express 

 the relationship of the various gradations to each other. 

 Criticisms of this method of disposing of these variations is 

 expected, but the writer is willing to endure it for the sake 

 of his opinions concerning the matter. If names for these 

 variants seem necessary, other students of the group are, of 

 course, at liberty to furnish them. 



Arrangement of Specific Descriptions. 



In this paper, the specific descriptions, within the various 

 groups, have been arranged as far as it has been possible, 

 to indicate the relative closeness of the relationship of the 

 species to one another. Such an arrangement for the Du- 

 moucheli and Fratermis groups has been far more difficult, 

 and will be found much less reliable, than for the other 

 groups. Of the other groups, the Pratorum group alone 



