18 BULLETIlir 86, UNITED STATES NATIONAL MUSEUM. 



subspecies. But birds from intermediate areas require more careful 

 examination, and often a series from a locality is necessary for 

 their proper allocation. Since, however, the certain breeding period 

 of Chordeiles is comparatively so short, most of the specimens col- 

 lected are either migrants or not surely breeding birds, and their 

 subspecific determination becomes often a matter of great difficulty, 

 and sometimes of more or less uncertainty as well. 



For obvious reasons an attempt has been made to name every 

 specimen handled in the present investigation, and the results are 

 contained in the lists of specimens under each subspecies. With 

 such a large proportion of specimens of unknown breeding origin, 

 in a group which presents so many puzzling and overlapping vari- 

 ations as do the subspecies of Chordeiles virgimanus and Chordeiles 

 acutipennis, it is too much to hope that I have correctly placed 

 every single individual of the large number that I have examined; 

 but I have carefully scrutinized and compared every one, so that the 

 final determination as printed represents, in each individual case, 

 the best opinion I can offer with the means at my present disposal. 

 Some doubtful specimens have been identified and reidentified dur- 

 ing the progress of the work as many as six or eight times, not 

 always, it must be admitted, with the same result ! In instances such 

 as those mentioned above, where a given variant specimen can be 

 matched by individuals belonging to two or even more races, the 

 specimen , has been assigned to that form' by the average subspecific 

 characters of which it is most closely approached; and, in deter- 

 mining this, both the average of single characters among individuals 

 of the same form, and an average of all the differential characters 

 of the given subspecies are taken into consideration. This seems to 

 be the only logical way to treat such cases, and by it we probably 

 much more closely approximate the truth than by any other method. 



Variaiion. — ^The species of this genus have a wide range of all 

 kinds of variation excepting seasonal, which is at a minimum, due 

 partly to the comparatively little abrasion to which the plumage 

 is subject; partly to the great similarity of freshly molted birds 

 to those in worn plumage, for there are, on the fresh feathers, 

 no differently colored tips or edges to wear off; and partly, in the 

 case of a semiannual molt, to the sameness of successive adult 

 plumages. 



Individual variation is often so great as seriously to complicate the 

 problem of geographic forms, as well as to render difficult the 

 identification of specimens. Furthermore it is just as great in the 

 sedentary as in the migrant forms. This is, however, fully treated 

 under the various species and subspecies. 



Sexual differences are about equally great in all the species. In 

 Chordeiles acutipennis and Chordeiles mrginicmus they are of very 



