654 BULLETIN 50, UNITED STATES NATIONAL MUSEUM. 



having the black on the forehead very narrow and that of the auricular 

 region more restricted. Both of these species, however, have the 

 outermost primary longer than the second (sometimes longer than the 

 fourth), and the tail shorter than the wing, though longer than the lat- 

 ter from the bend to the tips of the secondaries. 



The geographic variations in G. trichas are difficult to understand 

 satisfactorily, largely from lack of necessary material. Most of the 

 specimens contained in collections were taken during the seasons of 

 migration, and therefore it is in many cases impossible to determine 

 whether certain specimens would have remained to breed in the locali- 

 ties where they were secured, specimens which would undoubtedly pass 

 farther northward to their summer home often lingering in a more 

 southern locality until the birds there resident have commenced to 

 breed. The chief difficulty is with the birds of this species in the 

 Pacific coast district, where four subspecies have been supposed to 

 occur, three of which 1 have been able to verify, but in the case of one 

 only have been able to determine with any degree of accuracy the 

 breeding range. Some of the geographic forms of the species are 

 very strongly marked subspecies, while others are very slightly differ- 

 entiated; in fact they might more properly be termed "subraces" than 

 subspecies, and would, in the case of two of them at least, hardly be 

 worth recognizing by name were not their respective ranges separated 

 by many hundreds of miles, the intervening territory being occupied 

 by a form which, instead of being intermediate in its character, is more 

 different from either of the forms on opposite sides than these are 

 from one another. With the exception of these extreme western 

 forms, of which only two are strongly characterized, the extent and 

 character of geographic variation is very evident when a sufficient 

 series of specimens representing all parts of the very extensive area 

 inhabited by the species is examined, especially when those taken dur- 

 ing migration are eliminated. Altogether 1 have been able to diag 

 nose, more or less satisfactorily, eight forms, whose principal charac- 

 ters and ranges, so far as the material examined enables me to deter- 

 mine them, are as follows: 



(1) G. trichas triclms. (Type locality, Maryland.)— Characterized 

 by smallest size (with one exception among the Pacific coast forms'), 

 the adult male nearly always with the under parts of the body exten- 

 sively pale buffy or buffy whitish, the yellow being confined to the 

 throat, chest, breast, and under tail-coverts, and the upper parts of a 

 duller, more grayish, olive-green. The bi'eeding range of this form 

 comprises the States of Virginia (except the southern coast district) 

 and Maryland, the more southern portions of Pennsylvania, and the 

 District of Columbia; probably also Delaware and southern New Jer- 

 sey, and the upland districts of the Carolinas. The only extralimital 



' G. t. sinuosa, confined to the salt marshes skirting San Francisco Bay. 



