DETERMINANTS OF MATING SUCCESS IN FEMALE TREE SWALLOWS
(TACHYCINETA BICOLOR)
by
Lisha L. Berzins
B.Sc. (Hon.) Trent University 2007
M.Sc. Trent University 2009

DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
December 2016
© Lisha L. Berzins

Abstract
Females from a wide variety of taxa display elaborate ornaments and aggressive
behaviours that are similar to those expressed by males. Although recent empirical
investigation has demonstrated that ornamental traits and behaviours of females may function
by attracting mates or signalling competitive ability when competing against conspecifics for
access to mates and resources, less is known about how such traits influence the mating
success of free-living female birds. For my research, I experimentally examined how
variation in plumage brightness and behaviour of female tree swallows (Tachycineta bicolor)
influenced their mating success.
Plumage brightness of females had no influence on investment in parental care or
mating strategies of males, or the quality of social mate paired to the female. These results
suggest that bright plumage is not a signal of attractiveness preferred by male tree swallows.
In contrast, I report evidence that plumage brightness of female tree swallows is involved in
agonistic interactions with conspecifics. Females whose plumage brightness was enhanced to
signal high quality were less able to retain their nest site than females whose plumage
brightness was reduced to signal low quality. This suggests that females displaying bright
plumage may be challenged by conspecifics of high quality to test the quality signalled by
bright plumage, and is supported by the finding that females displaying enhanced plumage
brightness suffered social costs, such as delaying breeding and producing low-quality
nestlings. Despite these costs, females in the enhanced plumage brightness treatment mated
with extra-pair males that were higher quality than their social mate.
Behaviour of females, manipulated by elevating testosterone (T), lowered the
proportion of extra-pair offspring in the broods of T-treated females. Females manipulated so
that the androgenic and estrogenic actions of T were blocked also produced fewer extra-pair
ii

offspring, and suggest that the androgenic and/or estrogenic actions of T influence extra-pair
copulation behavior of female tree swallows. In conclusion, my results show that plumage
brightness and behaviour of female tree swallows influence their mating success, and
highlight the importance of studies experimentally manipulating ornamental and behavioural
traits of free-living female birds prior to breeding.

iii

Permission to Reproduce
Published manuscripts
Chapter 2 is a revised version of:
Berzins LL, Dawson RD (2016) Experimentally altered plumage brightness of female tree
swallows: a test of the differential allocation hypothesis. Behaviour 153:525-550
and is reproduced with the permission granted from Koninklijke Brill NV (permission
granted 16 November 2016)
Author contributions:
LLB and RDD designed the experiment. LLB performed the fieldwork. LLB analyzed the
data with input from RDD. LLB wrote the manuscript with input from RDD.

iv

Table of Contents
Abstract ................................................................................................................................... ii
Permission to Reproduce .......................................................................................................iv
Table of Contents ..................................................................................................................... v
List of Tables ........................................................................................................................ viii
List of Figures .......................................................................................................................... x
Acknowledgements ............................................................................................................... xii
Chapter 1: Introduction .......................................................................................................... 1
1.1. Function and evolution of ornamental traits of females ................................................. 1
1.2. Mechanisms maintaining the honesty of quality signals ................................................ 2
1.3. Personality and behavioural traits ................................................................................... 4
1.4. Study species ................................................................................................................... 5
1.5. Research objectives ......................................................................................................... 7
1.5.1. Female ornamentation as a signal of attractiveness to males ................................... 8
1.5.2. Female ornamentation as a signal assessed by conspecifics .................................... 9
1.5.3. Female ornamentation and behaviour as proximate causes of extra-pair
mating ............................................................................................................................. 111
Chapter 2: Experimentally altered plumage brightness of female tree swallows: a
test of the differential allocation hypothesis ........................................................................ 13
2.1. Abstract ......................................................................................................................... 13
2.2. Introduction ................................................................................................................... 13
2.3. Material and methods .................................................................................................... 17
2.3.1. Study area and general field methods .................................................................... 17
2.3.2. Parental provisioning.............................................................................................. 20
2.3.3. Spectral analysis ..................................................................................................... 20
2.3.4. Statistical analysis .................................................................................................. 22
2.4. Results ........................................................................................................................... 24
2.4.1. Parental care ........................................................................................................... 24
2.4.2. Nestling quality ...................................................................................................... 26
2.4.3. Fledging success ..................................................................................................... 27
2.5. Discussion ..................................................................................................................... 28
Chapter 3: Are there social costs of displaying bright plumage for female tree
swallows? ................................................................................................................................ 39
3.1. Abstract ......................................................................................................................... 39
3.2. Introduction ................................................................................................................... 39
3.3. Material and methods .................................................................................................... 44
v

3.3.1. Study area and general field methods .................................................................... 44
3.3.2. Spectral analysis ..................................................................................................... 45
3.3.3. Nest-site quality...................................................................................................... 46
3.3.4. Statistical analysis .................................................................................................. 47
3.4. Results ......................................................................................................................... 499
3.4.1. Nest site retention ................................................................................................. 499
3.4.2. Reproductive success ............................................................................................. 49
3.5. Discussion ..................................................................................................................... 50
Chapter 4: Does bright plumage enhance extra-pair mating success of female
passerines? An experiment with tree swallows ................................................................... 60
4.1. Abstract ......................................................................................................................... 60
4.2. Introduction ................................................................................................................... 61
4.3. Material and methods .................................................................................................... 65
4.3.1. Study area and general field methods .................................................................... 65
4.3.2. Spectral analysis ..................................................................................................... 67
4.3.3. Paternity analysis .................................................................................................... 68
4.3.4. Statistical analysis .................................................................................................. 70
4.4. Results ........................................................................................................................... 74
4.5. Discussion ..................................................................................................................... 76
Chapter 5: Is ornamentation of female passerines related to offspring quality? An
experimental alteration of plumage brightness in female tree swallows .......................... 90
5.1. Abstract ......................................................................................................................... 90
5.2. Introduction ................................................................................................................... 91
5.3. Material and methods .................................................................................................... 95
5.3.1. Study area and general field methods .................................................................... 95
5.3.2. Molecular sexing .................................................................................................... 96
5.3.3. Statistical analysis .................................................................................................. 98
5.4. Results ......................................................................................................................... 100
5.5. Discussion ................................................................................................................... 101
Chapter 6: Experimentally altering exposure to testosterone and its estrogenic
metabolites prior to breeding reduces extra-pair paternity in female tree swallows .. 1111
6.1. Abstract ..................................................................................................................... 1111
6.2. Introduction ............................................................................................................... 1122
6.3. Material and methods ................................................................................................ 1166
6.3.1. Study area and species........................................................................................ 1166
6.3.2. General field methods ........................................................................................ 1166
6.3.3. Hormone measurements ..................................................................................... 1199
6.3.4. Paternity analysis .................................................................................................. 121
vi

6.3.4. Statistical analysis .............................................................................................. 1233
6.4. Results ....................................................................................................................... 1266
6.5. Discussion ................................................................................................................... 130
Chapter 7: Synthesis ............................................................................................................ 141
8.1. References .................................................................................................................... 1555

vii

List of Tables
Table 2.1. Results of random intercept linear mixed models testing whether nestling
size or growth differed among broods where the plumage brightness of females was
experimentally reduced or enhanced, or remained unchanged (controls), and where
brood size was reduced by removing two nestlings or enlarged by adding two nestlings,
compared to control broods (see text for more details) ........................................................... 34
Table 3.1. Predictions for whether female tree swallows retain their nest site or are
usurped following a plumage brightness manipulation to reduce or enhance plumage
brightness compared to controls .............................................................................................. 57
Table 3.2. The number of female tree swallows whose plumage brightness was
experimentally reduced and enhanced, or remained unchanged (control), prior to
breeding in 2010 and 2011, and the number of females in each treatment that retained
or left their nest box, and successfully bred following manipulation. Nine females were
manipulated in both years and are only represented once in the data set (see text for
details)...................................................................................................................................... 58
Table 4.1. Variability of alleles for six microsatellite loci used to assign paternity in a
population of tree swallows from 2010–2012 ......................................................................... 83
Table 4.2. Results of random intercept linear mixed models comparing phenotypic
traits of social males paired to female tree swallows whose plumage brightness was
experimentally reduced or enhanced compared to controls. Presented are the mean
(± SE) for each phenotypic trait. Sample sizes are indicated by parentheses, and vary
because individuals with feathers that were damaged or broken were not measured. See
methods for details of calculating condition and plumage colour metrics .............................. 84
Table 4.3. Results of likelihood ratio tests, and general or generalized linear mixed
models examining whether the plumage brightness treatment (experimentally reduced
or enhanced, or remained unchanged) of female tree swallows influenced female extrapair mating success (i.e., presence and proportion of extra-pair offspring) or male
mating strategies (i.e., whether males gained extra-pair fertilizations, maintained
within-pair paternity, and their total number of offspring sired). Presented are raw
proportion means (± SE), and sample sizes are indicated by parentheses............................... 85
Table 4.4. Results of linear mixed models comparing phenotypic traits between the
social and extra-pair mate of female tree swallows whose plumage brightness was
experimentally reduced or enhanced compared to controls. See methods for details of
calculating condition and plumage colour metrics .................................................................. 86
Table 5.1. Results of random intercept linear mixed models comparing mass and size
at day 16, and growth rates of nestling tree swallows reared in broods where female
plumage brightness was experimentally reduced or enhanced, or remained unchanged
(controls) ................................................................................................................................ 108
viii

Table 6.1. Summary of the number of female tree swallows that received implants
containing testosterone (T) or 1,4,6-androstatrien-3, 17-dione and flutamide (ATD+F),
or empty implants (control). ‘Initiated a clutch’ refers to the number of females that
laid at least one egg. ‘Hatched eggs’ refers to the number of females that hatched at
least one egg. ‘Disappeared’ refers to females who were usurped or abandoned their
nesting attempt, or whose eggs were depredated................................................................... 136

ix

List of Figures
Figure 2.1. Reflectance spectra from the back and rump feathers of female tree
swallows measured before (solid black line) and after treatment with blue permanent
marker to enhanced plumage brightness (top grey line; N = 14) and silicone paste to
reduce plumage brightness (bottom grey line; N = 14). Presented are the means (± SE)
at every 50 nm interval from 300 – 700 nm. See methods for specific details regarding
plumage manipulations ............................................................................................................ 35
Figure 2.2. Mean (± SE) feeding rates (trips/hour/nestling) of male tree swallows
rearing broods where a) female plumage brightness was experimentally reduced or
enhanced, or remained unchanged (controls) and b) female plumage was
experimentally altered (enhanced and reduced treatments combined) or remained
unchanged (control). Sample sizes indicate the number of broods within each treatment
group and are given above error bars ...................................................................................... 36
Figure 2.3. Mean (± SE) feeding rates expressed as a) trips/hour/nestling and b)
trips/hour by male tree swallows rearing broods where brood size was reduced by
removing two nestlings, enlarged by adding two nestlings, or remained unchanged
(controls). Sample sizes indicate the number of broods within each treatment group and
are given above error bars........................................................................................................ 37
Figure 2.4. Mean (± SE) a) length of ninth primary flight feathers (mm) at day 16 and
b) fledging success (proportion of nestlings fledged per brood) for nestling tree
swallows according to brood size treatment (reduced by two nestlings, increased by
two nestlings, or controls) and female plumage brightness treatment (experimentally
reduced and enhanced compared to controls). Sample sizes indicate the number of
broods within each treatment group and are given above error bars ....................................... 38
Figure 3.1. Mean (± SE) clutch initiation date of female tree swallows according to a)
whether their plumage brightness was experimentally reduced or enhanced, or
remained unchanged (controls), and b) whether they retained or switched nest sites
following treatment. Clutch initiation dates were standardized to a mean of 0 and a
standard deviation 1 for experimental nests separately for each year. Negative values
indicate clutches initiated early, while positive values indicate clutches initiated later in
the breeding season. Sample sizes are given above error bars ................................................ 59
Figure 4.1. Mean (± SE) difference in a) body condition, b) wing length, c) ninth
primary length, and d) outer rectrix length between social and extra-pair mates of
female tree swallows whose plumage brightness was experimentally reduced or
enhanced compared to controls. See methods for details of calculating condition.
Sample sizes indicate the number of pairs in each treatment and are given above error
bars........................................................................................................................................... 87
Figure 4.2. Mean (± SE) difference in a) head-bill length, b) average brightness, c) UV
chroma-hue (PC1), and d) blue chroma (PC2) between social and extra-pair mates of
female tree swallows whose plumage brightness was experimentally reduced or
x

enhanced compared to controls. See methods for details of calculating plumage colour
metrics. Sample sizes indicate the number of pairs in each treatment and are given
above error bars ....................................................................................................................... 88
Figure 4.3. . Mean (± SE) blue chroma (PC2) of social and extra-pair mates of female
tree swallows. See methods for details of calculating blue chroma (PC2). Sample sizes
indicate the number of males and are given above error bars ................................................. 89
Figure 5.1. Mean (± SE) length of the combined head and bill of nestling tree swallows
at day 16 in broods where female plumage brightness was experimentally reduced or
enhanced compared to controls. Sample sizes indicate the number of broods in each
treatment, and are given above error bars .............................................................................. 109
Figure 5.2. Mean (± SE) growth rate constant for mass of nestling tree swallows reared
in broods where female plumage brightness was experimentally reduced or enhanced
compared to controls. Sample sizes indicate the number of broods in each treatment,
and are given above error bars. See Methods for calculation of growth rate constants ........ 110
Figure 6.1. Concentrations of a) androgens and b) 17β-estradiol in female tree
swallows measured prior to receiving implants during pre-breeding in May and after
receiving implants containing testosterone or sham implants during the nestling period
in June. Black circles represent control females, white circles represent testosterone
(T)-treated females, and grey circles represent females treated with 1,4,6 androstatrien3, 17-dione and flutamide (ATD+F). Repeated measures for individuals are connected
by a line, whereas individual circles indicate single measurements where repeated
measurements were not available. Data points were jittered for easier interpretation .......... 137
Figure 6.2. Mean (± SE) proportion of extra-pair offspring in nests of tree swallows
where females were either implanted with 1,4,6 androstatrien-3, 17-dione and
flutamide (ATD+F) or testosterone (T), and for controls (C; sham implanted). Sample
sizes are given above error bars ............................................................................................. 138
Figure 6.3. Relationship between clutch size and initiation date (1 January = 1) for
nests where female tree swallows were implanted with 1,4,6 androstatrien-3, 17-dione
and flutamide (dotted line), testosterone (dashed line), or received sham implants (solid
line). Data points for controls were jittered for easier interpretation .................................... 139
Figure 6.4. Mean (± SE) body mass at day 16 of nestling tree swallows raised in
broods where females were treated with 1,4,6 androstatrien-3, 17-dione and flutamide
(ATD+F) and sham implants. Sample sizes indicate the number of broods within each
treatment and are given above error bars ............................................................................... 140

xi

Acknowledgements
First, I would like to thank my supervisor Dr. Russell Dawson for answering
questions and accepting manuscripts on Fridays! Your guidance and support at all stages of
this dissertation helped me succeed. Thank you for listening to *all* my research ideas,
providing opportunities for research collaborations, being an exceptional mentor, and giving
me the encouragement and freedom to grow as a scientist.
I thank my committee members Drs. Dezene Huber, Brent Murray, Ken Otter, Kathy
Parker, and Paul Siakaluk for their valuable contributors to this dissertation, and providing
me access to their laboratories and equipment. I am especially grateful to Russ and my
committee for providing me with support to make me feel confident in my public speaking
abilities. I also thank my collaborator for Chapter 6, Dr. Mark Shrimpton, for his help and
expertise running hormone assays and for the valuable input he contributed to this chapter. I
thank Drs. Erin O’Brien and Nicole Sukdeo, as well as Caitlin Pit, for their advice with
various aspects of the paternity analysis, Dr. Stefanie LaZerte for discussions regarding data
analyses, Dr. Kim Rosvall at Indiana University for advice on the testosterone implant
construction, and Dr. Rebecca Safran for being my external examiner.
I was very fortunate to have field assistants that worked very hard, and I am very
grateful to Chelsea Coady, Jessica Milko, Amanda Lacika, Sarah Guest, Sophia Unger,
Kourtney Scott, Sarah Sparks, Jennifer Salokannel, and Maite Amat Valero for all their help
collecting data. I am forever indebted to all members of the Dawson lab (2010-2016) for all
of their support, thought-provoking discussions, and trips to The Thirsty Moose. I am very
grateful to Chelsea Coady, Erin O’Brien, Jeannine Randall, Aija White, Graham Fairhurst,
Stefanie LaZerte, Alex Koiter, and Anne Marie Florres for their friendship, and also letting
me carry on chatting about extra-pair paternity, genetics, hormone assays, or statistics.
For providing me with access to their property, I am thankful to Karen and Dale
Steward, the City of Prince George, Yvonne Wallace and Greg Sanders. I am grateful for all
the financial support I personally received from the Natural Sciences and Engineering
Research Council of Canada (NSERC) and the University of Northern British Columbia.
Additional funding for my dissertation was provided by NSERC, the University of Northern
British Columbia, Canada Foundation for Innovation, and the British Columbia Knowledge
Development Fund through grants to Dr. Russell Dawson.
For their never ending love and encouragement, I thank my mom and dad. I am truly
lucky to have parents that have provided so much love and support throughout my life, and
enabled me to follow my dreams. I am also indebted to my husband Andrew for not letting
me starve while writing my dissertation! Thank you for all your support and encouragement,
and our ‘science’ discussions; it truly amazes me to hear you explain sexual selection or
extra-pair paternity to our friend’s parents or people on airplanes.
Finally, I thank all the birds, for it has been a privilege to examine your mating
decisions.

xii

Chapter 1: Introduction
1.1. Function and evolution of ornamental traits of females
Brightly coloured and exaggerated feathers, weaponry, and aggressive or bizarre
behaviours displayed by male animals have intrigued biologists for centuries. Darwin (1871)
hypothesized that sexual selection was the evolutionary process underlying the development
of these secondary sexual traits, hereafter referred to as ‘ornamental traits’. Ornamental traits
are hypothesized to evolve because they enhance the mating success of males expressing the
most elaborate forms of these traits when competing for mates, which results in variation in
reproductive success among individuals (Darwin, 1871). Sexual selection may arise by two
non-mutually exclusive mechanisms: competition among males for access to mates (intrasexual selection) and mate choice where certain males are preferred as mates by females
(inter-sexual selection; Andersson, 1994). Sexual selection tends to act most strongly in the
sex with the greater variance in reproductive success, which is usually males (Andersson,
1994). For example, studies in avian species have demonstrated that elaborately ornamented
males are successful at maintaining paternity within their own brood, as well as gaining
extra-pair offspring in the broods of other females, thereby increasing their reproductive
success (Albrecht et al., 2007; Balenger et al., 2009). Consequently, studies exploring the
function and evolution of ornamental traits have generally focused on males, with similar
studies on females being less common (Amundsen, 2000b).
Females from a wide variety of species also display elaborate ornamental traits that
may evolve directly by sexual selection if they provide an advantage to females during
competition with conspecifics for access to mates (Johnson, 1988; Rosvall, 2008) or are
preferred by males (Amundsen et al., 1997; Griggio et al., 2009; Cotton et al., 2015). In
1

contrast, traits that appear to not be preferred by males or reflect female quality, or are
negatively related to offspring quality (e.g., Muma and Weatherhead, 1989; Cuervo et al.,
1996; Wolf et al., 2004; reviewed in Nordeide et al., 2013), are hypothesized to evolve as a
by-product of selection acting on ornamentation displayed by males (i.e., genetic correlation
hypothesis, Lande, 1980). The focus of studies on male mate choice, however, may be too
narrow to fully understand the evolution of ornamentation in females (Tarvin and Murphy,
2012), especially since ornamental traits, such as aggressive behaviour, may enable females
to outcompete conspecifics (Rosvall, 2008), but entail costs to offspring quality (Rosvall,
2011b). Since females may compete not only for access to mates, but also other resources,
some authors have proposed examining the evolution of female ornaments within the broader
framework of social selection, of which sexual selection is one part (West-Eberhard, 1983;
Lyon and Montgomerie, 2012; Tobias et al., 2012). Nevertheless, whether ornamental traits
displayed by females are merely a genetic correlation or evolve by selection remains
equivocal.
1.2. Mechanisms maintaining the honesty of quality signals
Ornamental traits that evolve by sexual and/or social selection should honestly signal
the quality of their bearer if they are to provide reliable information to potential mates and
conspecifics. Models of honest signalling propose that physiological and social costs incurred
by individuals enforce the honesty of signals and that these costs differentially affect highand low-quality individuals (reviewed in Tibbetts, 2014; Vitousek et al., 2014b).
Physiological costs are the costs incurred by individuals to produce and maintain elaborate
ornamentation (reviewed in Tibbetts, 2014; Vitousek et al., 2014b), whereas social costs, on
the other hand, are those associated with the possession of elaborate ornamentation and are
2

imposed on individuals by conspecifics (Maynard Smith and Harper, 1988). Models of social
cost predict that individuals dishonestly signalling high quality will be punished (i.e.,
incongruence hypothesis; Rohwer and Rohwer, 1978; Tibbetts and Izzo, 2010) or frequently
challenged by individuals that are truly of high quality (i.e., social control hypothesis,
Rohwer, 1977; Møller, 1987). Although models of social cost were originally formulated to
explain status signalling during social contests (Senar, 2006), the role of social costs as a
mechanism maintaining the honesty of elaborate ornamentation has been increasingly
acknowledged (Tibbetts, 2014; Vitousek et al., 2014b). Moreover, social interactions with
conspecifics may alter an individual’s physiological state, suggesting that social and
physiological costs may work together to maintain the honesty of elaborate ornamentation
(reviewed in Tibbetts, 2014; Vitousek et al., 2014b). For example, non-ornamented female
pied flycatchers (Ficedula hypoleuca), manipulated to display a patch on their forehead and
resemble ornamented females, had higher levels of blood malondialdehydes (indicating
increased oxidative damage), compared to non-ornamented controls, and the level of blood
malondialdehydes was similar to that observed in females naturally displaying a forehead
patch (Moreno et al., 2013). Moreno et al. (2013) hypothesized that social costs suffered by
females dishonestly displaying a forehead patch may have been mediated by social control.
Therefore, studies testing for social control of signal honesty in female birds during the
breeding season would prove valuable, as evidence for such a mechanism is scarce.
Social costs that maintain the honesty of ornamentation also may result in lower
reproductive success, increased energy expenditure (Kotiaho, 2001), or decreased parental
care (Qvarnström, 1997). For example, enlarging the size of forehead patch displayed by
male collared flycatchers (Ficedula albicollis) increased competition among males for
territories, and yearling males displaying an enlarged forehead patch provisioned their
3

nestlings at a lower rate compared to control males (Qvarnström, 1997). Lower investment in
parental care was hypothesized to arise because males with an enlarged forehead patch spent
more time involved in competition with conspecifics compared to provisioning their nestlings
(Qvarnström, 1997). Spending more time engaging in agonistic interactions with conspecifics
also may reduce resources available to invest in reproduction and parental care, especially in
female birds that often provide the bulk of resources for producing offspring (Fitzpatrick et
al., 1995). Such trade-offs may result in poorer quality offspring, but as far as I am aware, no
study has manipulated female ornamentation prior to breeding and explored the
consequences of dishonest signalling on offspring quality.
1.3. Personality and behavioural traits
Within populations of animals, individuals vary in the expression of behaviours, such
as aggression, boldness, exploration, activity and sociability (Réale et al., 2007). Because the
behaviour expressed by an individual often is consistent over time and across contexts, these
consistent individual differences are commonly referred to as animal ‘personalities’ (Carere
et al., 2005; Dingemanse et al., 2010; Réale et al., 2010a). Personalities are generally
composed of a suite of correlated behaviours, known as a behavioural syndrome (Sih et al.,
2004b; Réale et al., 2010a). For example, the aggression syndrome posits that some
individuals are generally more aggressive than other individuals within a population and
because of the consistency in behaviour across contexts, those individuals that are more
aggressive may have increased success in competition with conspecifics, but such behaviour
also may be suboptimal in the presence of predators or when providing parental care (Sih et
al., 2004b). Personality has been shown to be correlated with life-history traits (Biro and
Stamps, 2008; Réale et al., 2010b), as well as behaviours such as foraging, provisioning,
4

predator avoidance, social dominance, and courtship (e.g., Quinn and Cresswell, 2005; Logue
et al., 2009; Cote et al., 2010; Jones and Godin, 2010; David et al., 2011; Barnett et al.,
2012). Behaviours and/or life-history traits may be related if they are correlated in their
expression due to pleotropic effects of genes or hormones, or influenced by previous
experience (reviewed in Sih et al., 2004b).
1.4. Study species
Tree swallows (Tachycineta bicolor) are small, migratory, aerial insectivores that
breed throughout the central and northern parts of North America (Winkler et al., 2011). This
species is a secondary cavity nester, using pre-existing tree cavities, but will readily breed in
nest boxes, making them a model species for avian ecology research (Jones, 2003). Tree
swallows arrive at my study areas in north central British Columbia, Canada, during late
April to early May (Dawson, unpublished data), with males arriving prior to females to
compete for nest sites (Winkler et al. 2011). Females, upon arrival, will also compete with
conspecifics for a male with a nest box (i.e., nest site), since nest sites are a limiting factor for
female reproductive success (Leffelaar and Robertson, 1985; Rosvall, 2008). Consequently,
there is often a large floating population of female tree swallows that do not breed
(Stutchbury and Robertson, 1985). Females that acquire a nest site frequently experience
territory intrusions from conspecifics and must defend their nest to avoid take-overs (i.e.,
usurpation; Leffelaar and Robertson, 1985). Overall, competition among female tree
swallows is often so intense that it can lead to injury or death (Leffelaar and Robertson,
1985).
Egg laying at my study areas usually begins in mid to late May (Dawson, unpublished
data), and clutch-size is generally 4-7 eggs. Females incubate alone for approximately 13
5

days (Winkler et al. 2011). After the hatching of eggs, the nestling-rearing period is 18-22
days, after which nestlings fledge from the nest (Winkler et al. 2011). Both sexes of parent
provision and deliver similar amounts of food to nestlings (McCarty, 2002).
Tree swallows have a socially monogamous mating system and generally form a pairbond with a single member of the opposite sex, but both members of the pair may mate with
individuals outside of the social pair-bond. Consequently, this species has one of the highest
rates of extra-pair paternity reported, with up to 85 % of nests having at least one offspring
not sired by the social father (see O’Brien and Dawson, 2007; Whittingham and Dunn, 2016
and references therein). Empirical studies testing for benefits gained by females from extrapair mating in tree swallows have reported little evidence to suggest that females obtain highquality genes for their offspring (but see O’Brien and Dawson, 2007). Rather, studies have
demonstrated support for the genetic compatibility hypothesis and suggest that extra-pair
mating occurs so that females can increase the genetic diversity of their offspring (e.g.,
Whittingham et al., 2006; Stapleton et al., 2007; Dunn et al., 2009; Whittingham and Dunn,
2010). Evidence of increased genetic diversity is also supported by the number of extra-pair
males that sire offspring within a single brood (e.g., 1-4 males; Whittingham and Dunn,
2014), and that broods of experienced females paired to genetically similar social mates have
higher hatching success when sired by a greater number of extra-pair males (Whittingham
and Dunn, 2010).
Tree swallows are one of few North American species where second-year (SY)
females (i.e., in their second year of life, but first potential breeding season) have delayed
plumage maturation (Stutchbury and Robertson, 1987). As such, the dorsal plumage colour
of SY females is predominantly brown, although rarely some individuals display fully
iridescent blue-green structural plumage colour (Hussell, 1983). SY females are hypothesized
6

to display dull brown plumage to signal their poor competitive ability, thereby reducing intrasexual aggression from older, after-second-year (ASY) females (Coady and Dawson, 2013).
ASY females and males display structural dorsal plumage colour that is bright iridescent
blue-green, and the colour is produced by the physical interaction of light waves and the
nanostructure (e.g., keratin, melanin, and air) of the feathers (Prum, 2006). ASY females
displaying plumage that is brighter, with greater blue and ultraviolet (UV) chroma, and
reflecting light maximally at shorter wavelengths (bluer hue), are considered to be more
ornamented because they have greater nanostructure organization of their feathers, lay
heavier eggs and fledge more offspring (Bitton et al., 2008; Bentz and Siefferman, 2013).
Although plumage colour of ASY females appears to reflect aspects of quality, those females
that are more ornamented have been shown to have greater levels of nest parasitism, poorer
immune defences, and lower hematocrit levels, and they also produce nestlings that are
smaller or in poorer condition (Coady, 2011; Bentz and Siefferman, 2013). In males, those
with brighter plumage sire a greater number of extra-pair offspring, and as a result have
greater reproductive success than duller males, suggesting that bright plumage enhances male
extra-pair mating success (Bitton et al., 2007; Whittingham and Dunn, 2016). Positive
assortative mating for plumage brightness occurs, which may be due to mutual mate
preference for this trait or competition for nest sites (Bitton et al. 2008).
1.5. Research objectives
The objective of my dissertation was to examine how ornamental and behavioural
traits displayed by female tree swallows influence their mating success. Tree swallows are an
excellent study species to achieve my research objective because females aggressively
compete with conspecifics for access to a male with a nest site (Rosvall, 2008) and defend
7

their nest sites from intruding conspecifics (Leffelaar and Robertson, 1985). The ornamental
bright plumage of females appears to signal quality (Bitton et al., 2008; Bentz and
Siefferman, 2013) and potentially influence male mating preferences or intra-sexual
competition (Bitton et al., 2008), but the possession of ornamental bright plumage may entail
social costs that influence offspring quality (Bentz and Siefferman, 2013). Social interactions
resulting from the possession of bright plumage may alter the physiology of females
(reviewed in Tibbets, 2014, Vitousek et al., 2014) by increasing levels of testosterone (T),
and provide a mechanism to explain the high rates of extra-pair paternity reported in tree
swallows (e.g., O’Brien and Dawson, 2007; Whittingham and Dunn, 2016 and references
therein), but this has never been tested experimentally. For my research, I performed
experimental manipulations to alter the plumage brightness and T-mediated behaviour of
female tree swallows to specifically identify: (1) whether plumage brightness of females is a
signal of attractiveness to males; (2) whether plumage brightness of females is a signal
assessed by conspecific females and whether females dishonestly displaying bright plumage
experience social costs, such as lower reproductive success, or trade-offs with fecundity or
parental care that influence nestling quality; and (3) whether plumage brightness or Tmediated behaviour influences extra-pair mating by female tree swallows.
1.5.1. Female ornamentation as a signal of attractiveness to males
In many avian species, males and females show assortative mating for phenotypic
traits, such as plumage colour (e.g., Jawor et al., 2003; Rowe and Weatherhead, 2011; Jacobs
et al., 2015). One mechanism by which assortative mating may occur is mutual mate choice,
where both sexes have a preference for mates displaying similar attractive ornamentation
(Johnstone et al., 1996). Male choice of mates expressing ornamental traits may occur more
8

frequently than previously assumed (Edward and Chapman, 2011), especially in species
where males invest in parental care and females vary in their quality (Johnstone et al., 1996;
Kokko and Johnstone, 2002; Edward and Chapman, 2011). Mate choice can be exercised by
males allocating their investment in reproduction non-randomly among females (Edward,
2015). Consequently, the attractiveness of female ornamentation may influence pairing
success (Griggio et al., 2005), the quality of social or extra-pair mates, or how much parental
care males invest in offspring (Burley, 1988). Although tree swallows have been reported to
assortatively mate for plumage brightness (Bitton et al., 2008), it remains to be determined
whether plumage brightness displayed by females is a signal of attractiveness that is
preferred by males. In Chapter 2, I tested whether experimentally altering the plumage
brightness of females influenced male investment in parent care and the consequences for
offspring quality. In Chapter 4, I tested whether experimentally altered plumage brightness of
females influenced the quality of social and/or extra-pair mates acquired, and the mating
strategies (i.e., maintain within-pair paternity versus gain extra-pair fertilizations) of the
females’ social mate. Collectively, these experiments allowed me to determine whether
bright plumage of female tree swallows is a signal of attractiveness that is preferred by
males, thereby providing evidence that positive assortative mating may occur in tree
swallows as a result of male (or mutual) mate choice.
1.5.2. Female ornamentation as a signal assessed by conspecifics
Ornamental traits displayed by females have been shown to function in competitive
contests with conspecifics (Murphy et al., 2009b). As such, assortative mating for plumage
characteristics reported in a variety of species may be the outcome of competition among
females, as opposed to male mate choice (e.g., Houtman and Falls, 1994; Creighton, 2001).
9

In such cases, high-quality females, presumably with greater competitive ability, successfully
outcompete conspecifics for access to high-quality territories or mates (Creighton, 2001).
Since female tree swallows are limited by the number of available males with a nest site
(Leffelaar and Robertson, 1985), conspecific females may challenge the honesty of
ornamentation displayed by females who were successful at acquiring such a valuable
resource. In Chapter 3, I tested whether the plumage brightness treatment influenced a
female’s ability to retain her nest box, which would suggest a role for the plumage brightness
of females to function in agonistic interactions. Although female tree swallows displaying
bright plumage benefit by gaining high-quality males with bright plumage as mates (Bitton et
al., 2008), previous studies have shown that more ornamented females have greater levels of
nest parasitism, poorer immune defences, and lower hematocrit levels, and they also produce
low-quality offspring (Coady, 2011; Bentz and Siefferman, 2013). This may be mediated by
agonistic interactions with conspecific females, and suggest that ornamented females
experience social costs as a result of the mechanisms enforcing signal honesty, such as
physiological changes (Moreno et al., 2013), lower reproductive success (Kotiaho, 2001), or
reduced parental care (Qvarnström, 1997). In Chapters 3 and 5, I examined whether females
that retained their nest site following the plumage brightness manipulation suffered
reproductive costs or produced low-quality offspring. Such costs would indicate that females
displaying bright plumage experience trade-offs between ornamentation and fecundity
(Fitzpatrick et al., 1995) or parental care (Qvarnström, 1997), as a result of engaging in
agonistic interactions with conspecific females. Overall, the results of these chapters identify
whether plumage brightness is a signal assessed by conspecific females, and whether the
honesty of bright plumage displayed by females is enforced by nest-site intrusions from
conspecific females. My results presented in Chapters 3 and 5 also highlight the importance
10

of considering the social costs associated with displaying elaborate ornamentation when
examining the function and evolution of ornamental traits of females.
1.5.3. Female ornamentation and behaviour as proximate causes of extra-pair mating
In socially monogamous species, females commonly mate with males other than their
social mate, resulting in extra-pair offspring (Griffith et al., 2002). Empirical studies
examining extra-pair mating have generally focused on the benefits and costs of extra-pair
mating to females (reviewed in Patrick et al., 2012), but fewer studies have examined
whether the ornamentation or behaviour of females underlies a female’s success at gaining
extra-pair copulations (e.g., García-Vigón et al., 2008; van Oers et al., 2008; Grunst and
Grunst, 2014; Jacobs et al., 2015). Ornamental traits may influence a female’s opportunity to
engage in extra-pair copulations by attracting extra-pair mates (Torres and Velando, 2005) or
enabling females to invade the territories of other females in search of extra-pair mates
(Smith, 1988). Moreover, the attractiveness of female ornamentation may influence whether
females mate with extra-pair mates that are higher or lower quality relative to their social
mate, although as far as I am aware, this has never been investigated by manipulating female
ornamentation. In Chapter 4, I examined whether the plumage brightness of female tree
swallows influenced their opportunity to engage in extra-pair copulations and the quality of
their extra-pair mate. If possessing elaborate ornamentation influences social interactions
with conspecifics (e.g., Qvarnström, 1997; see above), then this may alter the physiology and
behaviour of females (Vitousek et al., 2014), and potentially influence whether or not they
solicit or pursue extra-pair copulations. Indeed, previous studies have shown that extra-pair
paternity is correlated with behaviour, such as aggression (While et al., 2009); however,
studies that have manipulated T have demonstrated that females treated with T produce fewer
11

or a similar proportion of extra-pair offspring in their broods compared to control females
(García-Vigón et al., 2008; de Jong, 2013, Gerlach and Ketterson, 2013). None of these
studies, however, have manipulated T by preventing its conversion to 17β-estradiol (E2). In
Chapter 6, I manipulated the aggressive and sexual behaviour of females, by altering the
exposure of T and its estrogenic metabolites, to examine whether these behaviours influence
extra-pair copulation behaviour of female tree swallows. This experiment allowed me to
discern whether extra-pair copulation behaviour is related to behavioural traits of females
because they are mediated by hormones such as T and/or E2 and consequently correlated in
their expression. Overall, the results of Chapter 4 and 6 provide insight on whether extra-pair
paternity is mediated by the ornamentation or T-mediated behaviour of females.

12

Chapter 2: Experimentally altered plumage brightness of female tree swallows: a test of
the differential allocation hypothesis
2.1. Abstract
The differential allocation hypothesis posits that individuals should invest in the
current reproductive attempt according to the attractiveness of their mate, but studies of
allocation by males when female traits are manipulated to be more attractive are lacking. In
the current study, I experimentally enhanced and reduced the plumage brightness of female
tree swallows (Tachycineta bicolor) relative to controls to examine whether males adjust
investment in parental care according to female attractiveness, while simultaneously
performing a brood size manipulation. Contrary to my predictions, I found no evidence that
males provisioned nestlings according to the plumage brightness of females. However,
nestling quality and fledging success were lowest when female plumage brightness was
reduced and brood size was enlarged. This may be due to the plumage brightness treatment
influencing agonistic interactions with other females, and may suggest that plumage
brightness is a signal assessed by females.
2.2. Introduction
It is widely recognized in a variety of taxa that females display elaborate ornaments,
such as brightly coloured plumage. Two mechanisms often used to explain the presence of
ornamental traits in females are a genetic linkage with selection on male traits or direct
selection on female traits (Lande, 1980; Amundsen, 2000b). Direct selection on female traits
can arise by male mate choice or female-female competition for mates or resources, and so
may be driven by sexual and/or social selection (LeBas, 2006; Clutton-Brock, 2009; Edward
and Chapman, 2011; Tobias et al., 2012). Ornamental traits may provide honest information
13

about the quality of their bearer (Zahavi, 1975; Kodric-Brown and Brown, 1984; Grafen,
1990), and an accumulating number of studies have demonstrated that ornamental traits
reflect aspects of female quality (see Doutrelant et al., 2008; Gladbach et al., 2010; MartinezPadilla et al., 2011; Henderson et al., 2013). For example, some studies suggest that
elaborately ornamented females transmit more resources to eggs (Midamegbe et al., 2013),
have higher fecundity (Jawor et al., 2004; Cornwallis and Birkhead, 2007), provide more
parental care to offspring (Linville et al., 1998; Weiss, 2006; Silva et al., 2008; García-Navas
et al., 2012), or produce offspring of higher genetic quality (e.g., offspring with greater
immune defences; Roulin et al., 2000; Matysioková and Remeš, 2013). As such, males may
benefit from assessing the attractiveness of female ornamentation when deciding how much
parental care to invest in offspring.
Investing in the current reproductive attempt according to mate attractiveness is
known as differential allocation (Burley, 1986; Sheldon, 2000), and is expected to occur
when there is a trade-off between current and future reproduction, and when the
attractiveness of one’s mate signals the reproductive value of the brood, thereby influencing
this trade-off (Sheldon, 2000). Positive differential allocation occurs when mates of attractive
individuals invest more in offspring, whereas negative differential allocation occurs when
mates of unattractive individuals invest more in offspring (Ratikainen and Kokko, 2010).
Providing greater investment in parental care when mated to an unattractive individual is a
form of reproductive compensation (reviewed in Harris and Uller, 2009), but compensation is
not restricted to negative differential allocation; mates may also increase investment as a
compensatory mechanism when paired to an attractive individual to improve a poor situation,
such as a lack of parental care from an attractive mate (Ratikainen and Kokko, 2010) or when
paired to partners that are not preferred (Gowaty, 2008).
14

Overall, the differential allocation hypothesis has received empirical support in a
variety of taxa (Sheldon, 2000), including birds (Horváthová et al., 2012); however, the
majority of studies have focused on differential allocation by females, whereas studies
assessing how female attractiveness influences male reproductive decisions are
comparatively less common (Ratikainen and Kokko 2010; Edward and Chapman, 2011).
Since the first experimental demonstration of differential allocation by male birds in relation
to female attractiveness (Burley, 1988), general support for positive differential allocation
has come from studies experimentally altering female ornamental traits to be less attractive.
These studies have generally demonstrated that males provide lower parental investment in
the form of brood defence or provisioning nestlings when paired to unattractive females
(Roulin, 1999; Pilastro et al., 2003; Matessi et al., 2009; Mahr et al., 2012). One study found
the opposite, where male provisioning increased in response to experimentally reduced
ultraviolet (UV) crown reflectance of female blue tits (Cyanistes caeruleus; Limbourg et al.,
2013). In contrast, the only study that I am aware of that experimentally altered the
phenotypic appearance of females to be more attractive reported no support for the
differential allocation hypothesis; males paired to female rock sparrows (Petronia petronia)
with increased breast patch size showed no difference in brood defence or nestling
provisioning compared to controls (Pilastro et al., 2003). Therefore, support for positive
differential allocation by males when paired to females whose phenotypic traits have been
manipulated to be more attractive, as demonstrated by Burley (1988), is lacking.
In the current study, my aim was to test how variation in female attractiveness
influences male investment in parental care, and how investment decisions affect offspring
quality and fledging success in tree swallows (Tachycineta bicolor). Tree swallows are an
appropriate species to test male differential allocation because females that are in at least
15

their second breeding season (after-second-year; ASY) display ornamented dorsal plumage
that is similar to males (i.e., iridescent blue-green), and males contribute to parental care by
defending territories, feeding nestlings, and performing nest sanitation (Winkler et al., 2011).
Previous studies examining plumage colour in females have reported that females displaying
brighter plumage with greater UV and blue chroma, and bluer hues are older, make greater
investments in reproduction and fledge more offspring (Bitton et al., 2008; Bentz and
Siefferman, 2013).
I tested whether the perceived attractiveness of females influences male investment in
parental care by experimentally enhancing and reducing the plumage brightness of females
relative to controls. I chose to manipulate plumage brightness because previous studies have
shown brighter female tree swallows lay heavier eggs, thereby suggesting that plumage
brightness may signal female quality (Bitton et al., 2008), and positive assortative pairing
occurs for plumage brightness where bright females pair with bright males, which may result
from mutual mate choice or competition among females for nest sites (Bitton et al., 2008).
Given that competition among females is intense (Leffelaar and Robertson, 1985; Rosvall,
2008) and can lead to injury or death (Leffelaar and Robertson, 1985; Robertson et al., 1986),
it may be more likely that bright plumage of females is related to competitive ability.
Nevertheless, ornamental traits that evolve by social or intra-sexual competition may still be
exploited by males when choosing mates (LeBas, 2006) or deciding how much effort to
invest in parental care. Therefore, if males invest in parental care according to female
attractiveness, I predicted that males paired to females with enhanced plumage brightness
would provision their brood at a higher rate and produce offspring that grew faster and were
larger at fledging than males paired to females with control or reduced plumage brightness.
To explore whether adjustments in feeding rates of males in response to varying plumage
16

brightness of females was contingent on demands of the brood, I also simultaneously
performed a brood size manipulation. I expected to see the greatest increase in feeding rates
(to brood as a whole and to each individual nestling) for males rearing enlarged broods where
female plumage brightness was enhanced, whereas the greatest decrease in feeding rates was
expected to occur for males rearing enlarged broods where female plumage brightness had
been reduced.
2.3. Material and methods
2.3.1. Study area and general field methods
I studied tree swallows breeding in nest boxes at three study areas within 30 km of
each other near Prince George, BC, Canada (53oN, 123oW) from May to August in 2010 and
2011. These sites are located in areas consisting of open agricultural fields and small
wetlands mixed with patches of both coniferous and deciduous trees (see Bitton et al., 2007;
Dawson, 2008 for more details). In early May I visited nest boxes every 1-2 days to
document the start of egg laying, at which time I began to check boxes daily until clutches
were complete. Beginning 12 days after clutch completion, I checked nests daily to determine
hatching date (where January 1 = 1), which was defined as the date the first egg in each nest
hatched (designated day 0 of the nestling period).
On day 2 of the nestling period, I captured adults in nest boxes while they were
feeding nestlings. All individuals received an individually numbered aluminum leg band, and
five feathers were collected from the rump, and stored in opaque envelopes at room
temperature until spectral analysis (details below). Males were marked with a small white dot
on their back using non-toxic acrylic paint for identification during provisioning trials (see
below). I determined the age of females using dorsal plumage colour (Hussell, 1983), and
17

ASY females were allocated to treatment groups (enhanced, reduced or control plumage
brightness) sequentially upon capture after determining the order of treatment randomly.
Prior to manipulation, females in each treatment did not differ in their plumage
characteristics (all P values > 0.88; see below for description of plumage characteristics) or
hatching date (P > 0.76).
I altered female plumage brightness by applying evenly to the head, nape, mantle, and
rump feathers, a non-toxic permanent blue marker (Prismacolor® PM-39: True blue) to
enhance brightness, or silicon paste (Mucilin®) to reduce brightness (Figure 2.1); control
females were treated in the same manner, but the marker used contained water. I was
confident that my treatment would produce the desired effect throughout the study since
these permanent markers and silicone paste have been used previously (Ballentine and Hill,
2003; Johnsen et al., 2005; Safran et al., 2005; Liu et al., 2007), and the effects can still be
detected on average 15 days after manipulation (Johnsen et al., 2005). I tested the
effectiveness of my treatment in the laboratory by measuring plumage characteristics (see
below for more details) before and after the application of permanent marker or silicon paste
to rump and back feathers collected from females that were not part of the present study.
Treatment of feathers with the blue permanent marker significantly increased the average
brightness of feathers (paired t13 = -8.10, P < 0.001) without altering hue (paired t13 = -1.50,
P = 0.16), whereas the application of silicon paste reduced the average feather brightness
(paired t13 = 7.97, P < 0.001) and slightly shifted hue (9.07 ± 1.79 standard error (SE) nm)
toward longer wavelengths (i.e., made females slightly greener; paired t13 = -5.08, P <
0.001). The phenotypes produced by my experimental manipulation fell within the natural
range of plumage brightness for female tree swallows (Berzins, unpublished data). To test
whether enhanced plumage remained bright throughout the study period, I exposed a subset
18

of feather samples treated with blue marker to sunlight for 15 days, and although plumage
brightness had faded somewhat, feathers were still significantly brighter than they were prior
to treatment (paired t8 = -3.94, P < 0.01). In total, I manipulated the plumage brightness of
females from 46 nests (enhanced = 22, reduced = 24) while 24 females served as controls.
Two days after hatching, I performed a brood size manipulation to determine whether
male investment in parental care when paired to females that differ in attractiveness is
contingent on the demand of the brood. Nests were matched for hatching date (± 1 day), and
randomly assigned to a brood size treatment (enlarged, reduced, or control) independent of
the plumage treatment of the female. Each nestling was uniquely marked with non-toxic
markers, and two nestlings were removed from reduced broods and added to the enlarged
brood size group. In addition, I also cross-fostered 1-4 nestlings among nests when possible
to ensure nests contained original and cross-fostered offspring, thereby controlling for the
potential genetic effects on offspring quality. For these manipulations, I weighed all nestlings
using a spring balance (nearest 0.125 g) and selected only intermediate-sized nestlings that
were similar in mass to transfer between nests to ensure that overall size hierarchies within
nests were not altered (Dawson and Bortolotti, 2003). Following my experimental
manipulation of brood size, average nestling mass of the brood on day 2 did not differ by
plumage treatment (F2, 60 = 2.60, P = 0.08) or brood size treatment (F2, 60 = 1.87, P = 0.16)
after controlling for hatching date (F1, 60 = 3.21, P = 0.08). The suggestion of a trend for the
average mass of nestlings to differ by plumage treatment on day 2 was due to nestlings in
control nests tending to be lighter (mean ± SE; 3.25 ± 0.15) than nestlings in reduced (3.66 ±
0.15) and enhanced broods (3.65 ± 0.16). Mass of nestlings was subsequently measured with
a spring balance (nearest 0.125 g) from day 4 to 16 and the length of the ninth primary flight

19

feather with a ruler (nearest 0.5 mm) from day 8 to 16. I visited nests on day 22 to determine
the number of young that successfully fledged.
2.3.2. Parental provisioning
Investment in parental care was estimated by quantifying provisioning rates of parent
birds during 40 min sessions on days 6, 8, and 10 of the nestling period. Males were
distinguished from females in provisioning observations by the white dot applied to the back
of males following capture earlier in the nestling period. Observations were performed
between 0830 and 1900 hrs PST. For each session, the observer was situated approximately
50 m from the nest and used binoculars to document every visit each parent made to the nest
box. Since parent tree swallows rarely visit the nest box without bringing food, the number of
visits to the nest provides an accurate measure of food delivery (McCarty, 2002). Moreover,
the load sizes of food delivered to the nests are similar between the sexes and consistent
throughout the feeding period (McCarty 2002). Observations were performed at 25 boxes in
2010 and 17 boxes in 2011.
2.3.3. Spectral analysis
To quantify plumage characteristics of females prior to my experimental
manipulation (hereafter referred to as original plumage colour in analysis), four rump
feathers were taped to a piece of cardboard in an overlapping fashion that mimicked the
natural arrangement of rump feathers on the bird, following Bitton et al. (2007). Samples
were then placed on a black non-reflective background and reflectance spectra were
measured using a JAZ-PX spectrometer with a xenon light source (Ocean Optics, Dunedin,
FL, USA) and a bifurcated probe enclosed in a black holder that excluded ambient light. The
probe was held 90o to the surface of the feather and three measurements were taken,
20

removing the probe between each measurement. Spectral data were recorded using
SpectraSuite software (Ocean Optics, Dunedin, FL, USA) and reflectance was calculated as
the proportion of light reflected in 1 nm intervals between 300-700 nm relative to the
reflectance of a WS-1 diffuse white standard (Ocean Optics, Dunedin, FL, USA).
I used the pavo package (Maia et al., 2013) for R (R Development Core Team, 2015)
to smooth the spectral curves and quantify plumage characteristics that summarized the
spectral curve for each individual. Specifically, I quantified measures of hue, UV chroma,
blue chroma and average brightness because these plumage characteristics have been shown
previously to be biologically relevant measures of quality in tree swallows (Bitton et al.,
2007; Bitton and Dawson, 2008; Bitton et al., 2008; Bentz and Siefferman, 2013). Hue was
calculated as the wavelength of maximum reflectance, UV and blue chroma were each
calculated as the relative proportion of light reflected in the UV (300-400 nm) and blue (400512 nm) range relative to the entire spectrum (300-700 nm), and average brightness was
calculated as the average amount of light reflected by the feather over the entire spectrum
(300-700 nm; see Montgomerie, 2006). To reduce these plumage characteristics to individual
colour scores that reflect variation in plumage colour among individuals, I used a principal
components analysis (Montgomerie, 2006) using all ASY females from which I collected
feather samples over a four-year period (2010-2013). The first principal component (PC1)
explained 55.8 % of the total variation and the factor loadings (hue: -0.98; UV chroma: 0.88;
blue chroma: 0.70; brightness: -0.07) suggest females with larger positive PC1 scores had
greater UV and blue chroma, and reflected at shorter wavelengths (i.e., were bluer). PC2
explained 26.2 % of the total variation and was heavily weighted by brightness (factor
loadings = 0.12, -0.18, 0.30, and 0.95, respectively), so large positive PC2 scores represent
females with brighter plumage. Overall, females displaying brighter plumage with greater
21

UV and blue chroma, and bluer hues (larger, positive PC scores) are considered more
ornamented because such females have been shown to make greater investments in
reproduction and have greater reproductive success (Bitton et al., 2008; Bentz and
Siefferman, 2013).
2.3.4. Statistical analysis
To determine whether the experimental manipulation of female plumage brightness
influenced male investment in parental care, I analyzed the rate of food delivered to the nest
by male tree swallows using linear mixed models (LMM). Brood identity was the subject
grouping variable and chick age defined the repeated effect to account for the feeding
observations that occurred on days 6, 8 and 10. I was interested both in assessing how much
males invested into each individual nestling and in the brood as a whole, so I ran models
using the number of feeding trips per hour per nestling and number of feeding trips per hour
as the dependent variables. Plumage brightness treatment and brood size treatment, as well as
study area were included as fixed factors, and I tested for an interaction between plumage
brightness and brood size treatments. I did not include year as a factor because it was
confounded with study area. As covariates, I included the feeding rate and original plumage
colour (PC1 and PC2) of the female, and the start time of the observation. Since climatic
variation can influence parental feeding rates in tree swallows (Rose, 2009), I also included
rain as a categorical variable to indicate whether the trial ended because it had started to rain,
as well as the average temperature and wind speed for the 3 hour period prior to each feeding
observation; these data were obtained from weather stations located at each study area.
To test for effects of female plumage brightness and brood size treatments on
offspring quality, I used LMM to examine length of the ninth primary feather and mass at
22

day 16 (just prior to fledging), as well as growth rates of nestlings. Growth rate constants of
individual nestlings were calculated using a linear model for ninth primary and a logistic
model for mass following Dawson et al. (2005). Growth rate constants were only calculated
for nestlings that had complete growth measures, i.e., day 8-16 for ninth primary feather and
day 4-16 for mass. In models, the size or growth rate of nestlings was the dependent variable,
and study area, plumage brightness treatment and brood size treatment were included as fixed
factors, as well as the interaction between plumage brightness and brood size treatments.
Original brood size was included as a covariate for models testing size and growth, and as an
additional covariate for offspring size I included the time of day when nestlings were
measured; this covariate was not included in models testing nestling growth since I corrected
for time when calculating growth rate constants. Hatching date was not included as a
covariate in these models because it was negatively related to the time of day that nestlings
were measured (r = -0.34, P = 0.01). Brood of rearing was included as a random factor to
account for the lack of independence among nestlings reared within the same environment,
and brood of origin as a random factor to account for any potential genetic or maternal
effects on nestling quality. Due to the lack of independence among nestlings reared together,
brood of rearing was always retained in the model, and I used the Wald Z statistic to assess
the inclusion of brood of origin in each model (Garson, 2012).
Fledging success was calculated as the proportion of nestlings that successfully
fledged relative to the total number of offspring in the nest following the brood size
manipulation on day 2. I tested whether fledging success differed by treatment using analysis
of covariance, with proportion of offspring that fledged successfully as the dependent
variable. As fixed factors, I included plumage brightness and brood size treatments, as well
as study area, and tested for an interaction between plumage brightness and brood size
23

treatments. Hatching date, standardized to a mean of 0 and standard deviation of 1 for
experimental nests within each study area, was included as a covariate.
For all analyses, I removed interactions, covariates and main effects that did not
approach significance (P ≥ 0.10) in a backwards, stepwise fashion, but always retained
plumage brightness and brood size treatments in models. Residuals of models were checked
for normality and heteroscedasticity (Cleasby and Nakagawa, 2011). Results were considered
significant at P ≤ 0.05, and the overall significance of omnibus tests was examined using
post-hoc tests with a Bonferroni adjustment for multiple comparisons. Where appropriate, I
report least squares means ± 1 standard error, parameter estimates (B), and effect sizes.
Effects sizes were calculated as partial eta squared (ηp2) for omnibus statistical tests (Lakens,
2013) and as the correlation coefficient (r) for post-hoc comparisons (Field et al., 2012). All
statistical analyses were performed using SPSS v20 statistical software (IBM Corp., 2011).
2.4. Results
2.4.1. Parental care
Feeding rates (trips/hour/nestling) of male tree swallows did not vary among plumage
brightness treatments (F2, 95.67 =1.61, P = 0.21; Figure 2.2a) or brood size treatments (F2, 94.37
= 2.11, P = 0.13, Figure 2.3a), but increased with higher feeding rates by females (B = 0.16 ±
0.07, F1, 104.70 = 4.61, P = 0.03) and decreased as original brood size increased (B = -0.11 ±
0.05, F1, 87.81 = 4.52, P = 0.04). When wind speeds were higher prior to the observation
period, males fed nestlings at a greater rate (B = 0.19 ± 0.09, F1, 85.96 = 4.07, P = 0.047).
Original plumage colour of the female (PC1 and PC2), start time, temperature, and rain had
no effect on male feeding rates (P values > 0.22). Performing the analysis using trips per
hour showed the same overall conclusions, but feeding rates differed by brood size treatment
24

(F2, 102.00 = 3.99, P = 0.02, Figure 2.3b). Post-hoc tests showed that the number of feeding
trips per hour by males was greater for enlarged versus both reduced (t97.80 = 2.56, P = 0.04, r
= 0.25) and control broods (t105.13 = 2.28, P = 0.07, r = 0.22), but no difference was detected
between control and reduced broods (t102.66 = 0.38, P = 1.00, r = 0.04).
Although my results showed no effects of plumage treatments on provisioning by
males (Figure 2.2a), I further investigated whether male behaviour was influenced simply by
their mates having altered plumage brightness, regardless of whether it was reduced or
enhanced. Results that compared males paired to control females with those where data from
both experimental treatments were combined showed a trend for feeding rates of males to be
higher in experimental than control treatments (F1, 95.00 = 3.38, P = 0.069; Figure 2.2b).
Manipulated brood size, female feeding rate and wind were still significant predictors of
male feeding rates (all P values < 0.046). To clarify whether the increase in male feeding
rates when paired to experimental females was due to differential allocation or compensation
by males, I additionally analyzed the feeding rates of females. Feeding trips per hour per
nestling did not differ by plumage brightness treatment (F2, 110.18 = 0.51, P = 0.60), but varied
by brood size treatment (F2, 109.76 = 13.82, P < 0.001). Post-hoc tests showed that females fed
nestlings in the enlarged treatment at a lower rate than nestlings in the reduced (t109.77 = 4.94,
P < 0.0001, r = 0.43) and control brood sizes (t108.80 = 4.07, P < 0.0001, r = 0.36). Female
feeding rates also differed by the age of nestlings (F2, 64.57 = 5.41, P = 0.007) and study area
(F1, 109.87 = 4.47, P = 0.04), and were related to the original size of the brood (B = -0.28 ±
0.06, F1, 109.82 = 20.31, P < 0.0001). When I analyzed how many times females provisioned
per hour, I similarly found that female feeding rate did not differ by plumage brightness
treatment (F2, 113.40 = 0.001, P > 0.99) and the number of feeding trips per hour also did not

25

differ according to brood size treatment (F2, 113.17 = 1.28, P = 0.28). Nestling age and original
brood size were still significant predictors of female feeding rates (both P values < 0.004).
2.4.2. Nestling quality
Analysis of length of ninth primary feathers of nestlings revealed a significant
interaction between female plumage brightness and brood size treatments (Table 2.1), and so
I analyzed the data separately by plumage brightness treatment. The analysis revealed that in
the reduced plumage brightness treatment, the length of ninth primaries of nestlings at day 16
differed by brood size treatment (F2, 13.43 = 23.59, P < 0.0001, Figure 2.4a). Post-hoc tests
showed that nestlings reared in enlarged broods had shorter ninth primary feathers compared
to nestlings reared in control (t11.30 = 5.90, P < 0.0001, r = 0.87) and reduced broods (t13.38 =
6.53, P < 0.0001, r = 0.87; Figure 2.4a), but there were no differences between control and
reduced broods (t16.20 = 1.36, P = 0.58, r = 0.32). Nestlings reared in nests where female
plumage brightness treatment was control or experimentally enhanced had similar sized ninth
primary feathers prior to fledging regardless of brood size treatment (P values > 0.51).
Nestling body mass at 16 days of age was similar among plumage brightness
treatments, but increased with time of day nestlings were measured (B = 0.23 ± 0.09) and
varied by brood size treatment (Table 2.1). Post-hoc tests indicated that nestlings raised in
broods that were experimentally reduced in size were heavier than nestlings reared in control
(t51.00 = 2.75, P = 0.02, r = 0.36) and enlarged (t40.80 = 4.44, P < 0.0001, r = 0.57) broods.
Nestlings raised in control broods also tended to be heavier than nestlings raised in enlarged
broods (t45.38 = 2.28, P = 0.08, r = 0.32). Brood of rearing and brood of origin also predicted
nestling mass among broods (covariance parameter estimate = 1.74 ± 0.50, Wald Z = 3.48, P

26

= 0.001, and covariance parameter estimate = 0.42 ± 0.24, Wald Z = 1.78, P = 0.08,
respectively).
Growth of nestling ninth primary feathers was not influenced by the plumage
brightness treatment of the female, but differed by brood size treatment (Table 2.1). Post-hoc
tests indicated that nestlings raised in enlarged broods grew ninth primary feathers slower
than nestlings in reduced broods (t40.28 = -2.65, P = 0.03, r = 0.39). No differences between
enlarged and control broods (t40.04 = -1.50, P = 0.42, r = 0.23) or reduced and control broods
(t42.00 = 1.61, P = 0.34, r = 0.28) were detected. Nestlings gained mass at a similar rate
regardless of the plumage brightness treatment of the female or brood size treatment;
however, the study area influenced how fast nestlings gained body mass (Table 2.1).
2.4.3. Fledging success
The proportion of nestlings that fledged per brood was similar among plumage
brightness treatments (F2,58 = 0.07, P = 0.93, ηp2 = 0.002), but differed among brood size
treatments (F2, 58 = 8.14, P = 0.001, ηp2 = 0.22) and study areas (F2, 58 = 3.62, P = 0.03, ηp2 =
0.11), and decreased with hatching date (B = -0.09 ± 0.04, F1,58 = 4.95, P = 0.03, ηp2 = 0.08);
however, there was also some suggestion of an interaction between plumage brightness
treatment and brood size treatment (F4, 58 = 2.29, P = 0.07, ηp2 = 0.14). Analyzing the data by
plumage treatment showed that within the reduced plumage brightness treatment fledging
success differed by brood size treatment (F2, 18 = 21.58, P < 0.0001, ηp2 = 0.71, Figure 2.4b),
but no effects of hatching date (F1, 18 = 2.79, P = 0.11, ηp2 = 0.13) or study area (F2, 18 = 0.32,
P = 0.73, ηp2 = 0.03) were detected. Post-hoc tests showed that fewer nestlings fledged from
enlarged broods compared to control (t18 = 5.20, P < 0.0001, r = 0.77) and reduced broods
(t18 = 6.02, P < 0.0001, r = 0.82; Figure 2.4b), but no differences in fledging success were
27

detected between control and reduced broods (t18 = 1.50, P = 0.46, r = 0.33). The low
proportion of fledging success for the enlarged treatment group was because 5 of 8 nests
failed prior to fledging, presumably because parents abandoned their nesting attempts. In
contrast, fledging success did not differ by brood size treatment in the control or enhanced
plumage brightness treatments (all P values > 0.35; Figure 2.4b). Fledging success differed
by study area (F2,18 = 3.82, P = 0.04, ηp2 = 0.30) and was negatively related to hatching date
in control broods (B = -0.15 ± 0.07, F1, 16 = 4.72, P = 0.04, ηp2 = 0.21); these effects were not
detected in broods where female plumage was experimentally enhanced (study area: F2, 16 =
1.28, P = 0.31, ηp2 = 0.14; hatching date: B = -0.04 ± 0.09, F1,16 = 0.20, P = 0.66, ηp2 = 0.01).
2.5. Discussion
My aim was to test whether female attractiveness influenced male investment in
parental care by experimentally altering the plumage brightness of female tree swallows. I
also simultaneously manipulated brood size to determine whether male investment in
parental care when mated to females that differed in attractiveness is contingent on the
demand of the brood. Contrary to my predictions, I found no evidence that males adjust their
feeding rates when paired to females that differed in plumage brightness (Figure 2.2a);
therefore, my experiment provides no support that differential allocation occurs in tree
swallows in response to variation in plumage brightness of females, although I recognize my
samples sizes were small and may have limited my ability to detect differences in male
feeding rates. However, when I compared feeding rates of males between all experimental
broods (i.e., reduced and enhanced treatments combined) and control broods, my results
showed that males tended to feed experimental broods at a higher rate than control broods
(Figure 2.2b).
28

Higher feeding rates by males rearing experimental broods may be due to males in
each treatment group responding differently to the change in plumage brightness of their
mate. For example, males in the enhanced plumage brightness treatment may have increased
their investment in parental care when mated to an attractive mate (positive differential
allocation), as previously demonstrated by Burley (1988), whereas males rearing broods
where female plumage brightness was reduced may have increased their feeding rate to
compensate for the perceived poor quality of their mate (Harris and Uller, 2009). This pattern
of negative differential allocation has been previously demonstrated in blue tits, where males
increased their provisioning rates in response to an experimental reduction of their mate’s
UV crown colour (Limbourg et al., 2013). Negative differential allocation was also recently
suggested to occur in tree swallows by a study showing that both sexes provisioned at higher
rates when mated to a partner with greener plumage (Dakin et al., 2016). However, it seems
unlikely that my results are the response of males adjusting their level of care according to
the attractiveness of female plumage brightness. Since my experiments were performed when
nestlings were two days old, a more likely explanation for this trend is that males perceived
the sudden change in their mates’ phenotype as an indication that something was amiss, and
increased provisioning to compensate. That female feeding rates did not differ in relation to
the plumage brightness treatment is consistent with the idea that males increased
provisioning to compensate for the changed appearance of females and not because of lower
parental investment by females with manipulated plumage brightness.
Previous studies testing male differential allocation have generally demonstrated that
males adjust investment in parental care according to mate attractiveness as predicted by
positive differential allocation (e.g., Burley, 1988; Roulin, 1999; Pilastro et al., 2003; Matessi
et al., 2009; Mahr et al., 2012, but see Limbourg et al., 2013). For example, in response to
29

reduced plumage spottiness in female barn owls (Tyto alba) and reduced UV crown
reflectance in female blue tits, males provisioned nestlings less than males paired to control
females (Roulin, 1999; Mahr et al., 2012). While provisioning rate of male rock sparrows
was not influenced by the reduced breast patch size of females, males did reduce their level
of nest defence when paired to females displaying the less attractive phenotype compared to
control females (Pilastro et al., 2003; Matessi et al., 2009). That male tree swallows did not
provision in relation to female attractiveness suggests that plumage brightness is not a signal
assessed by males, at least when making decisions on how much to invest in parental care.
Given that less ornamented female tree swallows receive less aggression from older
conspecific females (Coady and Dawson, 2013), it is possible that plumage brightness of
females is instead a signal assessed by conspecific females.
Since males did not provision nestlings less when paired to females whose plumage
brightness had been reduced and females fed nestlings at a similar rate regardless of their
plumage brightness treatment, it is difficult to reconcile why nestlings reared in the treatment
where female plumage brightness was reduced and brood size was enlarged grew shorter
ninth primary flight feathers and were less likely to fledge (Figure 2.4a, b). Because I
performed my observations during early to mid-brood rearing to ensure females retained their
manipulated plumage brightness during the feeding observations, I may not have detected an
effect of treatment if males or females adjusted their feeding rates later in the season. This
would be reflected in the size of nestlings at day 16 and fledging success, but not
provisioning rate. Alternatively, nestlings in this treatment may have performed poorly if the
plumage brightness treatment affected the behaviour of manipulated females. While it is
unlikely that the plumage treatment directly affected female behaviour since the application
of these markers and silicone paste have been used previously with no reported adverse
30

effects on behaviour or feather quality (Ballentine and Hill, 2003; Johnsen et al., 2005; Liu et
al., 2007), it is possible that the altered plumage brightness influenced social interactions
among females, and the feedback females received from these interactions influenced their
behaviour or physiology (reviewed in Vitousek et al., 2014b). This social mechanism has
been proposed to explain the increased and decreased levels of androgens in male and female
barn swallows (Hirundo rustica erythrogaster), respectively, the week following an
experimental darkening of their ventral plumage (Safran et al., 2008; Vitousek et al., 2013).
In my study, agonistic interactions with other females may have increased levels of the stress
hormone corticosterone in females with reduced plumage brightness, a response similar to
the increase in corticosterone of female rats experiencing social instability (Haller et al.,
1999), which when combined with the demands of rearing an enlarged brood may have
caused females to abandon their nesting attempt. Although I have no behavioural or
hormonal data to support this mechanism, previous studies have demonstrated that female
barn swallows with greater baseline and stress-induced corticosterone levels were more likely
to abandon their clutch prior to incubation (Vitousek et al., 2014a) and broods of tree
swallows were less likely to survive when corticosterone levels of females were elevated
experimentally (Ouyang et al., 2015).
When faced with changes in brood size, total provisioning rates (trips per hour) of
males rearing reduced broods were similar to males rearing control broods; however, males
rearing enlarged broods provisioned nestlings at a higher rate than males rearing control and
reduced broods (Figure 2.3b). Since total provisioning rate was greater in the enlarged brood
size treatment, individual nestlings received similar amounts of food in enlarged and control
broods (Figure 2.3a). That per-capita feeding rates (trips per hour per nestling) did not
decline with increased brood size is inconsistent with previous studies in this species (e.g.,
31

Leffelaar and Robertson, 1986; Leonard et al., 2000; Murphy et al., 2000; Shutler et al.,
2006; Hainstock et al., 2010). Food abundance at my study areas may be low in comparison
to other populations of tree swallows (e.g., Bortolotti et al., 2011; Harriman et al., 2013), and
so my results may differ from previous studies if males need to invest heavily in offspring
when rearing an enlarged brood to maintain their quality. This may also explain why total
provisioning rates of males rearing reduced broods were not lower compared to controls.
Despite per-capita feeding rates being similar among males (Figure 2.3a), nestlings reared in
enlarged broods were lighter than controls prior to fledging, which suggests that while male
tree swallows were willing to increase parental investment, they were unable to fully
compensate for the lower per capita feeding rates of females rearing enlarged broods.
Overall, my results showed that investment in parental care by male tree swallows
was influenced by factors other than female attractiveness, such as brood size (Ardia, 2007;
Bortolotti et al., 2011) and weather (Rose, 2009). My results also showed that male feeding
rates were positively related to female feeding rates (also see Dakin et al., 2016). This
relationship may exist if males and females similarly adjust their feeding rates to the
perceived needs or cues of nestlings, such as age or the number of nestlings (Leffelaar and
Robertson, 1986; Ardia, 2007), or the begging intensity of hungry nestlings (Leonard and
Horn, 1998; Leonard and Horn, 2001), which increases with brood size (Leonard et al., 2000;
Thomas and Shutler, 2001). In addition, each member of the pair may respond directly to the
level of provisioning effort exerted by their mate (Hinde, 2006). Future experiments that
carefully tease apart the relative contributions of brood size, nestling cues, and partner
behaviour (e.g., Hinde and Kilner, 2007) will improve our understanding of how parental
investment decisions in tree swallows are determined.

32

In conclusion, I found that males did not adjust their investment in parental care in
relation to the experimentally altered plumage brightness of female tree swallows. My results
also showed that offspring quality and fledging success were lowest when female plumage
brightness was reduced and brood size was enlarged. This may be due to social feedback
females with reduced plumage brightness received following agonistic interactions with other
females combined with the increased demands of rearing an enlarged brood. Collectively, the
findings of my study suggest that plumage brightness of females is not a signal of
attractiveness assessed by males, but may instead be a signal assessed by females. As such,
selection on female plumage brightness may arise if brightness signals competitive ability,
thereby influencing female-female competition for access to nest boxes (Bitton et al., 2008).
Future work that experimentally manipulates ornamental traits in females prior to breeding
would greatly improve our understanding of how sexual or social selection acts to promote or
maintain elaborate phenotypic displays in females.

33

Table 2.1. Results of random intercept linear mixed models testing whether nestling size or
growth differed among broods where the plumage brightness of females was experimentally
reduced or enhanced, or remained unchanged (controls), and where brood size was reduced
by removing two nestlings or enlarged by adding two nestlings, compared to control broods
(see text for more details).
F

df

P

Nestling size
Ninth primary feather (mm)
Plumage brightness treatment
Brood size treatment
Time of measurement
Plumage brightness x brood size

1.39
1.22
4.65
3.39

2, 39.49
2, 39.88
1, 39.31
4, 39.71

0.26
0.31
0.04
0.02

Mass (g)
Plumage brightness treatment
Brood size treatment
Time of measurement

0.03
9.99
6.53

2, 46.89
2, 45.65
1, 49.25

0.97
<0.0001
0.01

Nestling growth
Ninth primary
Plumage brightness treatment
Brood size treatment
Study area

2.14
3.58
2.96

2, 41.52
2, 40.87
2, 41.83

0.13
0.04
0.06

Mass
Plumage brightness treatment
Brood size treatment
Study area

0.42
1.25
9.49

2, 43.65
2, 43.03
2, 43.96

0.66
0.30
<0.0001

34

Figure 2.1. Reflectance spectra from the back and rump feathers of female tree swallows
measured before (solid black line) and after treatment with blue permanent marker to
enhanced plumage brightness (top grey line; N = 14) and silicone paste to reduce plumage
brightness (bottom grey line; N = 14). Presented are the means (± SE) at every 50 nm interval
from 300 – 700 nm. See methods for specific details regarding plumage manipulations.

35

Figure 2.2. Mean (± SE) feeding rates (trips/hour/nestling) of male tree swallows rearing
broods where a) female plumage brightness was experimentally reduced or enhanced, or
remained unchanged (controls) and b) female plumage was experimentally altered (enhanced
and reduced treatments combined) or remained unchanged (control). Sample sizes indicate
the number of broods within each treatment group and are given above error bars.
36

Figure 2.3. Mean (± SE) feeding rates expressed as a) trips/hour/nestling and b) trips/hour by
male tree swallows rearing broods where brood size was reduced by removing two nestlings,
enlarged by adding two nestlings, or remained unchanged (controls). Sample sizes indicate
the number of broods within each treatment group and are given above error bars

37

Figure 2.4. Mean (± SE) a) length of ninth primary flight feathers (mm) at day 16 and b)
fledging success (proportion of nestlings fledged per brood) for nestling tree swallows
according to brood size treatment (reduced by two nestlings, increased by two nestlings, or
controls) and female plumage brightness treatment (experimentally reduced and enhanced
compared to controls). Sample sizes indicate the number of broods within each treatment
group and are given above error bars.
38

Chapter 3: Are there social costs of displaying bright plumage for female tree swallows?
3.1. Abstract
The function of female ornamentation as signals of quality has gained recent
empirical support; however, our understanding of the social costs that maintain signal
honesty are lacking. Two hypotheses have been proposed for socially enforcing honest
signals: the incongruence (punishment) and social control (like-versus-like aggression)
hypotheses. In tree swallows (Tachycineta bicolor), females aggressively compete for and
defend nest sites (i.e., males with a nest box), and as such, nest site intrusions by conspecifics
may prevent dishonest signaling. I tested whether plumage brightness of females influences
nest site retention and reproductive success by experimentally enhancing and reducing the
plumage brightness of female tree swallows relative to controls prior to breeding. Females in
the enhanced and control plumage brightness treatments were less likely to retain their nest
site than females in the reduced plumage brightness treatment. Moreover, clutch initiation
date was later for females in the enhanced plumage brightness treatment compared to females
in the control and reduced plumage brightness treatments. Overall, my results for nest site
retention are consistent with the social control hypothesis. This is further supported by
evidence of a social cost of delayed breeding imposed on females whose signal quality is
repeatedly challenged. Since repeated agonistic interactions may be costly, studies examining
the trade-offs between female ornamentation and fecundity should consider the underlying
social costs that maintain signal honesty.
3.2. Introduction
Females from a wide variety of taxa display elaborate ornaments that are
hypothesized to evolve by sexual and/or social selection (Clutton-Brock, 2009; Tobias et al.,
39

2012), or by a correlated response to selection acting directly on traits displayed by males
(Lande, 1980). Competition among females for access to mates or resources is often intense
(e.g., Leffelaar and Robertson, 1985; Moreno, 2015), and ornamental traits displayed by
females may play a role during competitive interactions (LeBas, 2006; Clutton-Brock, 2009).
Such traits may act directly as weaponry (Watson and Simmons, 2010b) and/or signal the
competitive ability or status of females (Pryke, 2007; Midamegbe et al., 2013; Morales et al.,
2014).
Traits that function by signaling the quality of an individual to conspecific rivals
should convey honest information about the signaller. Signal honesty may be maintained by
the physiological costs of producing and maintaining elaborate ornaments (reviewed in
Tibbetts, 2014; Vitousek et al., 2014b) or social costs of possessing ornaments that are
imposed on signallers by conspecifics (Maynard Smith and Harper, 1988). Two mechanisms
have been hypothesized to explain how social costs can maintain signal honesty (reviewed in
Senar, 2006; Tibbetts, 2014). The ‘incongruence hypothesis’ posits that low-quality
individuals that dishonestly signal high quality will be punished by conspecifics through
increased aggression or attacks when a mismatch between an individual’s behaviour and
signal is detected during social interactions (Rohwer and Rohwer, 1978; Ferns and Hinsley,
2004; Tibbetts and Izzo, 2010). Alternatively, the ‘social control hypothesis’ predicts that
individuals displaying high-quality signals frequently engage in agonistic interactions with
other individuals of high quality to test the quality of signal displayed (‘like versus like’
aggression, Senar, 2006); consequently, low-quality individuals dishonestly signaling high
quality will incur costs by being challenged frequently and aggressively by conspecifics that
are truly high quality (Rohwer, 1977; Møller, 1987). Recently, it also has been suggested that
social and physiological costs may act in concert to maintain signal honesty (Tibbetts, 2014;
40

Vitousek et al., 2014b). For example, painting the shield of male pukekos (Porphyrio
porphyrio) to reduce the quality of signal displayed resulted in increased aggression from
conspecifics and a dynamic reduction in the natural size of the shield following manipulation
compared to controls (Dey et al., 2014).
The hypotheses for the social enforcement of quality signals were originally
formulated and studied in the context of status signaling during social contests among
conspecifics (Senar, 2006); however, it is becoming increasingly recognized that social costs
may play an important role in maintaining the honesty of ornamentation (reviewed in
Tibbetts, 2014; Vitousek et al., 2014b). Recent empirical studies have demonstrated that
social interactions among conspecifics may alter the physiological state, such as androgen
levels or oxidative stress, of individuals displaying dishonest signals (e.g., Safran et al., 2008;
Vitousek et al., 2013). Such changes in physiology of dishonest signallers may be mediated
by conspecifics challenging the quality of signal displayed; however, whether or not signals
are tested may depend on the value of the resource relative to the costs of conflict (Maynard
Smith and Harper, 1988; Tibbetts, 2008). For example, Vitousek et al. (2016) hypothesized
that one mechanism explaining the difference in physiological state between male and female
North American barn swallows (Hirundo rustica erythrogaster) following signal
enhancement (higher and lower oxidative stress, respectively) was that signals displayed by
males may be tested, while those of females may be trusted. Since barn swallows build their
own nest, females may be more likely to trust signals of quality compared to obligate cavitynesting species that require existing nest sites to breed, such as pied flycatchers (Ficedula
hypoleuca) that compete aggressively with conspecific females because nest sites are limited
(Moreno, 2015). In line with this, Moreno et al. (2013) reported that non-ornamented female
pied flycatchers manipulated to display a forehead patch had higher oxidative damage
41

compared to control females and hypothesized that competition for limited nest sites may
enforce the honesty of female ornamentation. Territorial defence by resident female birds
toward decoy intruders is strongly influenced by the ornamentation of both the resident and
intruding female (Midamegbe et al., 2013; Morales et al., 2014), which may suggest that nest
site intrusions by conspecific females provides a mechanism to enforce the honesty of quality
signals, but to my knowledge this has never been experimentally tested in free-living female
birds.
In the current study, my aim was to examine whether bright plumage of female tree
swallows (Tachycineta bicolor) is a signal of quality enforced by agonistic interactions with
conspecifics. Female tree swallows aggressively compete with conspecific females for access
to a male with a nest site (Leffelaar and Robertson, 1985), and females that are more
aggressive are more likely to acquire nest sites when their availability is experimentally
reduced (Rosvall, 2008). Because some females are competitively excluded from breeding,
there often is a large population of floater females that breed only when a nest site becomes
available (Stutchbury and Robertson, 1985). Females successful at acquiring a nest site
frequently experience territory intrusions from conspecifics and must aggressively defend
their nest to avoid usurpations (Leffelaar and Robertson, 1985). Overall, competition among
female tree swallows is often so intense that it can lead to injury or death (e.g., Leffelaar and
Robertson, 1985).
Dorsal plumage colour of female tree swallows is variable ranging from dull brown to
bright iridescent blue-green, the latter being similar to plumage displayed by males. Females
displaying dull brown plumage are generally in their first breeding season (second year of
life: SY; Hussell, 1983), and this delay in maturation of their plumage may signal low
competitive ability since SY females receive less intra-sexual aggression from ornamented
42

after-second-year (ASY) females (Coady and Dawson, 2013). Among ASY females, those
that are more ornamented display plumage that is brighter, with greater UV and blue chroma,
and reflects light maximally at shorter wavelengths (i.e., bluer hue; Bitton et al., 2008; Bentz
and Siefferman, 2013). Although it is currently unknown whether bright plumage of ASY
females similarly functions as a status signal, females that display bright plumage
assortatively mate with bright males (Bitton et al. 2008), and since bright plumage of females
does not appear to be a signal of attractiveness preferred by males (Berzins and Dawson,
2016; Chapter 4), assortative mating in this species likely occurs because females with bright
plumage have greater competitive ability.
I tested whether plumage brightness of females influenced nest site retention by
experimentally enhancing and reducing the plumage brightness of female tree swallows prior
to breeding relative to controls (see Berzins and Dawson, 2016). If plumage brightness is a
signal of quality that is socially enforced by conspecifics challenging females with
incongruent signals, I predicted that females whose ornamentation was enhanced and
reduced, so that their quality and behaviour mismatched, would be less likely to retain their
nests (Table 3.1). In contrast, if the honesty of quality signals is enforced by social control
whereby conspecific females aggressively challenge females displaying ornamentation of
similar quality, I predicted that females whose plumage brightness was enhanced to signal
higher quality than their true competitive ability would be less able to retain their nest sites,
whereas females displaying reduced plumage brightness whose quality of signal is lower than
their true competitive ability would be better able to defend and retain their nests (Table 3.1).
It is also possible that feedback from agonistic interactions could alter hormone levels of
females so that the behaviour of females becomes congruent with their manipulated plumage
signals (Vitousek et al., 2013, 2014b). If this were the case, I predicted that females
43

displaying enhanced plumage brightness would be more likely to retain their nest sites,
whereas reduced females may be more likely to be usurped from their nests (Table 3.1). I
also tested whether the quality of nest site occupied influences retention and the reproductive
success of females, since social costs imposed on dishonest signallers may result in lower
reproductive success (Kotiaho, 2001).
3.3. Material and methods
3.3.1. Study area and general field methods
I studied tree swallows breeding in nest boxes near Prince George BC, Canada (53ºN,
123ºW) from May to August in 2010 and 2011. The study area consisted of open agricultural
areas intermixed with small wetlands and patches of coniferous and deciduous trees (see
Dawson et al. 2005 for more details). Shortly after tree swallows arrived in my study area in
early May, I began to check nest boxes daily to document nest building. Once a nest box
contained a nest with a fully formed cup, but before any eggs had been laid, I captured the
resident female. Each female was banded with an individually numbered aluminum leg band
and weighed using a spring balance (nearest 0.25 g). Five rump feathers were collected for
spectral analysis (details below). I determined the age of females using dorsal plumage
colour (Hussell 1983), and ASY females were sequentially allocated to the enhanced,
reduced, or control plumage brightness treatments by capture order, after determining a
treatment order randomly. Details of the plumage brightness manipulation are described in
Berzins and Dawson (2016). Briefly, plumage brightness was enhanced by applying nontoxic permanent blue marker, reduced by applying silicon paste evenly to the dorsal feathers
of females, or remained unchanged by treating females with a marker containing water.
Permanent markers and silicone paste have previously been used to alter plumage colour
44

(e.g., Johnsen et al., 2005; Safran et al., 2008) and since the effects of these treatments have
been detected on average for 15 days after being applied to feathers (Johnsen et al., 2005), I
was confident that my treatment produced the desired effect throughout the experimental
period. The brightness of plumage following the manipulation fell within the natural range of
plumage brightness for female tree swallows (Berzins and Dawson, 2016). Females in each
treatment did not differ by age, body mass, plumage characteristics (see below), or capture
date prior to experimental manipulation (all P values > 0.29). In total, I enhanced the
plumage brightness of 30 females, reduced the plumage brightness of 31 females, and 28
females were in the control group.
To examine whether experimentally altering female plumage brightness prior to
breeding influenced reproductive success, I monitored all nests at my study area daily to
record the date of clutch initiation (where January 1 = 1) and clutch size. Freshly laid eggs
were numbered with a non-toxic marker for identification and weighed with a digital scale
(nearest 0.01 g). I captured all females in my study area after their eggs had hatched and
recorded the box and band number to determine whether females retained, left and/or
switched nest sites after being manipulated.
3.3.2. Spectral analysis
The spectral analysis for females used in this study is described in Berzins and
Dawson (2016). Briefly, I quantified plumage characteristics of female tree swallows by
measuring the reflectance of feathers using a JAZ-PX spectrometer with a xenon light source
(Ocean Optics, Dunedin, FL, USA). Four rump feathers were taped to a piece of cardboard so
that they resembled the natural arrangement of rump feathers on a bird following Bitton et al.
(2008). Feather samples were placed on a black non-reflective surface and spectral data were
45

recorded using SpectraSuite software (Ocean Optics, Dunedin, FL, USA). Three
measurements for each bird were taken, moving the probe between each measurement.
The R package pavo (Maia et al., 2013) was used to smooth spectral curves and
quantify plumage characteristics, specifically hue, ultraviolet (UV) and blue chroma, and
average brightness, as these variables have been shown previously to reflect aspects of
quality in female tree swallows (Bitton et al., 2008; Bentz and Siefferman, 2013). The three
measures of each plumage characteristic were averaged, and then a principal components
analysis (PCA) was used to reduce each bird’s plumage characteristics to a single colour
score (Montgomerie, 2006). PCA was performed using all ASY females from which I
collected feather samples over a four-year period (2010-2013) at my study area. The first
principal component (PC1) explained 55 % of the total variation and the factor loadings
suggested that PC1 was heavily weighted by hue (-0.98), UV chroma (0.88), and blue
chroma (0.70) and so females with larger positive values had plumage with greater UV and
blue chroma, and reflected light at shorter wavelengths (Berzins and Dawson, 2016). The
second principal component (PC2) explained 26.2 % of the total variation and the factor
loadings suggested that PC2 was heavily weighted by brightness (0.95) so females with
larger positive values had brighter plumage (Berzins and Dawson, 2016).
3.3.3. Nest-site quality
Following Potti and Montalvo (1991), I calculated ‘nest-site quality’ as the number of
years a nest site was occupied in relation to the number of years the nest site was available
for breeding by tree swallows. My study area was established in 2002, so I used occupancy
data from 2003 to 2009 in my calculations. I confirmed whether this measure of occupancy
reflected nest-site quality in tree swallows by testing whether the number of years a nest site
46

was occupied was related to average clutch initiation date at that site (standardized so that
day 1 represented the first clutch initiation within each year at the site) and the total number
of nestlings fledged from 2004 to 2006 (I restricted my analysis to these years as
experimental manipulations in other years may have influenced fledging success; e.g.,
Dawson et al. 2005). Overall, box occupancy was negatively correlated with clutch initiation
date (Spearman rank correlation: rs = - 0.27, N = 118 nests, P < 0.01) and positively
correlated to the number of nestling fledged (Spearman rank correlation: rs = 0.57, N = 100
nests, P < 0.0001). That nest boxes used more frequently are bred in earlier and have a
greater number of nestlings fledged indicates that occupancy provides a reliable proxy for
nest-site quality. Nest-site quality did not differ among treatments prior to manipulation
(Kruskal-Wallis analysis of variance: H = 0.49, df = 2, P = 0.78).
3.3.4. Statistical analysis
I first tested whether the plumage brightness treatment of female tree swallows
influenced nest site retention using a logistic regression with nest retention (retained, left) as
the binary dependent variable, plumage brightness treatment as a fixed factor and quality of
the nest site and original plumage colour (PC1 and PC2) as covariates; however, these
covariates were not important predictors of whether a female stayed or left her nest site (all P
values > 0.26), so a likelihood ratio test was subsequently performed. Nine females were
captured in both years and I randomly chose data either from 2010 or 2011 to use in analyses
to avoid the inclusion of multiple non-independent observations for females. For those
females that switched nest sites after being manipulated, I tested whether the quality of nest
site a female subsequently occupied differed before and after moving using a repeated
measures analysis of variance (ANOVA). In this analysis, I included time (before and after
47

moving) as the within-subject factor and plumage brightness treatment as the betweensubject factor, and the interaction between time and plumage brightness treatment.
I tested whether altering the plumage brightness of females influenced their
reproductive success using linear mixed models (LMM; lme4; Bates et al., 2015) that
included female identity as a random factor to account for the multiple observations of
females. I tested whether clutch initiation date, standardized to a mean of 0 and a standard
deviation of 1 for experimental nests in each year, differed by plumage brightness treatment.
In this model, I also included whether females retained or switched nest sites as a categorical
variable, and female body mass as a covariate. I also tested whether clutch size and average
egg mass differed by the plumage brightness treatment of the female. I included year as a
categorical variable and standardized clutch initiation date as a covariate to control for the
seasonal decline in clutch size that occurs in tree swallows (Winkler et al., 2014). For these
analyses, I used the R package lmerTest (Kuznetsova et al. 2016) to calculate P values for
fixed factors and covariates.
For all analyses, variables that did not approach significance (P ≥ 0.10) were removed
in a backwards, stepwise fashion. I checked the residuals of models for normality and
heteroscedasticity (Cleasby and Nakagawa, 2011). All statistical tests were two-tailed and
results were considered significant at P ≤ 0.05. The overall significance of omnibus tests was
examined using post-hoc tests with false discovery rate adjustment for multiple comparisons
(Verhoeven et al., 2005). Where appropriate, I report least squares means ± 1 standard error
(SE), parameter estimates (B), and effect sizes. Effects sizes are reported as general eta
squared (ηG2) for omnibus statistical tests (Lakens, 2013) and the correlation coefficient (r)
for post-hoc comparisons (Field et al. 2012). All statistical analyses were performed using

48

SPSS v20 (IBM Corp. 2011) and R v3.2.3 (R Development Core Team 2015) statistical
software.
3.4. Results
3.4.1. Nest site retention
Plumage brightness treatment influenced whether female tree swallows retained or
left their nest site (G2 = 7.67, N = 80, P = 0.02, Table 3.2). To further examine these results, I
performed sub-analyses between pairs of treatment groups and found that females with
enhanced plumage brightness were more likely to leave their nest site following manipulation
than females with reduced plumage brightness (G1 = 7.35, N = 53, P < 0.01), but not the
control treatment (G1 = 0.67, N = 54, P = 0.41). Control females also tended to be more
likely to leave their nest sites than females whose plumage was experimentally reduced (G1 =
3.69, N = 53, P = 0.055). In total, 92 % of reduced, 70 % of control, and 56 % of enhanced
females bred following plumage brightness treatment (Table 3.2).
For females that switched nest sites following manipulation, plumage brightness
treatment did not influence the quality of the new nest site occupied (F2, 13 = 0.53, P = 0.78,
ηG2 = 0.03); however, the newly acquired nest site was overall of lower quality than the nest
site occupied by the female prior to manipulation (F1, 15 = 5.01, P = 0.04, ηG2 = 0.11).
3.4.2. Reproductive success
Clutch initiation date differed by the plumage brightness treatment of the female (F2,
60.00 = 3.80, P = 0.03, Figure 3.1a), and post-hoc tests revealed that females in the enhanced

plumage brightness treatment initiated clutches later than control females (B = -0.70 ± 0.28,
t60.0 = -2.50, P = 0.03, r = 0.31) and females with experimentally reduced plumage brightness

49

(B = -0.63 ± 0.26, t60.00 = -2.37, P = 0.03, r = 0.29). Females that switched nest sites after
being manipulated also initiated clutches later than females that retained their nest site (B =
-0.88 ± 0.25, F1, 60.00 = 12.40, P < 0.001, Figure 3.1b), as did those females that had lower
body mass when captured pre-breeding (B = -0.13 ± 0.07, F1, 60.00 = 3.57, P = 0.06).
As expected, there was a significant decline in clutch size with later dates of clutch
initiation (B = -0.36 ± 0.12, F1, 60.91 = 9.41, P < 0.01), as well as a decline in average egg
mass with initiation date (B = -0.06 ± 0.02, F1, 47.91 = 9.58, P < 0.01). Neither clutch size nor
average egg mass differed by the plumage brightness treatment of the female (F2, 60.68 = 0.42,
P = 0.66; F2, 50.74 = 0.23, P = 0.80, respectively), but clutches were smaller in 2011 than 2010
(B = -0.52 ± 0.23, F1, 58.50 = 5.39, P = 0.02).
3.5. Discussion
I manipulated the plumage brightness of female tree swallows to test whether
plumage brightness treatment influenced a female’s ability to retain her nest site, and whether
plumage brightness is a signal of quality that is socially enforced by conspecifics. Females in
the enhanced plumage brightness treatment were less likely to retain their nest site and breed
than females with experimentally reduced plumage brightness (Table 3.2). A similar trend
was also observed for control females compared to those with reduced plumage brightness
(Table 3.2). There are three possibilities to explain why females manipulated to display
bright, presumably high-quality, plumage ornamentation were less likely to retain their nest
site than females with dull, low-quality plumage. First, lower nest site retention may have
occurred if females in the enhanced plumage brightness treatment were depredated as a result
of increasing the conspicuousness of their plumage (Huhta et al., 2003). Indeed, of the
females that did not retain their nest site and breed after being manipulated, only two were
50

observed breeding at my study area in subsequent years (Berzins and Dawson, unpublished
data). If my results for nest site retention were due to predation, then females in the enhanced
plumage brightness treatment should have been less likely to retain their nest site than control
females, but there were no difference between the enhanced and control treatments (Table
3.2). Second, females in the enhanced plumage brightness treatment may have left their nest
site to obtain a higher quality site or mate (Otter and Ratcliffe, 1996) after being manipulated
to signal high quality, but I find it unlikely that a female would relinquish a limited resource
essential for breeding after intense competition with conspecifics (Leffelaar and Robertson,
1985; Rosvall, 2008). This explanation also does not explain why control females also tended
to be more likely to leave their nest site than females in the reduced plumage brightness
treatment. Moreover, females that switched nest sites bred later (Figure 3.1b) and in poorer
quality sites than the nest site they originally occupied so there appears to be little advantage
to females that leave their nest site. Third, my experiment may have altered social
interactions among females, as similarly reported in male collared flycatchers (Ficedula
albicollis) following the experimental enlargement of their forehead patch size (Qvarnström,
1997). Since nest usurpation by intruding female conspecifics occurs frequently in tree
swallows (Leffelaar and Robertson, 1985), females that left their nest site after being
manipulated were likely unable to defend their nest site from conspecifics; male tree
swallows are rarely involved in defence against intruding females and readily breed with the
female that takes over the nest site (Leffelaar and Robertson, 1985). Since female tree
swallows defending nest sites often suffer serious injury or death (Leffelaar and Robertson,
1985), this would also provide an explanation for why so few females that were usurped were
never observed breeding at my study area, although I cannot rule out that they bred in
subsequent years in areas outside my study area. Overall, the finding that plumage brightness
51

treatment influenced nest site retention suggests that agonistic interactions among female tree
swallows may enforce signal honesty, and the cost of dishonestly signaling high quality may
be usurpation. Since the average life expectancy of tree swallows is 2.7 years (Winkler et al.,
2011), females unable to retain their nest site may have to forego breeding, thereby imposing
fitness costs on unsuccessful females.
I predicted that if my results were consistent with the incongruence hypothesis
(Rohwer and Rohwer, 1978; Tibbetts and Izzo, 2010), females in both the enhanced and
reduced plumage brightness treatments would be less likely to retain their nest site due to
increased agonistic interactions from conspecifics (Table 3.1); however, contrary to these
predictions, nest site retention did not differ between females in the enhanced and control
plumage brightness treatments, and females with reduced plumage brightness tended to be
more likely than control females to retain their nest site (Table 3.2). This suggests that
females signaling dishonestly high or low quality did not experience increased aggression or
attacks from conspecifics as a result of incongruent signals. Although it is possible that a
social feedback mechanism altered the behaviour of females so that it was congruent with the
quality signalled by their plumage brightness treatment (Vitousek et al., 2014b), I believe this
is unlikely to explain my results for tree swallows for two reasons. First, females in the
enhanced plumage brightness treatment were not more likely than controls to retain their nest
site, which is inconsistent with previous studies showing that subordinates whose traits were
manipulated to signal high status became dominant against rivals following manipulation
(Rohwer, 1985); however, I do recognise that females in my study could have been usurped
from their nest site before feedback from repeated interactions with conspecifics altered
hormone levels and subsequently their behaviour. Second, if the behaviour of females in the
reduced plumage brightness treatment did become congruent with that of a low-quality
52

signaller, then fewer females in the reduced plumage brightness treatment should have
retained their nest site compared to control females either because they were more likely to
be usurped or to abandon their nest if interactions with conspecifics were stressful (Berzins
and Dawson 2016); instead, I observed the opposite (Table 3.2).
An alternative mechanism that is hypothesized to maintain the honesty of signals is
social control (Rohwer, 1977). That fewer female tree swallows in the enhanced plumage
brightness treatment retained their nest site compared to females with reduced plumage
brightness (Table 3.2) is consistent with the social control hypothesis. If challenged
aggressively and frequently by naturally bright females, then females in the enhanced
plumage brightness treatment may have been less able to defend their nest site from intruders
of greater competitive ability. Since social control can only occur in species that exhibit likeversus-like aggression between high-quality individuals (Senar, 2006), the finding that
control females also tended to be less likely to retain their nest site than females in the
reduced plumage brightness treatment suggests that conspecifics challenge signal quality by
way of nest site intrusions. The results for females in the reduced plumage brightness
treatment may also support the social control hypothesis if reduced females were better able
to defend and retain their nest site against similarly low- or mid-quality conspecifics with
poor or equivalent competitive ability; however, this explanation of my results suggests that
dishonestly signaling low quality is advantageous to females because they are more likely to
breed. Without a cost imposed on dishonest signallers, such a signaling system would be
susceptible to dishonesty (e.g., Owens and Hartley, 1991). Therefore, it is also possible that
females in the reduced plumage brightness treatment were able to retain their nest site simply
because they were involved in fewer agonistic interactions overall with brighter females. For

53

example, less ornamented second-year female tree swallows receive fewer intrusions onto
their own territories by older, more ornamented females (Coady and Dawson 2013).
Whether or not signals of quality are challenged may depend on the value of the
resource and costs of conflict (Maynard Smith and Harper 1988). For example, conspecific
paper wasps (Polistes dominulus) challenged food guards when the value of the resource was
high (Tibbetts, 2008), but when the value of the resource was low, conspecifics only
challenged those food guards signaling low quality and avoided those signaling high quality
(Tibbetts and Lindsay, 2008). My results for tree swallows may similarly depend on the
value of the resource; conspecifics may challenge females to test signal quality prior to
breeding when the value of obtaining a nest site is greater than the cost of fighting.
Depreciation in the value of nest sites toward the end of the breeding season may explain
why my results differ from a previous study that altered the plumage brightness of female
tree swallows after the hatching of eggs; females in the reduced plumage brightness treatment
were more likely to abandon their nesting attempt when rearing an experimentally enlarged
brood compared to females in enhanced and control plumage brightness treatments (Berzins
and Dawson, 2016). These results may be due to enhanced females signaling high quality
receiving fewer intrusions or agonistic interactions than females in the reduced plumage
brightness treatment (Berzins and Dawson, 2016), as described above in paper wasps when
the value of the resource is low (Tibbetts and Lindsay, 2008). Collectively, the results of the
two studies suggest that bright plumage displayed by females is tested when the contested
resource is valuable (this study), but trusted when the value of the resource is reduced
(Berzins and Dawson, 2016). Dale and Slagsvold (1995) reported that the outcome of
contests over nest sites in female pied flycatchers was determined by the value of the
breeding opportunity to the female and not asymmetries in resource-holding potential; this
54

also has been suggested in other studies examining female contest behaviour (see Elias et al.,
2010 and references therein). That nest site retention by female tree swallows was not related
to nest-site quality or the original plumage colour of females provides further evidence that
signal quality is tested because breeding opportunities to female tree swallows are valuable.
Social costs imposed on dishonest signallers may include increased aggression from
conspecifics, lower reproductive success, or increased energy expenditure (Kotiaho, 2001). I
recognize that by measuring nest site retention I was unable to directly observe any
aggressive interactions among females; however, manipulating wild female birds prior to
breeding allowed me to examine whether there were social costs for female reproductive
success. Indeed, females in the enhanced plumage brightness treatment that dishonestly
signalled high quality suffered social costs since they initiated their clutches later than
females in the control and reduced plumage brightness treatments (Figure 3.1a). Delayed
breeding for females in the enhanced plumage brightness treatment may have been due to
defending their nest site from naturally bright females who are likely superior competitors
since they pair with bright, high-quality male tree swallows (Bitton et al., 2008). Defence of
nest sites against superior competitors may reduce resources available for reproduction and
self-maintenance. As such, social costs should be considered in future studies examining
trade-offs between ornamentation and fecundity (Fitzpatrick et al., 1995) because, as my
results show, social costs of possessing ornamental traits may also lower female reproductive
success. Whether such costs translate into lower offspring quality remains to be tested, but
correlative studies in tree swallows suggest that females that are more ornamented, and
presumably of higher quality (Bitton et al., 2008; Bentz and Siefferman, 2013), produce
offspring of lower quality (Coady, 2011; Bentz and Siefferman, 2013). Moreover,
interactions with conspecifics may increase social instability resulting in elevated levels of
55

the stress hormone corticosterone (Haller et al., 1999), which may also explain why more
ornamented female tree swallows have greater levels of nest parasitism, poorer immune
defences, and lower hematocrit levels (Coady, 2011; Bentz and Siefferman, 2013). This is
supported by a recent study in pied flycatchers that reported higher levels of blood
malondialdehydes, indicative of oxidative damage, in non-ornamented females manipulated
to display a forehead patch compared to control females (Moreno et al. 2013).
In conclusion, the results of my experiment are consistent with Moreno et al. (2013)
in demonstrating that the honesty of signals in female birds may be enforced by social control
during the breeding season when the value of breeding opportunities is high. Fewer females
in the enhanced plumage brightness treatment were able to defend and retain their nest site
compared to females in the reduced plumage brightness treatment (Table 3.2), and a similar
trend was observed for control females compared to those in the reduced plumage brightness
treatment, which supports previous studies demonstrating that individuals of similar signal
quality engage in more aggressive interactions (Møller, 1987). Although nest site retention
did not differ between the enhanced and control plumage brightness treatments, the delayed
breeding date for females in the enhanced plumage brightness treatment is consistent with a
social cost imposed on females whose elaborate ornamentation is challenged by high-quality
females. As such, social costs of displaying ornamented plumage may reduce resources
available for reproduction, as hypothesized for production or maintenance costs of
ornamental traits (Fitzpatrick et al., 1995). Additional studies that experimentally manipulate
the ornamental traits of females prior to breeding would greatly improve our understanding
of social interactions among females and the costs of displaying such traits on female
reproductive success.

56

Table 3.1. Predictions for whether female tree swallows retain their nest site or are usurped following a plumage brightness
manipulation to reduce or enhance plumage brightness compared to controls.

57

Plumage
brightness
treatment

Predictions
of nest
retention

Rationale for nest site retention prediction

Reduced

Usurp

Control

Retain

Enhanced

Usurp

Individuals manipulated to signal high quality receive more aggression

Reduced

Retain

Social control hypothesis
Individuals manipulated to signal low quality are challenged by lowquality signallers

Control

Usurp

Enhanced

Usurp

Reduced

Usurp

Control

Retain

Enhanced

Retain

Incongruence hypothesis
Individuals manipulated to signal low quality receive more aggression

Individuals manipulated to signal high quality are challenged by highquality signallers
Social feedback hypothesis
Individuals manipulated to signal low quality receive more aggression
and adopt the behaviour of a low-quality signaller

Individuals manipulated to signal high quality receive less aggression
and adopt the behaviour of a high-quality signaller

Reference

Ferns and
Hinsley 2004
Tibbetts and
Izzo, 2010
Rohwer 1977

Møller, 1987

Dey et al., 2014

Rohwer, 1985

Table 3.2. The number of female tree swallows whose plumage brightness was experimentally reduced and enhanced, or remained
unchanged (control), prior to breeding in 2010 and 2011, and the number of females in each treatment that retained or left their nest
box, and successfully bred following manipulation. Nine females were manipulated in both years and are only represented once in the
data set (see text for details).
Treatment

Manipulated

Retained box

Left box

Reduced
Control
Enhanced

26
27
27

20
14
11

6
13
16

Found new
box
4
5
4

Total number of breeding
females after manipulation
24
19
15

58

Figure 3.1. Mean (± SE) clutch initiation date of female tree swallows according to a)
whether their plumage brightness was experimentally reduced or enhanced, or remained
unchanged (controls), and b) whether they retained or switched nest sites following
treatment. Clutch initiation dates were standardized to a mean of 0 and a standard deviation 1
for experimental nests separately for each year. Negative values indicate clutches initiated
early, while positive values indicate clutches initiated later in the breeding season. Sample
sizes are given above error bars.
59

Chapter 4: Does bright plumage enhance extra-pair mating success of female
passerines? An experiment with tree swallows
4.1. Abstract
Recent empirical evidence suggests that ornamental traits displayed by females may
have a signalling function during competitive interactions and/or mate attraction; however,
less is known about how such traits influence the social and extra-pair mating success of
females. To examine whether the attractiveness of female ornamentation influences the
quality of social and extra-pair mates, and the proportion of extra-pair paternity in the broods
of females, I experimentally enhanced and reduced the plumage brightness of female tree
swallows (Tachycineta bicolor) relative to controls. Contrary to my predictions, the quality
of social mate acquired by females did not differ by plumage brightness treatment.
Moreover, the rate of extra-pair paternity was similar among females regardless of plumage
brightness treatment, and I found no evidence to suggest that this was due to social mates
influencing extra-pair mating opportunities of females in response to the their perceived
attractiveness. Consistent with previous studies in my population, pairwise comparisons
showed that social and extra-pair mates did not differ for any phenotypic traits measured in
the reduced and control plumage brightness treatments; however, the extra-pair mates of
females with enhanced plumage brightness had longer flight feathers than social mates. This
result may be due to social feedback from conspecifics about signal quality influencing
female choice or ability to pursue high-quality extra-pair mates. Overall, my study shows
that female tree swallows displaying bright plumage have greater extra-pair mating success,
and highlights the importance of manipulating the ornamental quality of females when
examining extra-pair mating decisions.
60

4.2. Introduction
Ornamental traits may evolve by sexual selection if such traits increase the mating
success of their bearer and result in differential reproductive success (Darwin, 1871).
Elaborately ornamented males may have greater mating success if their trait provides an
advantage in direct competition with conspecifics for mates or if females prefer traits that are
more attractive (Andersson, 1994). Data from empirical studies have demonstrated that
males displaying attractive ornaments have greater success at copulating with females or
acquiring mates (Johnsen et al., 1998; Loyau et al., 2007), maintaining paternity within their
brood (Estep et al., 2005; Safran et al., 2005), or gaining extra-pair fertilizations (Bitton et
al., 2007; Albrecht et al., 2009).
In species where both sexes display similar ornamentation, female traits may evolve
by sexual or social selection (Clutton-Brock, 2009; Tobias et al., 2012). Elaborate ornaments
displayed by females may function as signals of attractiveness that are preferred by males
(Amundsen et al., 1997) or as signals of competitive ability or social status during
competition with conspecific females (Murphy et al., 2009b). Ornamental traits of females
that evolve by competition with conspecifics may also function secondarily in attracting
mates (LeBas, 2006). Such traits are known to have dual utility (see review in Berglund et
al., 1996), and traits that function in both contexts have been demonstrated in female birds
(e.g., Griggio et al., 2010). Although ornamental traits of females have been shown to
function during competitive interactions and /or mate attraction, less is known about how
such traits influence the social and extra-pair mating success of free-living female animals.
Experimentally altering the attractiveness of female ornamental traits has been shown
to influence courtship and copulation frequency (Pilastro et al., 2003; Griggio et al., 2005;
Torres and Velando, 2005) and sperm allocation by males (Cornwallis and Birkhead, 2007).
61

For example, reducing the breast patch size of female rock sparrows (Petronia petronia)
lowered the rate of courtship and sexual chases received from males (Griggio et al., 2005).
Females with a reduced ornament also were less likely to pair with a social mate than control
females (Griggio et al., 2005); however, it still remains to be determined how altered
attractiveness of female ornamentation influences the quality of social mate they acquire. In
addition, ornamental traits of females may also influence extra-pair mating success if females
seek extra-pair copulations to obtain genetic benefits for their offspring, such as good genes,
compatible genes, or ‘sexy sons’ (reviewed in Birkhead and Møller, 1992). Although males
are predicted to be less discriminating in their choice of extra-pair than social mates (Trivers,
1972, also see Krebs et al., 2004; Hasegawa et al., 2015 for examples), a previous study in
blue-footed boobies (Sula nebouxii) demonstrated that females manipulated to display dull
foot colour received less courtship attempts not only from social mates, but potential extrapair mates as well (Torres and Velando, 2005). As such, greater opportunities to engage in
extra-pair copulations may arise for females with elaborate ornamentation if they are able to
attract extra-pair mates or if social feedback from mates following signal enhancement
increases a female’s motivation to pursue extra-pair copulations (Burley, 1988).
Additionally, females whose ornaments signal high social status or greater competitive
ability may be able to secure extra-pair copulations by invading the territories of other
females in search of extra-pair mates. In contrast, females with attractive ornaments may be
less likely to engage in extra-pair copulations if they are paired to high-quality mates (e.g.,
Safran et al., 2005). Indeed, more ornamented female yellow warblers (Setophaga petechial)
had fewer extra-pair offspring in their broods compared to less-ornamented females (Grunst
and Grunst, 2014). Finally, it is also possible that the ornamentation of females is unrelated

62

to the proportion of extra-pair paternity, as previously demonstrated in western bluebirds
(Sialia mexicana; Jacobs et al., 2015).
In addition to ornamentation, a female’s success at acquiring extra-pair copulations
also may depend on the paternity guards of her social mate; for example, males may mate
guard or copulate frequently with their mates to reduce cuckoldry (reviewed in Birkhead and
Møller, 1992). Since paternity guards may be costly (Komdeur, 2001) and males may face a
trade-off between guarding their mate and gaining extra-pair fertilizations themselves
(Chuang-Dobbs et al., 2001), factors such as a female’s opportunity to engage in extra-pair
copulations (Wilson and Swaddle, 2013) or the attractiveness of male neighbours (Estep et
al., 2005) may influence whether a male guards his mate. In cattle egrets (Bubulcus ibis),
males invest more time at the nest and copulate more frequently when paired to females with
more ornamented plumage (Krebs et al., 2004). Similarly, a study in humans (Homo sapiens)
reported that female attractiveness increases the frequency of within-pair copulations
(Kaighobadi and Shackelford, 2008). Male rock sparrows also increase the time spent at the
nest when paired with females manipulated to display larger breast patches, which may
enable them to gain fertilizations in the second brood produced by the female (Pilastro et al.,
2003). Therefore, the attractiveness of female ornaments may also influence the proportion
of extra-pair paternity within a brood if males employ strategies to protect their paternity and
consequently their success at gaining extra-pair fertilizations, but as far as I am aware, this
has never been tested experimentally.
In the current study, my aim was to test whether attractive ornamentation influences
social and extra-pair mating success and the opportunity for extra-pair mating in female tree
swallows (Tachycineta bicolor). Females that are in at least their second breeding season
(after-second-year; ASY) display dorsal plumage that is similar to males (bright iridescent
63

blue-green), and evidence from previous studies suggests that plumage of ornamented,
higher quality females is brighter, with greater ultraviolet (UV) and blue chroma, and reflects
light at shorter wavelengths (Bitton et al., 2008; Bentz and Siefferman, 2013). Females and
males assortatively mate for plumage brightness (Bitton et al., 2008), and previous studies
suggest that plumage brightness is a signal assessed by female conspecifics during agonistic
interactions (Berzins and Dawson, 2016; Chapter 3); nevertheless, plumage brightness of
females may be a trait that also functions by attracting social or extra-pair mates. Therefore, I
experimentally enhanced and reduced plumage brightness prior to breeding of female tree
swallows relative to controls (see Berzins and Dawson, 2016) to test whether female
attractiveness influences social and extra-pair mate quality, and the proportion of extra-pair
paternity in the brood of the female. Although male tree swallows do not guard their social
mates (Leffelaar and Robertson, 1984), males that frequently copulate with their social mate
sire more within-pair offspring (Crowe et al., 2009). If this behaviour is influenced by female
attractiveness, then males socially paired to females with varying plumage brightness may
differ in their extra-pair mating success. Therefore, I also tested whether the plumage
brightness treatment of a female influenced whether their mates sired extra-pair offspring in
the broods of other females, the number of within-pair offspring sired, and the total number
of offspring (within- and extra-pair combined) sired by males. I predicted that females in the
enhanced plumage brightness treatment would be perceived as more attractive to males and
thus obtain social mates with brighter plumage than females in the reduced and control
plumage brightness treatments (Bitton et al., 2008). I also predicted that females in the
enhanced plumage brightness treatment would have fewer extra-pair offspring in their brood
if their social mate copulated with them frequently to protect their paternity, but that
enhanced females would attract extra-pair mates of higher quality (i.e., larger body size or
64

brighter plumage) than the control and reduced plumage brightness treatments. In contrast, I
predicted that females in the reduced plumage brightness treatment would have a greater
proportion of extra-pair paternity in their brood, engage in extra-pair copulations with males
of poorer quality than their social mate, and that their social mate would be more likely to
pursue extra-pair copulations than invest in paternity assurance through frequent within-pair
copulations.
4.3. Material and methods
4.3.1. Study area and general field methods
My study was conducted on tree swallows breeding in nest boxes near Prince George
BC, Canada (53ºN, 123ºW) from May to August in 2010 and 2011. The study area consisted
of open agricultural areas and small wetlands with patches of coniferous and deciduous trees
(see Dawson et al., 2005 for more details). During early May, I began visiting nests daily to
document nest building, and captured females in nest boxes once the construction of a nest
cup was complete, but before any eggs had been laid. Each female was banded and weighed
using a spring balance (nearest 0.25 g). I measured the lengths of the ninth primary flight
feather, wing chord and outer rectrix feather using a ruler (nearest 0.5 mm) and combined
head and bill (hereafter, head-bill) with digital calipers (nearest 0.1 mm). After determining
the age of females using dorsal plumage colour (Hussell, 1983), I sequentially allocated each
ASY female to one of three plumage brightness treatments (enhanced, reduced, control),
with the treatment order determined randomly. Details of the plumage brightness
manipulation are described in Berzins and Dawson (2016). Briefly, I evenly applied to the
dorsal feathers a non-toxic permanent blue marker to enhance or silicon paste to reduce the
plumage brightness of females; control females were treated in the same manner, but the
65

marker used contained water. The markers and silicone paste I used in my study have been
used previously in other studies to manipulate plumage colour (e.g., Safran et al., 2005;
Johnsen et al., 2005), and I was confident that my treatment produced the desired effect
throughout the experimental period because the effects of these markers and paste have been
detected for on average 15 days after their application (Johnsen et al., 2005). After the
plumage brightness treatment, the brightness of female plumage fell within the natural range
for female tree swallows at my study area (Berzins and Dawson, 2016). Age, body mass, and
plumage characteristics of females did not differ by plumage brightness treatment prior to
manipulation (all P values > 0.29). Plumage brightness was enhanced for 30 females and
reduced for 31 females, whereas the plumage brightness of 28 females remained the same.
Following the manipulation of female plumage brightness, I monitored all nests at my
study area and recorded the dates of clutch initiation and hatching (where January 1 = 1). The
date the first egg in each nest hatched was designated as day 0 of the nestling period. I
captured adults when they were feeding nestlings, and measured each bird as described
above. The band number of previously manipulated females was recorded, and males were
banded (if not previously banded) with an individually numbered aluminum leg band. I also
collected five rump feathers from each male, which were stored in opaque envelopes at room
temperature until spectral analysis (details below).
For the paternity analysis, a small blood sample (20 µL) was collected from the
brachial vein of each adult and twelve-day old nestling using non-heparinized capillary
tubes. Blood samples were transferred into microfuge tubes containing Queen’s lysis buffer
(Seutin et al., 1991) and were stored at -20 oC until DNA extraction. Eggs that failed to hatch
and nestlings that died prior to blood sampling were collected and stored at -20 oC.

66

4.3.2. Spectral analysis
Plumage characteristics of tree swallow feathers were quantified by taping four rump
feathers to a piece of cardboard so that they mimicked the natural arrangement of rump
feathers on a bird (Bitton et al., 2007), and using a JAZ-PX spectrometer with a xenon light
source (OceanOptics, Dunedin, FL, USA) and a bifurcated probe enclosed in a black plastic
holder to quantify reflectance spectra. Three measurements were taken for each sample and
spectral data were recorded using SpectraSuite software (OceanOptics, Dunedin, FL, USA).
The R package pavo (Maia et al., 2013) was used to smooth spectral curves and quantify hue,
ultraviolet (UV) chroma, blue chroma and average brightness; these plumage characteristics
have previously been shown to reflect aspects of male quality in tree swallows (Bitton et al.,
2007; Bitton and Dawson, 2008, Bitton et al., 2008). The three measures for each of the
plumage characteristics were averaged, and then I used a principal components analysis
(PCA) to reduce UV chroma, blue chroma, and hue into individual colour scores that were
used in analyses (Montgomerie, 2006). Average brightness was not included as a variable in
the PCA because it was not correlated with UV chroma, blue chroma, or hue (all P values >
0.12). In this PCA, I included all male tree swallows over a four-year period (2010-2013)
from which I had collected feather samples. The first principal component (PC1) explained
64.9 % of the total variation and the factor loadings (hue: -0.92; UV chroma: 0.97; blue
chroma: 0.16) suggest males with larger positive PC1 scores have plumage with greater UV
chroma that reflects light at shorter wavelengths, i.e., were bluer. The second principal
component (PC2) explained 34.2 % of the total variation and was heavily weighted by blue
chroma (-0.06, -0.22, 0.99, respectively), so males with larger positive scores have plumage
with greater blue chroma. Previous studies have shown that males with brighter plumage sire
a greater number of extra-pair offspring than duller males (Bitton et al., 2007; Whittingham
67

and Dunn, 2016). Older males, based on recapture status (see below), have plumage that is
brighter and bluer than younger males, and among duller males those that are bluer are more
likely to return to my study area to breed (Bitton and Dawson, 2008). These relationships
suggest that males displaying plumage that is brighter, with greater UV and blue chroma, and
that reflects light maximally at shorter wavelengths (larger positive PC1 and PC2 scores) are
‘more ornamented’ because they have greater reproductive success and/or apparent return
rates.
4.3.3. Paternity analysis
The paternity analysis for this study was conducted as part of a larger three-year
study (2010-2012). DNA from blood and tissue samples was extracted using DNeasy Blood
& Tissue kits (Qiagen #69506), and amplified at either five or six microsatellite loci
developed for Tachycineta swallows (Tle19, TaBi34, TaBi6, Tal8, TaBi10 and TaBi 8;
Makarewich et al., 2009). I performed duplex polymerase chain reaction (PCR) with two
primer pairs, TaBi6/Tle19 and Tal8/TaBi34, and PCR for TaBi10 and TaBi8 as single
reactions. Amplification of DNA extracted from blood samples was performed in 10 L final
volume containing 1x PCR buffer, 0.2 mM of each dNTP (Invitrogen #10297-018), 1.75–
2.25 mM MgCl2, 0.4 g/L bovine serum albumin (BSA; New England Biolabs #B90015),
0.08–0.2 M of forward-labelled and reverse primers, 0.05 U/L Taq polymerase
(Invitrogen #10342-020), and 20–50 ng of genomic DNA. PCR was performed in a MJ
Research Peltier Thermal Cycler under the following PCR cycling protocol: an initial
denaturing step at 95 °C for 4 min, followed by 35 cycles of 95 °C for 50 sec, 56–60 °C for
60 sec (56 °C for the duplex reactions and 60 °C for TaBi10 andTaBi8), and 65-72 °C for 90
sec (65 °C for the duplex reactions and 72 °C for TaBi10 andTaBi8). A final extension cycle
68

at 65–72 °C for 5 min (65 °C for duplex reactions and 72 °C for TaBi10 andTaBi8)
completed the PCR. Amplification of DNA extracted from tissue was performed using a
Qiagen Multiplex PCR kit (Qiagen #206143) following manufacturer’s instructions, except I
used a final volume of 10 L and 3–5 L of Multiplex PCR Master Mix. Concentrations of
primers and annealing temperatures remained the same as described above.
Products from PCR were run on an ABI 3130xl automated sequencer (Applied
Biosystems), and data were analyzed blind to sample identity using Peak ScannerTM
Software v1.0 (Applied Biosystems). Originally, all individuals from 2010 and 2011 were
genotyped at five loci: Tle19, TaBi34, TaBi6, Tal8, and TaBi10; however, due to a high
frequency of null alleles for Tal8 in my study population, I also genotyped all males and
extra-pair offspring that were homozygous at Tal8 plus the genetic mother of those offspring
at a sixth locus (TaBi8) so that 5 loci were used to identify genetic sires (see below). Starting
in 2012, I genotyped individuals at all six loci. I calculated allele frequencies and exclusion
probabilities for all microsatellite loci in my study population using CERVUS 3.0.7
(Kalinowski et al., 2007; see Table 4.1). TaBi8 significantly deviated from Hardy-Weinberg
equilibrium (P = 0.002; Table 4.1), but this locus was never used as a single criterion to
determine parentage. Offspring were classified as within-pair if they matched the resident
male at five loci and extra-pair offspring if they mismatched the resident male at a minimum
of one locus, but to account for the presence of null alleles a single mismatch was not
included if either the offspring or resident were homozygous at the mismatching loci
(O’Brien and Dawson, 2007). For this experiment, no offspring mismatched the resident
male at only a single locus; all offspring identified as extra-pair mismatched the resident
male at two or more loci. Among nests used in this study, no germinal disk or embryonic

69

tissue was observed in ten eggs (from three enhanced, three reduced, and two control nests),
and these eggs were considered as unfertilized for the paternity analysis.
For this study, I was interested in identifying the genetic sires of extra-pair offspring
in experimental broods and identifying extra-pair offspring sired by a female’s social mate.
To assign parentage of extra-pair offspring to a genetic sire I compared the alleles of each
extra-pair offspring to the alleles of all males sampled at my study separately for 2010 and
2011, and males that matched an extra-pair offspring at all five loci were assigned as the
extra-pair sire (only one sire was ever identified for each extra-pair offspring). Based on this
criterion, I assigned parentage to 155 of the 255 extra-pair offspring identified at my study
area in 2010 and 2011, and for those offspring I calculated the probability of chance
inclusion following O’Brien and Dawson (2007). Overall, the probability that a randomly
chosen male would match an extra-pair offspring at all five loci ranged from 3.0 x 10-7 to
0.0091 (mean ± SE: 0.00071 ± 0.00012). There were 7 additional cases where an extra-pair
offspring and putative genetic sire matched at four loci, but mismatched at a fifth locus
because either the offspring or putative genetic sire was homozygous. For two of these cases,
the offspring shared alleles with a maternal half-sibling that matched the same genetic sire at
all five loci, and in two other cases, the genetically related offspring of the putative genetic
sire also mismatched at the same locus. Nevertheless, even at only four loci the chance that a
random male would match these extra-pair offspring at all four loci was low (0.00094 ±
0.00032; range 0.00025 to 0.0025).
4.3.4. Statistical analysis
I tested whether the quality of a female’s social mate differed by plumage brightness
treatment using random intercept linear mixed models (LMM; lme4 package, Bates et al.,
70

2015). In these models, I included females that were manipulated in both years, but only if
they paired with a different male in each year (N = 5); one female paired with the same male
in both years so only data from the first year of the experiment were included in the models.
Identity of females and males were both included as random factors to account for multiple
observations of the same individuals. Each phenotypic trait of males (body condition, lengths
of wing, ninth primary feather, outer rectrix, and head-bill, UV chroma-hue (PC1), blue
chroma (PC2), and brightness) was a dependent variable, and plumage brightness treatment
was included as a fixed factor. Body condition was estimated by calculating the residuals
from a regression between body mass and length of the head-bill for all males captured at the
study area from 2010 to 2013 (F1, 217 = 23.81, P < 0.00001; Brown 1996). The age of
nestlings at capture was included as a covariate for models examining body condition to
account for declines in body mass during the nestling provisioning stage (Boyle et al., 2012).
Hatching date was not included as a covariate in models because timing of breeding is
confounded with individual quality (e.g., Dawson, 2008). Since phenotypic traits of male tree
swallows, such as the length of ninth primary and plumage brightness, increase with age
(Bitton and Dawson, 2008; O’Brien, 2006), I also included the capture status (previously
banded or unbanded) of males as a proxy for age (see Bitton and Dawson, 2008; Bitton et al.,
2007). Previously banded males that return to breed are presumed to be older than unbanded
males breeding at my study area for the first time because male tree swallows rarely change
breeding sites between years (i.e., less than 4 %, see Winkler et al., 2004). Males banded as
nestlings in the previous breeding season were grouped with unbanded males. Year was not
included in models as it was confounded with capture status. Females that switched nest
boxes after being manipulated bred later and in poorer quality nest boxes than they had
occupied prior to switching (Chapter 3), so I also included female status (retain or switch
71

nest boxes) as a categorical variable. P values for these LMM were calculated using the R
package lmerTest (Kuznetsova et al., 2016), and I removed predictor variables that did not
approach significance (P ≥ 0.10) in a backwards, stepwise fashion, but always retained
plumage brightness treatment in models. Samples sizes differ among phenotypic traits
because in some cases, feathers were damaged or broken and so were not measured.
To determine whether variation in plumage brightness of females influenced their
opportunity for extra-pair mating, I first tested whether the presence of extra-pair offspring
(yes or no) within a female’s brood was related to plumage brightness treatment using a
likelihood ratio test. Females that were manipulated and bred in both years were randomly
assigned to either 2010 or 2011 for this analysis since I had multiple observations for some
females (N = 4). I also tested whether the number of extra-pair young within nests differed
by the plumage brightness treatment of the female using a generalized linear mixed model
(GLMM; lme4 package, Bates et al., 2015) fitted with binomial errors and logit link function
for proportion data (Crawley, 2013). The number of extra-pair offspring was the dependent
variable, the number of offspring in the brood was the binomial denominator, and plumage
brightness treatment and year were included as fixed factors. The number of days a female
was manipulated (i.e., clutch initiation date – manipulation date; range 1 to 21, mean = 9.7 ±
0.8 SE) prior to initiating a clutch was included as a covariate. To account for the multiple
observations of females represented in the data set in both years I included female identity as
a random factor. The model was tested for overdispersion using the RVAideMemoie package
(Hervé, 2016), and I included an observation-level random factor in the model to account for
overdispersion (Browne et al., 2005). The significance of fixed effects and covariates was
examined by assessing the change in deviance between models (Crawley, 2013), and I report
Wald χ2 tests calculated using the Car package (Fox and Weisberg, 2011). For both analyses,
72

I only included broods where the paternity data were complete (i.e., the paternity status was
known for every egg laid by the female).
I tested whether the plumage brightness treatment of females influenced the quality of
their extra-pair mates by comparing phenotypic traits between social and extra-pair mates
within each brood using a series of LMM. Male paternity status (social or extra-pair mate)
was defined as the repeated effect and pair identity was the subject grouping variable. Each
phenotypic trait (condition, lengths of wing, ninth primary, outer rectrix, and head-bill,
average brightness, UV chroma-hue (PC1), and blue chroma (PC2)) was the dependent
variable in individual models, and I included plumage brightness treatment and male
paternity status as fixed effects, and the interaction between the two factors. Since my
objective was to test whether the plumage brightness treatment of females influenced the
quality of their social mates, models always retained plumage brightness treatment, male
paternity status, and their interaction. Some males were missing measurements because they
had damaged or broken feathers, so in these analyses I included only those males where the
majority (or all) phenotypic traits were measured so that I could compare as many traits as
possible between a female’s social and extra-pair mate.
To determine whether altering the perceived attractiveness of females influenced
male mating strategies, I tested whether social males sired extra-pair offspring (yes or no) in
relation to the plumage brightness treatment of the female using a likelihood ratio test. I also
tested whether the number of within-pair and total number of offspring sired by social mates
differed according the plumage brightness treatment of the female using a generalized linear
model (GLM) fitted with a quasi-Poisson error structure and log link function for
overdispersed count data (Crawley, 2013). Since the brightness of male plumage and capture
status may influence male mating strategies (Bitton et al., 2007; Whittingham and Dunn,
73

2016), I included both plumage brightness and capture status of males in models. As
described above, the significance of fixed factors and covariates was examined by assessing
the change in deviance between models. For females that were captured and manipulated in
both years (N = 6), I randomly selected only data for one of their social mates to be included
in these analyses.
Results were considered significant at P ≤ 0.05, and means are presented + 1 standard
error. When appropriate, parameter estimates (B) and effect sizes are reported. Effects sizes
were calculated as the correlation coefficient (r) following Field et al. (2012). All statistical
analyses were performed using SPSS v20 statistical software (IBM Corp., 2011) and R
v3.1.1 (R Development Core Team, 2015).
4.4. Results
The quality of social mate paired to females did not differ by plumage brightness
treatment (all P values > 0.12; Table 4.2), and females that switched nest sites after being
manipulated did not socially pair with males of lower phenotypic quality than females that
retained their nest sites (all P values > 0.25). Although body condition of males was similar
among plumage brightness treatments (Table 4.2), it was negatively related to the age of
nestlings at capture (B = -0.07 ± 0.03, F1, 19.58 = 5.13, P = 0.04). No effect of capture status
on male phenotypic traits was detected (all P values > 0.11) and so this variable was
removed from models.
Plumage brightness treatment of the female had no effect on whether a nest contained
extra-pair offspring or the proportion of extra-pair offspring in a brood (Table 4.3). Year and
the number of days females were manipulated prior to initiating a clutch were unrelated to
levels of extra-pair paternity (P values > 0.15). The plumage brightness treatment of the
74

female also did not influence whether or not her social mate sired offspring in other broods
(Table 4.3), or the number of within-pair offspring or the total number of offspring her social
mate sired (Table 4.3). Capture status and plumage brightness of males were unrelated to the
number of within-pair and total number of offspring sired (all P values > 0.13) and were
removed from models.
Comparisons of social and extra-pair mates for length of ninth primary flight feathers
suggested that the effect of female plumage brightness treatment varied with male paternity
status (Table 4.4), so I analyzed the data separately by plumage brightness treatment. In the
enhanced plumage brightness treatment, extra-pair mates tended to have longer ninth primary
feathers than social mates (B = 5.50 ± 1.61, F1, 2.00 = 11.71, P = 0.076, r = 0.92, Figure 4.1c);
no differences were detected between extra-pair and social mates in either the control (B =
-2.50 ± 2.22, F1, 3.00 = 1.27, P = 0.34, r = -0.54) or reduced plumage brightness treatments (B
= -1.20 ± 1.83, F1, 4.00 = 0.43, P = 0.54, r = -0.31). The analysis of length of outer rectrix also
showed an interaction between plumage brightness treatment and male paternity status
(Table 4.3). Analyzing the data separately by plumage brightness treatment showed that
extra-pair mates had longer outer rectrices than social mates in the enhanced plumage
brightness treatment (B = 2.33 ± 0.82, F1, 9.36 = 8.03, P = 0.02, r = 0.68, Figure 4.1d), but did
not differ in the control (B = -0.71 ± 1.13, F1, 10.03 = 0.40, P = 0.54, r = -0.19) and reduced
plumage brightness treatments (B = -1.25 ± 0.78, F1, 13.64 = 2.06, P = 0.17, r = -0.36). Body
condition did not differ between social and extra-pair mates or plumage brightness treatment,
but was negatively related to the age of nestlings at capture (B = -0.081 ± 0.04; Table 4.4).
Although extra-pair mates overall had plumage with greater blue chroma than social mates
(B = 0.81 ± 0.32; Figure 4.3), the magnitude of this difference was unrelated to the plumage
brightness treatment of the female (Table 4.4). Length of the wing and head-bill, plumage
75

brightness, and UV chroma-hue (PC1) did not differ between social and extra-pair mates or
the plumage brightness treatment of the female (Table 4.4).
4.5. Discussion
In the current study my aim was to alter the brightness of female plumage to examine
whether female attractiveness influenced social and extra-pair mating success and a female’s
opportunity for extra-pair mating. Contrary to my predictions, the quality of social mate
paired to females was similar regardless of plumage brightness treatment (Table 4.2).
Moreover, a similar proportion of extra-pair offspring in the broods of females (Table 4.3)
indicates that the plumage brightness treatment did not influence a female’s success at
acquiring extra-pair copulations. This is also supported by the percentage of broods
containing extra-pair paternity (see Table 4.3, overall 84 %), which is consistent with the rate
of extra-pair paternity reported previously at my study area and others (see O’Brien and
Dawson, 2007 and references therein; Chapter 6). Despite signaling lower quality plumage,
females in the reduced plumage brightness treatment did not mate with extra-pair males of
lower phenotypic quality compared to their social mate (Figures 4.1, 4.2). In contrast, extrapair mates of females in the enhanced plumage treatment had longer ninth primary feathers
and outer rectrices than their social males in pairwise comparisons (Figures 4.1c, d). If such
traits are indicative of quality or age (O’Brien, 2006), then my results may suggest that
female tree swallows displaying bright plumage have greater extra-pair mating success
because they secure extra-pair copulations from high-quality extra-pair mates. This finding is
in line with previous studies in male tree swallows showing that males with bright plumage
have greater extra-pair mating success because they sire a greater number of extra-pair
offspring (Bitton et al., 2007; Whittingham and Dunn, 2016).
76

The attractiveness of female ornamental traits may influence female mating success if
such traits enable females to be approached or courted more frequently by males (Pilastro et
al., 2003; Griggio et al., 2005; Torres and Velando, 2005), pair successfully (Griggio et al.,
2005), or acquire high-quality mates. For example, female rock sparrows manipulated to
display a breast patch reduced in size were less likely to pair with a social mate (Griggio et
al., 2005), but the quality of social mate acquired by the female was unknown. Since
assortative mating for breast patch size was observed in another population of rock sparrows,
Griggio et al. (2005) hypothesized that females with reduced breast patches may have mated
with low-quality males. Tree swallows also have been shown to assortatively mate for
plumage brightness (Bitton et al., 2008), but my study provided no evidence of females
displaying enhanced or reduced plumage brightness socially pairing with brighter or duller
mates compared to control females (Table 4.2). In fact, there were no phenotypic traits of
social mates that differed in relation to the plumage brightness treatment of the female (Table
4.2). Since female plumage brightness was manipulated after pair formation, I may only have
detected an effect of female treatment on quality of social mates if males divorced females
after the manipulation to pair with a female they perceived to be of higher quality (review in
Choudhury, 1995) and manipulated females acquired new social mates. Given that divorce
can be costly and unpaired females of high quality may be limited (Choudhury, 1995), males
may have pursued extra-pair copulations with females of higher quality than their social
mate; however, I also detected no evidence that the plumage brightness treatment of the
female influenced whether her social mate sired offspring in the broods of other females
(Table 4.3). As such, the most likely explanation for my results is that plumage brightness is
not a signal of attractiveness that influences the behaviour of male tree swallows. This is
consistent with a previous study that reported no evidence of differential allocation by male
77

tree swallows in relation to plumage brightness treatment of females they were paired with
(Berzins and Dawson, 2016). Instead, bright plumage of female tree swallows appears to be a
signal assessed by female conspecifics (Berzins and Dawson, 2016, Chapter 3) and likely
functions predominately in a competitive context, as demonstrated previously in female
American goldfinches (Spinus tristis; Murphy et al., 2009b).
That the proportion of extra-pair offspring in the broods of females was similar
among plumage brightness treatments (Table 4.3) suggests that in tree swallows attractive
ornamentation does not increase a female’s opportunity to engage in extra-pair copulations.
My results are consistent with a previous study in western bluebirds that reported no
relationship between female coloration and the rate of extra-pair paternity within broods
(Jacobs et al., 2015). In contrast, female yellow warblers displaying plumage with greater
melanin coverage, but lower carotenoid pigmentation had a higher proportion of extra-pair
offspring in their broods compared to females with plumage that had both high melanin
coverage and carotenoid pigmentation (Grunst and Grunst, 2014). Grunst and Grunst (2014)
hypothesized that females whose plumage had low carotenoid pigmentation may have
produced broods with a higher proportion of extra-pair offspring if they paired with poorer
quality social mates or were mate-guarded less intensively because of their less ornamented
dull carotenoid-based plumage and had a greater propensity to seek or engage in extra-pair
copulations due to their melanin-based plumage. Such a relationship between extra-pair
paternity and melanin-based plumage may exist as a result of a correlated expression with
behavioural traits, such as aggression and sexual behaviour (Ducrest et al., 2008). Although
there was no evidence that the quality or mating strategy of social mates influenced extrapair paternity in the broods of females whose plumage brightness was manipulated in my
study (Tables 4.2, 4.3), a study in tree swallows that treated females with 1,4,6-androstatrien78

3, 17-dione (ATD) and flutamide (F) in combination (ATD+F) to reduce aggressive and
sexual behaviour demonstrated that ATD+F females had fewer extra-pair offspring in their
broods than control females (Chapter 6). Therefore, a similar correlated expression between
female behavioural traits and extra-pair mating behaviour (Forstmeier et al., 2014) may exist
in tree swallows and would suggest female behaviour is a more important determinant of
whether females engage in extra-pair copulations than ornamentation. This is in line with a
previous study in female Lake Eyre dragon lizards (Ctenophorus maculosus) demonstrating
that while males tended to increase their sexual interest toward females manipulated to be
more ornamented, ultimately the reproductive state of the female influenced how often males
copulated with females; males copulated more frequently with receptive than non-receptive
or gravid females (Stuart-Fox and Goode, 2014).
Despite identifying overall differences in plumage brightness, and lengths of ninth
primaries and outer rectrices between males that sired extra-pair offspring and those males
that only sired within-pair offspring within my population (O’Brien, 2006; Bitton et al.,
2007), no phenotypic traits have been shown to differ in pairwise comparisons between
social and extra-pair mates at my study area (O’Brien, 2006; Bitton et al., 2007; but see
Whittingham and Dunn, 2014 for differences in other populations). Consistently, my results
for the reduced and control plumage brightness treatments showed no differences in
phenotypic traits between social and extra-pair mates of females in pairwise comparisons
(Figures 4.1, 4.2), but I did find that females manipulated to display enhanced plumage
brightness mated with extra-pair males that tended to have longer flight feathers than the
social mate they cuckolded (Figure 4.1c, d). Since males that are experienced breeders, and
presumably older, have longer ninth primary and outer rectrix feathers (O’Brien, 2006), this
may suggest that extra-pair mates of females in the enhanced plumage brightness treatment
79

were older, and perhaps of higher quality, than their social mates. Unfortunately, with my
limited sample size I was unable to test whether extra-pair mates were in fact older than the
social mate they cuckolded. Nevertheless, male tree swallows with longer flight feathers may
be higher quality if flight performance is influenced by feather length (Swaddle et al., 1996).
If females in the enhanced plumage brightness treatment mated with extra-pair mates
that were of higher quality than their social mate, there are two possible mechanisms to
explain this result. First, previous studies in fowl (Gallus gallus) have shown that highquality males prefer more ornamented females and when copulating with such females
produce ejaculates containing not only more sperm, but also seminal fluid that increases the
swimming speed of the sperm (Cornwallis and Birkhead, 2007; Cornwallis and O’Connor,
2009). As such, females in the enhanced plumage brightness treatment in my study may have
received higher quality ejaculates from extra-pair mates that outcompeted the sperm of lowquality males; however, since bright plumage of female tree swallows does not appear to be a
signal of attractiveness that influences the behaviour of males (Berzins and Dawson, 2016;
this study), it is more likely that males mate with any potential extra-pair females if the
opportunity arises, as demonstrated in Japanese barn swallows (Hirundo rustica gutturalis;
Hasegawa et al., 2015) and other species (Krebs et al., 2004). Second, social feedback from
conspecific females about the ornament quality after the manipulation (reviewed in Vitousek
et al., 2014b) may have altered female choice of extra-pair mates. Several studies have
demonstrated condition-dependent female mate choice (reviewed in Cotton et al., 2006). For
example, female zebra finches (Taeniopygia guttata) reared in experimentally reduced
broods so that they perceived themselves as high-quality preferred the song of high-quality
males (Holveck and Riebel, 2009). Similarly, female tree swallows in the enhanced plumage
brightness treatment may have perceived themselves as high quality after defending and
80

retaining their nest site from conspecific intruders displaying high-quality signals (Chapter
3). Moreover, since females in the enhanced plumage brightness treatment had previously
defeated conspecifics in agonistic interactions, they may have pursued extra-pair copulations
from high-quality males on the territories of other females because experiencing ‘winning’
may influence an individual’s perceived competitive ability and aggressive behaviour, and
increase the likelihood of defeating conspecifics in future challenges (i.e., the winner effect;
reviewed in Hsu et al., 2006). Although I have no hormone data, this behaviour may be
facilitated by changes in physiology, such as androgen levels (Oliveira et al., 2009).
Therefore, further studies that employ ornament and/or androgen manipulations combined
with observations of extra-pair copulations may provide novel insights into the proximate
mechanisms mediating female mating behaviour.
Overall, mating strategies of males were not related to the plumage brightness
treatment of their social mate (Table 4.3). Although I genotyped almost all offspring
produced in each year of study, I recognize that males may have sired extra-pair offspring at
feeding or roost sites away from my study area that were not included in my analyses.
Nevertheless, the mating strategies of male tree swallows are likely better predicted by the
attributes of males themselves as opposed to the ornamentation of their social mate. For
example, male tree swallows with brighter ventral plumage sire a greater number of withinpair offspring (Whittingham and Dunn, 2016), while males with brighter dorsal plumage sire
a greater number of extra-pair offspring (Bitton et al., 2007; Whittingham and Dunn, 2016).
It is unclear why the results of my study did not similarly show that extra-pair mates had
brighter dorsal plumage than the social mate they cuckolded (Table 4.4), as previously
reported (Whittingham and Dunn, 2016), but extra-pair mates did have plumage with greater
blue chroma than social mates (Table 4.4, Figure 4.3). The difference in results may be due
81

to altering perceived quality of females and social interactions within the study area (see
Berzins and Dawson, 2016), which may have altered female choice of extra-pair mates.
In conclusion, the finding that quality of social mates and male mating strategies did
not differ in relation to the plumage brightness treatment of the female is consistent with
Berzins and Dawson (2016) and suggests that male mating and parental care decisions are
not influenced by elaborate ornamentation of female tree swallows. As such, positive
assortative mating in this species for plumage brightness (see Bitton et al., 2008) is likely due
to competition among females to secure a male with a nest site, rather than male mate choice
of bright females. Nevertheless, the results of my experiment suggest that bright plumage
may increase a female’s extra-pair mating success by enabling females to engage in extrapair copulations with high-quality males, and highlights the need for future studies to
manipulate the signal quality of females when examining extra-pair mating decisions.

82

Table 4.1. Variability of alleles for six microsatellite loci used to assign paternity in a
population of tree swallows from 2010–2012.
Locus
n
k
Pei
hobs
hexp
F(Null) HW
Tle19
295
11
0.67
0.85
0.83
-0.0113 NS
TaBi34 295
17
0.69
0.85
0.84
-0.0117 NS
TaBi6
295
26
0.72
0.80
0.85
0.0287 NS
Tal8
292
25
0.85
0.77
0.93
0.0899 NS
TaBi10 295
17
0.76
0.89
0.88
-0.0031 NS
TaBi8
232
11
0.66
0.77
0.82
0.0300 *
n is the number of unrelated adults genotyped at each locus, k is the number of alleles, Pei is
the probability of exclusion with one parent (female) known, hobs is the observed
heterozygosity, hexp is the expected heterozygosity, F(Null) is the frequency of null alleles for
each locus, and HW indicates whether each locus significantly deviates from HardyWeinberg equilibrium, denoted as *. The combined probability of parental exclusion for all
six loci was 0.9997.

83

Table 4.2. Results of random intercept linear mixed models comparing phenotypic traits of social males paired to female tree swallows
whose plumage brightness was experimentally reduced or enhanced compared to controls. Presented are the mean (± SE) for each
phenotypic trait. Sample sizes are indicated by parentheses, and vary because individuals with feathers that were damaged or broken
were not measured. See methods for details of calculating condition and plumage colour metrics.
Plumage brightness treatment
Reduced
Control
Enhanced
1
Condition
-0.10 ± 0.21 (23)
-0.23 ± 0.24 (18)
-0.40 ± 0.28 (15)
Wing (mm)
120.53 ± 0.69 (17) 121.58 ± 0.82 (12)
121.16 ± 0.86 (11)
Ninth primary (mm)
95.39 ± 0.61 (17)
95.89 ± 0.72 (12)
94.94 ± 0.80 (11)
Outer rectrix (mm)
57.49 ± 0.38 (19)
57.19 ± 0.42 (15)
56.71 ± 0.60 (9)
Head-bill (mm)
28.82 ± 0.09 (23)
28.54 ± 0.10 (17)
28.70 ± 0.12 (15)
Average brightness
15.16 ± 0.47 (23)
15.92 ± 0.54 (17)
15.85 ± 0.60 (15)
2
UV chroma-hue (PC1)
0.10 ± 0.22 (23)
-0.12 ± 0.25 (17)
0.64 ± 0.28 (15)
3
Blue chroma (PC2)
-0.06 ± 0.19 (23)
-0.32 ± 0.23 (17)
-0.05 ± 0.24 (15)
1
Larger, positive values indicate males in better condition.
2
Larger, positive PC1 values indicate males with greater UV chroma and bluer hue plumage.
3
Larger, positive PC2 values indicate males with greater plumage blue chroma.
Trait

84

F

df

P

0.38
0.50
0.40
0.68
2.05
0.72
2.16
0.44

2, 48.59
2, 35.32
2, 36.57
2, 34.34
2, 51.42
2, 51.68
2, 51.74
2, 46.13

0.69
0.61
0.67
0.52
0.14
0.49
0.12
0.64

Table 4.3. Results of likelihood ratio tests, and general or generalized linear mixed models examining whether the plumage brightness
treatment (experimentally reduced or enhanced, or remained unchanged) of female tree swallows influenced female extra-pair mating
success (i.e., presence and proportion of extra-pair offspring) or male mating strategies (i.e., whether males gained extra-pair
fertilizations, maintained within-pair paternity, and their total number of offspring sired). Presented are raw proportion means (± SE),
and sample sizes are indicated by parentheses.
Plumage brightness treatment
Reduced
Control
Enhanced

Statistic

P

94 % (16)

80 % (14)

79 % (10)

G2 = 1.78

0.41

Proportion of EPO1
Percent of social mates that sired
EPO1
Number WPO2 sired

0.42 ± 0.06 (17)

0.50 ± 0.09 (14)

0.37 ± 0.08 (13)

χ22 = 0.80

0.67

50 % (22)

47 % (17)

25 % (12)

G2 = 2.24

0.33

3.32 ± 0.24 (22)

3.12 ± 0.54 (17)

3.17 ±0.53 (12)

χ22 = 0.14

0.93

Total # offspring sired

4.95 ± 0.55 (22)

4.76 ± 0.65 (17)

3.83 ± 0.73 (12)

χ22 = 1.57

0.46

Variable
Percent of nests containing extrapair paternity

85
1
2

EPO refers to extra-pair offspring.
WPO refers to within-pair offspring.

Table 4.4. Results of linear mixed models comparing phenotypic traits between the social
and extra-pair mate of female tree swallows whose plumage brightness was experimentally
reduced or enhanced compared to controls. See methods for details of calculating condition
and plumage colour metrics.
Phenotypic trait

F

df

P

Body condition
Plumage brightness treatment
Nestling age
Paternity status
Plumage brightness x paternity status

0.88
4.58
0.03
0.69

2, 32.56
1, 27.50
1, 32.82
2, 32.61

0.42
0.04
0.86
0.51

Wing (mm)
Plumage brightness treatment
Paternity status
Plumage brightness x paternity status

0.17
0.24
1.17

2, 17.52
1, 17.52
2, 17.52

0.85
0.63
0.33

Ninth primary (mm)
Plumage brightness treatment
Paternity status
Plumage brightness x paternity status

1.15
0.26
3.92

2, 9.00
1, 9.00
2, 9.00

0.36
0.62
0.06

Outer rectrix (mm)
Plumage brightness treatment
Paternity status
Plumage brightness x paternity status

0.05
0.09
3.85

2, 34.28
1, 34.28
2, 34.28

0.95
0.76
0.03

Average brightness
Plumage brightness treatment
Paternity status
Plumage brightness x paternity status

1.11
0.28
0.35

2, 33.41
1, 33.41
2, 33.41

0.34
0.60
0.71

UV chroma-hue (PC1)
Plumage brightness treatment
Paternity status
Plumage brightness x paternity status

2.11
1.79
0.75

2, 18.00
1, 18.00
2, 18.00

0.15
0.20
0.49

Blue chroma (PC2)
Plumage brightness treatment
Paternity status
Plumage brightness x paternity status

2.12
7.17
1.99

2, 18.00
1, 18.00
2, 18.00

0.15
0.02
0.17

86

87
Figure 4.1. Mean (± SE) difference in a) body condition, b) wing length, c) ninth primary length, and d) outer rectrix length between
social and extra-pair mates of female tree swallows whose plumage brightness was experimentally reduced or enhanced compared to
controls. See methods for details of calculating condition. Sample sizes indicate the number of pairs in each treatment and are given
above error bars.

88
Figure 4.2. Mean (± SE) difference in a) head-bill length, b) average brightness, c) UV chroma-hue (PC1), and d) blue chroma (PC2)
between social and extra-pair mates of female tree swallows whose plumage brightness was experimentally reduced or enhanced
compared to controls. See methods for details of calculating plumage colour metrics. Sample sizes indicate the number of pairs in each
treatment and are given above error bars.

Figure 4.3. Mean (± SE) blue chroma (PC2) of social and extra-pair mates of female tree
swallows. See methods for details of calculating blue chroma (PC2). Sample sizes indicate
the number of males and are given above error bars.

89

Chapter 5: Is ornamentation of female passerines related to offspring quality? An
experimental alteration of plumage brightness in female tree swallows
5.1. Abstract
It is widely recognized that females in a variety of taxa display ‘male-like’ traits, such
as elaborate ornamentation, and determining whether such traits evolve by selection or
genetic correlation remains a challenge. In mutually ornamented species, females displaying
elaborate traits are sometimes found to produce low-quality offspring and this is viewed as
support for the genetic correlation hypothesis because males would not benefit from
preferring the most ornamented females. Female traits, however, may instead function in
competition with conspecifics and while females that are the most elaborately ornamented
may be superior competitors, the costs associated with the possession of such traits may
influence offspring quality. In tree swallows (Tachycineta bicolor), females in their second
breeding season and older display bright blue-green iridescent plumage that is similar to
males, and previous studies have shown that more ornamented females produce low-quality
nestlings. In the current study, I enhanced and reduced the plumage brightness of females
relative to controls to test whether the social costs of displaying ornamental plumage
influence nestling quality. Although my results showed no difference in nestling quality
between the reduced and control plumage brightness treatments, nestlings reared by females
in the enhanced plumage brightness treatment were structurally smaller than nestlings in the
control and reduced plumage brightness treatments. This result is consistent with my
predictions and previous studies, and suggests that while bright plumage of female tree
swallows may evolve by sexual selection, the social costs associated with possessing

90

elaborate ornamentation also may influence nestling quality and should be considered in
future studies.
5.2. Introduction
Female animals from a wide variety of taxa display ‘male-like’ traits, such as
elaborate or brightly coloured plumage, and despite increased recent empirical investigation,
explaining the function and evolution of such traits has remained a challenge (e.g., Tarvin
and Murphy, 2012; Tobias et al., 2012; Nordeide et al., 2013). In species where both sexes
display elaborate ornamentation, such traits may evolve by mutual sexual selection
(Kraaijeveld et al., 2007) if they provide an advantage to females in competition for access to
mates (Johnson, 1988), or are attractive to or preferred by males (e.g., Torres and Velando,
2005; Griggio et al., 2009). Mutual mate choice may evolve in species where both sexes
contribute to parental care and males may therefore benefit from choosing a mate whose
ornamentation signals high quality (reviewed in Kraaijeveld et al., 2007). Females whose
ornamental traits signal quality should make greater investments in reproduction (Linville et
al., 1998; García-Navas et al., 2012) and/or produce offspring of higher quality (i.e., with
better immune defences, larger size, and/or better performance, Hidalgo-Garcia, 2006;
Kekäläinen et al., 2010; Remeš and Matysioková, 2013). In contrast, when elaborate traits
displayed by females appear to not be preferred by males or are not related to the quality of
offspring (e.g., Wolf et al., 2004; reviewed in Nordeide et al., 2013), such traits are
hypothesized to evolve as the by-product of sexual selection acting on the elaborate traits of
males (Lande, 1980). The focus on mate choice, however, may be too narrow (Tarvin and
Murphy, 2012), and the evolution of female ornamentation may be best considered under the
broader context of social selection since females often compete not only for access to mates
91

and breeding opportunities, but also for non-sexual resources (West-Eberhard, 1983; Tobias
et al., 2012). Although outcompeting conspecifics for limited breeding opportunities or
resources increases the reproductive success of females, traits that enhance female
competition also may be costly in terms of lower offspring quality (Rosvall, 2011b).
Ornamental traits that evolve by sexual and/or social selection should honestly signal
the quality of their bearer. Signal honesty may be maintained by the physiological costs of
producing and maintaining ornamental traits so that only those individuals of high quality are
able to display the most elaborate form of such traits (reviewed in Tibbetts, 2014). Since
females usually provide the bulk of resources for the production of offspring, they may be
faced with resource trade-offs between producing ornamentation and offspring (Fitzpatrick et
al., 1995). For example, females that invest resources, such as carotenoids, into the
production of elaborate ornaments have been observed to produce poor-quality eggs
(Morales et al., 2009). Consequently, males of mutually ornamented species may not benefit
from choosing the most elaborately ornamented females as mates if their offspring are lower
quality, and thus a negative relationship between female ornamentation and offspring quality
is interpreted as support for the genetic correlation hypothesis (Nordeide et al., 2013). While
this explanation may be limited to species where allocation of resources to the production of
ornamentation and offspring coincide during the breeding season and/or male mate choice
has been shown to operate (reviewed in Nordeide et al., 2013), it is also important to
consider alternative explanations that may arise by social mechanisms that maintain signal
honesty. For example, enlarging the size of forehead patch of male collared flycatchers
(Ficedula albicollis) increased competition among males, and yearling males displaying
enlarged forehead patches spent more time involved in competition and as a result,

92

provisioned their nestlings less compared to control males (Qvarnström, 1997).
Consequently, social costs of dishonest signalling may lower parental investment
(Qvarnström, 1997) or reduce the resources available to invest in reproduction (Fitzpatrick et
al., 1995), and may result in poorer quality offspring; however, as far as I am aware, no study
has manipulated female ornamentation prior to breeding and explored whether social costs of
possessing elaborate ornamentation influence offspring quality.
In the current study, my aim was to test whether variation in plumage brightness of
female tree swallows (Tachycineta bicolor) influences offspring quality. Female tree
swallows display dorsal plumage that ranges from dull brown to bright iridescent blue-green.
Dull brown plumage is predominately displayed by young females in their first breeding
season (i.e., second year females (SY); Hussell, 1983), whereas older females (i.e., aftersecond-year; ASY) display ‘male-like’ iridescent blue-green plumage. The delayed plumage
maturation of SY females has been shown to function as a signal of low competitive ability
thereby reducing the amount of aggression received from ornamented ASY females (Coady
and Dawson, 2013). ASY females that are more ornamented, and presumably higher quality
because they lay heavier eggs and fledge more offspring, display plumage that is brighter,
with greater ultraviolet (UV) and blue chroma, and reflects light maximally at shorter
wavelengths (i.e., bluer; Bitton et al., 2008; Bentz and Siefferman, 2013). Although it is
currently unknown whether bright plumage of ASY females functions as a status signal in
tree swallows, Bitton et al. (2008) reported positive assortative mating for plumage
brightness, which may result either from mutual mate choice or competition among females
for males with a nest site. Since male tree swallows do not alter their investment in parental
care or mating strategies in relation to experimentally altered plumage brightness of females

93

(Berzins and Dawson, 2016; Chapter 4), it is unlikely that bright plumage is a signal of
attractiveness that is preferred by males and assortative mating likely results from bright
females having greater competitive ability. Previous studies in tree swallows also have
shown that more ornamented females have greater levels of nest parasitism, poorer immune
defences, lower hematocrit levels and produce offspring that are smaller or in poor condition
(Coady, 2011; Bentz and Siefferman, 2013). This negative relationship between female
ornamentation and offspring quality could be interpreted as evidence in support of the
genetic correlation hypothesis (Nordeide et al., 2013); however, more ornamented female
tree swallows may experience social costs as a result of conspecifics challenging their status
(Coady, 2011; Chapter 3), thereby resulting in poorer quality offspring produced by more
ornamented females.
I tested whether plumage brightness of female tree swallows influences offspring
quality by experimentally enhancing and reducing the plumage brightness of females relative
to controls (Berzins and Dawson, 2016). I previously reported evidence of social control of
signal honesty in tree swallows (Chapter 3); females in the enhanced and control plumage
brightness treatments were less likely to retain their nest site following the plumage
manipulation than females in the reduced plumage brightness treatment. Moreover, females
in the enhanced plumage brightness treatment initiated their clutches later than females in the
control and reduced plumage brightness treatments, which is consistent with social costs
imposed on females dishonestly signalling high quality and engaging in frequent and
aggressive agonistic interactions with naturally bright conspecific females (e.g., like-versuslike aggression, Senar, 2006). Consequently, females whose signals are challenged by
conspecifics may spend more time involved in agonistic interactions and/or have fewer

94

resources to invest in self-maintenance and parental care; therefore, I predicted that females
in the enhanced plumage brightness treatment would produce offspring of lower quality than
females in the control treatment. I also predicted that females in the reduced plumage
brightness treatment, since they may have been involved in agonistic interactions with
conspecifics signalling lower quality or fewer interactions altogether (Chapter 3), would
produce offspring of higher quality than females in the control and enhanced plumage
brightness treatments. Following the manipulation of plumage brightness, all nests at my
study area were monitored (Chapter 3) and nestling growth, size, and mass were measured to
estimate quality (see Dawson et al., 2005). I also sexed all nestlings so that differences in
size, mass and growth between males and females could be controlled statistically (Rosivall
et al., 2009).
5.3. Material and methods
5.3.1. Study area and general field methods
Tree swallows were studied while breeding in nest boxes near Prince George BC,
Canada (53ºN, 123ºW) from May to August in 2010 and 2011. The study area consisted of
open agricultural areas and small wetlands that were intermixed with patches of coniferous
and deciduous trees (see Dawson et al., 2005 for more details). I began visiting nest boxes
daily in early May to document nest construction; once a nest box contained a nest with a
completed nest cup, I captured females in the nest box (Chapter 3). ASY females were
sequentially assigned to a plumage brightness treatment by capture order (Chapter 3), and the
brightness of their plumage was experimentally altered as described in Berzins and Dawson
(2016). Briefly, I enhanced plumage brightness by applying non-toxic permanent blue
marker, and reduced by applying silicon paste evenly to the dorsal feathers; control females
95

were treated in a similar matter, except that I used a marker containing water. Prior to
manipulation I found no differences in age, body mass and plumage characteristics of
females among plumage brightness treatments (Chapter 3). In total, the plumage brightness
of 61 females was experimentally altered (enhanced = 30, reduced = 31), while the plumage
brightness of 28 females were in the control group.
Following the experimental alteration of female plumage brightness, I monitored nest
sites to document clutch initiation and completion dates (Chapter 3). Beginning twelve days
after clutch completion, I checked nests daily to record the hatching of eggs, designated as
day 0 of the brood-rearing period. On day 4, the legs of each nestling in a brood were
uniquely marked with a non-toxic marker for individual identification, and I measured mass
with a spring balance (nearest 0.125 g) and the combined length of the head and bill
(hereafter head-bill, nearest 0.1 mm). After the day 4 measurement, I subsequently measured
mass and length of the head-bill every two days from day 6 to 16. Starting on day 8, I
measured the length of the ninth primary flight feather with a ruler (nearest 0.5 mm) and
continued measurements every two days until day 16. On day 12, I collected a small blood
sample (20 L) from the brachial vein of nestlings with non-heparinized capillary tubes for
molecular sexing; blood samples were stored in Queen’s lysis buffer (Seutin et al., 1991) at
-20 oC until DNA extraction. Any eggs that failed to hatch or nestlings found dead prior to
day 12 were collected and stored at -20 oC until tissue collection and DNA extraction. I
visited nests on day 22 to record the number of young that successfully fledged.
5.3.2. Molecular sexing
To determine the sex of offspring, I amplified DNA extracted from blood and tissue
samples (see Chapter 4) using the P2 and P8 primers, which have been used previously to
96

determine the sex of nestling tree swallows (Whittingham and Dunn, 2000). DNA extracted
from blood samples was amplified using polymerase chain reaction (PCR) performed in a 10
µL final volume containing 1x PCR buffer, 0.2 mM of each dNTP (Invitrogen #10297-018),
2.5 mM MgCl2, 0.4 g/L bovine serum albumin (BSA; New England Biolabs #B90015),
0.3 M of P2 and P8 primers, 0.05 U/L Taq polymerase (Invitrogen #10342-020), and 20–
50 ng of genomic DNA. PCR was performed in a MJ Research Peltier Thermal Cycler under
the following PCR cycling protocol: an initial denaturing step at 94 °C for 5 min, followed
by 35 cycles of 94 °C for 30 sec, 47 °C for 60 sec. and 72 °C for 60 sec. A final extension
cycle at 72 °C for 5 min completed the PCR. For DNA extracted from tissue samples of
unhatched eggs (see Chapter 6), PCR reactions were performed using a Qiagen Multiplex
PCR kit (Qiagen #206143) following the manufacturer’s instructions, except I used a final
volume of 10 L and 5 L of Multiplex PCR Master Mix; primer concentrations and
annealing temperatures remained the same as described above.
The P2 and P8 primers amplify the CHD-W and CHD-Z genes located on the avian
sex chromosomes W and Z (Griffiths et al., 1998). Since females are the heterogametic sex,
the sex of each individual can be determined based on the number of PCR products amplified
(i.e., when visualizing the gel, females are indicated by the presence of two bands, ZW,
whereas males are indicated by a single band, ZZ; Griffiths et al., 1998). In tree swallows,
the two bands for females are difficult to separate by electrophoresis because they are similar
in size and so the PCR products first need to be digested using the restriction enzyme HaeIII
(Whittingham and Dunn, 2000). Following Whittingham and Dunn (2000), I digested 5 µL
of the PCR product with 10 U of HaeIII (Invitrogen # 15205-016), 1 µL of 10x buffer
(provided with the HaeIII), and 5 µL water. The digest mixture was incubated at 37 °C for 1
97

hour, and a 5.4 µL aliquot was added to 0.6 µL of 10x loading buffer (provided with the
HaeIII), and then added to a 1.5 % agarose gel stained with ethidium bromide. The digested
PCR products were separated by electrophoresis and the gel was visualized under UV light to
determine the sex of offspring. HaeIII cuts the CHD-Z fragment, but not the CHD-W
fragment specific to females so females are identified by an individual having two bands, and
males by a single band (Whittingham and Dunn 2000). I included an adult male and female
tree swallow of known sex in every PCR and HaeIII digestion, and on every gel; in all cases
the sex recorded when visualizing the gel matched the known sex of the adult.
5.3.3. Statistical analysis
I used random intercept linear mixed models (LMM; lme4 package, Bates et al.,
2015) to examine whether mass and lengths of ninth primary feathers and head-bill on day
16, as well as growth rates of nestlings, were influenced by the plumage brightness treatment
of the female. Growth rate constants of individual nestlings were calculated using a linear
model for ninth primary, Gompertz model for head-bill, and a logistic model for mass
following Dawson et al. (2005), using only nestlings that had complete growth measures
(i.e., nestlings measured every two days from day 4 to 16 (mass and head-bill) or day 8 to 16
(ninth primary feather)). For these models, either size or mass at day 16, or growth rate
constant was the dependent variable, and plumage brightness treatment, year, and nestling
sex (male or female) were included as fixed factors. Brood size on day 16 and time of
measurement were included as covariates in models examining nestling size or mass at day
16, but only brood size was included as a covariate in models examining growth rate
constants. Hatching date was not included in models because it was negatively related to
time of measurement (r = -0.36, P = 0.02). Brood identity was included as a random factor to
98

account for the lack of independence among nestlings reared within the same brood and
female identity as a random factor to account for the multiple observations for some females
(N = 2). P-values for these LMM were calculated using the R package lmerTest (Kuznetsova
et al., 2016), and I removed non-significant covariates and factors that did not approach
significance (P ≥ 0.10) in a backwards, stepwise fashion, but always retained plumage
brightness treatment.
To test whether fledging success differed by the plumage brightness of the female, I
used a generalized linear model (GLM) fitted with binomial errors and logit link function for
proportion data (Crawley, 2013). The response variable was the number of offspring that
fledged and the binomial denominator was the original size of the brood. I included plumage
brightness treatment and year as fixed factors, and hatching date, standardized to a mean of 0
and standard deviation of 1 for experimental nests separately for each year, as a covariate.
Since the residual deviance was greater than the degrees of freedom, I refitted the model with
a quasi-binomial error structure to account for overdispersion (Crawley, 2013). I tested the
significance of covariates and factors by assessing the change in deviance between models.
Results were considered significant at P ≤ 0.05, and means are presented + 1 standard
error (SE). The overall significance of omnibus tests was examined using post-hoc tests with
a false discovery rate adjustment for multiple comparisons (Verhoeven et al., 2005).
Parameter estimates (B) and effect sizes are reported, and effect sizes for post-hoc tests were
calculated as the correlation coefficient (r) following Field et al. (2012). All statistical
analyses were performed using SPSS (IBM Corp., 2011) and R (R Development Core Team,
2015) statistical software.

99

5.4. Results
There was a tendency for length of the head-bill of nestlings to differ by the plumage
brightness treatment of the female (F2, 36.91 = 3.10, P = 0.057) and by the sex of nestlings (F1,
169.83 = 26.51, P < 0.0001); however, the residuals of the model were not normal (Shapiro-

Wilk normality test, P < 0.0001) and appeared to be influenced by three nestlings (out of
197) with small lengths of head-bill. After removing these nestlings from the model, the
residuals were normal (Shapiro-Wilk normality test, P = 0.14); further, the results showed
that length of head-bill differed significantly by the plumage brightness treatment of the
female (Table 5.1; Figure 5.1). Post-hoc tests showed that nestlings reared in the enhanced
plumage brightness treatment had shorter head-bills than nestlings reared in the control (B =
-0.42 ± 0.20, t33.21 = -2.12, P = 0.05, r = 0.34, Figure 5.1) and reduced plumage brightness
treatments (B = -0.59 ± 0.19, t37.94 = -3.03, P = 0.007, r = 0.44, Figure 5.1). No difference in
length of head-bills of nestlings was detected between the reduced and control plumage
brightness treatments (B = 0.16 ± 0.19, t38.86 = 0.89, P = 0.38, r = 0.14). Male nestlings had
longer head-bills at day 16 than female nestlings (B = 0.43 ± 0.08, Table 5.1).
Length of ninth primary flight feathers of nestlings did not differ by female plumage
brightness treatment, but did vary annually (Table 5.1); nestlings had shorter ninth primaries
in 2011 compared to 2010 (B = -4.53 ±1.64). Nestlings were also of similar mass on day 16
regardless of the plumage brightness treatment of the female (Table 5.1), but male nestlings
were heavier than females (B = 1.00 ± 0.27; Table 5.1).
Growth of nestling ninth primary feathers and head-bills were similar among
plumage brightness treatments, but ninth primary feathers grew faster in 2010 than in 2011
(B = 0.50 ± 0.16; Table 5.1). There was some suggestion that mass growth rate constants of

100

nestlings differed by plumage brightness treatment (Table 5.1, Figure 5.2). Post-hoc tests
showed that nestlings reared in broods in the enhanced plumage brightness treatment tended
to gain mass at a slower rate than nestlings reared in control broods (B = -0.06 ± 0.03, t38.15 =
-2.37, P = 0.05, r = 0.34, Figure 5.2), but not nestlings reared in the reduced plumage
brightness treatment (B = -0.04 ± 0.02, t40.12 = -1.47, P = 0.21, r = 0.22, Figure 5.2). Mass
gain was similar among nestlings reared in reduced and control plumage brightness
treatments (B = -0.02 ± 0.02, t39.59 = -0.97, P = 0.33, r = 0.15, Figure 5.2).
Fledging success of nestlings tended to decline with later hatching dates (B = -0.62 ±
0.36, F1, 46 = 3.12, P = 0.08), and was lower in 2011 than 2010 (B = -1.79 ± 0.47, F1, 47 =
14.56, P = 0.0004), but did not differ by the plumage brightness treatment of the female (F2,
48 = 1.07, P = 0.35).

5.5. Discussion
I manipulated the plumage of female tree swallows to test whether variation in
plumage brightness influenced nestling quality. Consistent with my predictions, females in
the enhanced plumage brightness treatment produced nestlings with shorter head-bills, so
they were structurally smaller, compared to nestlings in both the control and reduced
plumage brightness treatments (Figure 5.1). There was also a tendency for nestlings reared
by females with enhanced plumage brightness to gain mass more slowly than controls
(Figure 5.2), but despite this, they were no lighter on day 16 than nestlings reared in the
control and reduced plumage brightness treatments (Table 5.1). I also predicted that females
in the reduced plumage brightness treatment would produce nestlings of higher quality
(structurally larger, heavier, or faster growth) than nestlings reared by control females if they
were involved in agonistic interactions with low-quality females or fewer interactions overall
101

(Chapter 3), but I observed no difference in offspring quality between the reduced and
control plumage brightness treatments (Figure 5.1). This may suggest that although females
in the reduced plumage brightness treatment were challenged less or by low-quality
conspecifics, engaging in agonistic interactions is costly and reduces the amount of time or
resources females can invest in reproduction (Fitzpatrick et al., 1995; Qvarnström, 1997).
Alternatively, females may have been involved in fewer agonistic interactions, but their
nestlings were not higher quality than controls because males invested less in parental care.
Such differential allocation has been demonstrated previously in other species (Roulin, 1999;
Mahr et al., 2012), but I believe this is unlikely to explain my results since male tree
swallows do not provision their nestlings less when mated to a female whose plumage
brightness was reduced (Berzins and Dawson, 2016).
Previous studies in tree swallows have shown that nestlings produced by more
ornamented females were smaller or in poorer condition (Coady, 2011; Bentz and
Siefferman, 2013). Coady (2011) hypothesized that if more ornamented females are
perceived as higher status than less-ornamented females, then social costs experienced by
more ornamented females whose signal quality was challenged repeatedly by conspecifics
may explain why they produced smaller offspring. Although it is currently unknown whether
plumage brightness of female tree swallows is a signal of status, females in the enhanced
plumage brightness treatment manipulated to display high-quality signals were less able to
retain their nest site than females in the reduced plumage brightness treatment, suggesting
they were challenged more frequently and aggressively by conspecifics (Chapter 3), and
females with enhanced plumage brightness that did retain their nest site, bred later in the
season (Chapter 3) and produced structurally smaller offspring (Figure 5.1). My results for

102

female tree swallows are similar to those previously reported for yearling male collared
flycatchers; males manipulated to dishonestly display an enlarged forehead patch were less
likely to establish a breeding territory, and those that did fed their nestlings less than controls
(Qvarnström, 1997). Although I have no data on parental provisioning rates, length of
nestling head-bills on day 4 did not differ by plumage brightness treatment (P = 0.96,
Berzins, unpublished data), which may suggest that females in the enhanced plumage
brightness treatment provisioned their nestlings at lower rates and/or provided lower quality
food. Lower investment in nestlings by females in the enhanced plumage brightness
treatment is consistent with ornament-fecundity (Fitzpatrick et al., 1995) or ornamentparental care (Qvarnström, 1997) trade-offs, although I am unable to determine which
mechanism led to the lower offspring quality observed in the enhanced plumage brightness
treatment. Nevertheless, females displaying enhanced plumage brightness likely had fewer
resources or less time to invest in rearing nestlings, without jeopardizing their own selfmaintenance, as a result of engaging in agonistic interactions with high-quality conspecifics.
Thus, although elaborate ornamentation may benefit females, for example, in competition
with conspecifics for quality mates or nest sites (Bitton et al., 2008), social costs associated
with the possession of ornamentation may lower female fecundity (Chapter 3) and offspring
quality (Figure 5.1) and should be considered in future studies.
Although my results for lower nestling quality are consistent with brighter females
experiencing social costs, there are also other non-mutually exclusive explanations for my
results. First, nestlings in the enhanced plumage brightness treatment may have been smaller
than nestlings in the control and reduced plumage brightness treatments simply because they
hatched later in the season (P < 0.045, Chapter 3). Later hatched nestlings have lower

103

survival prospects (Dawson, 2008), and so parents may strategically reduce their
reproductive effort to invest in their own self-maintenance and enhance future survival and
fecundity rather than investing in offspring of low value (Winkler and Allen, 1996). It is
important to note, however, that the delayed breeding by females in the enhanced plumage
brightness treatment appears to be a consequence of increased agonistic interactions with
conspecifics (Chapter 3). In addition, if females in the enhanced plumage brightness
treatment were perceived as more attractive to potential mates and had greater opportunities
to engage in extra-pair copulations, then males paired to attractive females may have been
less certain of their paternity and provisioned nestlings less. This explanation, however, is
also unlikely since females in the enhanced plumage brightness treatment had a similar
proportion of extra-pair offspring in their broods as did females in the control and reduced
plumage brightness treatments (Chapter 4), and there is also no evidence that male tree
swallows provide lower investment in parental care when their brood contains extra-pair
offspring (Whittingham et al., 1993). Finally, the smaller structural size of nestlings may
have been due to females in the enhanced plumage brightness treatment depositing a greater
concentration of testosterone in the yolks of their eggs as a result of nest-site intrusions
(Whittingham and Schwabl, 2002) since they likely experienced more frequent and
aggressive challenges by naturally bright females of greater competitive ability (see Chapter
3). Although elevated yolk testosterone levels have been shown to enhance the growth of
nestlings in some species (Eising et al., 2001; Müller et al., 2010), negative effects on
nestling structural size have also been demonstrated (Sockman and Schwabl, 2000), and the
effect of androgens on nestling development may depend on the concentration deposited by
the female (Muriel et al., 2015). Unfortunately, the only way to distinguish whether the effect

104

of plumage brightness treatment on nestling quality is due to the effects of androgen
deposited in egg yolks or lower provisioning rates by females would be to cross-foster
nestlings among the plumage brightness treatments so that nestlings were reared by females
other than their genetic mother.
Interestingly, I only detected an effect of plumage brightness treatment on nestling
structural size, but not body mass or length of ninth primary flight feathers of nestlings
(Table 5.1). It is possible that nestling mass and feather length was maintained because males
mated to females whose plumage brightness was enhanced increased their investment in
parental care (Burley, 1988) and this at least partially compensated for any negative effect of
plumage brightness treatment on nestling quality; however, given that a previous study
reported no evidence of differential allocation by male tree swallows in relation to
experimentally altered plumage brightness of females (Berzins and Dawson, 2016), this
seems doubtful. Alternatively, an effect of plumage brightness treatment only on nestling
structural size may provide support to the above suggestion that the costs of engaging in
agnostic interactions did in fact influence female provisioning behaviour. For example,
females may have provisioned nestlings at an adequate rate to maintain the energy
requirements for tissue development, but nestlings may have received insufficient dietary
minerals, such as calcium, for skeletal growth (Dawson et al., 2005). Previously, a study in
tree swallows demonstrated that calcium-fed nestlings were structurally larger, heavier and
had longer ninth primary feathers than control nestlings (Dawson and Bidwell, 2005). Such
results suggest that parents may face a trade-off between provisioning rate and searching for
food items containing calcium (reviewed in Dawson and Bidwell, 2005), and consequently,
females in the enhanced plumage brightness treatment may have had fewer resources or less

105

time available to invest in searching for shells and other sources of calcium. Nestlings that
are structurally smaller at day 16 may be smaller-sized adults or fledge without their skeletal
structure being fully developed, which may lower post fledging survival if they are weaker,
poorer fliers, or have a greater risk of predation (reviewed in Tilgar et al., 2004). Moreover,
dietary intake during development may influence the brightness of feathers grown, as
demonstrated in male nestling eastern bluebirds (Sialia sialis; Doyle and Siefferman, 2014);
however, it is unclear whether such effects would persist into adult life since feathers are
molted and regrown annually and structural plumage coloration may be more sensitive to
stress experienced during molt rather than nutritional condition (Peters et al., 2011).
Unfortunately, recruitment of nestling tree swallows in my study area is very low (e.g., 0.05
probability of recruitment, Weegman et al., unpublished data), and so I am unable to
determine whether apparent effects of plumage brightness treatment on nestling skeletal
growth also had long-lasting effects on their plumage ornamentation and reproductive
success as adults.
Overall, my finding that females manipulated to signal enhanced plumage brightness
produced lower quality nestlings is consistent with previous studies in tree swallows (Coady,
2011; Bentz and Siefferman, 2013). I argue that this finding is not consistent with the genetic
correlation hypothesis (sensu Nordeide et al., 2013) to explain the evolution of ornamental
plumage of female tree swallows, but rather suggests that ornamented females may
experience social costs by the mechanisms that ensure the honesty of quality signals. Despite
such costs, however, female tree swallows may benefit from their ornamented plumage
because they mate with males that have bright plumage (Bitton et al. 2008), which may
increase the reproductive value of sons. For example, male tree swallows with bright

106

plumage sire a greater number of offspring than dull males (Bitton et al., 2007; Whittingham
and Dunn, 2016), and if male offspring inherit the attractive traits of their father (e.g., ‘sexy
son’ hypothesis, Weatherhead and Robertson, 1979), then these ‘sexy sons’ would become
‘sexy fathers’ for which evidence of female preference exists (Bitton et al., 2007;
Whittingham and Dunn, 2016). Given that there is there is little evidence to suggest male
mate choice in this mutually ornamented species (e.g., Berzins and Dawson, 2016; Chapter
4), positive assortative mating for plumage brightness (Bitton et al., 2008) is likely due to
females with bright plumage having greater competitive ability. Future studies that examine
the evolution of plumage brightness of female tree swallows should evaluate if this trait is
also beneficial to females in a non-sexual context, such as a status signal in competition for
resources during the non-breeding season (e.g., Tarvin and Murphy, 2012).

107

Table 5.1. Results of random intercept linear mixed models comparing mass and size at day
16, and growth rates of nestling tree swallows reared in broods where female plumage
brightness was experimentally reduced or enhanced, or remained unchanged (controls).
F

df

P

Nestling size
Head-bill (mm)
Plumage brightness treatment
Sex

4.75
33.15

2, 36.49
1, 164.22

0.01
<0.0001

Ninth primary feather (mm)
Plumage brightness treatment
Year

0.21
7.58

2, 48.63
1, 50.80

0.81
0.008

Mass (g)
Plumage brightness treatment
Year
Time of measurement
Sex

0.14
3.49
3.57
13.48

2, 28.87
1, 26.08
1, 35.27
1, 169.07

0.87
0.07
0.07
<0.001

Nestling growth
Head-bill
Plumage brightness treatment
Sex
Brood size

1.65
3.39
3.31

2, 38.25
1, 171.51
1, 55.56

0.20
0.07
0.07

Ninth primary
Plumage brightness treatment
Year

0.13
9.38

2, 48.30
1, 51.28

0.88
0.003

Mass
Plumage brightness treatment

2.87

2, 39.29

0.068

108

Figure 5.1. Mean (± SE) length of the combined head and bill of nestling tree swallows at
day 16 in broods where female plumage brightness was experimentally reduced or enhanced
compared to controls. Sample sizes indicate the number of broods in each treatment, and are
given above error bars.

109

Figure 5.2. Mean (± SE) growth rate constant for mass of nestling tree swallows reared in
broods where female plumage brightness was experimentally reduced or enhanced compared
to controls. Sample sizes indicate the number of broods in each treatment, and are given
above error bars. See Methods for calculation of growth rate constants.

110

Chapter 6: Experimentally altering exposure to testosterone and its estrogenic
metabolites prior to breeding reduces extra-pair paternity in female tree swallows
6.1. Abstract
Extra-pair paternity occurs commonly in many socially monogamous animal species,
yet the reasons why females mate with males outside the social pair bond remain poorly
understood. Evidence from previous studies suggests that individual differences in
behavioural traits may underlie variation in the rate of extra-pair paternity; however, it
remains to be determined whether behavioural traits and promiscuity are mediated by the
same hormone(s). Since testosterone (T) and its estrogen metabolites mediate aggressive and
sexual behaviours, I tested whether such hormones also influence extra-pair copulation
behaviour by experimentally elevating T using implants containing T, and blocking the
estrogenic and androgenic actions of T using implants containing 1,4,6-androstatrien-3, 17dione in combination with flutamide (ATD+F) in female tree swallows (Tachycineta
bicolor). Females with sham implants served as controls. Consistent with my predictions,
ATD+F females produced fewer extra-pair offspring compared to control females. This
result suggests that blocking the estrogenic and/or androgenic actions of T influences extrapair copulation behaviour of females. Contrary to my predictions, broods of T-treated
females also contained fewer extra-pair offspring compared to controls. This result is
consistent with previous studies and suggests that increasing exposure to T does not increase
the propensity of females to seek extra-pair matings. Overall, my study provides evidence
that altering exposure to T and its estrogenic metabolites during pre-breeding influences
whether females engage in extra-pair copulations, and provides insight into the role of

111

proximate mechanisms, such as hormone-mediated behaviour, in driving variation in rates of
extra-pair paternity in birds.
6.2. Introduction
In many socially monogamous animal species, females commonly engage in extrapair copulations, which may result in offspring being sired by males other than a female’s
social mate (Griffith et al., 2002). Since females may incur costs by copulating with males
outside of the social pair bond, this behaviour is hypothesized to have evolved because
females obtain fitness benefits from extra-pair mating (i.e., direct and indirect benefits;
reviewed in Birkhead and Møller, 1992; Petrie and Kempenaers, 1998). Alternatively,
females may engage in extra-pair copulations as a result of genetic constraints, such as extrapair copulation behaviour of females evolving as a correlated response to selection for
fecundity or other behaviours expressed by females (reviewed in Forstmeier et al. 2014).
Rates of extra-pair paternity in socially monogamous birds vary widely among
species and populations, and even within populations there is considerable variation; some
females have mixed-paternity broods while others do not (Ryder et al., 2012; Patrick et al.,
2012; Varian-Ramos and Webster, 2012; Hsu et al., 2015). Variation in rates of extra-pair
paternity among females may be due to intrinsic differences in behavioural traits, such as
exploratory (van Oers et al., 2008; but see Patrick et al., 2012) or sexual (Forstmeier, 2007)
behaviour. These individual differences in behavioural traits, commonly referred to as animal
‘personalities’(Carere et al., 2005; Dingemanse et al., 2010; Réale et al., 2010a), are
generally consistent over time and across contexts, such that some individuals tend to be
consistently more aggressive, bold or exploratory than other individuals (Sih et al., 2004a).
As such, these traits may confer an advantage to females in some contexts, but may also be
112

maladaptive in other contexts. For example, aggressive female tree swallows (Tachycineta
bicolor) are more successful at acquiring limited nesting sites (Rosvall, 2008), but they also
produce offspring of poor quality because aggressive females provide lower levels of
maternal care (Rosvall, 2011b). Nevertheless, females that display highly aggressive
behaviour may have a greater propensity to engage in extra-pair copulations, as demonstrated
previously in lizards (Egernia whitii) for which aggressive females produced a greater
number of offspring sired by extra-pair males (While et al., 2009). Therefore, individual
differences in personality of females may underlie variation in extra-pair paternity.
Personality traits may be related to promiscuous behaviour if they are mediated by
the same hormone(s) and as a result are correlated in their expression (hormone pleiotropy;
(Ketterson and Nolan, 1999). In female birds, concentrations of circulating hormones such as
testosterone (T) and 17β-estradiol (E2) peak prior to egg laying (e.g., Ketterson et al., 2005;
Williams et al., 2004), which coincides with an increase in aggressive (Cristol and Johnsen,
1994) and copulatory behaviour (Barber and Robertson, 2007; Crowe et al., 2009).
Aggressive behaviour has been shown to increase a female’s success at obtaining limited
nest sites (Rosvall 2008) and as such may influence extra-pair copulation behaviour by
enabling females to invade neighbouring territories in search of extra-pair mates. In addition,
sexual behaviour of females may enhance the opportunity for engaging in extra-pair
copulations; for example, during the fertile period, female humans (Homo sapiens) have
greater desire or interest in extra-pair mates (e.g., Gangestad et al., 2002; Gangestad et al.,
2005). Experimentally elevating T concentrations increases aggressive behaviour in female
birds (Zysling et al., 2006; Sandell, 2007; Rosvall, 2013) and sexual interest, libido and
behaviour in humans (reviewed in Bolour and Braunstein, 2005); however, studies

113

investigating whether elevated T mediates extra-pair copulation behaviour in female birds
have produced mixed results. T-treated female dark-eyed juncos (Junco hyemalis) did not
differ from control females in the proportion of extra-pair offspring in their nests (Gerlach
and Ketterson, 2013), whereas extra-pair paternity was lower for T-treated female spotless
starlings (Sturnus unicolor) and blue tits (Cyanistes caeruleus) when compared to control
females (García-Vigón et al., 2008; de Jong, 2013, respectively). Furthermore, de Jong
(2013) treated a third group of female blue tits with flutamide (F), which blocks androgen
receptors, and found that the number of nests containing extra-pair young did not differ
between F-treated and control females. The results of these experimental studies suggest that
elevated levels of T may not play a role in mediating the mating behaviour of females;
however, none of these investigations manipulated T by inhibiting the conversion of T to its
estrogenic metabolites. For example, the effects of T on extra-pair copulation behaviour may
occur through an alternative mechanism, such as the aromatization of T to E2. Experimental
studies have demonstrated that decreasing levels of E2 lowers sexual behaviour of female
birds, such as receptivity and solicitation (reviewed in Riters and Alger, 2011), and blocking
the estrogenic actions of T, in addition to androgen receptors, is necessary to fully reduce
aggressive behaviour (Archawaranon and Wiley, 1988). Therefore, the use of an aromatase
inhibitor to suppress the conversion of T to E2 in combination with F is necessary to fully
understand the effects of T on the mating behaviour of females.
My aim was to examine whether T and its estrogen metabolites mediate behavioural
traits that influence a female’s propensity to engage in extra-pair copulations and the
consequences for their reproductive success, using tree swallows (Tachycineta bicolor) as a
study species. Tree swallows exhibit high rates of extra-pair paternity, with up to 85 % of

114

nests containing at least one extra-pair young (O’Brien and Dawson 2007). Females in this
species engage in extra-pair copulations with males from neighbouring territories (Venier et
al., 1993), and presumably at roost sites away from the nest (sensu Dunn and Whittingham,
2005; Stapleton and Robertson, 2006). Both males and females initiate extra-pair
copulations, but those initiated by the female are more likely to result in fertilization (Venier
et al. 1993). Forced extra-pair copulations from males are rare (e.g., see Venier et al. 1993),
and to protect their paternity within the brood, males copulate frequently with their social
mate (Crowe et al., 2009). Female tree swallows are also highly aggressive in competition
with conspecifics for, and in defence of, a nest site (Leffelaar and Robertson 1985; Rosvall
2008). As such, hormones that mediate the sexual and aggressive behaviour of female tree
swallows may also influence the pursuit and acceptance of extra-pair copulations.
I tested the effects of T and its estrogenic metabolites on extra-pair paternity in tree
swallows by experimentally elevating concentrations of T using implants containing
exogenous T (increasing T) and blocking the estrogenic and androgenic actions of T using
implants containing 1,4,6-androstatrien-3, 17-dione (ATD) and flutamide (F) in combination
(ATD+F; decreasing the effects of T and E2); control females received empty implants. As
far as I am aware, the combined use of ATD+F has not been employed previously to
examine how T-mediated behaviour influences extra-pair paternity in female birds. If extrapair copulation behaviour of female tree swallows is correlated with the expression of
aggressive or sexual behaviour and is mediated by T and its estrogenic metabolites, then I
predicted that T-treated females would have a greater proportion of extra-pair offspring in
their broods than females in the ATD+F and control treatments as a result of being more
aggressive (Rosvall 2013) or having greater sexual interest in extra-pair mates. In contrast, I

115

predicted that females treated with ATD+F would produce fewer extra-pair offspring than Ttreated and control females as a result of being less aggressive (e.g., Hau et al., 2000;
Canoine and Gwinner, 2002) or having lower sexual interest or receptivity towards extra-pair
mates. Following the experimental alteration of T and its estrogenic metabolites, I
subsequently used microsatellite markers to assign parentage to offspring and determined the
proportion of extra-pair paternity in nests. I also monitored the nests of females during the
breeding season to examine how the treatment influenced reproductive success.
6.3. Material and methods
6.3.1. Study area and species
I studied tree swallows breeding in nest boxes west of Prince George, BC, Canada
(53ºN, 123ºW), from May to August 2012. The study area consisted of open agricultural
areas intermixed with small wetlands and patches of coniferous and deciduous trees (see
Dawson et al., 2005 for more details). Tree swallows are small, migratory, aerial insectivores
that are socially monogamous and exhibit high rates of extra-pair paternity (e.g., up to 85 %
of broods contain at least one extra-pair offspring, see O’Brien and Dawson, 2007 and
references therein). Clutch size is typically 4-7 eggs, which is incubated only by the female,
and both sexes contribute to provisioning offspring (Winkler et al., 2011).
6.3.2. General field methods
During early May, I visited nests daily to document nest building, and captured
females in nest boxes once construction of the nest cup was complete, but prior to eggs being
laid. From a subset of females I collected a blood sample (approx. 50–80 µL) from the
brachial vein to measure pre-laying levels of T and E2, and recorded handling time (elapsed

116

time from closing the trap to collection of the blood sample) using a stopwatch (mean ± SE:
4.16 ± 0.35 min, range: 2.78–9.35, N = 22 females). Blood samples were collected in
heparinized capillary tubes and stored on ice until they were centrifuged for 5 minutes in the
lab. Plasma was transferred to micro-centrifuge tubes and stored at –80 oC until hormone
analysis (details below).
Age of females was determined according to plumage colour as being in either their
second year (SY) of life or older (after-second year; ASY; Hussell 1983); only ASY females
were used in this study because SY females tend to breed later than ASY females
(Stutchbury and Robertson, 1988) and also may exhibit different strategies than ASY
females in response to experimental manipulations (e.g., Ardia and Clotfelter, 2007). ASY
females were sequentially allocated by capture order to the T, ATD+F, or control treatment
after determining the treatment order randomly. Females received implants constructed from
Silastic® laboratory tubing (1.47 mm I.D., 1.96 mm O.D.; Dow Corning #508-006) sealed at
both ends with Silastic glue. For T implants, I used 6-mm implants packed with 3.5 mm of
testosterone (approx. 0.0048 g; Sigma–Aldrich # T1500), a lower dose than Rosvall (2013),
who reported hatching failure of eggs in the nests of female tree swallows implanted with 5
mm of T. I constructed ATD and F implants of equal size and these contained on average
0.0052 g of ATD (Steraloids, Inc #A4100-000) and 0.0035 g of F (Sigma–Aldrich #F9397).
Implants from all treatments were cleaned with ethanol, and incubated at 37 oC overnight in
tubes containing 0.9 % saline (Sigma-Aldrich #S8776) to begin the diffusion process of
hormones across the tubing and avoid the surge of hormone release that can occur after
implantation (Smith et al. 1977). Implants were then transferred to tubes containing fresh
saline and transported to the field. In the field, I treated all females with a topical anesthetic

117

and then made a small incision using surgical scissors above the left flank. Two implants (T
and sham, ATD and F, or two sham) were inserted subcutaneously under the skin along the
flank and the incision site was sealed with cyanoacrylate tissue adhesive (VetbondTM
#1469SB). I implanted a total of 12 T, 15 ATD+F, and 13 control females for this study (see
Table 6.1).
To examine whether experimentally altering exposure to T and its estrogenic
metabolites influenced reproductive success of females, all nests boxes were monitored daily
to record the date of clutch initiation (where January 1 = 1) and clutch size. Freshly laid eggs
were numbered with a non-toxic marker for identification and were weighed with a digital
scale (nearest 0.01 g). Once a clutch was complete, I checked nests daily to determine the
onset of incubation, at which time nests were left undisturbed. None of the T-treated females
incubated their clutches; these eggs were therefore collected 7 to 9 days after clutch
completion and stored at –20 oC until DNA extraction. Nests in the ATD+F and control
treatments were checked daily beginning 12 days after clutch completion to record the day of
hatch, designated as day 0 of the brood-rearing period. Any eggs that failed to hatch from
these nests were collected and stored as described above. On day 4 after hatching, I uniquely
marked the legs of each nestling in a brood with a non-toxic marker for individual
identification. I measured mass with a spring balance (nearest 0.125 g) and length of the
combined head and bill (hereafter, head-bill) with digital calipers (nearest 0.1 mm) every 2
days from day 4 to 16, and length of the ninth primary feather with a ruler (nearest 0.5 mm)
from day 8 to 16. On day 12, a small blood sample (20 L) was collected for paternity
analysis from the brachial vein of nestlings and stored in Queen’s lysis buffer at 4 oC (Seutin
et al., 1991) until DNA extraction (see details below). From nestlings found dead in the nest

118

prior to day 12, I collected brain tissue and stored it in the same manner as blood. On day 22,
I visited nests to determine the number of nestlings that successfully fledged.
I captured adult males while they were feeding offspring and collected a blood
sample for paternity analysis. Control- and ATD+F-treated females were recaptured on day
12 to 16 of the nestling period to remove implants and collect blood samples to measure
levels of T and E2 post-implants. I also recaptured T-treated females, and captured their
social mates when possible, before collecting the clutch. Unfortunately, I was unable to
capture the social mate at 5 broods where the female had been implanted with T. Blood
samples drawn from experimental females to measure T and E2 concentrations were collected
on average 3.58 minutes after capture (± 0.23 standard error (SE), range: 2.2–7.4 minutes, N
= 22 females).
6.3.3. Hormone measurements
I quantified concentrations of T in plasma samples of control and T-treated females
using a T enzyme immunoassay (EIA) kit (Expanded Range Salivary Testosterone EIA kit,
Salimetrics #1-2402). As ATD has a high cross-reactivity with the T assay kit (Vandermeer,
2013), I instead quantified E2 concentrations in plasma samples of females in the control and
ATD+F treatment using an E2 EIA kit (High Sensitivity Salivary 17β-estradiol EIA,
Salimetrics #1-3702). Both assay kits have been used previously to quantify T and E2
directly from avian plasma without extraction (Washburn et al., 2007; Hall and MacDougallShackleton, 2012) and were validated for use in tree swallows by demonstrating a
displacement curve from a serial dilution of pooled plasma samples that was parallel to the
standard curve for each assay.

119

For the T assay, plasma samples were diluted 1:1 with assay buffer (total volume of
plasma was 12.5 µL). Due to low plasma volumes only five samples (all T-treated females)
were run in duplicate; all other samples were assayed only once. The T kit had a high crossreactivity with other androgens, such as dihydrotestosterone (36.4 %), 19-nortestosterone
(21.01 %), 11-hydroxytestosterone (1.90 %), and androstenedione (1.16 %), and a low crossreactivity (<0.5 %) with other hormones and compounds (see manufacturer’s instructions for
more details). Given the high cross-reactivity of the T kit with the androgens listed above, I
report my results from the T assay as ‘androgen concentrations’ as opposed to T
concentrations.
The volume of plasma used in the E2 assay ranged from 4 to 20 µL due to the
availability of plasma, and E2 concentrations were corrected based on the dilution factor
(range of dilutions: 1:4 to 1:20 in assay buffer). Most samples were assayed only once; two
ATD+F and one control sample were assayed in duplicate. The sensitivity of the assay was
0.1 pg/mL (see manufacturer’s instructions) and the concentration of all samples was above
this minimum detection limit. This assay had a low cross reactivity with estrone (1.23 %),
estriol (0.23 %), ethynylestradiol (0.19 %), and prednisone (0.02 %), and had no crossreactivity with other hormones or compounds (see manufacturer’s instructions for more
details).
All samples were run in a single assay for each kit. The intra-assay coefficient of
variation calculated from high and low controls of known concentration provided with each
kit were 8.2 % and 1.4 % for T and 0.7 % and 1.2 % for E2 respectively.

120

6.3.4. Paternity analysis
I removed the germinal disk or embryonic tissue from eggs that failed to hatch and
stored it in Queen’s lysis buffer at 4 oC (Seutin et al. 1991) until DNA extraction. The
germinal disk or embryonic tissue was not identified in three eggs (one egg from two
clutches in the T treatment, and one egg from a clutch in the ATD+F treatment), which I
considered as unfertilized for the paternity analysis. Genomic DNA from blood and tissue
samples was extracted using DNeasy Blood & Tissue kits (Qiagen #69506), and from
germinal disks using QIAamp DNA Micro kits (Qiagen #56304) following the
manufacturer’s protocols. For 22 females in this experiment where blood was collected for
hormone assays using heparin, I extracted genomic DNA from feather samples collected
from the same birds as part of another study, following a slightly modified feather extraction
protocol (protocol nine; Nishiguchi et al., 2002). Extracted DNA was amplified at either five
or six microsatellite loci developed for Tachycineta swallows (Tle19, TaBi34, TaBi6, Tal8,
TaBi10 and/or TaBi 8; Makarewich et al., 2009). Duplex polymerase chain reaction (PCR)
was performed with two primer pairs TaBi6/Tle19 and Tal8/TaBi34, and PCR for TaBi10
and TaBi8 was carried out as single reactions. For DNA extracted from blood samples, the
final 10 L volume for PCR contained 1x PCR buffer, 0.2 mM of each dNTP (Invitrogen
#10297-018), 1.75–2.25 mM MgCl2, 0.4 g/L bovine serum albumin (BSA; New England
Biolabs #B90015), 0.08–0.2 M of forward-labelled and reverse primers, 0.05 U/L Taq
polymerase (Invitrogen #10342-020), and 20–50 ng of genomic DNA. PCR was performed
in a MJ Research Peltier Thermal Cycler under the following PCR cycling protocol: an initial
denaturing step at 95 °C for 4 min, followed by 35 cycles of 95 °C for 50 sec, 56–60 °C for
60 sec (56 °C for the duplex reactions and 60 °C for TaBi10 andTaBi8), and 65-72 °C for 90
121

sec (65 °C for the duplex reactions and 72 °C for TaBi10 andTaBi8). A final extension cycle
at 65–72 °C for 5 min (65 °C for duplex reactions and 72 °C for TaBi10 andTaBi8)
completed the PCR. For all other DNA, PCR reactions were performed using a Qiagen
Multiplex PCR kit (Qiagen #206143) following manufacturer’s instructions, except I used a
final volume of 10 L and 3–5 L of Multiplex PCR Master Mix. For DNA from germinal
disks only, I added 0.4 g/L of BSA to the DNA and incubated the samples at 95 °C for 5
min prior to PCR. Primer concentrations and annealing temperatures remained the same as
described above. PCR products were run on an ABI 3130xl automated sequencer (Applied
Biosystems) to obtain individual genotypes and assign paternity. Data were analyzed blind to
sample identity using Peak ScannerTM Software v1.0 (Applied Biosystems).
Allele frequencies and exclusion probabilities for microsatellite loci in my study
population were calculated for all adults using CERVUS 3.0 (Kalinowski et al., 2007; see
Table 4.1). I genotyped 295 adults (resident females and males, and non-breeding
individuals) and 1116 offspring as part of a larger three-year study (2010–2012). Individuals
were genotyped at five loci Tle19, TaBi34, TaBi6, Tal8, and TaBi10; however, due to a high
null allele frequency for Tal8 in my study population, I also genotyped all males and extrapair offspring that were homozygous at Tal8 plus the genetic mother at a sixth locus (TaBi8)
for individuals sampled in 2010–2011. Starting in 2012, individuals were genotyped at six
loci to ensure that parentage was assigned using five loci. TaBi8 significantly deviated from
Hardy-Weinberg equilibrium (P = 0.002; Table 4.1), but this locus was never used as a
single criterion to determine parentage. I classified offspring as within-pair if they matched
the resident male at five loci, and extra-pair offspring if they mismatched the resident male at
a minimum of 1 locus. For this experiment, only a single nestling mismatched the resident

122

male at only 1 locus, and this nestling shared alleles with a sibling that mismatched the
resident male at 4 loci. All other mismatches among nestlings and resident males occurred at
2 or more loci. I was able to assign parentage using 5 loci with no apparent null alleles to all
offspring except for four offspring genotyped at only four loci either due to null alleles or
poor amplification. For these four offspring, I calculated the probability of chance inclusion
following O’Brien and Dawson (2007). All four offspring were within-pair, and the
probability that a randomly chosen male would match these offspring at all four loci was low
(mean ± standard error, 0.0028 ± 0.0025; range: 0.0010 to 0.0064).
6.3.4. Statistical analysis
To determine whether treating females with implants altered concentrations of
androgens and E2, I used a repeated measures analysis of variance (ANOVA), with timing of
blood sample (pre-breeding or nestling rearing) as a within-subject repeated measure,
treatment as a between-subject factor, and an interaction between timing of blood sampling
and treatment. Since only a few females from each treatment were blood sampled both
before and after receiving an implant (N = 3-5 females sampled at both times per treatment),
I also used a one-way ANOVA to test for differences in T and E2 levels among treatment
groups after females were implanted.
To examine whether extra-pair paternity differed by treatment, I used a generalized
linear model (GLM) fitted with binomial errors and logit link function for proportion data
(Crawley 2013). The response variable was the number of within-pair offspring and the
binomial denominator was the number of offspring in the brood. Treatment was included as a
fixed factor, and the number of days a female was implanted prior to initiating a clutch
(mean ± SE, 9.63 ± 1.35 days, range: 1–30) was included as a covariate. Clutch initiation
123

date and number of days implanted were correlated (r = 0.93, P < 0.0001), so I did not
include clutch initiation date in the analysis. I tested the model for overdispersion by
determining whether the residual deviance was greater than the degrees of freedom, and I
refitted the model with a quasi-binomial error structure to account for overdispersion
(Crawley, 2013). Following Crawley (2013), I tested the significance of the interaction and
covariate by assessing the analysis of deviance between models. Four females initiated
clutches the day following the manipulation (see below); I had paternity data for two (one T
and one ATD+F) and included these data in the analysis. The T female laid the first egg
(extra-pair offspring) and then delayed egg laying for 4 days; all subsequent offspring were
within-pair offspring. The ATD+F treated female continued to lay eggs after being
manipulated and exclusion of this female did not change the overall conclusions of my
results.
To test whether clutch initiation dates differed among females after receiving
implants, I used a general least-squares model from the package nlme (Pinheiro et al., 2016)
to account for heteroscedasticity in variances following Cleasby and Nakagawa (2011). From
this analysis, I excluded three T and one ATD+F female that initiated a clutch the morning
after receiving implants because the physiological processes associated with egg formation
would have been initiated before these females were implanted, and so it is unlikely the
hormones influenced initiation dates in these cases. The T-implanted females skipped egg
laying for 2–4 days after initiating a clutch, while the ATD+F-implanted female continued to
lay eggs; nevertheless all four females were included in subsequent analyses because the
effects of silastic implants on T and aromatase activity can be detected within a day of being
implanted (see Smith et al., 1977; Balthazart et al., 1990). I tested whether clutch size and

124

average egg mass differed among females in each treatment using separate analysis of
covariance (ANCOVA) for each variable that included clutch initiation date as a covariate to
control for the seasonal decline in clutch size that occurs in tree swallows (Winkler et al.,
2014). I used a z-transformation to center and scale clutch initiation date so that the effect of
treatment on clutch size or average egg mass could be interpreted in the presence of a
treatment x initiation date interaction (Schielzeth, 2010).
To determine whether nestling performance differed among the nests of females in
the ATD+F and control treatments, I examined mass and size (length of ninth primary
feather and head-bill) at day 16, and the growth rates of nestlings. Growth rate constants
were calculated for each nestling using a logistic model for mass, Gompertz model for headbill, and linear model for ninth primary following Dawson et al. (2005). All nests had
complete growth data, but one nest was measured on day 17 and was not included in the
analysis of mass or size at day 16. Moreover, since I was interested in the effects of female
treatment on nestling growth and size, I excluded from these analyses two nests (one ATD+F
and one control) where the female was not observed after the hatching of eggs and a second
(helper) female was provisioning the nestlings. I used random intercept linear mixed models
(nlme, Pinheiro et al., 2016) with mass or size at day 16, or growth rate constant as
dependent variables, treatment as the fixed factor, and brood identity as a random factor. I
also z-transformed brood size at day 16 and time of measurement and included both variables
as covariates in models analyzing mass and size at day 16, but only included brood size in
models analyzing growth rate constants. I also included the interaction between treatment
and brood size at day 16. When the residuals of models were not homoscedastic, I
incorporated a variance function to account for heteroscedasticity (Cleasby and Nakagawa,

125

2011).
I tested whether the fledging success of nestlings, i.e., number fledged in relation to
clutch size, differed between the ATD+F and control treatments for broods that hatched at
least one nestling using a (GLM) fitted with quasi-binomial errors and logit link function as
described above. Treatment was included as a fixed factor, and brood size at day 16 and
hatch date as covariates.
For all analyses, I removed interactions and covariates that did not approach
significance (P ≥ 0.10) in a backwards, stepwise fashion. Means are presented ± 1 SE, and
results were considered significant at P ≤ 0.05. The overall significance of omnibus tests was
examined using post-hoc tests with a false discovery rate adjustment for multiple
comparisons (Verhoeven et al. 2005). Where appropriate, I report parameter estimates (B)
and effect sizes; effect sizes for the repeated measures analysis as general eta squared (ηG2)
following Lakens (2013) and for post-hoc comparisons as the correlation coefficient (r)
following Field et al. (2012). All statistical analyses were performed using SPSS (IBM Corp.
2011) and R (R Development Core Team 2015) statistical software.
6.4. Results
Within individual females, androgen concentrations differed before receiving
implants during pre-breeding and after receiving implants when measured during the nestling
period (F1, 6 = 31.95, P = 0.00l, ηG2 = 0.59, Figure 6.1a), and between the T and control
treatments (F1, 6 = 10.679, P = 0.02, ηG2 = 0.56, Figure 6.1a). The interaction between
treatment and implant time was also significant (F1, 6 =129.50, P < 0.0001, ηG2 = 0.85), and
when I analyzed the data separately by treatment, androgen concentrations decreased in
control females after receiving sham implants (F1, 3 = 15.39, P = 0.03, ηG2 = 0.48, Figure
126

6.1a), but increased in females after being implanted with T (F1, 3 = 155.31, P = 0.001, ηG2 =
0.87, Figure 6.1a). Since the number of females blood-sampled both before and after
receiving implants was small (N = 4 for each treatment), I also compared circulating levels of
androgens for all control and T-treated females that were captured, either pre-breeding or
during the nestling period. There was no difference in androgen concentration between
treatments prior to females receiving implants (F1, 7 = 12.84, P = 0.14, ηG2= 0.29, Figure
6.1a), but T-treated females had a higher concentration of androgens compared to control
females during brood rearing (F1, 11 = 86.56, P < 0.0001, ηG2= 0.90, Figure 6.1a).
The concentration of E2 decreased from pre-breeding to day 12 of the nestling period
in females treated with sham and ATD+F implants (F1, 6 = 12.95, P = 0.01, ηG2 = 0.52);
however, there was no effect of treatment on E2 concentration (F1, 6 = 0.07, P = 0.80, ηG2 =
0.006) and the interaction between time and treatment was also non-significant (F1, 6 = 0.44,
P = 0.53, ηG2 = 0.04). Since only three control females were blood-sampled both before and
after receiving implants, I also compared E2 concentrations between treatment groups postimplant. E2 concentration did not differ between ATD+F-treated and control females for
blood samples on day 12 of the nestling period in June (F1, 11 = 0.14, P = 0.71, ηG2 = 0.01,
Figure 6.1b).
The proportion of extra-pair paternity differed significantly among treatment groups
(F2, 24 = 4.09, P = 0.03, Figure 6.2). The number of days a female was implanted prior to
initiating a clutch was not a significant predictor of extra-pair paternity and was removed
from the model (B = 0.06 ± 0.06; F1, 23 = 1.11, P = 0.30). Post-hoc tests showed that females
treated with ATD+F had a lower proportion of extra-pair offspring in their nests compared to
controls (B = -1.74 ± 0.71, t24 = -2.43, P = 0.045, r = 0.44). Surprisingly, when compared to

127

control females, females treated with T also produced a lower proportion of extra-pair
offspring (B = -1.98 ± 0.91, t24 =

-2.16, P = 0.046, r = 0.40). There was no difference in

the proportion of extra-pair offspring between nests where females were treated with ATD+F
and T (B = 0.24 ± 0.90, t24 = 0.27, P = 0.79, r = 0.06). In the above analysis, I considered all
nests with available paternity data; however, paternity data were incomplete for five nests
due to the disappearance or breakage of a single egg in each of the nests. Performing the
analysis with the exclusion of these data did not alter the conclusion.
Clutch initiation date differed by treatment (F2, 28 = 3.52, P = 0.04). Post-hoc tests
showed that T-treated females tended to initiate clutches later than control females (B = -7.90
± 3.32, t28 = -2.38, P = 0.07, r = 0.44), but not ATD+F females (B = -5.25 ± 3.61, t28 = -1.46,
P = 0.16, r = 0.27), and no difference in clutch initiation date was observed between
ATD+F-treated and control females (B = -2.63 ± 1.83, t28 = -1.44, P = 0.16, r = 0.26).
However, the trend for later breeding in the T treatment appeared to be due to the influence
of three females that initiated their clutch approximately 25-34 days later than the other
females in this treatment group. Removing these three females from the analysis showed that
clutch initiation date no longer differed by treatment (F2, 25=1.41, P = 0.26).
After controlling for clutch initiation date (B = -1.90 ± 0.75, F1, 28 = 6.45, P = 0.02),
clutch sizes differed among treatment groups (F2, 28 = 3.31, P = 0.05). Post-hoc tests
suggested that females treated with ATD+F tended to lay larger clutches than females in the
T (B = 0.80 ± 0.39, t28 = 2.11, P = 0.065, r = 0.37) and control treatments (B = 0.96 ± 0.43,
t28 = 2.20, P = 0.065, r = 0.38). Clutch size of control and T-treated females did not differ by
treatment (B = 0.16 ± 0.44, t28 = -0.35, P = 0.73, r = -0.06). There also was a significant
interaction between treatment and initiation date (F2, 28 = 3.58, P = 0.04, Figure 6.3).

128

Analyzing data by treatment revealed a negative relationship between clutch size and
initiation date in control (B = -1.87 ± 0.73, F1, 9 = 6.61, P = 0.03, r = -0.63) and T treatment
groups (B = -1.09 ± 0.34, F1, 9 = 9.92, P = 0.01, r = -0.72), but no relationship was detected
in the ATD+F treatment group (B = 0.06 ± 0.26, F1, 10 = 0.05, P = 0.82, r = 0.07). As noted
above, three T-treated females initiated their clutches later than the other T-treated females.
Removal of these females from the analysis changed the results so that no effect of treatment
on clutch size was detected (F2, 25 = 1.21, P = 0.31), even after controlling for clutch
initiation date (B = -1.18 ± 0.43, F1, 25 = 7.70, P = 0.01); however, there was still a significant
interaction between treatment and clutch initiation date (F2, 25 = 3.42, P = 0.048). This
interaction was due to the decline in clutch size with later laying dates in control nests only
(see above) as clutch sizes no longer declined with later clutch initiation dates in the T
treatment once the three late-breeding birds were removed from the analysis (B = -0.24 ±
0.35, F1, 6 = 0.50, P = 0.51). Average egg mass did not differ among treatment groups (F2, 30
= 0.67, P = 0.52).
After controlling for brood size (B = -0.63 ± 0.18, F1, 13 = 3.73, P = 0.08), nestlings
were lighter at day 16 in the ATD+F than control treatments (B = -0.52 ± 0.29, F1, 13 = 6.05,
P = 0.02, Figure 6.4). There also was a significant interaction between treatment and brood
size (B = 0.80 ± 0.27, F1, 13 = 8.71, P = 0.01). Analyzing data by treatment indicated that
there was a negative relationship between brood size and mass at day 16 for nestlings reared
in control broods (B =

-0.62 ± 0.15, F1, 5 = 16.12, P = 0.01), but no relationship was

observed in ATD+F broods (B = 0.18 ± 0.18, F1, 8 = 1.00, P = 0.35). Nestlings reared in
ATD+F and control broods did not differ in length of ninth primary (F1, 14 = 0.14, P = 0.72),
even after controlling for time of measurement (F1, 14 = 3.89, P = 0.07), or length of head-bill

129

(F1, 15 = 2.02, P = 0.18). Nestlings grew at similar rates regardless of the treatment of the
female (ninth primary, head-bill, and mass, all P-values > 0.16). Fledging success of
nestlings did not differ by treatment (F1, 18 = 0.90, P = 0.35); 86.2 ± 0.06 % of nestlings in
the ATD+F and 94.0 ± 0.04 % of nestling in the control treatments fledged successfully.
6.5. Discussion
My aim was to examine whether T and its estrogen metabolites mediate behavioural
traits that influence a female’s propensity to engage in extra-pair copulations. I predicted that
if extra-pair copulation behaviour was correlated to the expression of aggressive or sexual
behaviour, then altering hormones that mediate behavioural traits of females would also
influence their extra-pair copulation behaviour. Females in the T treatment had elevated
levels of androgens after being treated with T (Figure 6.1a), and produced fewer extra-pair
offspring than control females (Figure 6.2). Although this result is contrary to my
predictions, it demonstrates that altering T of female tree swallows influenced extra-pair
copulation behaviour. In contrast, females treated with ATD+F did not have lower E2
concentrations than control females (Figure 6.1b); however, I suspect that this is due to
collecting post-implant blood samples late in the season when females were provisioning
nestlings. E2 concentrations decline seasonally in wild female passerines (e.g., Schwabl et al.,
2014; Figure 6.1b), and I sampled females when E2 levels would have been naturally low in
control females. While I recognize that a second blood sample collected at egg laying was
needed to confirm differences in E2 between control and ATD+F females, this was not
feasible as it almost certainly would have resulted in females abandoning their nesting
attempt (e.g., Veiga et al., 2004). Nevertheless, that females treated with ATD+F had a lower

130

proportion of extra-pair offspring in their broods when compared to control females (Figure
6.2) demonstrates that the ATD+F treatment successfully altered the behaviour of females.
Previous studies in a variety of taxa have reported that consistent individual
differences in behavioural traits, or personalities, of females are related to variation in their
extra-pair copulation behaviour. For example, female lizards that are more aggressive have a
greater number of offspring in their litter sired by extra-pair males (While et al. 2009).
Similarly, exploratory behaviour increased the probability of extra-pair offspring in the nests
of female great tits (Parus major), but only when mated to a male with a similar personality
type (e.g., fast-fast explorers; van Oers et al. 2008); however, a relationship between
exploratory behaviour and extra-pair paternity was not detected in females from a different
population of great tits (Patrick et al. 2012). Moreover, female zebra finches (Taeniopygia
guttata) that exhibited greater readiness to copulate with a male during their first sexual
encounter in life were more likely to engage in extra-pair copulations once socially paired
than females that did not copulate during their first encounter with a male (Forstmeier 2007).
The low proportion of extra-pair offspring in ATD+F-treated, and presumably less
aggressive or sexually receptive, females compared to control females in my study provides
support for the hypothesis that personality may underlie variation in the rate of extra-pair
paternity among females; however, the similarly low proportion of extra-pair paternity
displayed by T-treated females suggests that increased aggressive behaviour in female birds
does not appear to be associated with enhanced extra-pair copulation behaviour.
Although I was unable to directly observe how my experimental manipulation altered
the extra-pair copulation behaviour of females, due to the rarity of witnessing such behaviour
in wild birds (e.g., Venier et al. 1993), I propose two mechanisms to explain the lower

131

proportion of extra-pair offspring in females treated with ATD+F (Figure 6.2). First, extrapair paternity in tree swallows often occurs when females visit neighbouring territories of
other males to copulate (Venier et al. 1993), and being aggressive would facilitate a female’s
ability to invade these territories that are defended by resident females (see Kempenaers et
al., 1992). Therefore, by blocking the actions of T and its estrogenic metabolites to reduce
aggressive behaviour of females in the ATD+F treatment, these females may have
experienced limited opportunities for encountering extra-pair males in territories other than
their own. Females treated with T, however, had similarly low levels of extra-pair paternity,
which suggests that aggressive behaviour does not facilitate the pursuit of extra-pair
copulations, although it cannot be ruled out that T-treated females similarly encountered
fewer potential extra-pair mates if they spent more time involved in territory defence or were
aggressive to potential extra-pair mates (see below).
Second, the sexual behaviour of females may have been lowered in ATD+F treatment
compared to controls as a result of blocking the estrogenic and androgenic actions of T. For
example, T can mediate behaviour by being aromatized to E2 in the brain and acting on
estrogen receptors (reviewed in Ball and Balthazart, 2008). In females, E2 mediates
receptivity and solicitation behaviour (reviewed in Riters and Alger, 2011), and experimental
studies in birds and mammals have demonstrated that these behaviours can be delayed or
suppressed using an aromatase inhibitor (Rissman et al., 1990; Leboucher et al., 1998; Belle
et al., 2005; but see Tomaszycki et al., 2006). While the lower proportion of extra-pair
offspring in the nests of ATD+F females may be due to lower receptivity or solicitation
behaviour as a result of ATD suppressing E2 concentrations, previous studies in male song
sparrows (Melospiza melodia morphna) and both sexes of zebra finch reported ATD to be

132

ineffective at suppressing E2 production (Soma et al., 1999; Kabelik et al., 2010,
respectively) and my blood samples collected during the nestling period cannot confirm that
E2 concentrations in the ATD+F females were lower than control females during egg laying.
As such, it is also possible that the lower proportion of extra-pair offspring in the nests of
ATD+F females was due to F preventing T from binding directly to the androgen receptors.
For example, female goats (Capra aegagrus hircus) treated with F are less receptive to males
than control females (Imwalle and Katz, 2004); however, a previous study in blue tits
reported that treating females with F alone did not influence extra-pair copulation behaviour
as no difference in extra-pair paternity was observed between control and F-treated females
(de Jong, 2013). Although my study is unable to determine whether the effect of the ATD+F
on extra-pair copulation behaviour of female tree swallows is due to inhibiting aromatase
activity, androgen receptor binding, or both, future studies could use implants containing E2
and other aromatase inhibitors, such as fadrozole, to examine the proximate effects of
estrogenic actions of T on the mating behaviour of females.
Contrary to my predictions, but consistent with previous studies in European starlings
and blue tits (García-Vigón et al., 2008; de Jong, 2013), T-treated female tree swallows also
had lower levels of extra-pair paternity than control females (Figure 6.5). García-Vigón et al.
(2008) hypothesized that T-treated females may have produced fewer extra-pair offspring
because T reduced the attractiveness of females, T-treated females invested more time in
aggressive encounters and territory defence, or rejected copulations from extra-pair males
because they were more aggressive. T has been shown to have masculinizing effects on
female traits in a variety of taxa, although whether such changes reduce their attractiveness
to potential mates is unclear (Ketterson et al., 2005; Lahaye et al., 2013). Given that

133

exogenous T increases aggressive behaviour in female tree swallows (Rosvall, 2013), in my
study T-treated female tree swallows likely spent more time involved in aggressive
interactions and territory defence, as previously suggested for spotless starlings (GarcíaVigón et al., 2008), rather than seeking extra-pair copulations.
Implanting females with T prior to laying appeared to have negative consequences for
reproductive success since females implanted with T during pre-laying did not incubate their
eggs, even though they developed brood patches and remained active at their nest sites
during the incubation and nestling periods of the breeding season (Berzins, personal
observation). These results are similar to a previous study in tree swallows that
experimentally elevated T levels of females during incubation and reported cooler nest
temperatures and complete hatching failure (Rosvall, 2013). Female great tits implanted with
T during nest building also had lower incubation temperatures and hatching success (de Jong
et al., 2016), but such negative effects of T on incubation behaviour were not observed in
female dark-eyed juncos implanted prior to breeding (Clotfelter et al., 2004). The results of
my study and others suggest that T levels elevated after egg laying disrupt incubation
behaviour in at least some species. In contrast, ATD+F females did not differ from control
females in the initiation date of their clutches, clutch size or average egg mass. While female
zebra finches treated with ATD+F produced fewer eggs compared to control females
(Tomaszycki et al. 2006), I did not detect such effects of treating my birds with ATD+F,
which may be due to using a lower dose of ATD+F in my study. Females treated with
ATD+F did have lighter nestlings in their brood on day 16 compared to control females
(Figure 6.4). Although I have no data on parental feeding rates, average mass of eggs (see
Results), as well as mass of nestlings on day four did not differ between the ATD+F and

134

control treatments (P = 0.94; Berzins, unpublished data), suggesting the lower mass of
nestlings on day 16 may be due to ATD+F-treated females or their mates provisioning
nestlings less than in the control group. Nonetheless, despite nestlings reared in the ATD+F
treatment being lighter on day 16, nestlings were of similar structural size, grew at similar
rates, and were as likely to fledge as those raised by control females.
In conclusion, my study demonstrates that experimentally altering exposure to T and
its estrogenic metabolites prior to breeding in female tree swallows reduces the rate of extrapaternity in their broods. Although my results for ATD+F females were consistent with my
predictions, the finding that T-treated females similarly produced a lower proportion of
extra-pair offspring, although consistent with previous studies, was contrary to my
predictions and suggests that elevated T does not enhance extra-pair mating of females. If Ttreated females were less likely to seek extra-pair copulations as a result of being more
aggressive, then such behaviour may actually lower female fitness if females are seeking
extra-pair copulations to improve offspring quality (e.g., see O’Brien and Dawson, 2007;
Stapleton et al., 2007; Whittingham and Dunn, 2010). Overall my study suggests that extrapair paternity in tree swallows is driven partly by the behaviour of females.

135

Table 6.1. Summary of the number of female tree swallows that received implants containing
testosterone (T) or 1,4,6-androstatrien-3, 17-dione and flutamide (ATD+F), or empty
implants (control). ‘Initiated a clutch’ refers to the number of females that laid at least one
egg. ‘Hatched eggs’ refers to the number of females that hatched at least one egg.
‘Disappeared’ refers to females who were usurped or abandoned their nesting attempt, or
whose eggs were depredated.
Treatment

Manipulated

Initiated a clutch

Hatched eggs

Disappeared

ATD+F
Control

15
13

12
11

11
9

5
4

T

12

12

0

0

136

Figure 6.1. Concentrations of a) androgens and b) 17β-estradiol in female tree swallows
measured prior to receiving implants during pre-breeding in May and after receiving
implants containing testosterone or sham implants during the nestling period in June. Black
circles represent control females, white circles represent testosterone (T)-treated females, and
grey circles represent females treated with 1,4,6 androstatrien-3, 17-dione and flutamide
(ATD+F). Repeated measures for individuals are connected by a line, whereas individual
circles indicate single measurements where repeated measurements were not available. Data
points were jittered for easier interpretation.
137

Figure 6.2. Mean (± SE) proportion of extra-pair offspring in nests of tree swallows where
females were either implanted with 1,4,6 androstatrien-3, 17-dione and flutamide (ATD+F)
or testosterone (T), and for controls (C; sham implanted). Sample sizes are given above error
bars.

138

Figure 6.3. Relationship between clutch size and initiation date (1 January = 1) for nests
where female tree swallows were implanted with 1,4,6 androstatrien-3, 17-dione and
flutamide (dotted line), testosterone (dashed line), or received sham implants (solid line).
Data points were jittered for easier interpretation.

139

Figure 6.4. Mean (± SE) body mass at day 16 of nestling tree swallows raised in broods
where females were treated with 1,4,6 androstatrien-3, 17-dione and flutamide (ATD+F) and
sham implants. Sample sizes indicate the number of broods within each treatment and are
given above error bars.

140

Chapter 7: Synthesis
Ornamental and behavioural traits expressed by female animals may improve their
ability to attract mates and outcompete conspecifics for access to limited breeding
opportunities and resources (Clutton-Brock, 2009; Rosvall, 2011a; Tobias et al., 2012).
Although studies have demonstrated that such traits function as signals of quality that are
preferred by males (Griggio et al., 2005; Torres and Velando, 2005) and are important in
agonistic interactions (Griggio et al., 2010; Midamegbe et al., 2011; Morales et al., 2014),
less is known about how ornamentation and behaviour influence the mating success of freeliving female birds. For my research, I used an experimental approach that altered the
plumage brightness or testosterone (T)-mediated behaviour of female tree swallows to
examine three main objectives: (1) whether variation in female plumage brightness is a
signal of attractiveness that influences male investment in reproduction, including parental
care (Chapter 2), social and extra-pair mate choice (Chapter 4) and male mating strategies
(Chapter 4); (2) whether plumage brightness is a signal assessed by conspecific females that
influences nest box retention (Chapter 3) and whether females with bright plumage
experience social costs, such as lower reproductive success (Chapter 3) or nestling quality
(Chapter 5); and (3) whether variation in plumage brightness (Chapter 4) or T-mediated
behaviours (Chapter 6) influenced a female’s opportunity to engage in extra-pair copulations.
Ornamental traits of females that are preferred by males may influence their
investment in reproduction (Edward, 2015). For instance, variation in the attractiveness of
ornamentation displayed by females may influence how much parental care males invest in
the current reproductive attempt (Burley, 1986; Sheldon, 2000). Previous studies have
demonstrated that males adjust their investment in parental care according to the manipulated

141

attractiveness of female ornaments (Burley, 1988; Roulin, 1999; Pilastro et al., 2003; Mahr et
al., 2012); however, male tree swallows did not adjust their parental care according to the
plumage brightness of females (Figure 2.1a), which is in contrast to these previous studies.
My results did show that male tree swallows tended to provision nestlings in experimental
broods (i.e., reduced and enhanced treatments) at a higher rate compared to control broods
(Figure 2.2b). Although I performed my experiment after the hatching of eggs, as done in
previous studies (Roulin, 1999; Pilastro et al., 2003; Mahr et al., 2012), my results show that
male tree swallows only increased their provisioning rates in response to the sudden change
in their mate’s ornamentation. This suggests that male tree swallows perceive the plumage
brightness of females, but do not respond as would be predicted if such a trait was attractive
to them (i.e., high provisioning of nestlings only for females with enhanced plumage
brightness).
In addition to differential allocation, male preference for ornamental traits of females
can be investigated by examining time spent associating with or the courtship and copulation
frequency directed towards females that differ in their attractiveness (Edward, 2015), as
demonstrated by previous studies that have manipulated the ornamentation of females
(Pilastro et al., 2003; Torres and Velando, 2005). For example, female rock sparrows
(Petronia petronia) manipulated to display a breast patch reduced in size received less
courtship from potential mates and they were also less likely to pair with a social mate
(Griggio et al., 2005), although it is unclear whether less attractive female rock sparrows also
paired with lower quality social mates. My results for Chapter 4 showed that the quality of
social mate paired to females did not differ by plumage brightness treatment (Table 4.1), and
females that switched nest sites, and presumably social mates, also did not pair with mates of

142

lower quality. Moreover, most females in the reduced plumage brightness treatment bred
following the manipulation (Chapter 3), which is in contrast to the lower pairing success for
less attractive female rock sparrows (Griggio et al., 2005). It is possible, however, that male
tree swallows were constrained in their choice of a social mate since I performed the
plumage manipulation after pair formation, and divorce may be a strategy that is costly
(Choudhury, 1995). Nevertheless, if plumage brightness of female tree swallows is a trait
that is attractive to males, then I predicted that males would alter their own mating strategies
in relation to the plumage brightness treatment of the female, but I observed no evidence to
support this (Table 4.2). Females in the enhanced plumage brightness treatment mated with
extra-pair males that had longer flight feathers (Figure 4.1), but given that plumage
brightness does not appear to be a signal of attractiveness to males (Chapters 2, 4), this result
may instead be due to social feedback from conspecifics (see below). Overall, the results I
presented in Chapters 2 and 4 showing that investment in parental care, social mate choice,
and mating strategies by males were not influenced by the plumage brightness treatment of
the female indicate that positive assortative mating for plumage brightness in tree swallows
is unlikely to be due to male mate choice for bright plumage displayed by females.
Ornamental traits of females that are not preferred by males sometimes function as
weaponry or signals of competitive ability or status during competition with conspecific
females (e.g., Murphy et al., 2009b; Watson and Simmons, 2010a, b). Even in humans
(Homo sapiens), studies have demonstrated that females purchase luxury items to signal high
quality when competing against conspecific females for mates (Hudders et al., 2014). In tree
swallows, competition among females for access to, and in defence of, males with a nest site
is intense and can lead to injury or death (Leffelaar and Robertson, 1985; Rosvall, 2008), so

143

it is perhaps not surprising that plumage brightness of female tree swallows is a signal
assessed by conspecific females (Chapters 2, 3). Since tree swallows show positive
assortative mating for plumage brightness (Bitton et al., 2008), and my results provide no
evidence of male mate choice as the underlying mechanism (Chapters 2, 4), assortative
mating in tree swallows likely results from females with bright plumage having greater
competitive ability and outcompeting duller conspecific females for males with a nest site. A
previous study in tree swallows reported that the dull brown plumage of second-year (SY)
females is a signal of low competitive ability (Coady and Dawson, 2013), and so positive
assortative mating occurring as a result of females with bright plumage having greater
competitive ability is in line with this previous study.
While my results provide evidence that bright plumage of female tree swallows
functions in agonistic interactions with conspecific females, I was unable to directly observe
interactions among females to confirm such a signalling role. Nevertheless, the results of
Chapters 2 and 3 are consistent with plumage brightness of female tree swallows functioning
as a status signal. Status signals that honestly reflect quality should be used by individuals to
assess the competitive ability of conspecifics so that costly agonistic interactions over
contested resources can be avoided (reviewed in Senar, 2006); however, whether signals that
reflect competitive ability are trusted or tested may depend on the value of the resource
compared to the cost of fighting (Maynard Smith and Harper, 1988; Tibbetts, 2008). The
results of Chapter 2 are consistent with signals of competitive ability being trusted since I
only observed evidence to suggest that females in the reduced plumage brightness treatment
experienced social interactions with conspecifics. For instance, females whose plumage
brightness was manipulated after the hatching of eggs likely appeared as strangers to

144

neighbouring females, but only females with reduced plumage brightness signalling low
competitive ability may have been challenged by neighbouring females, leading to lower
offspring quality and fledging success (Figure 2.4). In contrast, the results of the nest-site
retention experiment presented in Chapter 3 show that the competitive ability of females in
all plumage brightness treatments were tested (Table 3.2). Signals of quality may be tested
during the pre-breeding period because males with a nest site are a valuable resource since
they are limited (Leffelaar and Robertson, 1985) and without a male and nest site, female
tree swallows are unable to breed. Murphy et al. (2009a) similarly reported that female
streak-backed orioles (Icterus pustulatus pustulatus) responded with greater intensity to
model intruders that were manipulated to signal high competitive ability instead of trusting
status signals and relinquishing a valuable resource. Overall, my results and the results of
previous studies (e.g., Murphy et al., 2009a) show that conspecifics may challenge females
regardless of the competitive ability signalled by their plumage when in possession of a
valuable resource, such as a nest site. Although the results I report in Chapters 2 and 4 are
consistent with a status signalling function of bright plumage displayed by female tree
swallows, testing this directly would prove difficult in an aerial insectivore species that
cannot easily be moved into aviaries to conduct dominance trials for access to resources
where signals of status should be trusted (e.g., Murphy et al., 2009b). While manipulating the
plumage brightness of model intruders and simulating territory intrusions would allow me to
directly observe how conspecifics respond to females with enhanced plumage brightness, and
corroborate my results for Chapter 3, such an experiment would only demonstrate a role for
female plumage brightness to function in agonistic interactions and not necessarily confirm
status signalling (Amundsen, 2000a).

145

Ornamental traits that function as signals of attractiveness to potential mates or
competitive ability to conspecifics should honestly reflect the underlying quality of their
bearer; however, few studies have examined whether the honesty of ornaments displayed by
free-living female birds is maintained by social enforcement. In Chapter 3, the results of the
nest-site retention experiment demonstrate that nest-site intrusions from conspecific females
may provide a mechanism that enforces honest ornamentation; females that dishonestly
signalled high quality were less able to retain their nest site than females displaying lowquality signals. Moreno et al. (2013) proposed that the honesty of ornamentation displayed
by females may be enforced by social control after providing experimental evidence that
non-ornamented female pied flycatchers (Ficedula hypoleuca) manipulated to dishonestly
display a forehead patch had higher levels of blood malondialdehydes, indicative of
oxidative damage, compared to control females. My results providing support for the social
control hypothesis (Chapter 3) are consistent with Moreno et al. (2013) and suggest that
social control may maintain the honesty of ornamental traits in female birds.
Social control of honest signalling should impose costs on dishonest signallers, such
as lower reproductive success, increased energy expenditure (Kotiaho, 2001), or lower
investment in parental care (Qvarnström, 1997). The results I present in Chapters 3 and 5
showing that females in the enhanced plumage brightness treatment dishonestly signalling
high quality initiated their clutches later and produced nestlings of smaller structural size
than females in the control and reduced plumage brightness treatments (Figure 3.1a; Figure
5.1) provide evidence of social costs imposed on females dishonestly signalling high quality.
My results for females in the enhanced plumage brightness treatment are consistent with
previous studies in tree swallows reporting that more ornamented females produced low-

146

quality nestlings (Coady, 2011; Bentz and Siefferman, 2013). Coady (2011) hypothesized
that more ornamented females may experience costs associated with increased aggression
from conspecific females challenging the status signalled by their plumage ornamentation. If
female tree swallows with bright plumage engage in agonistic interactions with one another
to enforce signal honesty, as my results for nest retention suggest (Chapter 3), then my
results provide evidence of social costs experienced by more ornamented female tree
swallows. Costs associated with possession of ornamental plumage also have been observed
in other species. For example, non-ornamented female pied flycatchers manipulated to
dishonestly signal a forehead patch had increased oxidative damage compared to control
females, and the level of oxidative damage was similar to that observed in females naturally
displaying a forehead patch. These results are in contrast to female barn swallows (Hirundo
rustica erthrogaster) manipulated to signal darker ventral plumage (Vitousek et al., 2013),
and may differ from tree swallows and pied flycatchers if signals of quality are trusted by
conspecific females (see Vitousek et al., 2016). Interestingly, female tree swallows in the
reduced plumage brightness treatment dishonestly signalling low quality did not appear to
suffer social costs since they were more likely to retain their nest sites and breed (Table 3.2)
and did not differ from control females for any measure of reproductive success or nestling
quality (Chapter 3, 5). While it is possible that females dishonestly signalling low quality
may pay a cost by acquiring a low-quality social (Bitton et al., 2008) or extra-pair mate,
females in the reduced plumage brightness treatment were not paired to lower quality social
mates (Table 4.2), although this is could be because they had already acquired mates when I
manipulated their plumage brightness, and they did not mate with low-quality extra-pair
males (Table 4.4). Without a cost imposed on dishonest signallers, the signalling system

147

would be susceptible to dishonesty (e.g., Owens and Hartley, 1991); therefore, it is possible
that females in the reduced plumage brightness treatment incurred costs that I did not
measure, such as physiological costs, or were involved in fewer agonistic interactions overall
with conspecifics if they were perceived as a low competitive threat to more ornamented
females (Coady and Dawson, 2013). Disadvantages of dishonestly signalling low quality for
female tree swallows also may occur during the non-breeding season, such as on wintering
roost sites, but this explanation requires further investigation. Clearly, further studies that
manipulate the ornamental traits of female birds so that they signal higher and lower quality
prior to breeding and during the non-breeding season, followed by the observation of
agonistic interactions among conspecifics, would be valuable.
Recently, several empirical studies and reviews have highlighted the importance of
social feedback from potential mates and conspecifics about ornament quality on the
physiology or behaviour of signallers (e.g., Safran et al., 2008; Rubenstein and Hauber,
2008; Vitousek et al., 2013; Dey et al., 2014; Vitousek et al., 2014b). Results from two of my
chapters add to this literature by showing that social feedback from conspecifics may alter
the behaviour or physiology of females with manipulated plumage brightness. In Chapter 2,
lower nestling quality and fledging success for broods reared by females in the reduced
plumage brightness and enlarged brood size treatment (Figure 2.4) suggest that females in
the reduced plumage brightness treatment may have experienced social interactions with
conspecifics that were stressful, which may have elevated levels of the stress hormone
corticosterone, and when combined with the demands of rearing an enlarged brood size may
have caused females to abandon their nesting attempt. Although I have no behavioural or
hormonal data to support such a mechanism, female barn swallows with elevated levels of

148

corticosterone prior to incubation were more likely to abandon their nesting attempt
(Vitousek et al., 2014a) and experimentally elevating corticosterone of female tree swallows
resulted in lower nestling survival (Ouyang et al., 2015). In Chapter 4, females in the
enhanced plumage brightness treatment mated with extra-pair mates that were of higher
quality than their social mates (Figure 4.1). Since females in the enhanced plumage
brightness treatment may have been challenged by naturally bright, high-quality females
testing the quality of their ornamentation (Chapter 3), those females that defeated
conspecifics in agonistic interactions and retained their nest site could have perceived
themselves as strong competitors and pursued extra-pair copulations from high-quality males
on the territories of other females. Such ‘winner effects’ on the behaviour of female tree
swallows may be facilitated by changes in physiology, such as increased androgen levels
(Oliveira et al., 2009). If females with experimentally enhanced plumage brightness had
higher T levels as a result of defeating conspecifics, this would provide further support that
these females experienced social costs, and suggest a mechanism for why they bred later
(Figure 3.1a) and produced low-quality nestlings (Figure 5.1) compared to females in the
reduced and control plumage brightness treatments. For example, experimental elevation of
T in females birds delays egg laying (Chapter 6; Clotfelter et al., 2004, Veiga and Polo,
2008) and increases aggressive behaviour (Rosvall, 2013), and females that are more
aggressive provision their nestlings less and produce low-quality nestlings (Rosvall, 2011).
This mechanism is consistent with Qvarnström (1997), who reported that yearling male
collared flycatchers manipulated to display an enlarged forehead patch spent more time
competing with conspecifics and provisioned their nestlings less (i.e., ornament-parental care
trade-off) compared to control males.

149

Although the results I present in Chapter 4 are consistent with females in the
enhanced plumage brightness treatment that defeated conspecifics in agonistic interactions
having elevated levels of T that facilitated the pursuit of extra-pair copulations from highquality extra-pair mates (see above; Figure 4.1), the proportion of extra-pair offspring in the
broods of females with enhanced plumage brightness did not differ from controls (Table 4.3).
Such a result could indicate that levels of T in female tree swallows are not related to extrapair paternity, consistent with a previous study of dark-eyed juncos (Junco hyemalis) that
reported no difference in the level of extra-pair paternity in broods of T-treated and control
females (Gerlach and Ketterson, 2013), but inconsistent with the results I report for female
tree swallows treated with T (Chapter 6).
A lower proportion of extra-pair offspring observed in the broods of T-treated female
tree swallows (Figure 6.2) may differ from the results I present for females in the enhanced
plumage brightness treatment in Chapter 4 for two reasons. First, if females in the enhanced
plumage brightness treatment experienced winner effects after repeated challenges from
conspecifics, and their signal and behaviour became congruent (Vitousek et al., 2014), then
they may have been perceived as high quality and challenged less by conspecifics, resulting
in lower levels of T. For example, levels of T decreased within one week in female barn
swallows manipulated to display darker ventral plumage (Vitousek et al., 2013), and this may
be due to high-quality signals of female barn swallows being trusted (Vitousek et al., 2016).
This mechanism seems unlikely, however, as signals of quality should be tested when the
value of the resource (i.e., breeding opportunity) is high (Maynard Smith and Harper, 1988;
Tibbetts, 2008) and female tree swallows are nest-site limited (Leffelaar and Robertson,
1985). Because I was interested in examining extra-pair paternity, I was unable to capture

150

females following the plumage brightness manipulation and collect blood samples to
examine whether androgen levels decreased or increased, as it would have likely resulted in
females abandoning their nesting attempt (e.g., Veiga et al., 2004).
Second, although levels of androgens in T-treated females appear to be at the upper
limit of pre-breeding levels of T (Figure 6.2.), it is possible that the dose of T used in my
study (Chapter 6) was too high and increased aggressive behaviour (Rosvall, 2013) beyond
levels that would occur naturally in female tree swallows. As a consequence, T-treated
females may have spent most of their time involved in agonistic interactions instead of
pursuing extra-pair copulations (García-Vigón et al., 2008). Nevertheless, whether females
displaying enhanced plumage brightness had altered levels of T as a result of a social
feedback mechanisms (Vitousek et al., 2014; Chapter 4) or T was elevated experimentally
(Chapter 6), my results suggest that T does not enhance extra-pair paternity in tree swallows.
In Chapter 6, female tree swallows treated with 1,4,6-androstatrien-3, 17-dione
(ATD) and flutamide (F) in combination (ATD+F) to block the estrogenic and androgenic
actions of T produced fewer extra-pair offspring compared to controls (Figure 6.2). A
previous study in blue tits that treated females only with F, reported no difference in the level
of extra-pair paternity between F-treated females and controls (de Jong, 2013).
Consequently, the lower proportion of extra-pair paternity in the broods of females treated
with ATD+F in my study may be due to ATD blocking the conversion of T to 17β-estradiol
(E2). Blocking the effects of E2 may have lowered the receptivity or solicitation behaviour of
females (Rissman et al., 1990; Leboucher et al., 1998; Belle et al., 2005; but see Tomaszycki
et al., 2006), leading to the production of fewer extra-pair offspring, although it is important
to note that my blood samples collected late in the breeding season (June) cannot confirm

151

that E2 was lower in ATD+F treated females during egg laying (Figure 6.2). Nevertheless,
studies in other taxa, such as goats (Capra aegagrus hircus), also have demonstrated that
receptivity is lower in females treated with F (Imwalle and Katz, 2004). Therefore, blocking
the actions of both T and E2 may influence the receptivity or solicitation behaviour of female
tree swallows, and suggest that extra-pair paternity is related to sexual behaviour of females
(Chapter 6). Forstmeier (2007) also reported that extra-pair paternity in female zebra finches
(Taeniopygia guttata) was related to sexual behaviour; females with extra-pair offspring in
their brood were more likely to have copulated with a male during their first sexual
encounter in life (Forstmeier, 2007). Although my results show that sexual and extra-pair
copulation behaviour may be mediated by the estrogenic and/or androgenic actions of T, it is
possible that such a relationship exists because they are part of a larger behavioural
syndrome. For instance, sexual behaviour, such as receptivity, is part of a behavioural
syndrome associated with the expression of melanin-based plumage and so females with
darker plumage may have greater sexual receptivity towards males (Ducrest et al., 2008).
Indeed, a recent study in female yellow warblers (Setophaga petechia) proposed that higher
levels of extra-pair paternity in broods of females displaying greater melanin-based, but less
carotenoid-based plumage, may be explained by such females pairing with low-quality mates
or being mate guarded less intensively because of their less ornamented dull carotenoid
plumage, but also having a greater propensity to seek or engage in extra-pair copulations due
to the high coverage of melanin-based plumage (Grunst and Grunst, 2014). Although female
tree swallows have structurally-based plumage, and experimentally altered plumage
brightness is not associated with the rate of extra-pair paternity in the broods of females
(Table 4.3), the high levels of extra-pair paternity in this species may be due to females with

152

different personalities having different mating strategies. Future studies should explore
whether females with personality traits that are more bold, exploratory, and receptive
towards males have a greater proportion of extra-pair paternity in their brood, and whether
such traits are also indicative of whether females engage in extra-pair copulations with
neighbouring males or with males from long distances away from the breeding colony at
feeding and roosting sites. If female receptivity underlies extra-pair copulation behaviour of
female tree swallows, it also may explain why offspring in the broods of some females are
sired by up to 4 males (e.g., Whittingham and Dunn, 2014).
The function of ornaments and behaviours expressed by female birds was historically
neglected by scientists (Amundsen, 2000a), and although the number of empirical studies
investigating how the traits of females influence their ability to acquire mates and compete
against conspecifics has increased in recent decades (reviews in Clutton-Brock, 2009;
Rosvall, 2011a; Tobias et al., 2012), our understanding of how such traits influence the
mating success of free-living wild female birds is limited. Moreover, in species that are
similarly ornamented, males may not prefer females displaying elaborate ornaments
(Chapters 2, 4) and so more studies are needed to examine how the ornamentation of females
functions as signals of competitive ability or status in competition among conspecifics
(reviewed in Tarvin and Murphy, 2012). My research in female tree swallows shows that
bright plumage functions in agonistic interactions because the honesty of bright plumage
appears to be socially enforced by conspecific females (Chapter 3), and females that
dishonestly signal high quality suffer social costs (Chapters 3, 5). Few studies have examined
whether the honesty of elaborate ornaments is maintained by social enforcement in freeliving female birds, but my results and those of Moreno et al. (2013) suggest that honesty

153

may be maintained by social control. This finding highlights the importance of manipulating
the ornamental traits of females not just during the nestling provisioning period, but also
prior to breeding. Extra-pair paternity is a phenomenon that has intrigued scientists for
decades, but no previous study that I am aware of has manipulated the ornamentation of
females to examine whether this influences a female’s opportunity to engage in extra-pair
copulations or female choice of extra-pair mates. Several studies have been unable to detect
differences in phenotypic traits between a female’s social and extra-pair mate (reviewed in
Hsu et al., 2015). My results showing that females in the enhanced plumage brightness
treatment mated with extra-pair mates that had longer flight feathers than their social mate
(Figure 4.1) suggest that social feedback regarding signal quality from conspecifics may
influence female choice of or ability to pursue extra-pair mates. Additionally, few studies
have manipulated the behaviour of females by altering their exposure to T to examine extrapair paternity (see García-Vigón et al., 2008; Gerlach and Ketterson, 2013; de Jong, 2013 for
examples). My research also highlights the importance of examining the estrogenic actions
of T on extra-pair copulation behaviour. Moreover, further studies that examine how the
receptivity of female birds influence extra-pair paternity and whether extra-pair paternity is
related to personality traits as part of a behavioural syndrome would prove valuable.

154

8.1. References
Albrecht T, Schnitzer J, Kreisinger J, Exnerová A, Bryja J & Munclinger P (2007) Extrapair
paternity and the opportunity for sexual selection in long-distant migratory passerines.
Behavioral Ecology 18: 477–486.
Albrecht T, Vinkler M, Schnitzer J, Poláková R, Munclinger P & Bryja J (2009) Extra-pair
fertilizations contribute to selection on secondary male ornamentation in a socially
monogamous passerine. Journal of Evolutionary Biology 22: 2020–2030.
Amundsen T (2000a) Female ornaments: genetically correlated or sexually selected? In
Espmark, Y., Amundsen, T., Rosenqvist, G. (editors) Animal Signals: Signalling and
Signal Design in Animal Communication. Tapir Academic Press, Trondheim, pp. 133–154.
Amundsen T (2000b) Why are female birds ornamented? Trends in Ecology and Evolution
15: 149–155.
Amundsen T, Forsgren E & Hansen LTT (1997) On the function of female ornaments: male
bluethroats prefer colourful females. Proceedings of the Royal Society of London B:
Biological Sciences 264: 1579–1586.
Andersson M (1994) Sexual selection. Princeton University Press, Princeton, N.J.
Archawaranon M & Wiley, RH (1988) Control of aggression and dominance in whitethroated sparrows by testosterone and its metabolites. Hormones and Behavior 22: 497–
517.
Ardia DR (2007) Site- and sex-level differences in adult feeding behaviour and its
consequences to offspring quality in tree swallows (Tachycineta bicolor) following broodsize manipulation. Canadian Journal of Zoology 85: 847–854
Ardia DR & Clotfelter ED (2007) Individual quality and age affect responses to an energetic
constraint in a cavity-nesting bird. Behavioral Ecology 18: 259–266.
Balenger SL, Johnson LS & Masters BS (2009) Sexual selection in a socially monogamous
bird: male color predicts paternity success in the mountain bluebird, Sialia currucoides.
Behavioral Ecology and Sociobiology 63: 403–411.
Ball GF & Balthazart J (2008) Individual variation and the endocrine regulation of behaviour
and physiology in birds: a cellular/molecular perspective. Philosophical Transactions of the
Royal Society of London B: Biological Sciences 363: 1699–1710.
Ballentine B & Hill GE (2003) Female mate choice in relation to structural plumage
coloration in blue grosbeaks. The Condor 105: 593–598.

155

Balthazart J, Schumacher M & Evrard L (1990) Sex differences and steroid control of
testosterone-metabolizing enzyme activity in the quail brain. Journal of
Neuroendocrinology 2: 675–683.
Barber CJ & Robertson R (2007) Timing of copulations and the pattern of paternity in
relation to laying order in tree swallows Tachycineta bicolor. Journal of Avian Biology 38:
249–254.
Barnett CA, Thompson CF & Sakaluk SK (2012) Aggressiveness, boldness and parental
food provisioning in male house wrens (Troglodytes aedon). Ethology 118: 984–993.
Bates D, Mächler M, Bolker B & Walker S (2015) Fitting linear mixed-effects models using
lme4. Journal of Statistical Software 67: 1–49.
Belle MDC, Sharp PJ & Lea RW (2005) Aromatase inhibition abolishes courtship
behaviours in the ring dove (Streptopelia risoria) and reduces androgen and progesterone
receptors in the hypothalamus and anterior pituitary gland. Molecular and Cellular
Biochemistry 276: 193–204.
Bentz AB & Siefferman L (2013) Age-dependent relationships between coloration and
reproduction in a species exhibiting delayed plumage maturation in females. Journal of
Avian Biology 44: 080–088.
Berglund A, Bisazza A & Pilastro A (1996) Armaments and ornaments: an evolutionary
explanation of traits of dual utility. Biological Journal of the Linnean Society 58: 385–399.
Berzins LL & Dawson RD (2016) Experimentally altered plumage brightness of female tree
swallows: a test of the differential allocation hypothesis. Behaviour153: 525–550.
Birkhead TR & Møller AP (1992) Sperm competition in birds: evolutionary causes and
consequences. Academic Press, London.
Biro PA & Stamps JA (2008) Are animal personality traits linked to life-history
productivity? Trends in Ecology and Evolution 23: 361–368.
Bitton P-P & Dawson RD (2008) Age-related differences in plumage characteristics of male
tree swallows Tachycineta bicolor: hue and brightness signal different aspects of individual
quality. Journal of Avian Biology 39: 446–452.
Bitton P-P, Dawson RD & Ochs CL (2008) Plumage characteristics, reproductive investment
and assortative mating in tree swallows Tachycineta bicolor. Behavioral Ecology and
Sociobiology 62: 1543–1550.
Bitton P-P, O’Brien EL & Dawson RD (2007) Plumage brightness and age predict extrapair
fertilization success of male tree swallows, Tachycineta bicolor. Animal Behaviour 74:
1777–1784.
156

Bolour S & Braunstein G (2005) Testosterone therapy in women: a review. International
Journal of Impotence Research 17: 399–408.
Bortolotti LE, Harriman VB, Clark RG & Dawson RD (2011) Can changes in provisioning
by parent birds account for seasonally declining patterns of offspring recruitment?
Canadian Journal of Zoology 89: 921–928.
Boyle WA, Winkler DW & Guglielmo CG (2012) Rapid loss of fat but not lean mass prior to
chick provisioning supports the flight efficiency hypothesis in tree swallows. Functional
Ecology 26: 895–903.
Brown ME (1996) Assessing body condition in birds. In Nolan V Jr, Ketterson ED (editors)
Current ornithology, Volume 13 Springer US, New York, pp.67-135.
Browne WJ, Subramanian SV, Jones K & Goldstein H (2005) Variance partitioning in
multilevel logistic models that exhibit overdispersion. Journal of the Royal Statistical
Society: Series A (Statistics in Society) 168: 599–613.
Burley N (1986) Sexual selection for aesthetic traits in species with biparental care. The
American Naturalist 127: 415–445.
Burley N (1988) The differential-allocation hypothesis: an experimental test. The American
Naturalist 132: 611–628.
Canoine V & Gwinner E (2002) Seasonal differences in the hormonal control of territorial
aggression in free-living European stonechats. Hormones and Behavior 41: 1–8.
Carere C, Drent PJ, Privitera L, Koolhaas JM & Groothuis TGG (2005) Personalities in great
tits, Parus major: stability and consistency. Animal Behaviour 70: 795–805.
Choudhury S (1995) Divorce in birds: a review of the hypotheses. Animal Behaviour 50:
413–429.
Chuang-Dobbs HC, Webster MS, Holmes RT (2001) The effectiveness of mate guarding by
male black-throated blue warblers. Behavioral Ecology 12: 541–546.
Cleasby IR & Nakagawa S (2011) Neglected biological patterns in the residuals. Behavioral
Ecology and Sociobiology 65: 2361–2372.
Clotfelter ED, O’Neal DM, Gaudioso JM, Casto JM, Parker-Renga IM, Snajdr EA, Duffy
DL, Nolan V Jr. & Ketterson ED (2004) Consequences of elevating plasma testosterone in
females of a socially monogamous songbird: evidence of constraints on male evolution?
Hormones and Behavior 46: 171–178.
Clutton-Brock T (2009). Sexual selection in females. Animal Behaviour 77: 3–11.
157

Coady CD (2011) The signaling function of plumage characteristics in female tree swallows
and implications for reproductive success and life-history decisions. MSc Thesis,
University of Norther British Columbia, Canada.
Coady CD & Dawson RD (2013) Subadult plumage color of female tree swallows
(Tachycineta bicolor) reduces conspecific aggression during the breeding season. The
Wilson Journal of Ornithology 125: 348–357.
Cornwallis CK & Birkhead TR (2007) Experimental evidence that female ornamentation
increases the acquisition of sperm and signals fecundity. Proceedings of the Royal Society
of London B: Biological Sciences 274: 583–590.
Cornwallis CK & O’Connor EA (2009) Sperm: seminal fluid interactions and the adjustment
of sperm quality in relation to female attractiveness. Proceedings of the Royal Society of
London B: Biological Sciences 276: 3467–3475.
Cote J, Clobert J, Brodin T, Fogarty S & Sih A (2010) Personality-dependent dispersal:
characterization, ontogeny and consequences for spatially structured populations.
Philosophical Transactions of the Royal Society B: Biological Sciences 365: 4065–4076.
Cotton AJ, Cotton S, Small J & Pomiankowski A (2015) Male mate preference for female
eyespan and fecundity in the stalk-eyed fly, Teleopsis dalmanni. Behavioral Ecology 26:
376–385.
Cotton S, Small J & Pomiankowski A (2006) Sexual selection and condition-dependent mate
preferences. Current Biology 16: R755–R765.
Crawley, M.J., 2013. The R book, 2 edition. John Wiley & Sons, Ltd., West Sussex, United
Kingdom.
Creighton E (2001) Mate acquisition in the European blackbird and its implications for
sexual strategies. Ethology Ecology and Evolution 13: 247-260.
Cristol DA & Johnsen TS (1994) Spring arrival, aggression and testosterone in female redwinged blackbirds (Agelaius phoeniceus). The Auk 111: 210–214.
Crowe SA, Kleven O, Delmore KE, Laskemoen T, Nocera JJ, Lifjeld JT & Robertson RJ
(2009) Paternity assurance through frequent copulations in a wild passerine with intense
sperm competition. Animal Behaviour 77: 183–187
Cuervo JJ, de Lope F & Møller AP (1996) The function of long tails in female barn swallows
(Hirundo rustica ): an experimental study. Behavioral Ecology 7: 132–136.

158

Dakin R, Lendvai ÁZ, Ouyang JQ, Moore IT, & Bonier F (2016) Plumage colour is
associated with partner parental care in mutually ornamented tree swallows. Animal
Behaviour 111: 111–118.
Dale S & Slagsvold T (1995) Female contests for nest sites and mates in the pied flycatcher
Ficedula hypoleuca. Ethology 99: 209–222.
Darwin C (1871) The descent of man, and selection in relation to sex. Princeton University
Press, Princeton, NJ.
David M, Auclair Y & Cézilly F (2011) Personality predicts social dominance in female
zebra finches, Taeniopygia guttata, in a feeding context. Animal Behaviour 81: 219–224.
Dawson RD (2008) Timing of breeding and environmental factors as determinants of
reproductive performance of tree swallows. Canadian Journal of Zoology 86: 843–850.
Dawson RD & Bidwell MT (2005) Dietary calcium limits size and growth of nestling tree
swallows Tachycineta bicolor in a non-acidified landscape. Journal of Avian Biology 36:
127–134.
Dawson RD & Bortolotti GR (2003) Parental effort of American kestrels: the role of
variation in brood size. Canadian Journal of Zoology 81: 852–860.
Dawson RD, Lawrie CC & O’Brien EL (2005) The importance of microclimate variation in
determining size, growth and survival of avian offspring: experimental evidence from a
cavity nesting passerine. Oecologia 144: 499–507.
de Jong B (2013) Testosterone, a female hormone: testing the function and evolution of
testosterone in female birds. PhD dissertation, Ghent University, Belgium.
de Jong B, Lens L, Amininasab SM, van Oers K, Darras VM, Eens M, Pinxten R, Komdeur J
& Groothuis TGG (2016) Effects of experimentally sustained elevated testosterone on
incubation behaviour and reproductive success in female great tits (Parus major). General
and Comparative Endocrinology 230–231: 38–47.
Dey CJ, Dale J & Quinn JS (2014) Manipulating the appearance of a badge of status causes
changes in true badge expression. Proceedings of the Royal Society of London B:
Biological Sciences 281: 20132680.
Dingemanse NJ, Kazem AJN, Réale D & Wright J (2010) Behavioural reaction norms:
animal personality meets individual plasticity. Trends in Ecology and Evolution 25: 81–89.
Doutrelant C, Grégoire A, Grnac N, Gomez D, Lambrechts MM & Perret P (2008) Female
coloration indicates female reproductive capacity in blue tits. Journal of Evolutionary
Biology 21: 226–233.

159

Doyle A & Siefferman L (2014) Supplemental food increases ornamentation of male nestling
Eastern Bluebirds. Journal of Field Ornithology 85: 31–39.
Ducrest A-L, Keller L & Roulin A (2008) Pleiotropy in the melanocortin system, coloration
and behavioural syndromes. Trends in Ecology and Evolution 23: 502–510.
Dunn P & Whittingham L (2005) Radio-tracking of female tree swallows prior to egg-laying.
Journal of Field Ornithology 76: 259–263.
Dunn PO, Lifjeld JT & Whittingham LA (2009) Multiple paternity and offspring quality in
tree swallows. Behavioral Ecology and Sociobiology 63: 911–922.
Edward DA (2015) The description of mate choice. Behavioral Ecology 26: 301–310.
Edward DA & Chapman T (2011) The evolution and significance of male mate choice.
Trends in Ecology and Evolution 26: 647–654.
Eising CM, Eikenaar C, Schwabl H & Groothuis TGG (2001) Maternal androgens in blackheaded gull (Larus ridibundus) eggs: consequences for chick development. Proceedings of
the Royal Society of London B: Biological Sciences 268: 839–846.
Elias DO, Botero CA, Andrade MCB, Mason AC & Kasumovic MM (2010) High resource
valuation fuels “desperado” fighting tactics in female jumping spiders. Behavioral Ecology
21: 868–875.
Estep LK, Mays H Jr, Keyser AJ, Ballentine B & Hill GE (2005) Effects of breeding density
and plumage coloration on mate guarding and cuckoldry in blue grosbeaks (Passerina
caerulea). Canadian Journal of Zoology 83: 1143–1148.
Ferns PN & Hinsley SA (2004) Immaculate tits: head plumage pattern as an indicator of
quality in birds. Animal Behaviour 67: 261–272.
Field A, Miles J & Field Z (2012) Discovering statistics using R. SAGE Publications,
Thousand Oaks, California.
Fitzpatrick S, Berglund A & Rosenqvist G (1995) Ornaments or offspring: costs to
reproductive success restrict sexual selection processes. Biological Journal of the Linnean
Society 55: 251–260.
Forstmeier W (2007) Do individual females differ intrinsically in their propensity to engage
in extra-pair copulations? PLoS ONE 2: e952.
Forstmeier W, Nakagawa S, Griffith SC, & Kempenaers B (2014) Female extra-pair mating:
adaptation or genetic constraint? Trends in Ecology and Evolution 29: 456–464.

160

Fox J & Weisberg S (2011) An R companion to applied regression, second edition. SAGE
Publications, Thousand Oaks, California.
Gangestad SW, Thornhill R & Garver CE (2002) Changes in women’s sexual interests and
their partner’s mate–retention tactics across the menstrual cycle: evidence for shifting
conflicts of interest. Proceedings of the Royal Society of London B: Biological Sciences
269: 975–982.
Gangestad SW, Thornhill R & Garver-Apgar CE (2005) Women’s sexual interests across the
ovulatory cycle depend on primary partner developmental instability. Proceedings of the
Royal Society of London B: Biological Sciences 272: 2023–2027.
García-Navas V, Ferrer ES, Sanz JJ (2012) Plumage yellowness predicts foraging ability in
the blue tit Cyanistes caeruleus. Biological Journal of the Linnean Society 106: 418–429.
García-Vigón E, Cordero PJ & Veiga JP (2008) Exogenous testosterone in female spotless
starlings reduces their rate of extrapair offspring. Animal Behaviour 76: 345–353.
Garson GD (2012) Hierarchical linear modeling: guide and applications. SAGE Publications,
Thousand Oaks, California
Gerlach NM & Ketterson ED (2013) Experimental elevation of testosterone lowers fitness in
female dark-eyed juncos. Hormones and Behavior 63: 782–790.
Gladbach A, Gladbach DJ, Kempenaers B & Quillfeldt P (2010) Female-specific
colouration, carotenoids and reproductive investment in a dichromatic species, the upland
goose Chloephaga picta leucoptera. Behavioral Ecology and Sociobiology. 64: 1779–1789.
Gowaty PA (2008) Reproductive compensation. Journal of Evolutionary Biology 21: 1189–
1200.
Grafen A (1990) Biological signals as handicaps. Journal of Theoretical Biology 144: 517–
546.
Griffith SC, Owens IPF & Thuman KA (2002) Extra pair paternity in birds: a review of
interspecific variation and adaptive function. Molecular Ecology 11: 2195–2212.
Griffiths R, Double MC, Orr K & Dawson RJG (1998) A DNA test to sex most birds.
Molecular Ecology 7: 1071–1075.
Griggio M, Devigili A, Hoi H & Pilastro A (2009) Female ornamentation and directional
male mate preference in the rock sparrow. Behavioral Ecology 20: 1072–1078.
Griggio M, Valera F, Casas A & Pilastro A (2005). Males prefer ornamented females: a field
experiment of male choice in the rock sparrow. Animal Behaviour 69: 1243–1250.

161

Griggio M, Zanollo V & Hoi H (2010). Female ornamentation, parental quality, and
competitive ability in the rock sparrow. Journal of Ethology 28: 455–462.
Grunst AS & Grunst ML (2014) Multiple sexual pigments, assortative social pairing, and
genetic paternity in the yellow warbler (Setophaga petechia). Behavioral Ecology and
Sociobiology 68: 1451–1463.
Hainstock MH, Smith MC, Carr J & Shutler D (2010) Parental investment and brood value
in tree swallows, Tachycineta bicolor. Behaviour 147: 441–464.
Hall ZJ & MacDougall-Shackleton SA (2012) Influence of testosterone metabolites on songcontrol system neuroplasticity during photostimulation in adult European starlings (Sturnus
vulgaris). PLoS ONE 7: e40060.
Haller J, Fuchs E, Halász J & Makara GB (1999) Defeat is a major stressor in males while
social instability is stressful mainly in females: towards the development of a social stress
model in female rats. Brain Research Bulletin 50: 33–39.
Harriman VB, Dawson RD, Clark RG, Fairhurst GD & Bortolotti GR (2013) Effects of
ectoparasites on seasonal variation in quality of nestling tree swallows (Tachycineta
bicolor). Canadian Journal of Zoology 92: 87–96.
Harris WE & Uller T (2009) Reproductive investment when mate quality varies: differential
allocation versus reproductive compensation. Philosophical Transactions of the Royal
Society of London B: Biological Sciences 364: 1039–1048.
Hasegawa M, Watanabe M & Nakamura M (2015) Promiscuous copulation attempts and
discriminate pairing displays in male barn swallows as revealed by model presentation.
Ethology Ecology and Evolution 28: 163–174.
Hau M, Wikelski M, Soma KK & Wingfield JC (2000) Testosterone and year-round
territorial aggression in a tropical bird. General and Comparative Endocrinology 117: 20–
33.
Henderson LJ, Heidinger BJ, Evans NP & Arnold KE (2013) Ultraviolet crown coloration in
female blue tits predicts reproductive success and baseline corticosterone. Behavioral
Ecology 24: 1299–1305.
Hervé M (2016) RVAideMemoire: diverse basic statistical and graphical functions. R
package version 0.9-53. http://CRAN.R-project.org/package= RVAideMemoire.
Hidalgo-Garcia S (2006) The carotenoid-based plumage coloration of adult blue tits
Cyanistes caeruleus correlates with the health status of their brood. Ibis 148: 727–734.
Hinde CA (2006) Negotiation over offspring care?—a positive response to partnerprovisioning rate in great tits. Behavioral Ecology 17: 6–12.
162

Hinde CA & Kilner RM (2007) Negotiations within the family over the supply of parental
care. Proceedings of the Royal Society of London B: Biological Sciences 274: 53–60.
Holveck M-J & Riebel K (2009) Low-quality females prefer low-quality males when
choosing a mate. Proceedings of the Royal Society of London B: Biological Sciences 277:
153–160.
Horváthová T, Nakagawa S & Uller T (2012) Strategic female reproductive investment in
response to male attractiveness in birds. Proceedings of the Royal Society of London B:
Biological Sciences 279: 163–170.
Houtman AM & Falls JB (1994) Negative assortative mating in the white-throated sparrow,
Zonotrichia albicollis: the role of mate choice and intra-sexual competition. Animal
Behaviour 48: 377–383.
Hsu Y, Earley RL & Wolf LL (2006) Modulation of aggressive behaviour by fighting
experience: mechanisms and contest outcomes. Biological Reviews 81: 33–74.
Hsu Y-H, Schroeder J, Winney I, Burke T & Nakagawa S (2015) Are extra-pair males
different from cuckolded males? A case study and a meta-analytic examination. Molecular
Ecology 24: 1558–1571.
Hudders L, Backer CD, Fisher M & Vyncke P (2014) The rival wears prada: luxury
consumption as a female competition strategy. Evolutionary Psychology 12: 570–587.
Huhta E, Rytkönen S & Solonen T (2003) Plumage brightness of prey increases predation
risk: an among-species comparison. Ecology 84: 1793–1799.
Hussell DJ (1983) Age and plumage color in female tree swallows. Journal of Field
Ornithology 54: 312–318.
IBM Corp., 2011. IBM SPSS Statistics for Windows, Version 20.0. IBM Corp., Armonk,
NY.
Imwalle DB & Katz LS (2004) Divergent roles for estrogens and androgens in the expression
of female goat sexual behavior. Hormones and Behavior 46: 54–58.
Jacobs AC, Fair JM & Zuk M (2015) Coloration, paternity, and assortative mating in western
bluebirds. Ethology 121: 176–186.
Jawor JM, Gray N, Beall SM & Breitwisch R (2004) Multiple ornaments correlate with
aspects of condition and behaviour in female northern cardinals, Cardinalis cardinalis.
Animal Behaviour 67: 875–882.

163

Jawor JM, Linville SU, Beall SM & Breitwisch R (2003) Assortative mating by multiple
ornaments in northern cardinals (Cardinalis cardinalis). Behavioral Ecology 14: 515–520.
Johnsen A, Andersson S, Örnborg J & Lifjeld JT (1998) Ultraviolet plumage ornamentation
affects social mate choice and sperm competition in bluethroats (Aves: Luscinia s. svecica):
a field experiment. Proceedings of the Royal Society of London B: Biological Sciences 265:
1313–1318.
Johnsen A, Delhey K, Schlicht E, Peters A & Kempenaers, B (2005) Male sexual
attractiveness and parental effort in blue tits: a test of the differential allocation hypothesis.
Animal Behaviour 70: 877–888.
Johnson K (1988) Sexual selection in pinyon jays II: male choice and female-female
competition. Animal Behaviour 36: 1048–1053.
Johnstone RA, Reynolds JD & Deutsch JC (1996) Mutual mate choice and sex differences in
choosiness. Evolution 50: 1382–1391.
Jones J (2003) Tree swallows (Tachycineta bicolor): a new model organism? The Auk 120:
591–599.
Jones KA & Godin J-GJ (2010) Are fast explorers slow reactors? Linking personality type
and anti-predator behaviour. Proceedings of the Royal Society of London B: Biological
Sciences 277: 625–632.
Kabelik D, Morrison JA & Goodson JL (2010) Cryptic regulation of vasotocin neuronal
activity but not anatomy by sex steroids and social stimuli in opportunistic desert finches.
Brain, Behavior and Evolution 75: 71–84.
Kaighobadi F & Shackelford TK (2008) Female attractiveness mediates the relationship
between in-pair copulation frequency and men’s mate retention behaviors. Personality and
Individual Differences 45: 293–295.
Kalinowski ST, Taper ML & Marshall TC (2007) Revising how the computer program
CERVUS accommodates genotyping error increases success in paternity assignment.
Molecular Ecology 16: 1099–1106.
Kekäläinen J, Huuskonen H, Tuomaala M & Kortet R (2010) Both male and female sexual
ornaments reflect offspring performance in a fish. Evolution 64: 3149–3157.
Kempenaers B, Verheyen GR, den Broeck MV, Burke T, Broeckhoven CV & Dhondt A
(1992) Extra-pair paternity results from female preference for high-quality males in the
blue tit. Nature 357: 494–496.
Ketterson ED & Nolan V Jr. (1999) Adaptation, exaptation, and constraint: a hormonal
perspective. The American Naturalist 154: S4–S25.
164

Ketterson ED, Nolan V Jr & Sandell M (2005) Testosterone in females: mediator of adaptive
traits, constraint on sexual dimorphism, or both? The American Naturalist 166: S85–S98.
Kodric-Brown A & Brown JH (1984) Truth in advertising: the kinds of traits favored by
sexual selection. The American Naturalist 124: 309–323.
Kokko H & Johnstone RA (2002) Why is mutual mate choice not the norm? Operational sex
ratios, sex roles and the evolution of sexually dimorphic and monomorphic signalling.
Philosophical Transactions of the Royal Society of London B: Biological Sciences 357:
319–330.
Komdeur J (2001) Mate guarding in the Seychelles warbler is energetically costly and
adjusted to paternity risk. Proceedings of the Royal Society of London B: Biological
Sciences 268: 2103–2111.
Kotiaho JS (2001) Costs of sexual traits: a mismatch between theoretical considerations and
empirical evidence. Biological Reviews 76: 365–376.
Kraaijeveld K, Kraaijeveld-Smit FJL & Komdeur J (2007) The evolution of mutual
ornamentation. Animal Behaviour 74: 657–677.
Krebs EA, Hunte W & Green DJ (2004) Plume variation, breeding performance and extrapair copulations in the cattle egret. Behaviour 141: 479–499.
Kuznetsova A, Brockhoff PB & Christensen RHB (2016) lmerTest: Tests in Linear Mixed
Effects Models. R pack version 2.0-30. http://CRAN.R-project.org/package-lmerTest.
Lahaye SEP, Eens M, Darras VM & Pinxten R (2013) Hot or not: the effects of exogenous
testosterone on female attractiveness to male conspecifics in the budgerigar. PLoS ONE 8:
e74005.
Lakens D (2013) Calculating and reporting effect sizes to facilitate cumulative science: a
practical primer for t-tests and ANOVAs. Frontiers in Psychology 4: 863.
Lande R (1980) Sexual dimorphism, sexual selection, and adaptation in polygenic characters.
Evolution 34: 292–305.
LeBas NR (2006) Female finery is not for males. Trends in Ecology and Evolution 21: 170–
173.
Leboucher G, Béguin N, Mauget R & Kreutzer M (1998) Effects of fadrozole on sexual
displays and reproductive activity in the female canary. Physiology and Behavior 65: 233–
240.

165

Leffelaar D & Robertson RJ (1984) Do male tree swallows guard their mates? Behavioral
Ecology and Sociobiology 16: 73–79.
Leffelaar D & Robertson RJ (1985) Nest usurpation and female competition for breeding
opportunities by tree swallows. The Wilson Bulletin 97: 221–224.
Leffelaar D & Robertson RJ (1986) Equality of feeding roles and the maintenance of
monogamy in tree swallows. Behavioral Ecology and Sociobiology 18: 199–206.
Leonard ML & Horn AG (2001) Begging calls and parental feeding decisions in tree
swallows (Tachycineta bicolor). Behavioral Ecology and Sociobiology 49: 170–175.
Leonard, M.L., Horn, A.G., 1998. Need and nestmates affect begging in tree swallows.
Behavioral Ecology and Sociobiology 42: 431–436.
Leonard ML, Horn AG, Gozna A & Ramen S (2000) Brood size and begging intensity in
nestling birds. Behavioral Ecology 11: 196–201.
Limbourg T, Mateman AC & Lessells CM (2013) Opposite differential allocation by males
and females of the same species. Biology Letters 9: 20120835.
Linville SU, Breitwisch R & Schilling AJ (1998) Plumage brightness as an indicator of
parental care in northern cardinals. Animal Behaviour 55: 119–127.
Liu M, Siefferman L & Hill GE (2007) An experimental test of female choice relative to
male structural coloration in eastern bluebirds. Behavioral Ecology and Sociobiology 61:
623–630.
Logue DM, Mishra S, McCaffrey D, Ball D & Cade WH (2009) A behavioral syndrome
linking courtship behavior toward males and females predicts reproductive success from a
single mating in the hissing cockroach, Gromphadorhina portentosa. Behavioral Ecology
20: 781–788.
Loyau A, Gomez D, Moureau B, Théry M, Hart NS, Jalme MS, Bennett ATD & Sorci G
(2007) Iridescent structurally based coloration of eyespots correlates with mating success in
the peacock. Behavioral Ecology 18: 1123–1131.
Lyon BE & Montgomerie R (2012) Sexual selection is a form of social selection.
Philosophical Transactions of the Royal Society of London B: Biological Sciences 367:
2266–2273.
Mahr K, Griggio M, Granatiero M & Hoi H (2012) Female attractiveness affects paternal
investment: experimental evidence for male differential allocation in blue tits. Frontiers in
Zoology 9: 14.

166

Maia R, Eliason CM, Bitton P-P, Doucet SM & Shawkey MD (2013) pavo : an R package
for the analysis, visualization and organization of spectral data. Methods in Ecology and
Evolution 4: 906–913
Makarewich CA, Stenzler LM, Ferretti V, Winkler DW & Lovette IJ (2009) Isolation and
characterization of microsatellite markers from three species of swallows in the genus
Tachycineta: T. albilinea, T. bicolor and T. leucorrhoa. Molecular Ecology Resources 9:
631–635.
Martinez-Padilla J, Vergara P, Perez-Rodriguez L, Mougeot F, Casas F, Ludwig SC, Haines
JA, Zeineddine M & Redpath SM (2011) Condition- and parasite-dependent expression of
a male-like trait in a female bird. Biology Letters 7: 364–367.
Matessi G, Carmagnani C, Griggio M & Pilastro A (2009) Male rock sparrows differentially
allocate nest defence but not food provisioning to offspring. Behaviour 146, 209–223.
Matysioková B & Remeš V (2013) Faithful females receive more help: the extent of male
parental care during incubation in relation to extra-pair paternity in songbirds. Journal of
Evolutionary Biology 26: 155–162.
Maynard Smith J & Harper DGC (1988) The evolution of aggression: can selection generate
variability? Philosophical Transactions of the Royal Society of London B: Biological
Sciences 319: 557–570.
McCarty JP (2002) The number of visits to the nest by parents is an accurate measure of food
delivered to nestlings in tree swallows. Journal of Field Ornithology 73: 9–14.
Midamegbe A, Grégoire A, Perret P & Doutrelant C (2011) Female–female aggressiveness is
influenced by female coloration in blue tits. Animal Behaviour 82: 245–253.
Midamegbe A, Grégoire A, Staszewski V, Perret P, Lambrechts MM, Boulinier T &
Doutrelant C (2013) Female blue tits with brighter yellow chests transfer more carotenoids
to their eggs after an immune challenge. Oecologia 173: 387-397.
Møller AP (1987) Social control of deception among status signalling house sparrows Passer
domesticus. Behavioral Ecology and Sociobiology 20: 307–311.
Montgomerie R (2006) Analyzing colors. In Hill GE, McGraw KJ (editors) Bird Coloration,
Volume 1: Mechanisms and Measurements. Harvard University Press, Cambridge,
Massachusetts, pp. 90–147.
Morales J, Gordo O, Lobato E, Ippi S, Puente JM la, Tomás G, Merino S & Moreno J (2014)
Female-female competition is influenced by forehead patch expression in pied flycatcher
females. Behavioral Ecology and Sociobiology 68: 1195–1204.

167

Morales J, Velando A & Torres R (2009) Fecundity compromises attractiveness when
pigments are scarce. Behavioral Ecology 20: 117–123.
Moreno J (2015) The incidence of clutch replacements in the pied flycatcher ficedula
hypoleuca is related to nest-box availability: evidence of female-female competition?
Ardeola 62: 67–80.
Moreno J, Velando A, Ruiz-de-Castañeda R, González-Braojos S & Cantarero A (2013)
Oxidative damage in relation to a female plumage badge: evidence for signalling costs. acta
ethologica 16: 65–75.
Müller W, Boonen S, Groothuis TGG & Eens M (2010) Maternal yolk testosterone in canary
eggs: toward a better understanding of mechanisms and function. Behavioral Ecology 21:
493–500.
Muma KE & Weatherhead PJ (1989) Male traits expressed in females: direct or indirect
sexual selection? Behavioral Ecology and Sociobiology 25: 23–31.
Muriel J, Pérez-Rodríguez L, Puerta M & Gil D (2015) Diverse dose–response effects of
yolk androgens on embryo development and nestling growth in a wild passerine. Journal of
Experimental Biology 218: 2241–2249.
Murphy M, Armbrecth B, Vlamis E & Pierce Á (2000) Is reproduction by tree swallows
(Tachycineta bicolor) cost-free? The Auk 117: 902–912.
Murphy TG, Hernández-Muciño D, Osorio-Beristain M, Montgomerie R & Omland KE
(2009a) Carotenoid-based status signaling by females in the tropical streak-backed oriole.
Behavioral Ecology 20: 1000–1006.
Murphy TG, Rosenthal MF, Montgomerie R & Tarvin KA (2009b). Female American
goldfinches use carotenoid-based bill coloration to signal status. Behavioral Ecology 20:
1348–1355.
Nishiguchi MK, Doukakis P, Egan M, Kizirian D, Phillips A, Prendini L, Rosenbaum HC,
Torres E, Wyner Y, DeSalle R & Giribet G (2002) DNA isolation procedures. In DeSalle
DR, Giribet DG, Wheeler DW (editors) Techniques in Molecular Systematics and
Evolution, Methods and Tools in Biosciences and Medicine. Springer Basel AG,
Switzerland, pp. 249–287.
Nordeide JT, Kekäläinen J, Janhunen M & Kortet R (2013) Female ornaments revisited – are
they correlated with offspring quality? Journal of Animal Ecology 82: 26–38.
O’Brien EL (2006) Social and extra-pair mate choice by female tree swallows: the
importance of male and nest-site quality, and consequences for offspring performance. MSc
thesis, University of Northern British Columbia, Canada.

168

O’Brien EL & Dawson RD (2007) Context-dependent genetic benefits of extra-pair mate
choice in a socially monogamous passerine. Behavioral Ecology and Sociobiology 61: 775–
782.
Oliveira RF, Silva A & Canário AVM (2009) Why do winners keep winning? Androgen
mediation of winner but not loser effects in cichlid fish. Proceedings of the Royal Society of
London B: Biological Sciences 276: 2249–2256.
Otter K & Ratcliffe L (1996) Female initiated divorce in a monogamous songbird:
abandoning mates for males of higher quality. Proceedings of the Royal Society of London
B: Biological Sciences 263: 351–355.
Ouyang J, Lendvai Á, Dakin R, Domalik A, Fasanello V, Vassallo B, Haussmann M, Moore
I & Bonier F (2015) Weathering the storm: parental effort and experimental manipulation
of stress hormones predict brood survival. BMC Evolutionary Biology 15: 219.
Owens IPF & Hartley IR (1991) “Trojan sparrows”: Evolutionary consequences of dishonest
invasion for the badges-of-status model. The American Naturalist 138: 1187–1205.
Patrick SC, Chapman JR, Dugdale HL, Quinn JL & Sheldon BC (2012) Promiscuity,
paternity and personality in the great tit. Proceedings of the Royal Society B: Biological
Sciences 279: 1724–1730.
Peters A, Kurvers RHJM, Roberts ML & Delhey K (2011) No evidence for general
condition-dependence of structural plumage colour in blue tits: an experiment. Journal of
Evolutionary Biology 24: 976–987
Petrie M & Kempenaers B (1998) Extra-pair paternity in birds: explaining variation between
species and populations. Trends in Ecology and Evolution 13: 52–58.
Pilastro A, Griggio M & Matessi G (2003) Male rock sparrows adjust their breeding strategy
according to female ornamentation: parental or mating investment? Animal Behaviour 66:
265–271.
Pinheiro J, Bates D, DebRoy S, Sarkar D & R Core Team (2016) nlme: Linear and nonlinear
mixed effects models. R package version 3: 1-124.
Potti J & Montalvo S (1991) Male arrival and female mate choice in pied flycatchers
Ficedula hypoleuca in Central Spain. Ornis Scandinavica 22: 45–54.
Prum RO (2006). Anatomy, physics, and evolution of structural colors. In Hill GE, McGraw
KJ (editors) Bird Coloration Volume I: Mechanisms and Measurements. Harvard
University Press, Cambridge, Massachusetts, pp. 295–353.
Pryke SR (2007) Fiery red heads: female dominance among head color morphs in the
Gouldian finch. Behavioral Ecology 18: 621–627.
169

Quinn JL & Cresswell W (2005) Personality, anti-predation behaviour and behavioural
plasticity in the chaffinch Fringilla coelebs. Behaviour 142: 1377–1402.
Qvarnström A (1997) Experimentally increased badge size increases male competition and
reduces male parental care in the collared flycatcher. Proceedings of the Royal Society of
London B: Biological Sciences 264: 1225–1231.
R Development Core Team (2015) R: A language and environment for statistical computing.
R Foundation for Statistical Computing, Vienna, Austria.
Ratikainen II & Kokko H (2010) Differential allocation and compensation: who deserves the
silver spoon? Behavioral Ecology 21: 195–200.
Réale D, Dingemanse NJ, Kazem AJN & Wright J (2010a) Evolutionary and ecological
approaches to the study of personality. Philosophical Transactions of the Royal Society of
London B: Biological Sciences 365: 3937–3946.
Réale D, Garant D, Humphries MM, Bergeron P, Careau V & Montiglio P-O (2010b)
Personality and the emergence of the pace-of-life syndrome concept at the population level.
Philosophical Transactions of the Royal Society B: Biological Sciences 365: 4051–4063
Réale D, Reader SM, Sol D, McDougall PT & Dingemanse NJ (2007) Integrating animal
temperament within ecology and evolution. Biological Reviews 82: 291–318.
Remeš V & Matysioková B (2013) More ornamented females produce higher-quality
offspring in a socially monogamous bird: an experimental study in the great tit (Parus
major). Frontiers in Zoology 10: 14.
Rissman EF, Clendenon AL & Krohmer RW (1990) Role of androgens in the regulation of
sexual behavior in the female musk shrew. Neuroendocrinology 51: 468–473.
Riters LV & Alger SJ (2011) Hormonal regulation of avian courtship and mating behaviors.
In Norris DQ, Lopez KH (editors) Hormones and Reproduction of Vertebrates. Academic
Press, Burlington, Massachusetts, pp. 153–180.
Robertson RJ, Gibbs HL & Stutchbury BJ (1986) Spitefulness, altruism, and the cost of
aggression: evidence against superterritoriality in tree swallows. The Condor 88: 104–105.
Rohwer S (1985) Dyed birds achieve higher social status than controls in Harris’ sparrows.
Animal Behaviour 33: 1325–1331.
Rohwer S (1977) Status signaling in Harris sparrows: some experiments in deception.
Behaviour 61: 107–129.

170

Rohwer S & Rohwer FC (1978) Status signalling in Harris sparrows: experimental
deceptions achieved. Animal Behaviour 26: 1012–1022
Rose AP (2009) Temporal and individual variation in offspring provisioning by tree
swallows: a new method of automated nest attendance monitoring. PLoS ONE 4: e4111.
Rosivall B, Szöllősi E, Hasselquist D & Török J (2009) Effects of extrapair paternity and sex
on nestling growth and condition in the collared flycatcher, Ficedula albicollis. Animal
Behaviour 77: 611–617.
Rosvall KA (2008) Sexual selection on aggressiveness in females: evidence from an
experimental test with tree swallows. Animal Behaviour 75: 1603–1610.
Rosvall KA (2011a) By any name, female–female competition yields differential mating
success. Behavioral Ecology 22: 1144–1146.
Rosvall KA (2011b) Cost of female intrasexual aggression in terms of offspring quality: a
cross-fostering study. Ethology 117: 332–344.
Rosvall KA (2013) Life history trade-offs and behavioral sensitivity to testosterone: an
experimental test when female aggression and maternal care co-occur. PLoS ONE 8:
e54120.
Roulin A (1999) Nonrandom pairing by male barn owls (Tyto alba) with respect to a female
plumage trait. Behavioral Ecology 10: 688–695.
Roulin A, Jungi TW, Pfister H & Dijkstra C (2000) Female barn owls (Tyto alba) advertise
good genes. Proceedings of the Royal Society of London B: Biological Sciences 267: 937–
941.
Rowe KMC & Weatherhead PJ (2011). Assortative mating in relation to plumage traits
shared by male and female American robins. The Condor 113: 881–889.
Rubenstein DR & Hauber ME (2008) Dynamic feedback between phenotype and physiology
in sexually selected traits. Trends in Ecology and Evolution 23: 655–658.
Ryder TB, Fleischer RC, Shriver WG & Marra PP (2012) The ecological–evolutionary
interplay: density-dependent sexual selection in a migratory songbird. Ecology and
Evolution 2: 976–987.
Safran RJ, Adelman JS, McGraw KJ & Hau M (2008) Sexual signal exaggeration affects
physiological state in male barn swallows. Current Biology 18: R461–R462.
Safran RJ, Neuman CR, McGraw KJ & Lovette IJ (2005) Dynamic paternity allocation as a
function of male plumage color in barn swallows. Science 309: 2210–2212.

171

Sandell MI (2007) Exogenous testosterone increases female aggression in the European
starling (Sturnus vulgaris). Behavioral Ecology and Sociobiology 62: 255–262.
Schielzeth H (2010) Simple means to improve the interpretability of regression coefficients.
Methods in Ecology and Evolution 1: 103–113.
Schwabl H, Lindsay WR, Barron DG & Webster MS (2014) Endocrine correlates of mate
choice and promiscuity in females of a socially monogamous avian mating system with
alternative male reproductive phenotypes. Current Zoology 60: 804–815.
Senar JC (2006). Color displays as intrasexual signals of aggression and dominance. In Hill
GE, McGraw KJ (editors) Bird Coloration Vol II: Function and Evolution. Harvard
University Press, Cambridge, Massachusetts, pp. 87–136.
Seutin G, White BN & Boag PT (1991) Preservation of avian blood and tissue samples for
DNA analyses. Canadian Journal of Zoology 69: 82–90.
Sheldon BC (2000) Differential allocation: tests, mechanisms and implications. Trends in
Ecology and Evolution 15: 397–402.
Shutler D, Clark RG, Fehr C & Diamond AW (2006) Time and recruitment costs as
currencies in manipulation studies on the costs of reproduction. Ecology 87: 2938–2946.
Sih A, Bell A & Johnson JC (2004a) Behavioral syndromes: an ecological and evolutionary
overview. Trends in Ecology and Evolution 19: 372–378.
Sih A, Bell AM, Johnson JC & Ziemba RE (2004b) Behavioral syndromes: an integrative
overview. The Quarterly Review of Biology 79: 241–277.
Silva N, Avilés JM, Danchin E & Parejo D (2008) Informative content of multiple plumagecoloured traits in female and male European Rollers. Behavioral Ecology and Sociobiology
62: 1969–1979.
Smith ER, Damassa DA & Davidson JM (1977) Hormone administration: peripheral and
intracranial implants. In Myers RD (editor) Methods in Psychobiology Vol III: Advanced
Laboratory Techniques in Neuropsychology and Nurobiology. Academic Press, New York,
New York, pp. 259–279.
Smith SM (1988) Extra-pair copulations in black-capped chickadees: the role of the female.
Behaviour 107: 15–23.
Sockman KW & Schwabl H (2000) Yolk androgens reduce offspring survival. Proceedings
of the Royal Society of London B: Biological Sciences 267: 1451–1456.

172

Soma KK, Sullivan K & Wingfield J (1999) Combined aromatase inhibitor and antiandrogen
treatment decreases territorial aggression in a wild songbird during the nonbreeding season.
General and comparative endocrinology 115: 442–453.
Stapleton MK, Kleven O, Lifjeld JT & Robertson RJ (2007) Female tree swallows
(Tachycineta bicolor) increase offspring heterozygosity through extrapair mating.
Behavioral Ecology and Sociobiology 61: 1725–1733.
Stapleton MK & Robertson RJ (2006) Female tree swallow home-range movements during
their fertile period as revealed by radio-tracking. The Wilson Journal of Ornithology 118:
502–507.
Stuart-Fox DM & Goode JL (2014) Female ornamentation influences male courtship
investment in a lizard. Frontiers in Ecology and Evolution 2:1–9.
Stutchbury BJ & Robertson RJ (1985) Floating populations of female tree swallows. The Auk
3: 651–654.
Stutchbury BJ & Robertson RJ (1987) Signaling subordinate and female status: two
hypotheses for the adaptive significance of subadult plumage in female tree swallows. The
Auk 104: 717–723.
Stutchbury BJ & Robertson RJ (1988) Within-season and age-related patterns of
reproductive performance in female tree swallows (Tachycineta bicolor). Canadian Journal
of Zoology 66: 827–834.
Swaddle JP, Witter MS, Cuthill IC, Budden A & McCowen P (1996). Plumage condition
affects flight performance in common starlings: implications for developmental
homeostasis, abrasion and moult. Journal of Avian Biology 127: 03–111.
Tarvin KA & Murphy TG (2012) It isn’t always sexy when both are bright and shiny:
considering alternatives to sexual selection in elaborate monomorphic species. Ibis 154:
439–443.
Thomas K & Shutler D (2001) Ectoparasites, nestling growth, parental feeding rates, and
begging intensity of tree swallows. Canadian Journal of Zoology 79: 346–353.
Tibbetts EA (2008) Resource value and the context dependence of receiver behaviour.
Proceedings of the Royal Society of London B: Biological Sciences 275: 2201–2206.
Tibbetts EA (2014) The evolution of honest communication: integrating social and
physiological costs of ornamentation. Integrative and Comparative Biology 54: 578–590.
Tibbetts EA & Izzo A (2010) Social punishment of dishonest signalers caused by mismatch
between signal and behavior. Current Biology 20: 1637–1640.

173

Tibbetts EA & Lindsay R (2008) Visual signals of status and rival assessment in Polistes
dominulus paper wasps. Biology Letters 4: 237–239.
Tilgar V, Ots I & Mänd R (2004) Bone alkaline phosphatase as a sensitive indicator of
skeletal development in birds: a study of the great tit nestlings. Physiological and
Biochemical Zoology 77: 530–535.
Tobias JA, Montgomerie R & Lyon BE (2012) The evolution of female ornaments and
weaponry: social selection, sexual selection and ecological competition. Philosophical
Transactions of the Royal Society of London B: Biological Sciences 367: 2274–2293.
Tomaszycki ML, Banerjee SB & Adkins-Regan E (2006) The role of sex steroids in
courtship, pairing and pairing behaviors in the socially monogamous zebra finch.
Hormones and Behavior 50: 141–147.
Torres R & Velando A (2005) Male preference for female foot colour in the socially
monogamous blue-footed booby, Sula nebouxii. Animal Behaviour 69: 59–65.
Trivers RL (1972) Parental investment and sexual selection. In Campbell BG (editor) Sexual
selection and the descent of man, Aldine de Gruyter, New York, New York, pp. 136–179.
van Oers K, Drent PJ, Dingemanse NJ & Kempenaers B (2008) Personality is associated
with extrapair paternity in great tits, Parus major. Animal Behaviour 76: 555–563.
Varian-Ramos CW & Webster MS (2012) Extrapair copulations reduce inbreeding for
female red-backed fairy-wrens, Malurus melanocephalus. Animal Behaviour 83: 857–864.
Veiga JP, Polo V (2008) Fitness consequences of increased testosterone levels in female
spotless starlings. The American Naturalist 172: 42-53.
Veiga JP, Viñuela J, Cordero PJ, Aparicio JM & Polo V (2004) Experimentally increased
testosterone affects social rank and primary sex ratio in the spotless starling. Hormones and
Behavior 46: 47–53.
Venier LA, Dunn PO, Lifjeld JT & Robertson RJ (1993) Behavioural patterns of extra-pair
copulation in tree swallows. Animal Behaviour 45: 412–415.
Verhoeven KJF, Simonsen KL & McIntyre LM (2005). Implementing false discovery rate
control: increasing your power. Oikos 108: 643–647.
Vitousek MN, Jenkins BR & Safran RJ (2014a) Stress and success: Individual differences in
the glucocorticoid stress response predict behavior and reproductive success under high
predation risk. Hormones and Behavior 66: 812–819.
Vitousek MN, Stewart RA & Safran RJ (2013) Female plumage colour influences seasonal
oxidative damage and testosterone profiles in a songbird. Biology Letters 9: 20130539.
174

Vitousek MN, Tomášek O, Albrecht T, Wilkins MR & Safran RJ (2016) Signal traits and
oxidative stress: a comparative study across populations with divergent signals. Frontiers in
Ecology and Evolution 4: 56.
Vitousek MN, Zonana DM & Safran RJ (2014b) An integrative view of the signaling
phenotype: Dynamic links between signals, physiology, behavior and social context.
Current Zoology 60: 739–754.
Washburn BE, Millspaugh JJ, Morris DL, Schulz J &, Faaborg J (2007) Using a
commercially available enzyme immunoassay to quantify testosterone in avian plasma. The
Condor 109: 181–186.
Watson NL & Simmons LW (2010a). Mate choice in the dung beetle Onthophagus
sagittarius: are female horns ornaments? Behavioral Ecology 21: 424–430.
Watson NL & Simmons LW (2010b) Reproductive competition promotes the evolution of
female weaponry. Proceedings of the Royal Society of London B: Biological Sciences 277:
2035–2040.
Weatherhead PJ & Robertson RJ (1979) Offspring quality and the polygyny threshold: “the
sexy son hypothesis.” The American Naturalist 113: 201–208.
Weiss SL (2006) Female-specific color is a signal of quality in the striped plateau lizard
(Sceloporus virgatus). Behavioral Ecology 17: 726–732.
West-Eberhard MJ (1983) Sexual selection, social competition, and speciation. The
Quarterly Review of Biology 58: 155–183.
While GM, Sinn DL & Wapstra E (2009) Female aggression predicts mode of paternity
acquisition in a social lizard. Proceedings of the Royal Society B: Biological Sciences 276:
2021–2029.
Whittingham LA & Dunn PO (2000) Offspring sex ratios in tree swallows: females in better
condition produce more sons. Molecular Ecology 9: 1123–1129.
Whittingham LA & Dunn PO (2010) Fitness benefits of polyandry for experienced females.
Molecular Ecology 19: 2328–2335.
Whittingham LA & Dunn PO (2014) Extra-pair mating and sexual selection on male traits
across populations. The Wilson Journal of Ornithology 126: 9–18.
Whittingham LA & Dunn PO (2016) Experimental evidence that brighter males sire more
extra-pair young in tree swallows. Molecular Ecology 25: 3706–3715.

175

Whittingham LA, Dunn PO & Robertson RJ (1993) Confidence of paternity and male
parental care: an experimental study in tree swallows. Animal Behaviour 46: 139–147.
Whittingham LA, Dunn PO & Stapleton MK (2006) Repeatability of extra-pair mating in
tree swallows. Molecular Ecology 15: 841–849.
Whittingham LA & Schwabl H (2002) Maternal testosterone in tree swallow eggs varies
with female aggression. Animal Behaviour 63: 63–67.
Williams TD, Kitaysky AS & Vézina F (2004) Individual variation in plasma estradiol-17β
and androgen levels during egg formation in the European starling Sturnus vulgaris:
implications for regulation of yolk steroids. General and Comparative Endocrinology 136:
346–352.
Wilson LC & Swaddle JP (2013) Manipulating the perceived opportunity to cheat: an
experimental test of the active roles of male and female zebra finches in mate guarding
behavior. Behavioral Ecology and Sociobiology 67: 1077–1087.
Winkler DW & Allen PE (1996) The seasonal decline in tree swallow clutch size:
physiological constraint or strategic adjustment? Ecology 77: 922–932.
Winkler DW, Hallinger KK, Ardia DR, Robertson RJ, Stutchbury BJ & Cohen RR (2011).
Tree Swallow (Tachycineta bicolor). In Poole A (editor) The Birds of North America
Online. Ithaca: Cornell Lab of Ornithology; retrieved from the Birds of North American
online: http://bna.birds.cornell.edu/bna/species/011
Winkler DW, Ringelman KM, Dunn PO, Whittingham L, Hussell DJT, Clark RG, Dawson
RD, Johnson LS, Rose A, Austin SH, Robinson WD, Lombardo MP, Thorpe PA, Shutler
D, Robertson RJ, Stager M, Leonard M, Horn AG, Dickinson J, Ferretti V, Massoni V,
Bulit F, Reboreda JC, Liljesthröm M, Quiroga M, Rakhimberdiev E & Ardia DR (2014)
Latitudinal variation in clutch size–lay date regressions in Tachycineta swallows: effects of
food supply or demography? Ecography 37: 670–678.
Winkler DW, Wrege PH, Allen PE, Kast TL, Senesac P, Wasson MF, Llambías PE, Ferretti
V, Sullivan PJ (2004) Breeding dispersal and philopatry in the tree swallow. The Condor
106: 768–776.
Wolf WL, Casto JM, Nolan V Jr & Ketterson ED (2004) Female ornamentation and male
mate choice in dark-eyed juncos. Animal Behaviour 67: 93–102.
Zahavi A (1975) Mate selection—a selection for a handicap. Journal of theoretical Biology
53: 205–214.
Zysling DA, Greives TJ, Breuner CW, Casto JM, Demas GE & Ketterson ED (2006)
Behavioral and physiological responses to experimentally elevated testosterone in female
dark-eyed juncos (Junco hyemalis carolinensis). Hormones and Behavior 50: 200–207.
176