RUFOUS-WINGED SPARROW 903 



brushy bottomlands south of town. About 1951 the species began to 

 spread out in the valley, and in 1956 an extraordinary expansion of 

 range carried the species into areas of almost solid grassland and even 

 into one wooded mountain canyon to the south, well above its usual 

 desert or Lower Sonoran Zone haunts. Few species have shown such 

 drastic fluctuations in habitat and numbers. It is small wonder that 

 A. carpalis was considered extinct in the United States by some 

 authors, including the American Ornithologists' Union (1931) espe- 

 cially as no one imderstood the bird's requirements and everyone 

 looked for it in the wrong places (Swarth, 1929; van Rossem, 1931, 

 1936b; PhiUips, Marshall, and Monson, 1964). 



The genus Aimophila, with certain close allies, forms a large group 

 of spaiTows centered in the southern half of Mexico. A few species 

 extend northward well into the United States, but only one, A. aesti- 

 valis, is limited to this country. In their life histories, most or all of 

 these sparrows react markedly to factors other than the calendar and 

 the photoperiod. Thus, the customary way of presenting life his- 

 tories must, in these cases, be modified to obtain a clear understanding. 



Habitat and phenological background. — The essential parts of the 

 normal habitat of the rufous-winged sparrow, as already stated 

 (Phillips, 1955), are gi-ass and brush (thorny, dense, or preferably 

 both). The grasses favored in Arizona are of bunching types, not 

 plain low Bouteloua, though in recent years areas with very sparse 

 grass and low weeds have been inhabited. The brush-grass combi- 

 nation and the birds are common and widespread over much of 

 Sonora; they give way southward to taller, denser thorn forest in 

 Sinaloa, and northward to deserts devoid of grass and to unbroken, 

 brushless desert gi'asslands in Arizona. In both Sinaloa and Arizona 

 the birds thus seem to be in rather restricted, isolated colonies. Even 

 these, however, are not uniform ecologically, at least near Tucson 

 (the only point whence we have detailed data on the life history). To 

 discuss these data, we must distinguish at least five types of habitat: 



Swale habitat: Areas where flood waters seep into desert soils, 

 not yet badly gullied, producing wide stands of grass in the bottom, 

 chiefly tobosa (Hilaria mutica). Around these stands the better 

 drained edges support leguminous brush and low trees, and usually 

 desert hackberry (Celfis tala var. pallida) bushes which may be of 

 considerable density and circumference, and over 2 meters tall, when 

 the soil is still healthy. Moisture comes entirely from rains over the 

 local drainage area; no underground water is within reach of the roots. 



Desert wash habitat: Similar, but gently sloping, without seepage 

 or the consequent stands of tobosa; the bottom is plain sand, com- 

 pletely drained and usually gullied shallowly. Dominant trees are 

 usually palo verde and mesquite, and there is an understory of brush. 



