6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I46 



Females (12 specimens), wing 471-495 (482), tail 231-251 (241), 

 culmen from cere 22.7-25.9 (24.1), tarsus 58.6-66.5 (62.5) mm. 



Range. — Breeds from the lower Colorado River Valley (Riverside 

 Mountain, Calif.), southern Arizona (Yarnell, Yavapai County; 

 V^ail, Pima County; Bisbee, Cochise County), southern New Mexico 

 (San Luis Mountains), and southern Texas (Delaware Creek; Lim- 

 pia Creek; Chisos Mountains; Starr County) south into Sonora, 

 Chihuahua, and Coahuila (southern limits on the Mexican tableland 

 uncertain), through the Caribbean slope of tropical and subtropical 

 Mexico, and Central America to Honduras, probably to central Costa 

 Rica; the Bahama Islands (Mangrove Cay, Grand Bahama, Abaco, 

 Andros), Cuba, Isle of Pines, Jamaica, and southwestern Puerto 

 Rico (introduced) : On the mainland intergrades on the north with 

 C. a. meridionalis. 



In migration and winter south through Panama to Darien (Jaque). 

 including the Pearl Islands (Isla San Jose). 



Remarks. — It is evident that the division in two races, meridionalis 

 and typical aura, appears arbitrary, with a considerable area of over- 

 lap. However, such separation seems required in view of the dis- 

 parity between the populations with small size of the far south and 

 those decidedly larger of the north, the range in the wing in males 

 being from 462 to 525 mm. and in females from 471 to 526 mm. 

 Smaller size is coupled with tropical and lower subtropical zone 

 range, against the mainly Sonoran and Temperate Zone distribution 

 of the larger, northern birds. It is probable that there are other 

 intangible factors of difference involved that have not been evident 

 in examination of museum specimens. 



The criterion for the size limits assigned to the southern sub- 

 species has been fixed through measurements of specimens from the 

 resident birds of the Greater Antilles, where there is no confusion 

 through the periodic invasion of northern migrants as is the case in 

 the continental breeding range. The dimensions of the Antillean 

 group then have been the yardstick used to outline the breeding range 

 assigned to the subspecies aura on the mainland. Division between 

 this group and meridionalis comes near the boundary between Mexico 

 and the United States, with the smaller southern form penetrating a 

 short distance to the north of this line. The southern limit indicated 

 is tentative since no specimens definitely known to be on their breed- 

 ing grounds have been seen from southern Central America. I have 

 taken winter migrants in eastern Darien so that this race may range 

 to Colombia. 



