|
Article Excerpt
Savannah Sparrows (Passerculus sandwichensis) are among the most widespread and geographically variable New World songbirds. They breed from the Aleutian Islands east across Alaska and mainland Canada and the United States, north to the Arctic Ocean and Hudson Bay, and south to Pennsylvania, northern New York, and northwestern Georgia (in the Appalachian Mountains), central Ohio, Indiana, Illinois, and Iowa, northern Nebraska, western Colorado and New Mexico, and in the west from Alaska south to California. They are resident in the interior of Mexico and south in the highlands to southern Mexico (and perhaps to Guatemala), and along the Pacific coast in saltmarshes from California to southern Baja California (Rising and Beadle 1996). The American Ornithologists' Union (AOU 1957) recognized 16 subspecies of Savannah Sparrows in addition to the Ipswich Sparrow (P. princeps) which has more recently has been treated as a subspecies of the Savannah Sparrow (AOU 1998). Paynter and Storer (1970) recognized 21 subspecies (including P. s. princeps). Hellmayr (1938) listed 15 subspecies (including P. s. princeps) for the world, and Miller et al. (1957) listed 15 subspecies from Mexico. This large number of subspecies is a reflection of the species' considerable geographic variation in size, proportions, and, most especially, coloration (Peters and Griscom 1938, Aldrich 1940, Hubbard 1974). Rising (2001) summarized geographic variation in size of this species, but discussed variation in coloration only in passing. Rising (2007) proposed that only six subspecies be recognized. The objectives of this paper are to: (1) provide a quantitative account of geographic variation among populations of Savannah Sparrows throughout their range, and (2) describe the relationship of patterns of plumage variation to local climatic environments, as well as to latitude, longitude, and elevation at each sampling site.
METHODS
Collections.--JDR collected and analyzed samples of Savannah Sparrows from sites coveting nearly the entire breeding range of the species (Fig. 1, Appendix; Rising 2001). All specimens were collected just prior to, or during, the breeding season and are assumed to have been collected near their breeding site. Adults of this species undergo a complete Pre-basic molt (partial in first-year birds) in the fall that occurs, at least in part, in breeding areas after breeding (Pyle 1997). The molt may be suspended during migration if it is not completed in breeding areas, although this has not been studied. The timing of the molt also may vary geographically. There is a limited Prealternate molt in spring that involves principally head feathers, but may also result in the replacement of some tertials and central rectrices. Thus, birds collected in the breeding season have worn plumage, whereas sparrows collected in autumn have unworn plumage. Fall- and winter-taken specimens are of unknown geographic origin and are not useful for analyses or geographic variation in color. Both males and females were collected at each site, and each was prepared as a skin-and-skeletal specimen (Rising 2001). The skins were all prepared in the same way: roughed out in the field, and then cleaned and stuffed in the museum. These specimens are housed in the Royal Ontario Museum in Toronto, Ontario, Canada.
[FIGURE 1 OMITTED]
Plumage Quantification.--Savannah Sparrows are streaked and spotted birds, and we decided it would be difficult to obtain repeatable estimates of spectral reflectance. Thus, JDR selected 14 quantitative multistate characters to quantify variation in plumage pattern and coloration. Each was scored on a scale of 1 to 6 or 1 to 4 with 1 being the lowest value for each feature (size or extent of spotting, palest color, etc.) and 4 or 6 being the highest. Each specimen scored was compared to a reference series of 18 birds, selected to show most of the variation in the species. JDR scored all of the specimens once, and then each was scored a second time in an effort to decrease variation in the scoring. Only 11 features were selected for the second scoring as three features could not be scored in a consistent manner; all colors follow Smithe (1975). These 11 features are:
spot size (the size of the ventral spots; 1 = small and 6 = large),
spot color (the color of the ventral spots; 1 = fawn [color 25] and 6 = dusky brown [19]),
spot extent (the extent of the ventral spotting; 1 = few, restricted to upper breast with few on the flanks, and 6 = extensive streaking),
background color (the ventral background color; 1 = whitish and 6 = pale horn [92]),
throat color (1 = white and unspotted and 4 = spotted with dark spots),
undertail covert color (1 = white and no beige patches and 4 = with large patches),
supercilium color (the amount of yellow in the supercilium; 1 = none and 6 extensively, bright spectrum yellow [55]),
mantle color (the color of the centers of the mantle feathers and scapulars; 1 = pale tawny [38] and 6 = dusky brown or black),
margin color (brightness) (the color of the edges of the mantle feathers; 1 = hoary and 6 = dark),
crown stripe (the distinctiveness of the median crown stripe; 1 = distinct, and 6 = absent), and
spot width (the width of the lateral crown stripe spots; 1 = small and 6 = broad).
The scores were selected a priori to cover the possible range of variation. Only specimens that could be scored for all 11 characters were used in the analyses.
Statistical Analyses.--We calculated the Euclidean distances among all pairs of birds to examine the extent of similarity among specimens for the combination of characters. We performed principal coordinates analysis (PCoA) using NTSYS-pc...
|
