14 BULLETIN 86, UNITED STATES NATIONAL MUSEUM. 



course, necessary for the sake of convenience, and as it is more satis- 

 factory for a list or technical treatise to begin the catalogue of sub- 

 species under each species with the " typical," or rather the first-de- 

 scribed race, the name of which the species as a whole bears, the 

 writer has tried to arrange the subspecific forms in as natural an order 

 as is possible under these circumstances^ and this sequence he has fol- 

 lowed in the detailed treatment of the forms of the genus. 



A linear arrangement more satisfactory from the standpoint of re- 

 lationship may, of course, be obtained by disregarding the purely 

 arbitrary placing of the first-described form of a species at the head 

 of the list of subspecies. Such a list, with due regard for the phy- 

 logeny and present-day affinities of both species and subspecies, and 

 one which seems to the writer to express, as eloquently as a mere list 

 can, their known or assumed relationships, is as follows : 



Chordeiles virginianus minor. 

 Chordeiles virginianus vicimis. 

 Chordeiles virginianus chapniani. 

 Chordeiles virginianus aserriensis. 

 Chordeiles virginianus virginianus. 

 Chordetles virginianus hesperis. 

 Chordeiles virginianus sennetti. 

 Chordeiles virginianus howelli. 

 Chordeiles virginianus henryi. 

 Chordeiles acutipennis acutipennis. 

 Chordeiles acutipennis exilis. 

 Chordeiles acutipennis microTneris. 

 Chordeiles acutipennis texensis. 

 Chordeiles acutipennis inferior. 

 Chordeiles impestris rupestris. 

 Chordeiles rupestris xyostictus. 

 Chordeiles rupestris zaleucus. 



The necessity of beginning in each species with the oldest or lowest 

 form brings into juxtaposition, in a linear arrangement, the lowest 

 or oldest form of any one species with the highest or most recent 

 form of the preceding species, but there seems to be no better way. 

 In the above list this is most noticeable with regard to Chordeiles 

 virgi7iia7ius and Chordeiles acutipennis, for, of the races of the 

 former, Chordeiles virginianus minor is undoubtedly the one most 

 closely allied to Chordeiles acutipennis acutipennis, though of neces- 

 sity placed farthest from it serially. 



The sequence of subspecies under each species is also difficult to ex- 

 press in a single linear list, owing to the fact that some races are 

 about equally related to two or three others. Thus in the case of 



