SAN BENITO SAVANNAH SPARROW 719 



tall cactus and a few shrubs. During a stay of two days, no mammals 

 and only a few lizards were seen. Only five species of small land birds 

 were found, and only one of these was at all abundant — the "large- 

 billed" Savannah sparrow. Young were just out of the nest at the 

 time of the visit, about Apr. 24, 1906. One young male, in post- 

 juvenal molt, that was possibly a straggler from San Benito Island, 

 was found on Cerros Island on April 21. William Brewster (1902) 

 points out that this habitat is "essentially different from that of any 

 of the salt marsh Ammodrami [now Passerculus]." 



This subspecies was formerly considered a race of the species 

 rostratus, which is now regarded as a well-marked subspecies of the 

 sandwichensis complex. In color and size it more nearly resembles the 

 races rostratus and atratus of the Sonora coast than it does the popula- 

 tions of the "guttatas" group resident on the adjacent coast of Baja 

 California. As van Rossem (1947) points out, "The obviously close 

 relationship of sanctorum, rostratus, and atratus can easily lead to the 

 speculation that the colony on the San Benito Islands is a remnant 

 population. At any rate, it is obvious that the three are more closely 

 interrelated than are any of them to the present-day occupants of the 

 intervening peninsula." 



Nesting. — WiUiam Brewster (1902) quotes R. C. McGregor who 

 found three nests of sanctorum on the San Benito Islands. All the 

 nests were placed on the ground under small bushes. A nest found 

 March 30 was sunk level with the ground, which served to support 

 the thin walls. The outside was composed of large grass straws while 

 the lining was of finer grass and a few feathers. The three eggs were 

 shghtly incubated. 



A. W. Anthony (1906) mentions a nest of sanctorum about a foot 

 from the ground in a low bush. Other nests were well hidden in 

 shallow depressions in the soil and overhung with vegetation. The 

 nests were very similar, in fact, to those of beldingi. 



Plumages. — Van Rossem's (1947) exhaustive treatment furnishes 

 the following subspecific characters: 



Bill large, stout, and deep at base as in Passerculus sandwichensis rostratus 

 and Passerculus sandwichensis atratus, but culnoen outline normally straight or 

 nearly so rather than convex. Tarsi slightly shorter tlian in those races, but 

 notably stout and, together with the feet, horn color or plumbeous brown rather 

 than flesh color or light brown, a distinction which persists in most dried speci- 

 mens. Wing slightly shorter and tail decidedly so, the latter relatively as well 

 as actually. Dorsal pattern moderately variegated or contrasted as in rostratus 

 and atratus, but differs in the presence of a more or less extensive intermixture 

 of light gray or grayish white edgings in the inter-scapular area. Brown phase 

 with tones tending to chestnut rather tlian pinkish. Ventral streaking relatively 

 narrow as in rostratus but black, or nearly so, instead of brown. 



64ft-737 — 68 — pt. 2 9 



