774 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 123, No. 4. December 2011 
nestlings were 5 days of age. Nest holes were first testosterone and corticosterone were quantified 
blocked for 30 min prior to recording behavior to using a competitive binding radioimmunoassay 
ensure that no nestlings were satiated. Nestlings following Wingfield and Farner (1975). Testos- 
were removed from the nest box and placed in a terone recoveries averaged 60%; however, testos- 
cup with nesting material. A researcher stimulated 
the nestlings every 30 sec for 5 sec over the 
duration of 3.5 min by gently rattling the cup and 
whistling. 
One researcher (NS) quantified begging behav¬ 
ior of individual nestlings through measurements 
of begging intensity. We scored the maximum 
posture to measure intensity during each trial as 0 
= not begging; | = mouth open; 2 = mouth open, 
head back; 3 = mouth open, head back, neck 
stretched; 4 = mouth open, head back, neck 
stretched, back vertical. We summed the intensity 
rankings for each nestling to assess the overall 
begging intensity during the trial; thus, begging 
intensity includes both frequency and degree of 
begging. 
Reflectance Spectrometry. —One researcher 
(LS) recorded spectral data with an Ocean Optics 
S2000 spectrometer (range - 250-880 nm: Du¬ 
nedin, FL, USA) and a light source illuminated 
with both a deuterium bulb (IJV light source) and 
a tungsten-halogen bulb (visible light source). Wc 
generated reflectance data relative to a WS-2 
white standard (Labsphere Inc., North Sutton, NH, 
USA) and placed a micron fiber-optic probe at a 
90-degree angle 1 mm from the feather surface. 
Reflectance data were summarized by calculating 
two standard descriptors; brightness and UV- 
chroma. Brightness was calculated as the mean of 
the summed reflectance from 300 to 700 nm and 
UV-chroma as the proportion of the total 
reflectance that is in the ultraviolet range (1300— 
400/(300-700). 
Hormone Analyses.— Serum corticosterone and 
testosterone were extracted and separated using 
celite column chromatography following methods 
modified from Schwabl (1993). Briefly, 20 pi of 
serum was mixed with 3 ml diethyl ether, 
vortexed, and allowed to settle for 20 min' 
Samples were snap frozen and the liquid portion 
containing the steroid hormones was reserved and 
dried using a Ni stream. Samples were resus¬ 
pended in 1 ml of 10% ethyl acetate in isooctane 
(VanWatersRogers, Suwanec, GA, USA) after 
which steroid hormones were eluted through the 
columns in the following fractions of ethyl acetate 
in isooctane: dihydrotestosteronc 10%, testos¬ 
terone -20%, and corticosterone -30%. Samples 
were further dried using a N 2 stream, and 
terone was below detectable levels in all samples, 
and testosterone concentrations were not used in 
analyses. Corticosterone recoveries averaged 
70%, and intra-assay variation was 5.19. 
Statistical Analyses .—We tested whether laying 
order predicted hatching order and whether the 
sex ratio within broods differed between early- 
and latc-hatched nestlings using Generalized 
Linear Models with binomial error distribution 
and logit link. Wc performed Linear Mixed Effect 
Models to analyze the effects of early- versus late- 
hatching on begging behavior, corticosterone 
level, structural size, and plumage coloration. 
All models include gender of the nestlings and 
early-versus latc-hatching as fixed factors, brood 
size as a covariate, and nest as the random factor. 
Wc performed a mixed-effect model to analyze 
the effects of the early- versus late-hatching on 
juvenile body mass (measured at ages 2, 5. 8,11. 
and 14 days). The model included gender of 
nestlings and early- versus late-hatching as fixed 
lactors, age and brood size as covariales, and nest 
as the random factor. We used a stepwise 
backward procedure for simplification of the 
mixed models and tested two-way interactions 
between the covariate and the fixed factors. Data 
were normally distributed. None of the interaction 
terms was significant (P > 0.05) and interactions 
were removed from models. SPSS (2006: Version 
15.0) software was used to analyze data and all 
tests were two-tailed. 
RESULTS 
Brood Size and Incidence of Brood Reduction.— 
Seven of 31 clutches had a total of two nestlings 
each, nine had three nestlings, nine had four 
nestlings, and six had five nestlings. Seventeen 
percent of the 208 Eastern Bluebird nests we 
monitored experienced total nest failure. Eight 
percent of the 172 successful nests (at least I 
nestling fledged) had one or more nestlings die 
from apparent starvation. We were able to identify 
whether the dead was an early- or late-hatched 
nestling in nine nests; 44% were early-hatched 
nestlings and 56% W'ere late-hatched nestlings. 
Laying Order, Hatching Order. Hatching Span, 
and Nestling Gender. —Laying order predicted 
hatching order (Wald X 2 = 14.46, n = 31, P < 
0.001); 90% of late-laid eggs resulted in late- 
