908 U.S. NATIONAL MUSEUM BULLETIN 237 part 2 



weather turns hot and dry again, as Paul Schwartz and E. T. Gilliard 

 believe may occur normally with Venezuelan birds (GiUiard, 1959). 



Spring. — As the rufous-winged sparrow neither migrates nor often 

 gathers into flocks that are ob^dously larger than family size, the 

 changes in its behavior and local distribution in spring are not striking. 

 By mid- or late March most of the birds seem to be in pairs, but this 

 is not universal. For one thing, the apparent pairs may not be true 

 pairs. Thus on Mar. 24, 1940, I collected one bird from an apparent 

 pair which on skinning later, I found to be a male (testes somewhat 

 enlarged, left 2.8X1.8, right 1.7X1.7 millimeters); yet while I was 

 preparing it to take home its supposed ''mate" sang. Therefore, 

 either the female may sing or (more likely) some apparent pairs are 

 not of opposite sexes, a condition that was proved, also in March, 

 in the case of a wren, Uropsila leucogaster brachyurus in Yucatdn 

 (Chapman, 1896b). 



For another thing, even in wet springs some birds remain in small 

 flocks through March. On Mar. 29, 1958, my ornithology students 

 saw two groups of three bii'ds each near Tucson in the swale habitat. 

 More exceptional was a flock of about five or six sparrows that I saw 

 in the desert wash habitat on Apr. 27, 1958; at least four birds, if not 

 the whole flock, were rufous-winged sparrows; yet, in the same area, 

 some pairs already had eggs in the nests. Farther south in southern 

 Sonora according to A. J. van Rossem (1945a) the birds were not yet 

 paired in late May and even up to June 22 ,1937 ; while Joe T. Marshall, 

 Jr. (MS.) could find no nests there from May 8 to 12, 1958, on which 

 dates some pairs near Tucson had young nearly ready to leave the 

 nest (in desert wash) or actually on the wing (in riparian habitat). 

 In northwestern and central western Sonora I found at least some 

 birds paired, though not yet nesting, Apr. 20 and 28, 1947. 



Territory. — As might be expected, the size of the territory evi- 

 dently varies inversely to the desirability of the habitat. In the 

 farmland habitat Patrick J. Gould (MS.) found a pair holding about 

 % hectare (2 acres) ; most of this was not contiguous to the territories 

 of other pairs. The favored swale habitat may support as many as 

 8 to 10 pairs per % kilometer (K mile), the swale being perhaps 50-100 

 meters wide on an average (A. R. Phillips, 1955); probably most 

 territories here do not greatly exceed % hectare. On Mar. 29, 1958, 

 flights here seldom exceeded 15 meters; and in a choUa cactus flat I 

 saw two "pairs" perched in chollas not over 10 meters apart with no 

 apparent friction. Whether these birds had yet established territories 

 at that date, however, is uncertain. 



In the original riparian habitat, their density must have been 

 rather spectacular. In 1872 in a strip of mesquite trees, brush, and 

 grass measuring about 1500X400 meters, C. E. Bendke (1882) 

 "found not less than forty- three of then- nests with eggs and a still 



