Behav Ecol Sociobiol (2012) 66:419–431
DOI 10.1007/s00265-011-1288-x

ORIGINAL PAPER

Delayed maturation of multiple signals
in a migratory songbird
Ryan R. Germain & Matthew W. Reudink &
Peter P. Marra & Peter T. Boag & Laurene M. Ratcliffe

Received: 24 August 2011 / Revised: 24 October 2011 / Accepted: 31 October 2011 / Published online: 12 November 2011
# Springer-Verlag 2011

Abstract Many animals use multiple signals in sexual
communication, but our understanding of the interactions
between multiple signals, particularly in inexperienced
breeders, is limited. In birds, delayed plumage maturation
(DPM) is well documented; young birds appear duller than
adults, despite reaching sexual maturity. Age-related
changes in song structure are also common in songbirds,
though the extent to which songs of yearling males differ
from those of adults (delayed song maturation, DSM) and
its prevalence in species with DPM is unknown. We tested
for DSM in the mate-attraction (repeat) song of a species
with dramatic DPM, the American redstart (Setophaga
ruticilla). Repeat song structure of territorial yearling and
adult males differed significantly, based on discriminant
Communicated by T. Bakker
R. R. Germain : M. W. Reudink : P. T. Boag : L. M. Ratcliffe
Department of Biology, Queen_s University,
Kingston, ON K7L 3N6, Canada
M. W. Reudink : P. P. Marra
Smithsonian Conservation Biology Institute,
Migratory Bird Center, National Zoological Park,
3001 Connecticut Ave. NW,
Washington, DC 20008, USA
Present Address:
R. R. Germain (*)
Centre for Applied Conservation Research,
University of British Columbia,
2424 Main Mall,
Vancouver, BC V6T 1Z4, Canada
e-mail: r.germain@alumni.ubc.ca
Present Address:
M. W. Reudink
Department of Biological Sciences, Thompson Rivers University,
900 McGill Rd.,
Kamloops, BC V2C 5N3, Canada

analysis of nine non-collinear song features. Combined
with previous evidence of delayed maturation in territorial
(serial) song, we provide the first evidence for DSM in
different song types used in different behavioral contexts
during the breeding season of a migratory songbird. Within
adults, variation in repeat song was associated with pairing,
earlier onset of breeding, and number of offspring sired,
suggesting a potential benefit for expressing more adult-like
song. We found no relationship between the expression of
adult-like repeat song and plumage in either age class.
These results indicate that delayed maturation of repeat
song in yearling redstarts is not due to a generalized delay
in signal development, and suggest further work is needed
to determine whether mate attraction (for both yearling and
adult males) provides greater pressure to sound more
“adult-like” than to look more “adult-like”.
Keywords Delayed song maturation . Delayed plumage
maturation . Multiple signals . Mate choice . American
redstart

Introduction
Many animals display and perceive more than one signaling
trait in sexual communication. Because animals communicate using multiple sensory systems, careful consideration
of multiple aspects of sensory emission is necessary to
understand the role of sexual signals (Partan and Marler
2005). Several hypotheses attempt to explain the evolution
of multiple signals, including the “redundant signal”
(multiple ornaments signal overall individual quality),
“multiple message” (different ornaments signal different
properties), and “unreliable signal” (uncorrelated signals
become exaggerated by runaway selection) hypotheses

420

(reviews in Møller and Pomiankowski 1993; Hebets and
Papaj 2005). While the interactions between multiple
signaling traits may differ across species, multiple signals
may also evolve via selection on multi-trait displays acting
as functional units (Hebets and Papaj 2005).
In birds, female preference for the elaborate songs and
plumage of males is well described (Andersson 1994).
However, because most studies focus on a single sensory
domain (vocal, visual), the processes affecting the relative
expression of song and plumage are poorly understood
(Shutler and Weatherhead 1990; de Repentigny et al. 2000;
Badyaev et al. 2002; Ornelas et al. 2009). Some studies
indicate that female selection for one trait will eventually
displace the other trait in a given environment (Shutler and
Weatherhead 1990; Badyaev et al. 2002), whereas others
suggest that such traits may be redundant (Møller and
Pomiankowski 1993; Møller et al. 1998; de Repentigny et
al. 2000). Alternatively, song and plumage signals may
reflect different parameters of individual condition (Read
and Weary 1992; Schluter and Price 1993; Iwasa and
Pomiankowski 1994), or be unrelated (Ornelas et al. 2009).
In North American passerine birds, sub-adults from more
than 30 species from at least eight families exhibit delayed
plumage maturation (DPM), a delay in the attainment of
their definitive adult plumage (Rohwer et al. 1980; Studd
and Robertson 1985). Despite numerous hypotheses for
DPM, there is no consensus about its function and adaptive
benefits (Cucco and Malacarne 2000; Karubian et al. 2008).
Young birds (usually males) in species with DPM have dull
or female-like plumage throughout their first year (Rohwer
et al. 1980). In most instances, such young males are
sexually mature and able to reproduce during their first
breeding season, although their reproductive success is
typically much lower than adults (Ficken and Ficken 1967;
Rohwer et al. 1980; Omland and Sherry 1994; Payne 1982;
Muehter et al. 1997). Although delayed expression of
sexual signals in otherwise sexually mature males is found
in a number of taxa including lizards (Martín and Forsman
1999) and primates (Setchell 2003), results from bird
studies contribute to the majority of our understanding of
how sexual signals change over an individual's lifetime.
In a review of age-related signaling in birds, Cucco and
Malacarne (2000) suggested that DPM may reflect a more
generalized delay in sexual signaling traits. In a comparison
of 137 Western Palearctic passerines, the authors found that
for the 29 species where data were available, eight
exhibited a distinct song in their first breeding season and
underwent significant changes (predominantly through
additions/deletions to their song repertoire) between their
first and second years, whereas 21 did not. They termed
those species in which the songs of first-year birds differ
significantly from those of older birds, but adult song does
not change with age, as undergoing delayed song matura-

Behav Ecol Sociobiol (2012) 66:419–431

tion (DSM). Cucco and Malacarne (2000) noted that of the
eight species exhibiting DSM, seven also exhibited DPM.
Although the observed co-occurrence between song and
plumage signaling traits is intriguing, the authors cautioned
that studies demonstrating a clear difference in song
between yearling and adult males are relatively scarce.
More research is needed to understand both the proximate
basis and adaptive/evolutionary significance of DSM.
While the mechanisms underlying DSM are unclear,
learning almost certainly plays a central role. Here, we
use the term DSM in a manner consistent with previous
work (Cucco and Malacarne 2000) to describe species
exhibiting a significant difference between the songs of
first-year and older birds, with little or no change to song
features in subsequent years. Our usage of the term does
not imply that observed song changes must necessarily
result from maturation; such changes may also be the
outcome of processes such as vocal production learning.
The majority of studies considering age-based changes
in song examine additions or deletions to song repertoire
between the first and second breeding seasons, particularly
in large-repertoire species (e.g., Kiefer et al. 2006, 2009).
Recent studies have begun to accumulate examples of agebased changes in performance-related structural song traits
(reviewed in Kipper and Kiefer 2010); however, very few
have included species with marked DPM (but see Cucco
and Malacarne 1999). If DPM and DSM do co-occur, and
the delayed maturation of one signaling trait is due to a
generalized delay in signal development, as theorized by
Cucco and Malacarne (2000), we would predict a relationship between the relative expression of each signaling trait.
The most obvious way to test for such a relationship is a
comparison of the song and plumage traits associated with
delayed maturation. This could help determine if birds of
either age class with a more adult-like expression of one
trait show similar levels of development in the other
(positive association), or if expression of one signaling
trait comes at the expense of the other (trade-off).
Alternatively, comparing each signaling trait with pairing
and reproductive success may offer insight into the
selective pressures (costs/benefits) of expressing a more
developed signal and may help elucidate the targets of
female mate choice in species with multiple sexual signals.
Although several studies have investigated plumage and
reproductive success in yearling males with DPM (e.g.,
Greene et al. 2000; Karubian 2002), our understanding of
how differences in adult/yearling song might be reflected in
the reproductive success of species with delayed signaling
development is limited. Furthermore, because song functions in dual roles (mate choice and territorial defense) in
many species, past studies examining age differences in
song changes may not have been able to distinguish the
significance of such changes to neighboring males and

Behav Ecol Sociobiol (2012) 66:419–431

females. Therefore, it is useful to examine the relation
between DPM and DSM in a species where songs directed
at females are clearly discernable from those directed at
conspecific males.
The American redstart (Setophaga ruticilla) offers an
ideal model for comparing age-related development in male
plumage and song-based sexual signals. American redstarts
are a highly ornamented, sexually dichromatic species of
Parulid warbler, which exhibits marked DPM. Adult males
(after-second-year, ASY) are predominantly black (melanin
pigment-based plumage), with orange coloration on their
wings, tail, and flanks (Sherry and Holmes 1997). Yearling
(sub-adult, second-year, SY) males appear female-like
during their first winter and breeding seasons and are
greenish-gray with yellow plumage coloration instead of
the orange of adult males on the wings, flanks, and tail. The
orange/yellow plumage of American redstarts is carotenoid
pigment-based (McGraw et al. 2005), and recent work
indicates that color variation in adult males is associated
with mate choice (Reudink et al. 2009a; Germain et al.
2010a) and winter territory acquisition (Reudink et al.
2009b). Individual variation in yearling male plumage can
also be pronounced; some yearling males exhibit various
degrees of adult-like orange coloration in their normally
yellow plumage regions and/or exhibit irregular patches of
adult-like black plumage (primarily on the head and breast)
as a result of adventitious feather loss (Rohwer et al. 1983;
Sherry and Holmes 1997).
Song may also differ between yearling and adult male
American redstarts. The repertoire of this species, like
many closely related Parulid warblers, contains two distinct
song categories that are used in different social contexts:
repeat and serial song (Ficken and Ficken 1965; Lemon et
al. 1994; Staicer et al. 2006). Repeat song (one song sung
continuously in repetitive manner) is widely thought to
function in mate attraction for Parulid warblers (Kroodsma
1981; Kroodsma et al. 1989). Repeat song in American
redstarts is sung more commonly early in the breeding
season before and during female arrival, whereas serial
songs are more commonly used later in the breeding season
during male–male aggressive territorial interactions (Ficken
and Ficken 1965; Lemon et al. 1985, 1987; Staicer et al.
2006). Serial songs can be comprised of all notes and
remaining (other than repeat) song types in an individual's
repertoire and are sung with immediate variety. Although
adult male redstarts can change their serial song repertoire
by adding and deleting new notes and songs from year to
year to match those of neighbors (Ficken and Ficken 1965;
Lemon et al. 1985, 1987, 1994); Lemon et al. (1994)
determined that birds were more likely to make changes to
their repertoires after their first year than in subsequent
years. Age-related changes in repeat song of redstarts,
however, have not been documented. Most male American

421

redstarts conform to the typical repeat song described in
Lemon et al. (1987); however, the composition (here
defined as the distinct notes which comprise an individual's
song) of repeat song may vary considerably between
individuals (Fig. 1), and some males may incorporate notes
into their repeat song that are used in the serial repertoire of
others (Lemon et al. 1985, 1987). Once males acquire
territories during their first breeding season, the composition of repeat song typically remains unchanged from year
to year (i.e., males do not add or delete repeat song notes
across years, Lemon et al. 1994). Some evidence suggests
that yearling males sing repeat songs at a higher rate than
adults (Tourangeau 2002; but see Procter-Gray and Holmes
1981; Staicer et al. 2006), but no study has investigated the
structural aspects of repeat song in the American redstart
across age classes or determined the aspects of repeat song
which may be involved in pairing and reproductive success.
In this study, we compared the structure and delivery of
repeat (mate-attraction) songs in yearling and adult males to
determine whether American redstarts exhibit delayed song
maturation. No differences in repeat song across age class
would indicate that age-based differences in song are
limited to the territorial (serial) song type in this species.
We then tested for a correlation between the relative
expression of “adult-like song” and “adult-like plumage”
for males in each age class. If a more adult-like expression
of one signaling trait corresponds to a more adult-like
expression in the other, this would be consistent with a
generalized delay in signal development, whereas a
negative correlation would suggest there is a trade-off.
Finally, we examined individual variance in both song and
plumage signals to determine predictors of pairing success
and onset of breeding in both age classes as well as realized
reproductive success in adult males.

Methods
Field work was conducted during the 2007–2008 breeding
seasons (May through July) at the Queen's University
Biological Station in south-eastern Ontario, Canada (44º34′N,
76º19′W). We walked daily transects across the 100-ha study
site to determine arrival date and breeding status for all
individual color banded adult (2007, n=54, 2008, n=47) and
yearling (2007, n=15, 2008, n=14) males. Arrival date was
calculated as the number of days after the arrival of the first
male each season to standardize across years (Reudink et al.
2009c). We monitored each male for 20–30 min/day to
determine pairing status and onset of breeding (date of social
partner's first egg). Date of first egg was standardized across
years in the same manner as arrival date. We recorded songs
of adult (2007, n=19, 2008, n=24) and yearling (2007, n=9,
2008, n=14) males for 20 min within 24–36 h after arrival

422

Behav Ecol Sociobiol (2012) 66:419–431

Fig. 1 Repeat song examples from four adult and four yearling male
American redstarts, illustrating repeat song diversity between and
within age class. Numbers represent phone categories: 1 =first
category syllable (FCS), 2=high frequency sweep (HFS), 3=terminal
accent note (TAN). Although most males exhibit repeat songs similar

to either the first and second song examples for each age class, the two
males W/P, Y/Al and Bk/W, Bk/Al illustrate extreme examples of
males incorporating notes into their repeat song typically used as
components of serial song in other males

on their territory, using a Marantz PMD670 digital recorder
(recordings saved as .wav files) and an Audiotechnica
directional microphone. All songs were recorded within a
distance of 6–8 m from the focal male. Recording sessions
took place between 06:00 and 10:00 EDT during the period
of male arrival, from May 4th–June 18th of each year. If a
newly arrived male was first discovered later in the day, we
recorded songs the following morning. For a subset of males
(adult, n=5, yearling, n=5), multiple 20-min recording bouts
were sampled opportunistically over a 4-week interval
(mean±SD=4.2±1.03 recordings/male, 5.83±2.59 days between recordings from same male) to test for repeatability in
repeat song structure across the breeding season.
In 2008, we also conducted weekly transects across a
series of 30–40 ha study plots unconnected to the main
study site (separated by ≥3 km), where yearling males (n=
10) were recorded opportunistically. In these distant plots,
we could not accurately determine arrival date for any of
the ten males and could only determine mating status for
one. Following song recording, we captured all males (both
main and peripheral plots) on their territories using mist
nets accompanied by a decoy and song playback, and

marked each bird with both a Canadian Wildlife Service
aluminum band and a unique combination of three color
bands. Males were aged following Pyle (1997), where the
rumps of yearling males are dark gray and contrast with the
pale gray back, while uniformly black adult males have no
distinct color contrast between the back and rump.
From each bird, we recorded unflattened wing chord length
(millimeters) as a measure of relative body size (Reudink et al.
2009a, b) and plucked both a single tail feather (rectrix 3)
and 12–15 flank feathers. We classified adult male bib size in
hand using a 1–5 scale developed by Lemon et al. (1992). In
addition, in 2008, we photographed yearling male breast and
head plumage (4–5 pictures/individual) on a gridded
background in a series of standard poses to quantify the
amount of black plumage present in each region.
While several previous studies of American redstarts
have used genetic paternity as the main measure of
reproductive success in adult males (e.g., Perreault et al.
1997; Reudink et al. 2009c), a large proportion of yearling
males fail to attract mates (Ficken 1962; Lemon et al. 1987;
Lozano et al. 1996). Of 21 yearling males with known
mating status in our study population (2007–2008), only 7

Behav Ecol Sociobiol (2012) 66:419–431

(33%) attracted females, only 4 (19%) had dates of first egg
that were recorded, and of these only one achieved
paternity for at least one chick (M. Reudink, unpubl. data).
Therefore, following Lozano et al. (1996), we used pairing
status (presence/absence of a female building a nest on the
male's territory), rather than date of first egg or paternity, as
the measure of “success” for yearling males.
In the subset of adult males included in this study (n=9,
all 2007) for which full measures of realized (genetic)
reproductive success were previously assessed (Reudink et
al. 2009a, c), we used total within-pair paternity and total
genetic offspring fledged as measures of reproductive
success. Paternity analyses (storage, extraction, amplification, assignment) follow Reudink et al. (2009c). Briefly,
paternity analysis was conducted using five microsatellite loci
(Dpμ01, Dpμ03, Dpμ05, Dpμ15, Dpμ16) originally isolated
from yellow warblers, Dendroica petechial (Dawson et al.
1997). Over the 4 years of study in which paternity was
assessed, Reudink et al. (2009c) analyzed DNA from 265
offspring from 75 nests and all putative parents. Total withinpair paternity was calculated as the number of offspring sired
at the primary and secondary (if polygynous) nests, and total
genetic offspring fledged as the number of fledged offspring
from both within-pair nesting attempts and extra-pair young.
Song analysis
Song recordings were digitized at a sampling rate of
22,050 Hz (16 bit format), and analyzed using the Syrinx
sound analysis program (Burt 2005). For each song
measured, we used a fixed resolution of 1.5 ms, and we
recorded frequency measures to ±0.01 kHz and time
measures to ±0.01 s. All variables of fine song structure
were measured manually using the time and frequency

423

cursor (following Foote et al. 2008) by the same observer
(RRG), and information about singer's age was masked
during analysis by using band number to code the sound
files. Each recording contained on average 87.3% (±7.6%)
vocalizations classified as repeat songs, where the remaining vocalizations were either single-note calls or “chips”
(vocalizations made while fighting with an intruding male).
Songs were considered repeat songs if they met all three of
the following criteria: (1) were those sung early in the
breeding season, before and during female arrival (Lemon
et al. 1985, 1987), (2) were sung repeatedly and continuously throughout the focal recording, and (3) matched the
focal recording (both by ear and sound spectrogram) on
successive days when the birds were visited. We analyzed
20-min recording bouts from a total of 74 individuals for
which we had both plumage and vocal data (adult, 2007=
19, 2008=24; yearling, 2007=9, 2008=22). Any deviations
from these sample sizes in further analyses are due to
missing data points for given variables and are described
throughout. We recorded and analyzed only one 20-min
singing interval from each male, as repeat song structure in
this species has been shown to be consistent throughout the
breeding season for adult males (Lemon et al. 1994). We
confirmed this result across age class using a subset of
individuals by comparing all song structural variables
(Table 1) from two recordings (>14 days apart) during the
same season (adult, n=5, yearling, n=5, two-tailed t test, all
p>0.34). From each 20-min recording bout, we calculated
song rate (number of songs/min) and subsequently analyzed
the highest quality (i.e., least amount of background or
overlap noise) songs (mean±SD=33±13 songs/bird out of a
mean 138±52 songs/recording) for seven measures of song
structure (Table 1). These measures included duration
(seconds), number of first category syllables (#FCS),

Table 1 Mean±SD of song rate, consistency, and ten structural measures of male American redstart repeat song (adult=43, yearling=31)
Variable

Abbreviation

Adult males

Yearling males

Song rate (songs/min)
Song duration (s)
Coefficient of variation duration
Number of first category syllables used in most frequent repeat song
Inverse proportion of total repeat songs that are non-standard in lengtha
Bandwidth of last first category syllable (kHz)
Bandwidth of high frequency sweep (kHz)
Bandwidth of terminal accent note (kHz)
Bandwidth of whole song (kHz)
Coefficient of variation-bandwidth of whole song
Frequency at maximum amplitude (kHz)
Coefficient of variation-frequency at maximum amplitude

Rate
Duration
CV duration
FCS
Consistency
BW-FCS
BW-HFS
BW-TAN
BW-whole
CV-whole
FMA
CV-FMA

6.45±2.69
1.10±0.14
0.10±0.12
4.55±1.13
29.13±52.95
2.41±0.53
0.78±0.66
2.99±1.42
4.63±0.53
0.06±0.05
7.11±0.46
0.07±0.03

7.70±2.36
1.03±0.15
0.06±0.05
4.37±0.85
30.31±44.46
2.74±0.62
0.63±0.55
2.92±1.44
4.66±0.64
0.06±0.03
7.36±0.59
0.07±0.03

a

Calculated as the inverse ratio of songs which had additional/fewer first category syllables (than the most frequent version of the song) to the
total number of songs

424

bandwidth (kilohertz) of last first category syllable (BWFCS), bandwidth (kilohertz) of high frequency sweep (BWHFS), bandwidth (kilohertz) of terminal accent note
(BW-TAN), bandwidth (kilohertz) of the entire song (BWwhole), and the frequency (kilohertz) at maximum amplitude
(FMA) (Fig. 1). In addition to mean measures of individual
song characteristics, we quantified the ability of males to
deliver stereotyped songs by calculating mean coefficients of
variation (CV) for song duration, frequency at maximum
amplitude, and bandwidth of the entire song (Table 1).
Finally, we calculated the overall consistency of individual
male repeat song by dividing the number of songs which had
additional/fewer first category syllables than the normal
(most frequent) version of the song by the total number of
songs. These ratios were then divided by 1 to produce the
reported “consistency” values. We considered song rate and
consistency as behavioral measures of repeat song, and
therefore did not include them in discriminant function
analysis of repeat song structure.
Plumage analysis
Standardized photos of 21 yearling males were uploaded
into Adobe® Photoshop® CS3 (v. 10.0) and analyzed
following Germain et al. (2010b). Briefly, we measured
the area of black (adult-like) plumage on the head (from
behind bill to nape, including lores, hereafter called head)
and breast (chin, throat, and breast, hereafter called breast)
and calculated the total area (square millimeter) of black
plumage visible on the breast and head of each yearling
male, controlling for body size.
Preparation and spectrophotometric analysis of collected
feathers from both yearling and adult males followed
Reudink et al. (2009a), using an Ocean Optics USB4000
spectrometer (Dunedin, FL, USA) attached to a PX-2
pulsed xenon light source. We took 25 readings throughout
the orange (adult) or yellow (yearling) region of both flank
and tail feathers, and gathered reflectance data into 1 nm
bins using CLR 1.03 (Montgomerie 2008). We also
calculated the extent of tail color patch size by measuring
the area (square millimeter) of yellow/orange color patches
on both sides of the rachis of collected tail feathers using
digital calipers (±0.01 mm). This measure was then divided
by the total area of the tail feather to control for feather size.
In instances where the color patch on tail feathers was too
small to gather accurate reflectance readings (adult, n=5,
yearling, n=6), or not enough flank feathers were collected
(adults only, n=2), we excluded that feather region from
spectrophotometric analysis. This yielded a final sample of
35 adult and 24 yearling males with song and tail
reflectance spectra data, and 38 adult and 30 yearling
males with song and flank reflectance spectra data. To
calculate the brightness of each plumage region, we

Behav Ecol Sociobiol (2012) 66:419–431

averaged the percent reflectance from 320–700 nm. We
then calculated the standard variables of hue, UV chroma,
and red chroma (saturation) using color equations described
in Germain et al. (2010b), adapted from Montgomerie
(2006).
Statistical analysis
All statistics were performed using JMP 8.0.1 (SAS
Institute 2009) and R 2.10.1 for Windows (R Development
Core Team 2009). In instances where data were collected
for the same individual across 2 years (n=4 adults in both
years, n=0 yearling-to-adult resamples), we randomly
excluded data from 1 year (determined by coin flip) to
avoid pseudoreplication. Data on morphology and song
were pooled across both years of study and tested for year
effects using t tests. Feathers collected in 2007 had
previously undergone spectrophotometric analysis as part
of several other studies on American redstart ornamentation. To remove any variance associated with year that
might be due to feather degradation from repeated measurements, we centered plumage color values by year with a
series of one-way ANOVAs (each color variable versus
year) and used the residuals from these tests in subsequent
analysis. All variables were tested for equal variance
(Levene's homogeneity of variance test) and normality
(Shapiro–Wilks test) before use in further analysis.
Age-based song and plumage separation
We used Pearson's correlation analysis to examine all song
structure and all plumage variables for multicollinearity.
For variables found to be highly collinear (i.e., r≥0.7:
McGarigal et al. 2000), we calculated univariate ANOVAs
for each variable and compared their F values. Variables
with the highest F value were retained while those with
lower F values were excluded from further analysis (Noon
1981; Herring et al. 2008). In addition, we used Mahalanobis distance outlier analysis to identify multivariate
outliers (song n=3, plumage n=0), which may represent
birds with unusual song features unlikely to be part of the
population of song features we wish to describe (Quinn and
Keough 2002) and were subsequently excluded from
discriminant analysis.
Song variables with non-normal distributions (CVduration, consistency, BW-FCS, BW-HFS, BW- TAN,
BW-whole, and CV-whole) were transformed (log10, third
root, log10, square root, square root, square root, log10,
respectively). Four of these transformed variables (log10
CV-duration, √BW-HFS, √BW-TAN, √BW-whole) still
failed to meet the assumptions of normality. However,
because non-normality in each of these cases was due to
skewness and not outliers, these transformed variables may

Behav Ecol Sociobiol (2012) 66:419–431

still be entered into discriminant analysis and the resultant
significance tests are reliable (Tabachnick and Fidell 1996;
Quinn and Keough 2002). Linear discriminant function
analysis (DFA) was then used to determine the best single
or combination of song structural characteristics that separated
males by age, using age class as the classification variable, and
nine non-collinear measures of song structure (duration, CVduration, FMA, CV-FMA, BW-FCS, BW-HFS, BW-TAN,
BW-whole, CV-whole) as predictor variables.
Most plumage variables fit the assumptions of normality,
except the standardized area of black head and breast
plumage on yearling males and flank hue (yearling and
adult males). We transformed black head and black breast
plumage by taking the square root of each variable.
However, flank hue was not transformed, as (1) it was
centered by year (residuals from univariate ANOVA, year
as independent variable) and thus contains a proportion of
individuals with negative values for this measure, and (2)
the non-normal distribution was once again due to
skewness and not significant outliers. For plumage-based
DFA of all males, we used age class as the classification
variable and six non-collinear measures of carotenoid
coloration for all males (tail color patch size, tail brightness,
tail red chroma, flank brightness, flank hue, flank red
chroma) as predictor variables.

425

ical variate scores) between paired and unpaired males of
each age class. We then compared date of first egg for
adults with individual plumage and song variables using a
series of multiple regressions with backward stepwise
deletion and canonical variate scores with linear regressions, respectively. Finally, we compared the individual
song and plumage characteristics of a small subset of adult
males (n=9, all from 2007) for which we had full measures
of genetic reproductive success (total within-pair paternity
and total genetic offspring fledged) using a series of linear
regressions. Because of the small sample size for adult male
realized reproductive success, we reduced the number of
variables in this analysis by including only the song and
plumage characteristics found to differ between yearling
and adult males (those significantly correlated with canonical scores obtained from discriminant analysis).
All tests were two-tailed (equal variance), and all models
were tested using Akaike's information criterion (AIC) to
ensure that final models were the most parsimonious.
Table-wide sequential Bonferroni corrections were applied
to control for the number of tests used.

Results
Year effects and age differences

Interactions of song and plumage
We performed an ANCOVA between song and plumage
canonical variate scores (obtained from DFA of plumage
and song variables, respectively) with age as a co-variate to
determine if, overall, individuals with more “adult-like”
song structure also had more “adult-like” plumage. In
addition, we conducted a series of linear regressions to
determine if song canonical score was related to the
melanin-pigment based plumage of adult (bib score) and
yearling (extent of adult-like black plumage) males, and if
plumage canonical score was related to singing behavior.

Adult male American redstarts captured in 2008 had larger
relative body size (wing length, R2 =0.16, t37 =2.63, p=
0.01) than those captured in 2007. Yearling males captured
in 2007 had higher song rates (R2 =0.13, t28 =−2.09, p=
0.05) and tended to have earlier standardized arrival on the
breeding grounds than those captured in 2008, although the
difference was not statistically significant (R2 =0.18, t18 =
2.00, p=0.06). Across age classes, adult males had longer
wing length (ANCOVA, year as covariate, F2,63 =28.65, p<
0.0001) and arrived earlier (R2 =0.47, t58 =7.18, p<0.0001)
than yearling males.

Predictors of pairing and reproductive success

Age-based song and plumage separation

Predictors of pairing success (binary response of mated/
unmated) were determined using a series of nominal
logistic regressions, with backwards stepwise elimination,
that were separated based on (1) age, morphology, and
arrival date, (2) song structure and singing behavior, and (3)
plumage coloration. Because yearling black breast and head
plumage data were collected only in 2008, we excluded
these variables from multivariate analysis. Instead, we used
t tests to determine if there was a significant difference in
the amount of black breast and head plumage between
paired and unpaired yearling males. We used t tests to
compare summary measures of song and plumage (canon-

A DFA significantly separated the structure of adult and
yearling American redstart repeat song based on canonical
variate scores (R2 =0.21, t65 =4.16, p<0.0001; Fig. 2a) and
predicted age class with 79.1% accuracy (29/37 adult
males, 24/30 yearling males). Canonical variate scores
were significantly negatively correlated with song duration
(R2 =0.13, F1,65 =9.68, p=0.003), CV duration (R2 =0.10,
F1,65 =6.96, p=0.01), and BW-HFS (R2 =0.06, F1,65 =4.02,
p=0.05), and significantly positively correlated with both
FMA (R2 =0.27, F1,65 =23.53, p<0.0001), and BW-FCS
(R2 =0.43, F1,65 =49.50, p<0.0001). In other words, the
repeat songs of individuals with higher (more “yearling-like”)

426

Behav Ecol Sociobiol (2012) 66:419–431

(R2 =0.67, F1,28 =56.36, p <0.0001). Yearling males with
more consistent songs tended to have lower CV-whole,
although this relationship was not statistically significant
(R2 =0.12, F1,28 =3.95, p =0.06). In other words, both
yearling and adult males who were most consistent in the
number of first category syllables included in their songs
also had the least variation in song duration, as expected.
Adult males with higher song output rate had larger BWWHOLE (R2 = 0.17, F1,38 = 7.63, p =0.009). No other song
structure variables were related to either measure of
singing behavior (all p >0.14).
The plumage coloration of adult and yearling males
differed significantly, based on DFA canonical variate
scores for all males (R2 =0.87, t57 =19.38, p<0.0001;
Fig. 2b). These scores predicted age class with 100%
accuracy (35/35 adult males, 24/24 yearling males). Across
age class, plumage canonical variate scores increased with
increasing tail brightness (R2 =0.08, F1,57 =4.77, p=0.03),
flank brightness (R2 =0.28, F1,57 =22.60, p<0.0001), and
flank hue (R2 = 0.91, F 1,57 = 566.12, p < 0.0001), and
decreased with tail color patch size (R2 =0.17, F1,57 =
11.58, p=0.001), tail red chroma (R2 =0.61, F1,57 =89.47,
p<0.0001), and flank red chroma (R2 =0.59, F1,57 =80.61,
p<0.0001). Thus, the overall carotenoid coloration of
individuals with higher (more “yearling-like”) canonical
variate scores were characterized by: (1) a smaller relative
area of coloration on the tail feather, (2) tail feathers that
were brighter and less saturated with red/orange pigment,
and (3) flank feathers that were brighter, less saturated
with red/orange pigment, and more “yellow-like” in hue.
There were no significant relationships between carotenoid canonical score and either bib score in adult males or
the area of black plumage on the breast or head of yearling
males (all p>0.35).
Interactions of song and plumage
Fig. 2 Box plots of (a) repeat song canonical variate score for adult
male (n=37) versus yearling male (n=30) repeat song, and (b)
plumage canonical variate scores for adult male (n=35) versus
yearling male (n=24) American redstarts. Horizontal lines represent
the median, 10th, 25th, 75th, and 90th percentiles, respectively

canonical scores were characterized by: (1) shorter mean
song duration, and more stereotyped song duration, (2)
higher mean FMA, (3) larger mean FCS bandwidth, and
(4) smaller mean HFS bandwidth. Yearling males tended
to sing more songs per minute than adult males, although
this relationship was not significant (ANCOVA, year as
covariate, F2,67 =2.74, p=0.07).
For song behavior within each age class, males with
more consistent songs had lower CV-duration in both adult
(R2 =0.33, F1,38 = 18.49, p=0.0001) and yearling males

While plumage and song canonical variate scores were
significantly positively correlated in the pooled sample of adult
and yearling males (Fig. 3), this relationship was predominantly due to age, which was the only significant predictor of
plumage and song score (ANCOVA, age as covariate, F3,52 =
7.25, pplumage =0.10, page =0.002, page*plumage =0.41). Variance
of both song and plumage scores did not differ across age
class, and variance did not differ between signaling
traits within each age class (all p>0.17). We found no
significant relationships between plumage canonical score
and measures of singing behavior (song rate, consistency)
in either age class or between song canonical score and
bib score in adult males (all p > 0.38). Similarly, in
yearling males, we found no relationship between song
canonical score and the extent of adult-like black on head
or breast plumage (all p>0.17).

Behav Ecol Sociobiol (2012) 66:419–431

Fig. 3 Song and plumage canonical scores for adult (solid circle) and
yearling (open circle) male American redstarts. Each axis extends from
lower “adult-like” scores to higher “yearling-like” scores. While song
and plumage scores are correlated across age class (R2 =0.15, F1,54 =
9.82, p=0.003), there is no linear relationship between the two traits
within adult (R2 =0.02, F1,30 =0.48, p=0.5) or yearling males (R2 =0.10,
F1,22 =2.34, p=0.14)

Predictors of pairing and reproductive success
Logistic regression of pairing success using all males
revealed, as expected, that age class was the strongest
predictor of pairing success (Table 2). Because of the strong
relationship between age and wing length (“year and age
differences”, above), relative body size was not entered into
the model of age, morphology, and arrival date. Due to the
strong effect of age on pairing success (Table 2), we also
separated models by age class to determine whether any of
the tested predictors of pairing success differed across age.
When separated by age, both BW-FCS and FMA song

427

variables were associated with pairing success in adult, but
not yearling, males (Table 2). Post-hoc analysis of these
variables revealed that paired adult males had lower mean
FMA (two-tailed t test; R2 =0.11, t37 =−2.18, p=0.04), and
larger BW-FCS (two-tailed t test; R2 =0.11, t37 =2.10, p=
0.04) than unpaired adult males, where the latter result is
inconsistent with the previous result (Table 2) of larger
BW-FCS being a more yearling-like song trait. After tablewide Bonferroni correction (α=0.001), only age remained a
significant predictor of pairing success. T tests revealed no
significant difference in plumage or song canonical scores
between paired and unpaired males of either age class
(all p>0.32). In addition, there was no significant difference
between paired (n=4) and unpaired (n=8) yearling males in
the amount of either black breast plumage or black head
plumage (two-tailed t tests, both p>0.35).
Stepwise regressions comparing date of first egg with
individual measures of adult plumage coloration and song
revealed that adult males with duller and less saturated tail
feathers began nesting earlier (Table 3), although only tail
brightness was significant after table-wide correction (α=
0.003). There were no significant predictors of date of first
egg with individual measures of song structure or singing
behavior. Comparisons of date of first egg with overall
plumage and song development (canonical scores) revealed
that song score was a significant predictor of onset of
breeding, where adult males with more adult-like song
scores began nesting earlier (R2 =0.26, F1,16 =5.59, p=
0.03), but that there was no relationship between plumage
score and date of first egg (p=0.77).
Linear regressions of adult realized reproductive success
with plumage and song variables that significantly contributed to age-based separation of each trait (six correlates of
plumage canonical score, song rate, and five correlates of
song canonical score) revealed that males which fledged a
greater number of total genetic offspring had lower mean
FMA (R2 =0.57, F1,7 =9.37, p=0.02). This relationship,

Table 2 Final models of nominal logistic regressions examining pairing success in relation to age, arrival date, plumage coloration, and song
structure/behavior using stepwise backward deletion
Predictor variables
All males

Age, arrival date (final model, χ2 =14.25, n=60, p=0.0002)

Adult

Song variables (final model, χ2 =11.15, n=39, p=0.004)

Yearling
Adult
Yearling

Song variables (final model, n=21)
Plumage coloration (final model, n=39)
Plumage coloration (final model, n=21)

β

Intercept
−0.41
Age
−1.11
Intercept
−12.74
BW-FCS
−20.70
FMA
2.54
No significant predictors
No significant predictors
No significant predictors

SE

χ2

p

0.31
0.31
8.82
10.31
1.24

1.76
12.61
2.08
4.03
4.16

0.19
0.0004
0.15
0.04
0.04

Adult males had significantly greater pairing success than yearlings. Within adults, birds with lower frequencies at maximum amplitude and
higher bandwidths of the first category syllable paired more successfully

428

Behav Ecol Sociobiol (2012) 66:419–431

Table 3 Final models of multiple regressions examining standardized date of first egg in relation to male plumage coloration and song structure/
behavior, using stepwise backward deletion

Song variables (final model n=21)
Plumage variables (final model, R2 =0.41, F2,16 =5.52, p=0.02)

however, was not significant after sequential Bonferroni
correction (α=0.002). In addition, adult males which sired
more within-pair offspring tended to have lower mean
FMA (R2 =0.35, F1,7 =3.77, p=0.09) and decreased flank
red chroma (R2 =0.34, F1,7 =3.57, p=0.10), although these
differences were not significant. Comparisons of adult
realized reproductive success with overall plumage and
song scores revealed no significant relationship between
realized success and repeat song canonical score (p=0.84),
and a marginally significant relationship with plumage
canonical score (R2 =0.44, F1,7 =5.55, p=0.051).

Discussion
In this study, we analyzed the structure of mate-attraction
(repeat) songs across age class in the American redstart and
found that the repeat songs of young males differed
significantly from those of adults. We determined that a
more adult-like song was correlated with an earlier onset of
breeding in adult males and that two of the structural
features of repeat song that most significantly contributed to
its separation across age class (frequency at maximum
amplitude and bandwidth of the first category syllable)
were correlates of adult pairing and genetic reproductive
success (before Bonferroni correction). Individual variables
of adult and yearling plumage were not linked with pairing
or reproductive success but were associated with onset of
breeding in adult males. Although the exact nature of the
relationships between the structural aspects of repeat song
and reproductive success are unclear (i.e., adult pairing
success correlated with one adult-like and one yearling-like
song trait) and require experimental tests, our results
provide the first evidence, to our knowledge, of delayed
maturation of a song type used principally during mate
attraction in a species with delayed plumage maturation.
We are not aware of any study since Cucco and
Malacarne's (2000) investigation that has specifically
addressed the relationship between plumage ornamentation
and song in yearling males of a species with DPM.
Combined with previous evidence of delayed maturation
in serial song of American redstarts (Lemon et al. 1994),
our finding of a small yet significant age-based separation
in repeat song indicates that DSM can occur in different

Predictor variables

Β

F

p

No significant predictors
Tail brightness

1.85

9.76

0.007

Tail red chroma

455.69

5.96

0.03

song types used in different contexts during the breeding
season. Although playback experiments are needed to
determine whether male and female redstarts perceive these
differences between yearling and adult repeat songs, our
results concur with the prediction that DSM is more likely
to occur in species with DPM (Cucco and Malacarne 2000).
However, contrary to the predictions of Cucco and
Malacarne (2000), we did not find evidence that the
delayed maturation of one signaling trait may represent a
more generalized delay in the development of secondary
sexual signals.
The development of repeat song in American redstarts
has not been studied and so the roles of learning from
others and practice in song maturation is unknown. Adult
males do not change their repeat song from year-to-year to
copy those of neighbors (Lemon et al. 1994), and neither
adults (Lemon et al. 1994) nor yearlings (this study) appear
to change their repeat song throughout the breeding season.
In contrast to this lack of information on repeat song
learning, detailed work by Lemon et al. (1994) shows that
male redstarts make significant changes to their serial song
repertoire between their first and second breeding seasons,
but not as frequently in subsequent years, mainly as a result
of males copying (through additions or deletions of two or
more songs) the serial songs of their neighbors, forming
song neighborhoods. Longitudinal studies of repeat song
recorded from individual males as yearlings and adults
could provide useful insights into the song development
process. Additionally, such studies could clarify whether
the marginally significant higher song rates of yearlings that
we detected were due to breeding status (mostly unpaired
subjects) rather than age. For example, while Staicer et al.
(2006) reported higher repeat song rates in unpaired than
paired male American redstarts of both age classes, neither
that study nor Procter-Gray and Holmes (1981) found
significant age effects on song rate.
In comparing how song and plumage varied with
reproductive success, we found that adult male redstarts
with more adult-like repeat songs began breeding earlier
each season than those with more yearling-like songs. In
addition, adult males with lower mean frequencies at
maximum amplitude (a more adult-like song characteristic)
had greater pairing success and fledged more genetic
offspring (though the latter result was not significant

Behav Ecol Sociobiol (2012) 66:419–431

following Bonferroni correction). The emergence of this
pattern despite a small sample size suggests that variation in
repeat song is associated with breeding success and
provides support for the role of repeat song in female mate
choice, which while historically predicted, has yet to be
explicitly tested in American redstarts. Interestingly, while
age-based song score was correlated with timing of breeding
in adult males, we found no relationship between song
canonical score and genetic reproductive success, despite
associations between this measure of success and individual
adult male song characteristics. These patterns suggest that
while repeat song may function in mate attraction, its role in
extra-pair behavior may be limited, compared to that of
plumage coloration (Reudink et al. 2009a).
With respect to plumage, both adult male tail brightness
and tail red chroma were associated with first-egg date,
where adult males with duller and less saturated feathers
bred earlier. Earlier onset of breeding is typically found to
have a strong positive influence on annual reproductive
success, both in this population (Norris et al. 2004) and
many passerine species (reviewed in Dunn 2004). While
less saturated tail feathers in adults may be indicative of
older, more experienced adult breeders (Reudink et al.
2009a), previous studies have found adult males with
brighter tails arrive earlier on the breeding grounds and
are more likely to be polygynous and that flank coloration
is more strongly associated with paternity in American
redstarts, based on larger sample sizes (Reudink et al.
2009a, b).
We did not sample the territorial (serial) songs of males
to compare with reproductive success, and it is important to
note that territory quality may also play an important role in
female mate choice. The acquisition of high-quality
territories, presumably influenced by earlier arrival dates,
may be more strongly related to male age, plumage
coloration, and winter territory quality (Reudink et al.
2009a, b, c) than to serial song types sung later in the
breeding season (Lozano et al. 1996), but additional studies
are needed to make direct comparisons. Future work should
also seek to determine if individuals differentially delay the
development of different song types (serial vs. repeat song)
during their first year, as the benefits of sounding more
adult-like to females through repeat song may be augmented by sounding more yearling-like (and less threatening) to
conspecific males through serial song.
Overall, we detected a stronger coupling of pairing and
genetic reproductive success with individual measures of
song, rather than plumage, in this age-based set of
comparisons (e.g., Table 2), and that canonical scores of
adult-like song and plumage are each associated with some
measure of “success” during the breeding season (song
scores with onset of breeding, plumage scores with realized
reproductive success). Adult and yearling male American

429

redstarts are easily identified based on their plumage
characteristics, which are very distinct (Fig. 2b). In this
age-biased mating system, ornamental adult plumage
coloration has previously been shown to be a correlate of
individual male quality (Reudink et al. 2009a). While
repeat songs are significantly different between age classes
(Fig. 2a), there is less separation between yearling and adult
males in this signal and a greater potential for individuals of
one group to be statistically misidentified to the other based
on their song. Therefore, there may be a greater pressure for
males of each age class to sound more adult-like than to
look more adult-like. By expressing a more adult-like song,
yearling males could provide a more current indicator of
individual condition during the timing of mating (as
opposed to timing of molt with plumage), allowing females
to better assess male quality based on several signaling traits
(Cucco and Malacarne 2000). We were unable to directly test
these ideas in this study, due to the relatively small number
of successfully paired yearlings (n=7 over 2 years of
monitoring), and the relatively low return rates of males
banded as yearlings in this system (none in this study,
approx. 9% from 2001 to 2008, L. Ratcliffe, unpubl. data).
Monitoring yearling male success over additional years,
especially in years of lower adult density, will be required
to better determine the influence of song and/or plumage on
female mate choice decisions as they influence yearling
reproductive success. For example, in a long-term study of
indigo buntings (Passerina cyanea), first-year males that
matched the individual songs of local adults and were able
to maintain territories near their song models had greater
mating success, were in higher-quality habitats, and were
better able to defend their territories from intruders (Payne
1982; Payne et al. 1988). Further, Payne (1982) showed
that first-year males with more adult-like blue plumage also
had greater pairing/nesting success, and those which
expressed both bluer plumage and matched the songs of
local adults could achieve the success levels of adult males.
The potential benefits of adult-like signaling are echoed in
several studies focusing on only one signaling trait, which
demonstrate that having more adult-like song is advantageous for young males (Hasselquist et al. 1996), as is
plumage that appears more adult-like (Karubian 2002).
Alternatively, several theories predict that reduced aggression by adult males leads to greater access to females for
subadult males with less-developed secondary sexual traits
(Rohwer et al. 1980; Lyon and Montgomerie 1986; Foster
1987). Additionally, in some cases, there is evidence of
disruptive selection for yearling appearance, as in lazuli
buntings (Passerina amoena), where yearling males with
the brightest (adult-like) and dullest (female-like) plumage
were more successful (i.e., defending high-quality territories,
pairing, siring offspring) than those with intermediate plumage (Greene et al. 2000).

430

We did not detect any relationship between the degree of
adult-like song and adult-like plumage expressed by males
of either age class (Fig. 3), suggesting that song and
plumage signaling traits may be unrelated in this species.
These findings are similar to those of Ornelas et al. (2009)
who detected no evidence of either an association or tradeoff between acoustic and visual signaling in trogons
(Trogonidae). Of these two potential alternative outcomes,
trade-offs between song and plumage traits are more likely
than positive correlations within wild populations, as
energetic resources are likely to be normally distributed
(Shutler 2011). Although several studies have tested for
trade-offs between song and plumage, there is limited
evidence for their prevalence (Shutler and Weatherhead
1990; Badyaev et al. 2002; Ornelas et al. 2009; Shutler
2011). Plumage coloration has been advocated as an
important mate choice signal in this species, particularly
with respect to achieving polygyny and retaining withinpair paternity (Reudink et al. 2009a); however, the role of
repeat song in mate choice is less well known. Our results
suggest that individual variation in American redstart repeat
song does play a role in female mating decisions, but the
specific targets of female choice for this signaling trait
remain unknown. While further investigations are necessary
to elucidate the interactions between song and plumage
signaling traits, and experimental evidence is needed to
directly test the response by females to variation in each
trait across age groups, our results illustrate an age-based
separation in the mate-attraction song of the species, and
support the notion that investment in sounding more adultlike may be beneficial for yearling males.
Acknowledgments We thank R. Montgomerie and T. Murphy for
providing color-analysis equipment and support, C. Scott for paternity
analysis, J. Foote for support in song analysis, and C. Toth, M.
Osmond, and many capable assistants in the field. We thank K.
Delmore, R. Robertson, P. Martin, W. Plaxton, and K. Munhall for
useful discussion and helpful comments on this manuscript. Funding
for this work was provided by Queen's University, Natural Sciences
and Engineering Research Council, National Science Foundation,
Ontario Government (MTCU), American Ornithologists' Union,
Society of Canadian Ornithologists, Sigma Xi, and the American
Museum of Natural History.
Ethical standards Experiments in this study comply with the
current laws of the countries in which they were performed.
Conflict of interest The authors declare that they have no conflict of
interest.

References
Andersson M (1994) Sexual selection. Princeton University Press,
Princeton
Badyaev AV, Hill GE, Weckworth BV (2002) Species divergence in
sexually selected traits: increase in song elaboration is related to

Behav Ecol Sociobiol (2012) 66:419–431
decrease in plumage ornamentation in finches. Evolution
56:412–419
Burt J (2005) Syrinx, version 2.6. University of Washington, Seattle.
Available at http://www.syrinxpc.com
Cucco M, Malacarne G (1999) Is the song of black redstart males an
honest signal of status? Condor 101:689–694
Cucco M, Malacarne G (2000) Delayed maturation in passerine birds:
an examination of plumage effects and some indications of a
related effect in song. Ethol Ecol Evol 12:291–308
Dawson RJG, Gibbs HL, Hobson KA, Yezerinac SM (1997) Isolation
of microsatellite DNA markers from a passerine bird, Dendroica
petechia (the yellow warbler), and their use in population studies.
Heredity 79:506–514
de Repentigny Y, Ouellet H, McNeil R (2000) Song versus plumage in
some North American oscines: testing Darwin's hypothesis.
Ecoscience 7:137–148
Dunn PO (2004) Breeding dates and reproductive performance. In:
Møller AP, Fiedler W, Berthold P (eds) Birds and climate
change. Advances in ecological research, vol. 35. Academic,
London, pp 69–87
Ficken MS (1962) Agonistic behavior and territory in the American
redstart. Auk 79:607–632
Ficken MS, Ficken RW (1965) Comparative ethology of the chestnutsided warbler, yellow warbler, and American redstart. Wilson
Bull 77:363–375
Ficken MS, Ficken RW (1967) Age-specific differences in breeding
behavior and ecology of American redstart. Wilson Bull 79:188–
199
Foote JR, Fitzsimmons LP, Mennill DJ, Ratcliffe LM (2008) Male
chickadees match neighbors interactively at dawn: support for the
social dynamics hypothesis. Behav Ecol 19:1192–1199
Foster MS (1987) Delayed maturation, neoteny, and social system
differences in two manakins of the genus Chiroxiphia. Evolution
41:547–558
Germain RR, Marra PP, Kyser TK, Ratcliffe LM (2010a) Adult-like
plumage coloration predicts winter territory quality and timing of
arrival on the breeding grounds of yearling male American
redstarts. Condor 112:676–682
Germain RR, Reudink MW, Marra PP, Ratcliffe LM (2010b)
Carotenoid-based male plumage predicts parental investment in
the American redstart. Wilson J Ornithol 122:318–325
Greene E, Lyon BE, Muehter VR, Ratcliffe L, Oliver SJ, Boag PT
(2000) Disruptive sexual selection for plumage coloration in a
passerine bird. Nature 407:1000–1003
Hasselquist D, Bensch S, von Schantz T (1996) Correlation between
male song repertoire, extra-pair paternity and offspring survival
in the great reed warbler. Nature 381:229–232
Hebets EA, Papaj DR (2005) Complex signal function: developing a
framework of testable hypotheses. Behav Ecol Sociobiol 57:197–
214
Herring G, Gawlik DE, Beerens JM (2008) Sex determination for the
great egret and white ibis. Waterbirds 31:298–303
SAS Institute (2009) JMP version 8.0.1. Duxbury, Pacific Grove
Iwasa Y, Pomiankowski A (1994) The evolution of mate preferences
for multiple sexual ornaments. Evolution 48:853–867
Karubian J (2002) Costs and benefits of variable breeding plumage in
red-backed fairy-wren. Evolution 56:1673–1682
Karubian J, Sillett TS, Webster MS (2008) The effects of delayed
plumage maturation on aggression and survival in male redbacked fairy-wrens. Behav Ecol 19:508–516
Kiefer S, Spiess A, Kipper S, Mundry R, Sommer C, Hultsch H, Todt
D (2006) First-year common nightingales (Luscinia megarhynchos)
have smaller song-type repertoire sizes than older males. Ethology
112:1217–1224
Kiefer S, Sommer C, Scharff C, Kipper S, Mundry R (2009) Tuning
towards tomorrow? Common nightingales Luscinia megarhynchos

Behav Ecol Sociobiol (2012) 66:419–431
change and increase their song repertoires from the first to the
second breeding season. J Avian Biol 40:231–236
Kipper S, Kiefer S (2010) Age-related changes in birds' singing styles:
on fresh tunes and fading voices? Adv Study Behav 41:77–118
Kroodsma DE (1981) Variation and functions of song types in
warblers (Parulidae). Auk 98:743–751
Kroodsma DE, Bereson RC, Byers BE, Minear E (1989) Use of song
types by the chestnut-sided warbler: evidence for both intra- and
inter-sexual functions. Can J Zool 67:447–456
Lemon RE, Cotter R, MacNally RC, Monette S (1985) Song repertoires
and song sharing by American redstarts. Condor 87:457–470
Lemon RE, Monette S, Roff D (1987) Song repertoires of American
warblers (Parulinae): honest advertising or assessment? Ethology
74:265–284
Lemon RE, Weary DM, Norris KJ (1992) Male morphology and
behavior correlate with reproductive success in the American
redstart (Setophaga ruticilla). Behav Ecol Sociobiol 29:399–403
Lemon RE, Perreault S, Weary DM (1994) Dual strategies of song
development in American redstarts, Setophaga ruticilla. Anim
Behav 47:317–329
Lozano GA, Perreault S, Lemon RE (1996) Age, arrival date and
reproductive success of male American redstarts Setophaga
ruticilla. J Avian Biol 27:164–170
Lyon BE, Montgomerie RD (1986) Delayed plumage maturation in
passerine birds: reliable signaling by subordinate males. Evolution
40:605–615
Martín J, Forsman A (1999) Social costs and development of nuptial
coloration in male Psammodromus algirus lizards: an experiment.
Behav Ecol 10:396–400
McGarigal K, Cushman S, Stafford S (2000) Multivariate statistics for
wildlife and ecology research. Springer, New York
McGraw KJ, Hudon J, Hill GE, Parker RS (2005) A simple and
inexpensive chemical test for behavioral ecologists to determine
the presence of carotenoid pigments in animal tissues. Behav
Ecol Sociobiol 57:391–397
Møller AP, Pomiankowski A (1993) Why have birds got multiple
sexual ornaments? Behav Ecol Sociobiol 32:167–176
Møller AP, Saino N, Taramino G, Galeotti P, Ferrario S (1998)
Paternity and multiple signaling: effects of a secondary sexual
character and song on paternity in the barn swallow. Am Nat
151:236–242
Montgomerie R (2006) Analyzing colors. In: Hill GE, McGraw KJ
(eds) Bird coloration, vol 1, Mechanisms and measurements.
Harvard University Press, Cambridge, pp 90–147
Montgomerie R (2008) CLR, version 1.03. Queen's University,
Kingston. Available at http://post.queensu.ca/~mont/color/
Muehter VR, Greene E, Ratcliffe L (1997) Delayed plumage
maturation in lazuli buntings: tests of the female mimicry and
status signalling hypotheses. Behav Ecol Sociobiol 41:281–290
Noon BR (1981) The distribution of an avian guild along a temperate
elevational gradient: the importance and expression of competition.
Ecol Monogr 51:105–124
Norris DR, Marra PP, Kyser TK, Sherry TW, Ratcliffe LM (2004)
Tropical winter habitat limits reproductive success on the
temperate breeding grounds in a migratory bird. Proc R Soc
Lond B 271:59–64
Omland KE, Sherry TW (1994) Parental care at nests of two age
classes of male American redstart: implications for female mate
choice. Condor 96:606–613
Ornelas JF, Gonzalez C, Espinosa de los Monteros A (2009)
Uncorrelated evolution between vocal and plumage coloration
traits in the trogons: a comparative study. J Evol Biol 22:471–484
Partan SR, Marler P (2005) Issues in the classification of multimodal
communication signals. Am Nat 166:231–245

431
Payne RB (1982) Ecological consequences of song matching:
breeding success and intraspecific song mimicry in indigo
buntings. Ecology 63:401–411
Payne RB, Payne LL, Doehlert SM (1988) Biological and cultural
success of song memes in indigo buntings. Ecology 69:104–
117
Perreault S, Lemon RE, Kuhnlein U (1997) Patterns and correlates of
extrapair paternity in American redstarts (Setophaga ruticilla).
Behav Ecol 8:612–621
Procter-Gray E, Holmes RT (1981) Adaptive significance of delayed
attainment of plumage in male American redstarts: tests of two
hypotheses. Evolution 35:742–751
Pyle P (1997) Identification guide to North American birds, part I.
Braun-Brumfield Inc, Ann Arbor
Quinn GP, Keough MJ (2002) Experimental design and data analysis
for biologists. Cambridge University Press, New York
R Development Core Team (2009) R: a language and environment for
statistical computing. R Foundation for Statistical Computing,
Vienna
Read AF, Weary DM (1992) The evolution of bird song: comparative
analyses. Philos Trans R Soc Lond B 338:165–187
Reudink MW, Marra PP, Boag PT, Ratcliffe LM (2009a) Plumage
coloration predicts paternity and polygyny in the American
redstart. Anim Behav 77:495–501
Reudink MW, Marra PP, Kyser TK, Boag PT, Langin KM, Ratcliffe
LM (2009b) Non-breeding season events influence sexual
selection in a long-distance migratory bird. Proc R Soc Lond B
276:1619–1626
Reudink MW, Studds CE, Marra PP, Kurt Kyser T, Ratcliffe LM
(2009c) Plumage brightness predicts non-breeding season territory quality in a long-distance migratory songbird, the American
redstart Setophaga ruticilla. J Avian Biol 40:34–41
Rohwer S, Fretwell SD, Niles DM (1980) Delayed maturation in
passerine plumages and the deceptive acquisition of resources.
Am Nat 115:400–437
Rohwer S, Klein WP, Heard S (1983) Delayed plumage maturation
and the presumed prealternate molt in American redstarts. Wilson
Bull 95:199–208
Schluter D, Price T (1993) Honesty, perception and population
divergence in sexually selected traits. Proc R Soc Lond B
253:117–122
Setchell JM (2003) Behavioural development in male mandrills
(Mandrillus sphinx): puberty to adulthood. Behaviour
140:1053–1089
Sherry TW, Holmes RT (1997) American redstart (Setophaga
ruticilla). In: Poole A (ed) The birds of North America. No.
277. Cornell Lab of Ornithology, Ithaca. Retrieved from the
Birds of North America Online: http://bna.birds.cornell.edu/bna/
species/277. doi:10.2173/bna.277
Shutler D (2011) Sexual selection: when to expect trade-offs. Biol Lett
7:101–104
Shutler D, Weatherhead PJ (1990) Targets of sexual selection: song
and plumage of wood warblers. Evolution 44:1967–1977
Staicer CA, Ingalls V, Sherry TW (2006) Singing behavior varies with
breeding status of American Redstarts (Setophaga ruticilla).
Wilson J Ornithol 118:439–451
Studd MV, Robertson RJ (1985) Life-span, competition, and delayed
plumage maturation in male passerines: the breeding threshold
hypothesis. Am Nat 126:101–115
Tabachnick B, Fidell L (1996) Using multivariate statistics. Harper &
Row, New York
Tourangeau L (2002) Correlates of age and pairing success in male
American redstarts (Setophaga ruticilla). B.Sc. Honour's thesis,
Queen's University