# Field methods
From 17 October to 4 November 2015 (just after prebasic molt; McGraw and Hill 2004), we captured 95 (n = 45 female and n = 50 male) house finches from 4 different sites in the Phoenix, Arizona metropolitan area: 2 urban (Arizona State University [ASU]— Tempe campus, downtown Phoenix residence) and 2 rural (Estrella Mountain and South Mountain regional parks), which differ significantly in both land-use/land-cover metrics and human population density (see Giraudeau et al. 2014 for additional site details). We maintained wire feeders with black oil sunflower seeds for 1 week prior to trapping, and birds were captured using basket traps hung around feeders. At capture, each bird was banded with a numbered US Geological Survey (USGS) tag for identification and examined visually (by P.H. and B.E.S.) to determine sex based on plumage coloration (Hill 2002). Age of these birds could not be determined at the time of this study because all finches were in adult plumage.
Body mass was determined to the nearest 0.1 g using a digital scale (Smart Weigh, Chestnut Ridge, NY), and tarsal bone length was measured to the nearest 0.1 mm using digital calipers (Neiko Tools, Homewood, IL) so that we could also evaluate body condition as the residuals saved from a linear regression between tarsus length and body mass (F1,93 = 10.15, P = 0.002; sensu Toomey and McGraw 2009; Burtka and Grindstaff 2015; Trigo and Mota 2014). Birds were inspected visually (by P.H. and B.E.S.) and scored for poxvirus infections, but poxvirus prevalence was very low in this study (n = 3 birds), so pox infection prevalence and severity were not included in further analyses. All birds were processed at the ASU-Tempe campus and photographed to obtain color metrics. A Sony Cyber-Shot DSC-W800 was used to take photos in a dark room and with consistent flash settings, subject distance, camera presets, and background (gray scale photography card; Hasegawa et al. 2014). Each plumage patch of carotenoid coloration (crown, breast, and rump in males; same patches in females when red/orange/yellow color was present there) was photographed twice, and images were analyzed in Adobe Photoshop CS6 (Adobe Systems, San Jose CA, USA). We selected carotenoid color patches using the quick selection tool in Photoshop and obtained patch size (in pixels) and red-green-blue (RGB) values using the Histogram window (Giraudeau et al. 2013). Hue, saturation, and brightness values were computed from corresponding RGB values using the “Color Picker” function. A size standard (of known area) was included in each photograph in order to convert number of pixels to square millimeters. Females that lacked visible patches of red/orange/yellow plumage pigmentation were not photographed or analyzed for hue, saturation, and brightness.
# Quantification of coccidian endoparasitism
After being photographed, birds were individually housed in small cages with ad libitum access to seed and water within an indoor vivarium on the ASU-Tempe campus. Due to the diel shedding cycle of coccidian gametes (i.e., in the late afternoon and early evening; Brawner et al. 2000), we housed finches until 16:30 h, at which point we changed the paper in each bird’s housing cage and returned to the housing room at 17:30 h to scrape small (ca. 0.25 g) fecal samples from the cage paper into 1.5 mL screw cap Eppendorf tubes containing 1 mL of 2.2% potassium dichromate. Samples were stored in the dark at room temperature until analysis (1–2 months later). We measured presence and severity of coccidian infection using standard fecal float and microscope slide preparations (Brawner and Hill 1999; Giraudeau et al. 2014). Following prior work (Brawner et al. 2000; Surmacki and Hill 2014), coccidian oocyst number was estimated under a light microscope using a logarithmic scale (presence of no oocysts = a score of 0; 1–10 occysts = 1; 11–100 occysts = 2; 101–1,000 oocysts = 3; 1,001–10,000 occysts = 4; > 10,000 oocysts = 5). Samples were prepared and scored at random by 2 observers (B.E.S. and Emily Hopkins) and identified only by individual band number to avoid observer bias.
