.1
.1193.

a

1.. 3...‘ y.

.. .Muwandw..n

.sz Hun:

 

 

 

at i».

 

flung. uni. $3..

 

.533

 

 

This is to certify that the
dissertation entitled

SEXUAL SELECTION AND REPRODUCTIVE INVESTMENT
IN HOUSE WRENS (T ROGLODYTES AEDON)

presented by

Natalie Suzanne Dubois

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Zoology

 

 

film W

MajorvProfessbr’s Signature
‘-l [‘1 8/04

Date

 

MSU is an Affinnative Action/Equal Opportunity Institution

 

 

. LIBRARY
M'Chigan State
University

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DatoDuepss-pJS

77* *7 . .7 —— — ——» ——

 

SEXUAL SELECTION AND REPRODUCTIVE INVESTMENT IN HOUSE WRENS
(TROGLODYTES AEDON)

By

Natalie Suzanne Dubois

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Zoology
Program in Ecology, Evolutionary Biology & Behavior
WK. Kellogg Biological Station

2004

ABSTRACT
SEXUAL SELECTION AND REPRODUCTIVE INVESTMENT IN HOUSE WRENS
(TROGLODYTES AEDON)
By

Natalie Suzanne Dubois

An individual’s fitness depends on both the quality and quantity of offspring it is
able to produce over its lifetime. For iteroparous species, the optimal allocation of
investment to a particular breeding attempt should be mediated by its reproductive value
relative to the costs to and benefits from future reproduction. Reproductive value might
be influenced by mate or territory quality, environmental variables such as food
abundance, or opportunities to secure additional matings. Within a breeding attempt,
individuals can maximize reproductive success by choosing a high quality mate,
investing more in offspring with high reproductive values, and taking advantage of
additional mating opportunities if the benefits of doing so exceed costs. Individual trade-
offs between investment in current and future reproduction may result in conflicts of
interest between parents if, for example, a parent cannot simultaneously provision young
and attract mates.

I investigated mate choice and reproductive investment decisions in the house
wren (T roglodytes aedon), a facultatively polygynous cavity nester. In contrast to what is
observed in a congener (winter wren, T. troglodytes), I found that the presence of empty
(or ‘cock’) nests had no effect on mate choice or maternal investment in reproduction, but

the availability of cavities within territories did affect male and female reproductive

investment. Females mated to males with surplus nest boxes added to their territories
laid larger clutches at early season nests and produced more males per clutch than
females mated to males with single cavities in their territories. These results suggest that
female house wrens might use cavity availability to assess male quality. Males with
surplus nest boxes added to their territories sang more than males with Single nest boxes,
but this increased song rate had no effect on paternal provisioning efforts. Both males
with surplus cavities and males with single cavities in their territories were less likely to
provision at early season nests than at late season nests. Costs of provisioning at early
season nests might be high for males, regardless of mate attraction opportunity, if males

that provision at early season nests are less likely to rear second broods.

ACKNOWLEDGMENTS

This endeavor would not have been possible without the help of many people. I
would like to thank the members of my graduate committee, Dr. Jeffrey Conner, Dr. Kay
Holekamp, and Dr. Catherine Lindell, for their advice and guidance throughout my
graduate career. My graduate advisor, Dr. Thomas Getty, has been a supportive mentor
and a friend. His creativity, integrity, and passion for science have both inspired and
sustained my academic pursuits. It has been my pleasure to have been his student.

I could never have completed my fieldwork without the help of several wonderful
REUS and field assistants: Robin Siewers, Laura Vogel, Tara Poloskey, Chris Noud,
Theresa Grattan, Jenny Glazer, and Erica Zontek. I thank them for their commitment to
their work and for making long days in the field fun. My lab-mates, Rob Olendorf, Katie
Wharton, and Lindsey Walters have all contributed to this project. I thank my friends,
Dale Kennedy and Doug White, professors from my undergraduate days at Albion
College, for introducing me to field research and for their continued support as I pursued
my graduate degree.

The KBS community is a very special place, and I thank all of those who have
made it my home, through dinner parties, game nights, and summer cookouts. Tara
Darcy-Hall and Stefanie Whitmire have Shown me the truest meaning of friendship. The
members of my academic cohort, Eric Thobaben, Nate Dom, and Jeremy Wojdak, as
well as Laura Broughton and Stacy (Epp) Umstatter, have been an important part of my
life at MSU. Above all, my family has always provided constant support. My parents

have been my loving field assistants, from the first installation of the nest boxes to the

iv

final take-down. And it is with love that I thank my husband, Michael Homish, who has

always believed in me, even during those moments when I was not so sure myself.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... viii
LIST OF FIGURES ......................................................................................................... x
CHAPTER 1
INTRODUCTION ........................................................................................................... 1
Literature Cited ................................................................................................................ 5
CHAPTER 2
EMPTY NESTS DO NOT AFFECT FEMALE MATE CHOICE OR MATERNAL
INVESTMENT IN HOUSE WRENS ............................................................................. 8
Abstract ................................................................................................................ 9
Introduction .......................................................................................................... 9
Methods ................................................................................................................ 11
Experimental manipulation ...................................................................... 12
Behavior and reproductive success .......................................................... 14
Statistical analyses ................................................................................... 15
Results .................................................................................................................. 1 7
Male response to artificial nest supplementation ..................................... 17
Settlement patterns ................................................................................... 17
Reproductive investment ......................................................................... 17
Discussion ............................................................................................................ 1 8
Acknowledgments ................................................................................................ 20
Literature cited ..................................................................................................... 22
CHAPTER 3
DOES CAVITY AVAILABILITY AFFECT TRADE-OFFS BETWEEN CURRENT
AND FUTURE REPRODUCTION IN MALE HOUSE WRENS? ................................ 27
Abstract ................................................................................................................ 27
Introduction .......................................................................................................... 28
Methods ................................................................................................................ 3 1
Experimental manipulation ...................................................................... 32
Provisioning effort and mate attraction .................................................... 34
Statistical analyses ................................................................................... 35
Results .................................................................................................................. 37
Parental investment .................................................................................. 37
Mate attraction ......................................................................................... 38
Discussion ............................................................................................................ 39
Acknowledgments ................................................................................................ 43
Literature cited ..................................................................................................... 44

vi

CHAPTER 4
MATERNAL INVESTMENT BY FEMALE HOUSE WRENS DEPENDS ON
TERRITORY CHARACTERISTICS RELATED TO MALE MATE ATTRACTION

OPPORTUNITY .............................................................................................................. 56
Abstract ................................................................................................................ 56
Introduction .......................................................................................................... 57
Methods ................................................................................................................ 59

Cavity manipulation ................................................................................. 59
Statistical analyses ................................................................................... 60
Results .................................................................................................................. 61
Female settlement .................................................................................... 61
Clutch Size ................................................................................................ 62
Site fidelity ............................................................................................... 64
Discussion ............................................................................................................ 64
Timing of female settlement .................................................................... 67
Site fidelity between broods and late season nests .................................. 69
Acknowledgments ................................................................................................ 71
Literature cited ..................................................................................................... 72

CHAPTER 5

NESTLING SEX RATIOS VARY WITH A MALE TERRITORY CHARACTERISTIC

IN HOUSE WRENS ........................................................................................................ 80
Abstract ................................................................................................................ 80
Introduction .......................................................................................................... 80
Methods ................................................................................................................ 82

Cavity manipulation ................................................................................. 82
Offspring sex identification ..................................................................... 83
Statistical analysis .................................................................................... 84
Results .................................................................................................................. 86
Offspring sex ratio ................................................................................... 86
Sex differences in nestling size ................................................................ 87
Maternal provisioning effort .................................................................... 87
Discussion ............................................................................................................ 88
Acknowledgments ................................................................................................ 93
Literature cited ..................................................................................................... 94

vii

LIST OF TABLES

Table 2.1. Observed effect sizes (i 95% CI) and power to detect a medium effect size
(f = 0.25, Cohen 1988) for dependent variables for which we failed to find a Significant
treatment effect or treatment X season interaction between house wrens with empty nests
added to their territories and house wrens with egg-containing nests added to their
territories. The upper limit of the observed effect size has been back-transformed into
original units where appropriate. ..................................................................................... 24

Table 3.1. Proportion of males provisioning nestlings in territories containing surplus
and single cavities (post—laying treatment) at early and late season nests. ...................... 46

Table 3.2. Repeated measures analysis of paternal provisioning rates per nestling on
brood days 4, 8, and 12 for males in territories containing surplus nest boxes and single
nest boxes during the nestling period (post-laying treatment) at early and late season
nests. Only males that provisioned for the entire nestling period were included. .......... 47

Table 3.3. Number of male house wrens retaining the same territory for first and second
broods, switching territories between broods, and losing territories in relation to paternal
provisioning at first brood nests. Only males with successful first brood nests were

included. ........................................................................................................................... 48

Table 3.4. Repeated measures analysis of male song rate during the incubation and
nestling periods for males in territories containing surplus nest boxes and single nest
boxes (post-laying treatment) at early season nests. ........................................................ 49

Table 3.5. Repeated measures analysis for the length of the introduction sequence of
male songs during the incubation and nestling periods for male house wrens in territories
containing surplus nest boxes and Single nest boxes during the pre—pairing treatment
(PPT) and post-laying treatment (PLT) periods at early season nests. ............................ 50

Table 4.1. Morphological measurements for females settling in territories containing
Single and surplus cavities at the beginning of the breeding season. ............................... 75

Table 4.2. Numbers of females in “young” and “old” age classes that settled in territories
containing surplus cavities versus single cavities. There were no significant differences
among treatments (X2 = 0.14, df= 1, P > 0.6). ................................................................ 76

Table 4.3. Numbers of females rearing successful first broods that retained the same
territory versus left the territory for the second brood. Females leaving the territory
attempted a second brood elsewhere on the study site or left the study site. There were no
significant differences among treatments (X2 = 2.3, df = 3, P > 0.5). ............................. 77

Table 5.1. Size of male and female offspring on brood day 12. ..................................... 96

viii

Table 5.2. Repeated measures analysis of total maternal provisioning rates on brood days
4, 8, and 12 for females that settled in territories containing surplus nest boxes and single
nest boxes (pre-pairing treatment) at early season nests .................................................. 97

Table 5.3. Repeated measures analysis of maternal provisioning rates per nestling on

brood days 4, 8, and 12 for females that settled in territories containing surplus nest boxes
and single nest boxes (pre-pairing treatment) at early season nests. ............................... 98

ix

LIST OF FIGURES

Figure 2.1. Average provisioning rates (both parents combined) at early and late season
nests for house wren pairs with empty nests (unfilled squares) and pairs with egg-
containing nests (solid squares) added to their territories. The treatment X season
interaction was significant (P < 0.02). During the early season, pairs with empty nests
added to their territories provisioned at higher rates than pairs with egg-containing nests
added to their territories (P < 0.01). There were no differences in provisioning rates for
late season nests. Back-transformed means i: 95% confidence intervals are shown ...... 25

Figure 2.2. Average (a) male and (b) female provisioning rates per nestling at early and
late season nests for house wren pairs with empty nests (unfilled circles) and pairs with
egg-containing nests (solid circles) added to their territories. The treatment X season
interaction was Si gnificant (P < 0.05) for males but not for females. For males, pairwise
comparisons of treatments within a brood were not significantly different. Seasonal
means were significantly different for males (P < 0.01) and females (P < 0.05). Back-
transformed means i 95% confidence intervals are shown. ............................................ 26

Figure 3.1. Frequency distribution of nest lining dates in 2002. In both years of the
study, 6 June was used as the division between early and late season nests. The first nest
was lined on 30 April. ...................................................................................................... 51

Figure 3.2. Spectrogram of a male house wren song. The softer introduction sequence
consists of complex elements and is followed by a louder terminal trill consisting of
repeated elements. Amplitude is indicated in the box above the spectrogram. .............. 52

Figure 3.3. Average paternal provisioning rates per nestling during 30-minute
observation periods for males in territories containing surplus nest boxes (solid symbols)
and single nest boxes (open symbols) during the nestling period. Early season nests
indicated by circles, late season nests indicated by diamonds. Only males that
provisioned for the entire nestling period were included. Nests with incomplete data
were excluded, therefore samples sizes (indicated) are the same across brood days.
There was a significant brood day effect (P < 0.01 ). ....................................................... 53

Figure 3.4. Song rate index for males in territories containing surplus cavities (solid
symbols) and single cavities (open symbols) during the incubation and nestling periods at
early season nests. The treatment main effect was Significant (P < 0.02). The interaction
was marginally significant (P = 0.05). Nests with incomplete data were excluded.
Sample sizes are indicated. .............................................................................................. 54

Figure 3.5. Length of the introduction sequence for songs of male house wrens in
territories containing surplus (solid symbols) and single (open symbols) cavities in their
territories during the incubation and nestling periods. Pre-pairing treatment is indicated
by circles (surplus cavities) and squares (single cavities). Means were not significantly
different. Nests with incomplete data were excluded, therefore samples sizes (indicated)
are the same for the incubation and nestling periods ....................................................... 55

Figure 4.1. Number of days male house wrens with surplus and Single cavities in their
territories remained unmated (data log transformed and standardized within years) at the
start of the breeding season. Means were not significantly different (P > 0.9). Sample
sizes are indicated. ........................................................................................................... 78

Figure 4.2. The effect of cavity availability on standardized clutch size at early season
nests. Means were significantly different (P < 0.02). Sample sizes are indicated ......... 79

Figure 5.1. Average offspring sex ratios (proportion males) of clutches produced by
females mated to males with surplus and Single cavities in their territories. Means were
significantly different at the P < 0.02 level. Sample sizes are indicated above the
symbols. ........................................................................................................................... 99

Figure 5.2. Average total maternal nestling provisioning rates during 30-minute
observation periods for females that settled in territories containing surplus nest boxes
(solid symbols) and single nest boxes (open symbols) through the laying period. Only
the time effect was significant (P < 0.0001). Sample sizes are indicated ....................... 100

Figure 5.3. Average maternal provisioning rates per nestling during 30-minute
observation periods for females that settled in territories containing surplus nest boxes
(solid symbols) and single nest boxes (open symbols) through the laying period. Only the
time effect was significant (P < 0.0001). Sample sizes are indicated ............................. 101

xi

CHAPTER 1

INTRODUCTION

Reproductive success depends on both the quantity and quality of offspring an
individual is able to produce over his or her lifetime. For iteroparous Species, investment
in any given reproductive attempt should depend on the costs to future reproduction as
well as the benefits of the current reproductive attempt (Steams 1989, HOrak 2003). An
important factor affecting reproductive success is the quality of the mate or breeding
situation. Individuals can maximize the value of a current reproductive attempt by
choosing a high quality mate. The characters involved in mate choice may Si gnal fitness
benefits such as increased provisioning effort or increased viability of young (reviewed in
Kirkpatrick and Ryan 1991, Kokko et al. 2003). Additionally, individuals can further
increase the value of a reproductive attempt by investing more in reproduction when mate
quality or the benefits provided by a mate are high (Burley 1986, reviewed by Sheldon
2000). Such trade-offs between current and future reproduction are a central component
of life history theory (Steams 1989). Of particular interest are Situations characterized by
sexual conflict in which the fitness interests of male and female parents diverge (Davies
1989)

Darwin (1871) identified sexual selection as the result of differences in
reproductive success arising from competition over mates, either through intrasexual
contests, generally males competing for access to females, or through intersexual mate
choice. Sexual selection theory offered an evolutionary explanation for the prevalence of
conspicuous secondary sexual traits such as the widowbird’s tail (Andersson 1982), the

elaborate constructions of bowerbirds (Borgia 1985), and the complex song repertoire of

the great reed warbler (Catchpole 1980). Individuals might be able to use these signals to
assess mate quality. For example, plumage coloration signals male capacity for parental
care in house finches (Hill 1991), tail length signals offspring viability in barn swallows
(Moller 1994), and song rate signals nest site quality in blackcaps (Hoi-Leitner et al.
1995)

Intersexual variance in reproductive success has traditionally been viewed as most
apparent in polygamous systems, where some individuals mate with several individuals
while others mate with few or none. Polygyny threshold models predicted that females
settling with already-mated males are compensated for the costs of polygyny by acquiring
a high quality breeding Situation (e.g., Orians 1969). Emlen and Oring (1977) proposed
that the environmental potential for polygamy would depend on ecological factors such
as the spatial distribution of resources. Polygyny would be favored when males could
monopolize critical resources required by mates. Advances in our understanding of
genetic versus social mating systems (e. g., Westneat et al. 1990) have revealed female
control over “resources” such as genetic parentage, and emphasize the importance of both
sexes in shaping genetic and social breeding systems. For example, males might face
trade-offs between mate- guarding and extra-pair copulations (Hasselquist and Bensch
1991), or, if females use male parental care at first brood nests in future reproductive
investment decisions, males might use provisioning effort to guard parentage in
subsequent broods (F reeman-Gallant 1996).

Preferred mates may Sire more offspring through extra-pair copulations in
monogamous as well as polygamous systems (sperm competition reviewed in Birkhead

and Moller 1992). Even in the absence of sexual selection through extra-pair matings,

differences in quality among mates could lead to variation in the quality and quantity of
offspring produced. Preferred mates might produce more or higher quality offspring if
they are able to attract higher quality mates (Parker 1983, Moller 1991), or if their mates
invest more in reproduction (Burley 1986). For example, females mated to high quality
males might increase provisioning rates (de Lope and Moller 1993), decrease rates of
extra-pair copulation (Burley et al. 1994), lay larger eggs (Cunningham and Russell
2000), or increase levels of testosterone in eggs (Gil et al. 1999).

Conflicts of interest between the sexes over investment in mating opportunities
and parental care may shape mating systems if the reproductive payoffs of monogarny
and polygamy differ for males and females (Davies 1989). Gowaty (1996) suggested that
facultative expression of male parental care might occur when the ecological opportunity
for males to seek additional matings (either through extra—pair copulations or by
attracting secondary females) depends on the ability of females to rear offspring alone
within the environmental constraints of the given breeding situation. That is, males with
mates of high social quality (able to rear young without assistance) would suffer less
from a trade-off between mating effort and parental care. The relative importance of
paternal care might depend on the ability of females to compensate for reduced care
(Dunn and Harmon 1992, Sejberg et al. 2000), on environmental conditions (Bart and
Tornes 1989, Kuitunen et al. 1996), or on the changing energetic demands of young
(Johnson et al. 1992).

In my dissertation, I examine factors affecting reproductive investment decisions
in house wrens (T roglodytes aedon). In this Species, males are facultatively polygynous

and are limited in their opportunity to attract additional females by the number of cavities

(or nest sites) they can defend in their territory. Since males provide parental care by
provisioning offspring, this sets up a potential conflict of interest between males and
females in terms of investment in parental care. For males, investment in a given
reproductive attempt may limit opportunities to attract additional mates, whereas for
females, mate attraction effort by males (or the acquisition of a secondary female) may
limit male contribution to parental care. Thus there are potential trade-offs in terms of
the costs and benefits associated with allocation to current reproduction versus firture
reproduction for both males and females. Females make decisions regarding how much
to invest in each reproductive attempt within the context of male quality and/or benefits
received from the male. Males make similar decisions, but perhaps on a shorter time
scale when opportunities to attract additional mates are high. Specifically, I have taken
an experimental approach and addressed mate choice and parental investment decisions
in several contexts. In Chapter 2, I examine whether empty nests, the result of egg
removal from nests by house wrens, affect female mate choice and subsequent
reproductive investment decisions by females. In Chapters 3 — 5, I focus on the effects of
cavity availability on male and female investment in reproduction. I examine (a) whether
the opportunity for males to acquire additional mates (i.e., the presence of surplus cavities
in a male’s territory) affects a trade-off between mate attraction and paternal provisioning
and (b) whether cavity availability affects mate choice and reproductive investment

decisions by females.

LITERATURE CITED

Andersson, M. 1982. Female choice selects for extreme tail length in a widowbird.
Nature 299: 818-820.

Bart, J. and A. Tomes. 1989. Importance of monogamous male birds in determining
reproductive success. Evidence for house wrens and a review of male-removal
studies. Behavioral Ecology and Sociobiology 24: 109—116.

Birkhead, TR. and AP. Moller. 1992. Sperm Competition in Birds: Evolutionary Causes
and Consequences. Academic Press, London.

Borgia, G. 1985. Bower quality, number of decorations and mating success of male satin
bowerbirds (Ptilonorhynchus violaceus): an experimental analysis. Animal
Behaviour 33: 266-271.

Burley, N. 1986. Sexual selection for aesthetic traits in species with biparental care.
American Naturalist 127: 415-445.

Burley, N.T., D.A. Enstrom, and L. Chitwood. 1994. Extra-pair relations in zebra finches:

differential male success results from female tactics. Animal Behaviour 48: 1031-
1041.

Catchpole, CK. 1980. Sexual selection and the evolution of complex songs among
European warblers of the genus Acrocephalus. Behaviour 74: 149-166.

Cunningham, E.J.A. and AF. Russell. 2000. Egg investment is influenced by male
attractiveness in the mallard. Nature 404: 74-77.

Darwin, C. 1871. The Descent of Man, and Selection in Relation to Sex. Murray,
London.

Davies, NE. 1989. Sexual conflict and the polygamy threshold. Animal Behaviour 38:
226-234.

de Lope, F. and AP. Moller. 1993. Female reproductive effort depends on the degree of
ornamentation of their mates. Evolution 47: 1152-1160.

Dunn, PO. and SJ. Harmon. 1992. Effects of food abundance and male parental care on
reproductive success and monogamy in Tree Swallows. Auk 109: 488-499.

Emlen, ST. and L.W. Oring. 1977. Ecology, sexual selection, and the evolution of
mating systems. Science 197: 215-223.

Freeman-Gallant, CR. 1996. DNA fingerprinting reveals female preference for male
parental care in Savannah Sparrows. Proceedings of the Royal Society of London
Series B 263: 157-160.

Gil, D., J. Graves, N. Hazon, and A. Wells. 1999. Male attractiveness and differential
testosterone investment in zebra finch eggs. Science 286: 126-128.

Gowaty, PA. 1996. Field studies of parental care in birds: new data focus questions on
variation among females. Pp. 47 7-53 1 in Advances in the Study of Behavior, vol.
25 (Rosenblatt, J .S. and CT. Snowdon, eds.). Academic Press, San Diego.

Hasselquist, D. and S. Bensch. 1991. Trade-off between mate guarding and mate
attraction in the polygynous great reed warbler. Behavioral Ecology and
Sociobiology 28: 187-193.

Hill, GE. 1991. Plumage coloration is a sexually selected indicator of male quality.
Nature 350: 337-339.

Hoi-Leitner, M., H. Nechtelberger, and H. Hoi. 1995. Song rate as a signal for nest site
quality in blackcaps (Sylvia atricapilla). Behavioral Ecology and Sociobiology
37: 399-405.

HOrak, P. 2003. When to pay the cost of reproduction? A brood Size manipulation
experiment in great tits (Parus major). Behavioral Ecology and Sociobiology 54:
1 05-1 12.

Johnson, L.S., M.S. Merkle, and L.H. Kermott. 1992. Experimental evidence for
importance of male parental care in monogamous House Wrens. Auk 109: 662-
664.

Kirkpatrick, M. and M.J. Ryan. 1991. The evolution of mating preferences and the
paradox of the lek. Nature 350: 33-38.

Kokko, H., R. Brooks, MD. J ennions, and J. Morley. 2003. The evolution of mate choice
and mating biases. Proceedings of the Royal Society of London Series B 270:
653-664.

Kuitunen, M., A. J antti, J. Suhonen, and T. Aho. 1996. Food availability and the male's
role in parental care in double-brooded Treecreepers Certhia familiaris. Ibis 138:
638-643.

Moller, AP. 1991. Preferred males acquire mates of higher phenotypic quality.
Proceedings of the Royal Society of London Series B 245: 179-182.

Moller, AP. 1994. Male ornament size as a reliable cue to enhanced offspring viability in
the barn swallow. Proceedings of the National Academy of Sciences of the USA
91: 6929-6932.

Orians, G.H. 1969. On the evolution of mating systems in birds and mammals. American
Naturalist 103: 589-603.

Parker, GA. 1983. Mate quality and mating decisions. Pp. 141-166 in Mate Choice
(Bateson, P., ed.). Cambridge University Press, Cambridge.

Sejberg, D., S. Bensch, and D. Hasselquist. 2000. Nestling provisioning in polygynous
great reed warblers (Acrocephalus arundinaceus): do males bring larger prey to

compensate for fewer nest visits? Behavioral Ecology and Sociobiology 47: 213-
219.

Sheldon, B.C. 2000. Differential allocation: tests, mechanisms and implications. Trends
in Ecology and Evolution 15: 397-402.

Steams, SC. 1989. Trade-offs in life history evolution. Functional Ecology 3: 259-268.

Westneat, D.F., P.W. Sherman, and ML. Morton. 1990. The ecology and evolution of
extra-pair copulations. Current Ornithology 7: 331-369.

CHAPTER 2

Dubois, NS. and T. Getty. 2003. Empty nests do not affect female mate choice or
maternal investment in House Wrens. Condor 105: 382-387.

EMPTY NESTS DO NOT AFFECT FEMALE MATE CHOICE OR MATERNAL
INVESTMENT IN HOUSE WRENS

with Thomas Getty

ABSTRACT

House Wrens (T roglodytes aedon) remove eggs from the nests of other birds,
including conspecifics and heterospecifics and both cup and cavity nests. Egg removal
by males before females arrive increases the number of empty nests in and around a male
House Wren’s territory, and females might use this trait in mate choice. We manipulated
the presence of empty nests in House Wren territories prior to female settlement by
adding artificial nests with or without plastic eggs. We used the timing of female
settlement as an index of mate preference. Our manipulation had no effect on the timing
of female settlement or on variables related to maternal investment such as clutch Size,
egg volume, or provisioning effort. Differential investment in offspring was based on the
timing of a reproductive attempt, which was unrelated to the experimental manipulation.

Key words: egg removal, empty nests, House Wren, mate choice, nest destruction,
parental investment, Troglodytes aedon.

INTRODUCTION

The propensity for egg removal and nest destruction by House Wrens
(T roglodytes aedon) has been reported for over a century (Hill 1869, Creaser 1925,
Kendeigh 1941, Kennedy and White 1996). Both sexes puncture and remove eggs from
the nests of conspecifics and heterospecifics, but the behavior is suppressed in males after
pairing and in females after the onset of egg laying (Belles-Isles and Picman 1986). Egg

removal might accrue negligible costs, in which case the trait could be maintained as a

vestige in the absence of current benefits. But there are several plausible hypotheses
suggesting benefits of egg removal by House Wrens. House Wrens might destroy nests
to reduce competition for resources such as nest cavities, food, or mates (Belles-Isles and
Picman 1986) or to reduce predation risk by changing the ratio of active to inactive nests
(Finch 1990).

There have been few critical tests of these hypotheses, due in part to the
difficulties inherent in designing large-scale manipulations of the relevant factors.
Several authors have documented cavity takeovers following nest destruction by male
House Wrens (Quinn and Holroyd 1989, Pribil and Picman 1991), providing some
support for a nest-cavity—limitation hypothesis, but these observations fail to explain the
destruction of noncavity nests. Only the food competition and predation hypotheses can
explain attack of both conspecific and heterospecific nests. Regardless of the ultimate
function of egg removal, the major hypotheses focus on potential benefits accrued by a
nesting pair in terms of increased reproductive success, and they also suggest direct
benefits that choosy females might gain if mate choice were based on nest destruction (or
its result, the presence of empty nests). Here, we take a different approach from previous
studies, which have focused on the ecological function of egg removal, and ask instead
whether females use empty nests as a mate-choice cue.

Female House Wrens might prefer males with empty nests in their territories if
the presence of empty nests signals male or territory quality. We might expect females
mated to preferred males to invest more in reproduction. This could occur if preferred
males pair with higher quality females, or if females mated to preferred males have

access to high quality territories. Additionally, females might allocate resources to

10

reproduction in response to the perceived quality of their mates (Burley 1986). Examples
of differential allocation by females mated to high quality males include increased
provisioning effort (de Lope and Moller 1993, but see Witte 1995), increased clutch size
(Petrie and Williams 1993), and larger eggs (Cunningham and Russell 2000).

Male House Wrens control the number of empty nests in their territories not only
by egg removal but also by placing sticks in unoccupied cavities (Kendeigh 1941). The
evidence that female House Wrens attend to nest site characteristics is mixed. Secondary
females in a polygynous population of House Wrens preferred males with front-entrance
nest boxes over males with roof-entrance boxes (Johnson and Searcy 1993), but there is
no evidence that the Size of a male’s nest influences female choice or reproductive
success (Alworth 1996). Neither of these studies addressed the effect of male
manipulation of multiple nest sites on female mate choice. Here we present the results of
a nest-site manipulation study in a primarily monogamous, double-brooded population of
House Wrens. We test the hypothesis that female House Wrens prefer males with empty
nests over males with egg-containing nests in their territories and examine subsequent
patterns of reproductive investment.

METHODS

House Wrens are small (10-12 g), migratory, secondary cavity nesters (Johnson
1998). Male House Wrens in our population arrive in late April, and the first females
arrive several days to a week later. Upon arrival, males establish territories and begin
constructing nests by placing a few to hundreds of small sticks in available cavities.
Females inspect cavities before selecting a nest site, and once the nest cup is completed,

line the nest with soft material (Kendeigh 1941 ). House Wrens in our population

11

frequently double-brood, with early season clutches generally initiated in May and late
season clutches initiated mid June through July. We considered nest lining the date of
nest initiation.

We conducted this study April through August of 1999 and 2000 at the Kellogg
Biological Station’s Lux Arbor Reserve (42°29’N, 85°28’W) in Barry County, Michigan.
Lux Arbor Reserve is a 529-ha Site consisting of fragmented habitats including hardwood
and softwood forests, open fields, wetlands, and agricultural areas. We installed nest
boxes in clusters (hereafter referred to as Sites) composed of four nest boxes spaced 5 m
apart. Boxes were placed along forested edges wherever possible. We did not control
the orientation of the box entrance, but this characteristic appears to have little effect on
nest site preferences in House Wrens (Lumsden 1986). In March 1999, 36 sites were
established approximately 100 m apart. Four Sites were added in April 2000. All nest
boxes were installed on metal poles and placed approximately 1.2 m above the ground.
To minimize losses due to predation, the poles were greased at the beginning of the
breeding season in 2000, and a 0.75-m conical predator guard was installed on the poles
of all active nests.

Experimental manipulation

We checked nest boxes at least every other day during male and female settlement
periods and at least every 4 days during the nesting cycle (usually more often). We set up
the experiment within 1-3 days of finding a nest box with at least 20 sticks in it and
always prior to female nest lining. We designated the nest box with the most sticks in it
as the primary nest. This nest was not manipulated. Experimental manipulations were

set up in the remaining (dummy) nest boxes. In one half of sites, empty artificial nests

12

(made of dried grass) were placed in dummy nest boxes to Simulate nests from which
eggs had been removed. In the remainder of sites, artificial nests containing a clutch of
artificial plastic eggs were placed in dummy nest boxes to simulate the presence of active
nests. Artificial nests were placed in nest boxes at a height approximating that of a fully
lined House Wren nest (approximately 3 cm below the entrance hole). Dried Spanish
moss (T illandsia usneoides) was used to disguise the materials supporting the artificial
nest in the nest box. The artificial eggs were slightly smaller than wren eggs (15 mm in
length) and cream in color with brown speckling, but House Wrens are not choosy about
the characteristics of the eggs they remove (Belles-Isles and Picman 1986). We wired
artificial eggs into nests to prevent removal by House Wrens. We removed sticks placed
over the artificial nests by House Wrens, except in cases where lining material was also
present.

We paired territories by male settlement date (which we assumed was the date we
first observed sticks in a nest box at a site) and randomly assigned treatments within
those pairs. In 1999, only a subset of early season nests was manipulated (n = 19). In
2000, both early (n = 36) and late season nests (n = 18) were manipulated. Early nests
were defined as nests lined before 11 June 2000 (the median lining date). Nests lined
after that date were classified as late broods regardless of whether one or both members
of a pair were known to have fledged an earlier brood. Dummy nest boxes were cleaned
out several days before nestlings fledged at the primary box in order to allow males to
begin building late season nests. Primary nest boxes were cleaned out after nestlings

fledged.

l3

an

We randomly reassigned treatments to males building late season nests. Nests
initiated after a successful first brood were almost always built in former dummy boxes.
We included each House Wren pair only once in analysis by excluding late season nests
of double-brooded pairs from the data set (except for provisioning data from one pair for
which we only recorded data during the second brood). Late season nests of novel pairs
(either as a result of mate switching or the appearance of new breeders to our study site)
were included in our data set.

Behavior and reproductive success

In order to assess whether the experimental treatment affected male behavior prior
to female settlement, we conducted 20-min focal observations 1-2 days after installation
of the experimental treatments on a subset of males (n = 23) during the early nesting
period in 2000. We recorded at 15-sec intervals whether a male was singing, and
whether a male was in or on a primary or dummy nest box. We initiated an observation
only when we located a male on the territory either by Si ght or sound. We calculated
male time budgets from these data.

The number of days a male remained unmated was calculated as the lag between
the date we first observed sticks in a box at a site and the date of female nest lining
(which we used as an index of pairing date). In cases where we missed the first Sign of
lining and found a lined nest containing an egg on a subsequent visit, we used the day
between our nest checks as the nest-lining date.

We used clutch size, egg volume, nestling mass and number, and parental
provisioning rates as our measures of reproductive investment. We measured egg length

and breadth to the nearest 0.1 mm using dial calipers and estimated egg volume (in mL)

14

using the formula, length x breadthz >< 0.000491 (formula in Johnson 1998). In 2000,
nestlings were counted and weighed with a spring scale on brood day 12 (brood day 0 =
hatch). Nests that failed partway through the nesting period were included in analyses
based on data collected before the failure.

We observed provisioning trips for 30-min periods in the morning (between 06:00
and 11:00 EST) on brood days 4, 8, and 12. Any parental visit to the entrance of the nest
box was considered a provisioning trip, regardless of whether we actually observed food.
We captured adults prior to provisioning observations using a mist net placed in the
territory or a box trap, and marked each adult with an aluminum US. Fish and Wildlife
Service leg band and a unique combination of three color bands. Trapping and banding
was conducted under the appropriate federal, state, and university permits. Provisioning
observations were conducted during fair weather using 10x binoculars mounted on a
tripod, taking care to minimize disturbance. We excluded from analysis any observations
during which neither parent resumed provisioning.

Statistical analyses

Statistical analyses were performed using SAS v8.1 (SAS Institute Inc. 1999).
We used Wilcoxon rank sum tests to compare premating male time budgets. All other
analyses were performed using parametric tests, transforming data when necessary to
improve normality. Settlement data were log transformed and provisioning data (counts)
were square-root transformed. We included only complete clutches in our analysis of
clutch size and egg volume. We excluded renests from our analyses for two reasons: (1)

females might invest differently in egg number and size after a nest failure, and (2)

15

 

IfllCSlS

expert

sures
1999

ri'hicl

strum
also 1
same
mOdr
mea:
Opllt'

[real

We c
also
(Tat
Unit:

Valu

renests were often built in dummy boxes over the artificial nests, thus disrupting the
experimental manipulation.

The effects of the experimental manipulation on settlement and reproductive
success data were analyzed using ANOVAS with season as a blocking factor (3 levels:
1999 early brood, 2000 early brood, 2000 late brood, except for nestling mass models for
which we had no 1999 measurements). Provisioning rates were analyzed using the
MIXED procedure in SAS (repeated measures with a compound symmetry covariance
structure). In instances where we report only data on provisioning rates per nestling, we
also conducted these analyses per nest and results from both analyses were essentially the
same. In addition to treatment, the following factors were included in provisioning rate
models: season (2 levels: 2000 early brood, 2000 late brood), observation day (repeated
measure, 3 levels: brood days 4, 8, and 12), and all interactions. We used the SLICE
option in SAS to test simple effects of treatment within each level of brood for significant
treatment X season interactions.

Estimates for power were based on Cohen (1988) and calculated using PASS
(Hintze 2001). In cases where we failed to demonstrate a significant nest treatment effect
we calculated 95% confidence intervals on the effect size (Steidl and Thomas 2001) and
also back-transformed the upper limit into the measured variable where appropriate
(Table 2.1). For comparison, means for transformed data are presented in untransformed
units with 95% confidence limits (based on analysis of the transformed data). All other

values are presented as means :E SE.

16

RESULTS
Male response to artificial nest supplementation

The addition of empty nests or egg-containing nests to male House Wren
territories had minor effects on male behavioral time budgets prior to mating. We did not
observe significant differences in singing activity (empty vs. egg-containing nests:
54.8 i 2.1, vs. 45.7 i 2.3 observations out ofa possible 81; S = 113.0, P > 0.2) nor in
time spent in or on the primary box (0.6 i 0.2 vs. 3.3 :1: 0.5 observations; S = 89.0,
P > 0.5), although the power to detect a medium effect Size was low for both tests
(power = 0.2). Males with empty nests added to their territories spent significantly more
time in or on dummy boxes than males with egg-containing nests added to their territories
(3.3 :t: 0.5 vs. 0.3 :1: 0.04 of81 observations respectively; S = 124.5, P < 0.03).

Settlement patterns

Of first-brood pairs included in the experiment, more than 50% of males had
established territories by the time the first nest was lined in the study population (1999:
13 of 19 males, 2000: 21 of 36 males). Therefore, we assumed that females could choose
among males. The experimental manipulation did not affect the number of days a male
remained unmated (back-transformed means [95% CI] for empty vs. egg-containing
nests: 7.4 [5.7-9.6] vs. 9.4 [7.5-11.7] days; F1 ‘63 = 1.9, P > 0.1).

Reproductive investment

The experimental manipulation had no effect on our measures of reproductive
success: clutch size (empty vs. egg-containing nests: 6.2 i 0.2 vs. 6.2 :l: 0.1 eggs;
F155 < 0.01, P > 0.9), average egg volume per nest (empty nests = egg-containing nests =

1.3 i 0.02 mL; F150: 0.1, P > 0.7), number of nestlings in the nest on brood day 12

17

(empty vs. egg-containing nests: 5.6 :1: 0.3 vs. 5.5 i 0.2 nestlings; F136 = 0.06, P > 0.8), or
average nestling mass per nest on brood day 12 (empty vs. egg-containing nests:
10.0 i 0.2 vs. 10.2 :1: 0.1 g; F138 = 0.5, P > 0.5).

Combined provisioning rates. There was no significant treatment effect on the
total number of provisioning visits per 30 min made by both parents, but there was a
significant treatment X season interaction (treatment, F1,“ = 0.2, P > 0.6, treatment X
season, F 1 ,3}?- 6.7, P < 0.02). Comparisons of treatments within a brood indicate that
pairs with empty nests added to their territories made more provisioning visits than pairs
with egg-containing nests added to their territories during early broods (P < 0.01) but not
late broods (P > 0.2, Figure 2.1). Provisioning rates per nestling showed similar patterns
(treatment, F132 = 0.01, P > 0.9, treatment X season, F132 = 4.0, P < 0.06).

Individual provisioning rates. There was no significant treatment effect on
provisioning rates per nestling by males, but there was a significant treatment X season
interaction as well as a strong season effect on male provisioning rates per nestling
(Figure 2.2a; treatment, F132 = 0.1, P > 0.7, season, F132 = 10.5, P < 0.01, treatment X
season, F132 = 4.5, P < 0.05). The experimental manipulation had no effect on
provisioning rates per nestling by females, but there were significant seasonal effects on
female provisioning (Figure 2.2b; treatment, F133 = 0.1, P > 0.7, season, F133 = 4.4,

P < 0.05, treatment X season, F133 = 0.9, P > 0.3). Seasonal differences in female
provisioning rates were opposite those observed for males (Figure 2.2).
DISCUSSION
We found no evidence that the addition of empty or egg-containing nests to male

House Wren territories affected female mate choice. Male Winter Wrens (T roglodytes

18

troglodytes) build empty nests, which females use as a mate choice cue (Evans and Burn
1996). However, multiple nest building in Marsh Wrens (Cistothorus palustris) does not
affect female mate choice but may affect reproductive success (Leonard and Picman
1987). Depredated Marsh Wren nests had fewer empty nests around them than
successful nests (Leonard and Picman 1987). Before we installed predator guards,
predators in a section of our study area destroyed nests regardless of their contents or the
contents of surrounding dummy nests (some baited with quail eggs). This pattern
suggests that dummy nests in our study area have no effect on nest predation from the
most common predator, raccoons (Procyon lotor).

We found no differences in our measures of reproductive investment for female
House Wrens mated to males with egg-containing nests in their territories and those
mated to males with empty nests. Although the power of these tests was relatively low,
we would have been able to detect an effect Size roughly half the observed decrease in
clutch Size between broods. In addition, we observed extremely small effect sizes for our
measures of reproductive success. Combined with the lack of a consistent pattern in the
direction of the treatment effect on these variables, we consider it unlikely that the
presence of empty nests has a biologically significant effect on maternal investment.
We found a significant treatment X season interaction on combined parental provisioning
rates to early nests. In early broods, pairs with egg-containing nests added to their
territories made fewer nestling provisioning visits than pairs with empty nests added to
their territories. Even though individual parental provisioning rates tended to Show the
same pattern, the interaction was Significant only for males, and comparisons of

treatments within a brood were not significant. Time budgets from the premating period

19

provide no evidence that males with egg-containing nests added to their territories wasted
a lot of time trying to remove eggs (in fact, they spent less time in dummy boxes than
males with empty nests added to their territories). Although it seems unlikely, we cannot
rule out the possibility that these males (or pairs) perceived increased competition or
predation risks as a result of the manipulation and allocated time to other activities such
as territorial defense at the expense of provisioning effort. In any case, the differences in
provisioning rates observed among treatments had no effect on nestling mass on brood
day 12.

We did observe seasonal differences in parental investment and reproductive
success. We documented increased provisioning rates by male House Wrens late in the
season when the potential to secure additional mates was likely reduced. These results
are consistent with the idea that male House Wrens demonstrate adjust their investment
in reproduction based on the possibility of attracting additional mates. We are currently
testing this hypothesis.

ACKNOWLEDGMENTS

NSD was supported under a National Science Foundation Graduate Research
Fellowship, the MSU Distinguished Fellowship Program, the Research Training Group
(NSF) at the Kellogg Biological Station, the Departments of Zoology and Ecology,
Evolutionary Biology and Behavior, a Lauff Research Award, and a Wallace Scholarship
Award. Purchase and construction of nest boxes was made possible by M. Klug and
B. Stovall. Field assistance was provided by R. Siewers, L. Vogel, S. Dubois, E. Dubois,

M. Hornish, J. Johnson, and several volunteers. We thank J. Conner, K. Holekamp,

20

C. Lindell, and two anonymous reviewers for comments on previous versions of this

manuscript. This is Kellogg Biological Station contribution number 991.

21

 

.1111 01

ECHO

Burl:

Colic-

Crea

Curr:

del

E\ at

Pine

Hill
Hill

10111

Rem

LEOn‘

LITERATURE CITED

Alworth, T. 1996. An experimental test of the function of sticks in the nests of House
Wrens. Condor 98: 841-844.

Belles-Isles, J -C. and J. Picman. 1986. House Wren nest-destroying behavior. Condor 88:
190-193.

Burley, N. 1986. Sexual selection for aesthetic traits in species with biparental care.
American Naturalist 127: 415-445.

Cohen, J. 1988. Statistical power analysis for the behavioral sciences, second edition.
Lawrence Erlbaum Associates, Hillsdale, NJ.

Creaser, CW. 1925. The egg-destroying activity of the House Wren in relation to
territorial control. Bird-Lore 27: 163-167.

Cunningham, E.J.A. and AF. Russell. 2000. Egg investment is influenced by male
attractiveness in the mallard. Nature 404: 74—77.

de Lope, F. and AP. Moller. Female reproductive effort depends on the degree of
ornamentation of their mates. Evolution 47: 1152-1160.

Evans, MR. and J .L. Burn. 1996. An experimental analysis of mate choice in the wren: a
monomorphic, polygynous passerine. Behavioral Ecology 7: 101-108.

Finch, D.M. 1990. Effects of predation and competitor interference on nesting success of
House Wrens and Tree Swallows. Condor 92: 674-687.

Hill, MS. 1869. The house wren. American Naturalist 3: 49.
Hintze, J. 2001. NCSS and PASS. Number Cruncher Statistical Systems, Kaysville, UT.

Johnson, LS. 1998. House Wren (T roglodytes aedon) in The Birds of North America,
No. 380 (Poole, A. and F. Gill, eds.). The Birds of North America, Inc.,
Philadelphia, PA.

Johnson, LS. and WA. Searcy. 1993. Nest site quality, female mate choice, and
polygyny in the house wren T roglodytes aedon. Ethology 95: 265-277.

Kendeigh, SC. 1941. Territorial and mating behavior of the house wren. Illinois
Biological Monographs 18: 1-120.

Kennedy, ED. and D.W. White. 1996. Interference competition from House Wrens as a
factor in the decline of Bewick's Wrens. Conservation Biology 10: 281-284.

Leonard, M.L., and J. Picman. 1987. The adaptive Significance of multiple nest building
by male marsh wrens. Animal Behaviour 35: 271-277.

22

 

lun

Pelt

Pril

Qu

SA

\\"

Lumsden, H.G. 1986. Choice of nest boxes by Tree Swallows, T achycineta bicolor,
House Wrens, T roglodytes aedon, Eastern Bluebirds, Sialia sialis, and European
Starlings, Sturnus vulgaris. Canadian F ield-Naturalist 100: 343-349.

Petrie, M. and A. Williams. 1993. Peahens lay more eggs for peacocks with larger trains.
Proceedings of the Royal Society of London Series B 251: 127-131.

Pribil, S. and J. Picman. 1991. Why House Wrens destroy clutches of other birds: a
support for the nest site competition hypothesis. Condor 93: 184-185.

Quinn, MS. and G.L. Holroyd. 1989. Nestling and egg destruction by House Wrens.
Condor 91 : 206-207.

SAS Institute Inc. 1999. SAS/STAT User’s Guide. Version 8. SAS Institute Inc., Cary,
NC.

Steidl, RI. and L. Thomas. 2001. Power Analysis and Experimental Design Pp. 14-36 in
Design and Analysis of Ecological Experiments, second edition (Scheiner, SM.
and J. Gurevitch, eds.). Oxford University Press, Oxford.

Witte, K. 1995. The differential-allocation hypothesis: does the evidence support it?
Evolution 49: 1289-1290.

23

.8323 .8 3:5 83:23 So: Soc :03? m<m 5 SEE 85 was: BSA—men 0.53 SE meweoom a
onoMmSE Hooroszcm a

comboMmemb m3 a

 

3:: on you weaamo: cog 3E3

 

 

. So ammo a sod S 3 mm .2 ea maaemsea 283
was -- 93 a 2.0 no no mm ._ E as: mass: swam;
3o -- a2 a 25 we 3 am .2 .282 3:82
38 -- So a :3 3 3 R .a 33 0822, wmm
was -- 93 a :3 as Sov mm .2 9me 9% €56
m2 5 .35 a 2.0 3 3 8 .2 cusses. as: man

530a as: 5%: mmmfiwww A m a 03%;

 

gained? 82:» SE: Efiwto BE eoEoMmewbiomn zoos mm: 36 Bebe
33890 2: mo :8: bag: 05. .motoHEB :05 8 323 $8: memfifieoomwo £39 323 3:0: 28 motowEB :2: 3 323 88:
.395 HEB EBB 8:0: c8252— :ocoEBE Emma x 60585 .8 Soto 308305 Edemfiwmm a can 8 3:8 25 553 do.“ 83mg.»

“523% do.“ Ammo“ eonoo .36 n .3 ONE “8ch 8388 a 88% 8 630a 98 CO 33 av momwm “cote 833.20 .mm 033.

24

_x
01
r

 

 

 

 

 

 

.E'

E

O

CO

3 12 ~ 3

52 ‘i

I; i
E 9 :

.5 T [3
(D

E

o. 5 ‘

9

(D

E

.9.)

E 3‘

g El Empty nests

To I Egg-containing nests
2 0

Early season Late season

Figure 2.1. Average provisioning rates (both parents combined) at early and late season
nests for house wren pairs with empty nests (unfilled squares) and pairs with egg-
containing nests (solid squares) added to their territories. The treatment X season
interaction was Significant (P < 0.02). During the early season, pairs with empty nests
added to their territories provisioned at higher rates than pairs with egg-containing nests
added to their territories (P < 0.01). There were no differences in provisioning rates for

late season nests. Back-transformed means i 95% confidence intervals are Shown.

25

“mp" "est? A 6‘
1.6 ‘ OEgg-contarnrng nests

1.24 O

 

.0
oo

 

.0
.rs
_.__

 

 

O

A
O)
[D
-+o

A
N
I
’L
V

Provisioning visits per nestling per 30 min.

 

O
(I)
_._

 

 

.0
is

 

 

 

0

Early season Late season

Figure 2.2. Average (a) male and (b) female provisioning rates per nestling at early and
late season nests for house wren pairs with empty nests (unfilled circles) and pairs with
egg-containing nests (solid circles) added to their territories. The treatment X season
interaction was significant (P < 0.05) for males but not for females. For males, pairwise
comparisons of treatments within a brood were not significantly different. Seasonal
means were significantly different for males (P < 0.01) and females (P < 0.05). Back-

transformed means i 95% confidence intervals are Shown.

26

CHAPTER 3
DOES CAVITY AVAILABILITY AFFECT TRADE-OFFS BETWEEN
CURRENT AND FUTURE REPRODUCTION IN MALE HOUSE WRENS?
ABSTRACT
Parental investment in a given reproductive effort is generally constrained by

trade-off costs to future reproduction. For polygynous and multi-brooded species these
future costs can be almost immediate if investment in a current reproductive effort
decreases the opportunity to attract additional mates. AS a result, increased opportunity
to attract additional mates might shift the balance towards less investment in early
broods. I examined this trade-off in a polygynous, double-brooded passerine, the house
wren (T roglodytes aedon), by experimentally manipulating opportunities for males to
attract additional mates. Since cavity availability limits a male’s opportunity to acquire
additional mates, I manipulated male mate attraction opportunity by controlling male
access to additional cavities. Newly settled territorial males with one nest box were
randomly assigned zero or two additional cavities. I recorded male song as an indicator
of mate attraction effort and monitored paternal provisioning as a measure of parental
investment. Although males with surplus cavities in their territories sang more, paternal
provisioning was not affected by cavity availability. Paternal provisioning rates were
observed during the middle to late nestling periods, and reproductive trade-offs might
have occurred earlier in the nestling period when females were brooding young.
Regardless of treatment, males provisioned less at early season nests than at late season

nests. The costs of investing in parental care at early season nests might be high,

27

regardless of mate attraction opportunity, if paternal care towards the end of the nestling
period affects the probability of rearing a second brood.
INTRODUCTION

For iteroparous Species, parental investment in a given reproductive attempt is
generally constrained by trade-off costs to future reproduction and survival (Steams
1989). These trade-offs can be difficult to study using correlational approaches because
uncontrolled variables among individuals, such as resources, may result in positive
relationships between traits across individuals, thereby masking negative trade-offs
within individuals (van Noordwijk and de Jong 1986). For example, an individual in
good condition might allocate more to reproduction and also have higher survival than an
individual in poor condition, even though each individual faces a negative trade-off
between investment in reproduction and survival. Experimental manipulations of
phenotype (Sinervo and Basolo 1996, also “phenotypic engineering” Ketterson et al.
1996) can randomize phenotypes across individuals, thus breaking up naturally co-
occurring traits and exposing underlying trade-offs.

Phenotypic manipulation might be achieved by (a) direct manipulation of a trait,
(b) manipulation of the environment in order to produce a change in a trait, or (c)
manipulation of a physiological mechanism that links multiple traits (Sinervo and Basolo
1996). To examine trade-offs between current and future reproduction, investment in
current reproduction might be manipulated by increasing clutch size (i. e., an
environmental manipulation affecting parental investment in current reproduction). For
example, female collared flycatchers rearing experimentally enlarged clutches produce

fewer offspring in subsequent reproductive attempts, demonstrating a trade-off between

28

current and future reproduction (Gustafsson and Part 1990). Another possibility is to
manipulate a proximate physiological mechanism, such as a hormone, that underlies the
trade-off. For instance, testosterone affects trade-offs between territorial defense, a
component of mate attraction, and survival in male mountain Spiny lizards (Sceloparus
jarrovi, Marler and Moore 1988). In dark-eyed juncos (Junco hyemalis), males
implanted with testosterone decrease provisioning effort (Ketterson et al. 1992) but
produce more extra-pair young than control males (Raouf et al. 1997).

In the case of polygamous or multi-brooded species, trade-off costs to future
reproduction can be almost immediate if investment in a current reproductive attempt
(e. g., parental care) decreases the opportunity to attract additional mates. Under this
trade-off, we would expect increased opportunities for additional breeding attempts to
result in decreased investment in current reproduction. Male two-spotted gobies
(Gobiusculusflavescens) reduce levels of parental care when fertile females are present
(Bj elvenmark and Forgren 2003), and fairy martins (Hirundo arieI) decrease parental
care when opportunities for extra-pair matings are high (Magrath and Elgar 1997). When
opportunities to attract additional mates are limited by nest Sites, the availability of nest
sites might affect trade-offs between investment in parental care and mate attraction.
This trade-off has been observed in European starlings (Sturnis vulgaris), where already-
mated males with an additional nest site added to their territory traded-off investment in
parental care for increased mate attraction effort (Smith 1995).

Most studies that have manipulated trade-offs between investment in current and
fiiture reproduction have focused primarily on investment decisions of one parent and

disregarded potential responses by the mate. For species with bi-parental care, the

29

optimal level of investment by one parent may depend on the level of investment of its
mate (Chase 1980, Winkler 1987, Lazarus 1990). If individuals are able to assess the
liklihood that a potential mate will invest in parental care, this information might be used
in mate choice and reproductive investment decisions. For example, if nest Site
availability mediates a trade-off between investment in mate attraction and parental care
for males, then females might be able to use cavity availability to assess potential mates.

I investigated reproductive investment in a secondary cavity nester, the house
wren (T roglodytes aedon). Song is an important component of mate attraction in house
wrens (Johnson and Searcy 1996). Unmated males produce ‘high volume spontaneous
song’ that ceases immediately after pairing but resumes during incubation, presumably to
attract additional mates (Johnson and Kerrnott 1991). Previous studies have found that
polygynous males have greater reproductive success than monogamous males (Soukup
and Thompson 1997), and that paternal care is important to female reproductive success
(Johnson et al. 1992, Johnson et al. 1993). These potentially competing interests set up a
possible trade-off between investment in current and future reproduction if males cannot
simultaneously invest in mate attraction and care for young. Male opportunities to attract
additional mates are limited by the number of defendable cavities in their territory, and
therefore, females might be able to use cavity availability in male territories to assess
aspects of male quality such as parental investment.

In earlier experiments, I documented a Significant difference in paternal
provisioning effort between early and late season nests in house wrens (Chapter 2).
Males provision late season broods at higher rates than early season broods. This

increase in provisioning effort coincides with reduced mating opportunity late in the

30

breeding season and suggests that male house wrens might increase paternal effort when
opportunities to attract additional mates are low. I experimentally tested this hypothesis
by manipulating cavity availability in male house wren territories. Since cavity
availability limits a male’s opportunity to acquire additional mates, I manipulated male
mate attraction opportunity by controlling male access to additional cavities. 1 evaluated
the effect of increased mate attraction opportunity on parental investment by both males
and females. In this chapter, I examine whether cavity availability affects a trade-off
between male investment in mate attraction and parental care.
METHODS

I conducted this study April through August of 2001 and 2002 at the Kellogg
Biological Station’s Lux Arbor Reserve (42°29’N, 85°28’W) in Barry County, Michigan.
Lux Arbor Reserve is a 529-ha Site consisting of fragmented habitats including hardwood
and softwood forests, open fields, wetlands, and agricultural areas. Cedar nest boxes
were installed in clusters (hereafter referred to as sites) composed of three boxes spaced
5-10 meters apart. All nest boxes were installed on metal poles and placed approximately
1.2 m above the ground. To minimize losses due to predation, a 0.75-m conical predator
guard was installed on the poles of all active nests. F iffy-three sites were established, and
each site was separated by a minimum distance of approximately 100m, which limited
virtually all males to a single site per territory.

Male house wrens arrive at the study area in mid to late April and immediately
begin establishing territories. Males begin placing sticks in available cavities shortly
after arrival, and the first females arrive several days later. Once paired, females may

help build the stick platform of the nest (Alworth and Scheiber 2000) and then line the

31

nest cup with soft material (Kendeigh 1952, McCabe 1965). On our study Site,
approximately 5% of males are polygynous.

In Michigan, most house wrens returning at the beginning of the breeding season
are double-brooded, resulting in a bimodal distribution for the date of nest lining (Figure
3.1). I classified nests with line dates in the first peak as early season nests (those lined
prior to 6 June). Nests lined after 6 June were considered late season nests. Most late
season nests were second broods for one or both members of the pair, but some new
breeders did appear on the study Site during the late season nesting period.

Experimental manipulation

Prior to male arrival, I plugged two of the three nest boxes in each territory with
rubber stoppers. Therefore, all males settled in single box territories. I checked nest
boxes at least every other day during the male and female settlement periods. A territory
was considered settled after a male was observed singing near an available nest box for
two consecutive nest checks or if nesting material was observed in a nest box.

I manipulated cavity availability in male territories twice during the nesting
period. First, I applied a pre-pairing treatment within two days of male settlement in
order to randomize male quality across treatments, while leaving open the possibility for
males to opportunistically vary investment in mate attraction and/or paternal care. I
randomly assigned territories to either single cavity or surplus cavity treatments. In
single cavity territories, stoppers were maintained in the two plugged nest boxes, whereas
in surplus cavity territories, stoppers were removed from the two plugged nest boxes. To
control for time effects of male settlement, I paired sites by male settlement date and

randomly applied treatments within those pairs. By randomizing treatments across males

32

prior to female settlement, 1 was able to examine the effect of cavity availability
(independent of male quality) on female mate choice and reproductive investment
(Chapters 4—5).

In order to look at the effect of cavity availability on male trade-offs between
mate attraction and provisioning effort during the incubation and nestling periods, I also
needed to randomize differences in female quality across the experimental treatments.
Therefore I re-randomized experimental treatments at the onset of incubation (the post-
laying treatment), again assigning territories to either single or surplus cavity treatments.
This second experimental manipulation allowed me to look for an effect of cavity
availability on male trade-offs between mate attraction and provisioning effort
independent of differences in female condition. For example, suppose that females
preferred monogamous males, then higher quality females might preferentially settle in
single cavity territories. As a result, males in territories with single cavities might invest
less in provisioning, not as a result of cavity availability, but because they were mated to
females in better condition that could provision more. By re-randomizing treatments at
the onset of incubation, I de-coupled the possible correlation between female condition
and cavity availability during the incubation and nestling periods.

After nestlings fledged, I cleaned nesting material out of nest boxes and
maintained the ‘post-laying’ configuration as the ‘pre-pairing’ treatment for late season
nests. At the onset of incubation, I again re-randomized the experimental treatments, just

as for early broods.

33

Provisioning eflort and mate attraction

I captured adults prior to provisioning observations using a mist net or box trap,
and marked each adult with an aluminum US. Fish and Wildlife Service leg band and a
unique combination of three color bands. I observed parental provisioning visits for 25-
30 minutes in the morning (between 7:00 and 11:00 am) when nestlings were 4-5 days
old, 8-9 days old, and 11-12 days old. Brood day (BD) 0 is defined as the day the first
nestling in a given nest hatched. Provisioning observations were conducted during fair
weather using 10x binoculars mounted on a tripod, taking care to minimize disturbance.
Any parental visit to the entrance of the nest box was considered a provisioning trip,
regardless of whether food was actually observed.

I recorded male song using a Sony TCM-SOOOEV or a Marantz PMD221 cassette
recorder and a parabolic reflector (Sony PBR-330) with a Calrad super cardioid
condenser microphone. Songs consist of a relatively low amplitude introduction section
made up of acoustically complex elements followed by a louder terminal section of
repeated trill notes (Platt and Ficken 1987; Figure 3.2). Song was recorded in the
morning for 20-45 minutes during the incubation period (4-5 days after the post-laying
treatment was applied) and again when nestlings were 6-8 days old. If recording
conditions were poor, or no song was recorded, males were recorded again on a following
day (total recording time never exceeded 45 minutes). I calculated an index of song rate
by dividing the number of 15-sec intervals during which a male sang at least once by the
total number of recording intervals. This index is a conservative estimate of song rate
because it underestimates high song rates (males that sing frequently sing 3-4 times per

15-sec interval). In addition, males that sang little tended to be recorded for shorter

34

periods of time, in which case, the index would overestimate song rates of males that
sang little. I counted any vocalization during an interval as a “song.” By using counts
instead of time spent singing, I avoided problems associated with variable recording
conditions in the field, e.g., wind interference or poor recording quality. Hence, songs
could be counted from the tape even when precise measurements of song length were not
possible.

Avisoft SASLab Pro was used to analyze digitized song recordings. For each
recording, I measured the duration of the introductory section for a randomly selected
subset of five songs containing a terminal trill section (except for one male that only sang
three songs). Song elements separated by less than 0.2 seconds were considered part of
the same song. I excluded songs consisting only of “introduction notes” and songs with
“introduction notes” inserted into an otherwise continuous trill sequence. Songs for
which the beginning or end of the introduction sequence could not be distinguished due
to poor recording quality were omitted from the analysis. Males rarely sang during
recording sessions at late season nests, and I included only early season nests in song
analyses.

Statistical analyses

I used standard tests to analyze frequency data (Sokal and Rohlf 1995): Cochran’s
Q test for changes in the proportion of males provisioning over time (repeated measures)
and Chi-square tests of independence for treatment effects on the proportion of males
provisioning. In one case, analysis of treatment effects at late season nests, Chi-square
tests of independence were inappropriate due to small sample sizes (i. e., the expected

values were less than 5 for some categories). Therefore, I calculated 95% confidence

35

intervals around the log odds ratio for the difference in the proportion of males
provisioning in surplus versus single cavity territories in order to determine whether the
log odds ratio differed from zero (Sokal and Rohlf 1995). I analyzed provisioning rates
and song variables (repeated measures) using profile analysis with MANOVA (PROC
GLM) in SAS v.8.1 (SAS Institute Inc. 2000).

Data were transformed where necessary to improve normality of residuals and
reduce heteroscedasticity for parametric analyses. Provisioning rates were standardized
to a 30-minute observation period for analysis. The pre-pairing treatment and
pre-pairing X post-laying treatment interaction were non-significant in analyses of
provisioning and song rates (P > 0.2), indicating that prior treatment history had no effect
on these variables after the post-laying treatment was applied. Therefore, I included only
the post-laying treatment in those analyses. There was a tendency for the pre-pairing
treatment to affect introduction length of songs, so both treatments and the interaction
were included in that particular analysis.

Polygynous males and males that lost control of a “surplus” nest box to intruding
males were excluded from analyses because both occurrences disrupted the surplus cavity
treatment (i. e., such males were unlikely to have opportunities to attract additional
females). Polygyny occurred seven times over the two years of the study. Split
territories occurred five times in 2002 only. Males that returned between years were
assumed to be independent data points for analysis. Eleven males produced at least one
clutch during both years. Analyses including returning males are presented, but I also ran
these analyses with each male included only once in the data set (randomly selecting

which entry to include). Unless noted, these more conservative tests yielded the same

36

results. Means 2t se are presented (means for transformed data are back-transformed into
original units).
RESULTS
Parental investment

At early season nests, the proportion of males provisioning was affected by
nestling age (surplus cavities: Q = 18.0, df= 2, P < 0.001; single cavities: Q = 24.7,
df= 2, P < 0.0001; Table 3.1). Almost all males provisioned on BD 4 (97%, the two
males that did not occurred in surplus cavity territories). Fewer males provisioned when
nestlings were older, but there were no differences between territories containing surplus
and single cavities (BD 8: X2 = 0.012, df= 1, P > 0.9; BD 12: X2 = 0.05, df= 1, P > 0.8).
At late season nests, the proportion of males that provisioned was high (> 90%) and was
not related to age of nestlings (surplus cavities: all males provisioned on all days; single
cavities: Q = 0.5, df = 2, P > 0.7). There was no difference in the proportion of males
provisioning on BD 12 between males with surplus and single cavities in their territories
at late season nests (log odds ratio with 95% confidence limits: 2.2 :l: 2.98). Combining
treatments, more males provisioned at late season nests than early season nests on BD12
(X2 = 24.0, df= l, P < 0.0001). Of males that provisioned young for the entire nestling
period, there were no differences in per-nestling provisioning rates between treatments or
between broods, but there was a Si gnificant effect of nestling age on paternal provisioning
rates (Table 3.2; Figure 3.3).

Males delivered only a single prey item during virtually all of the provisioning
trips we observed. We did not measure prey items, but the vast majority of prey items

were small (less than one bill length long) and varied little between territories. It is

37

unlikely that there were any significant differences between treatments in the amount of
food males delivered per provisioning trip.

Investment in paternal care at first brood nests might have affected territory
retention or the likelihood of rearing a second brood. Males that provisioned first brood
young on brood day 12 were not more likely to retain the same territory for second
broods than males that did not provision young on brood day 12 (X2 = 1.1, df= 1,

P > 0.3; Table 3.3). Seventy-six percent of males that did not provision first brood young
reared second broods, whereas only 56% of males that provisioned first brood young
reared second broods, but these differences were not significant (X2 = 2.6, df= l,
P < 0.11; Table 3.3). Polygynous males were included in this analysis, but all
unsuccessful first brood nests were excluded.

Mate attraction

At early season nests, males with surplus nest boxes in their territories had a
higher song rate index than males with single nest boxes in their territories (Table 3.4;
Figure 3.4). This difference was primarily due to males in territories containing single
nest boxes reducing the amount of time spent singing between the incubation and nestling
periods, whereas males in territories containing surplus nest boxes maintained similar
song rates between periods. In the nestling period, males with surplus cavities sang
during approximately 40% of recording intervals versus approximately 20% of recording
intervals for males with single cavities in their territories.

Cavity availability during the incubation and nestling periods had no effect on the
introduction length of male songs (Table 3.5; Figure 3.5), although there was a tendency

towards an historical effect of the pre-pairing treatment. Males with surplus nest boxes in

38

their territories before the incubation period tended to sing longer introduction sequences
during the incubation period than males with single nest boxes in their territories before
the incubation period. The repeated measures analysis included only nests for which I
was able to record song during both the incubation and nestling periods, and therefore
excluded many nests that failed after the incubation recording. In order to utilize the
entire data set, I restricted the analysis to the incubation period. Again, there was a non-
significant increase in introduction length for males with surplus cavities in their
territories before the incubation period (AN OVA: pre-pairing treatment, F1,“ = 3.0,
P < 0.09, overall model 115; means: surplus cavities = 0.85 i 0.069 sec, single cavities =
0.67 i 0.076 sec).
DISCUSSION

Whether a male should invest in parental care or mate attraction depends on the
relative fitness benefits of current and future reproduction. For cavity nesters,
opportunities to attract additional mates are limited by both mate availability and the
availability nest Sites. I found that already-mated males with surplus cavities in their
territories sang more after pairing than mated males with a single (occupied) nest site in
their territories. Adding surplus cavities to a male’s territory not only increased a male’s
opportunity to attract additional mates, but might also have increased the perceived
quality of the territory, resulting in a shift in investment towards mate attraction effort.

The observed differences in song rate were much greater during the nestling
period than the incubation period. Males with surplus cavities sang at high rates during
both the incubation and nestling periods, whereas males with single cavities sang at high

rates during the incubation period, but decreased song rates during the nestling period.

39

Since male house wrens do not incubate and the risk of cuckoldry is reduced during the
incubation period, opportunities for extra-pair matings might be high during this time
regardless of cavity availability. If opportunities for extra-pair matings decrease as the
nesting cycle progresses (e. g., if fewer neighboring females are fertile after eggs hatch),
then males with single nest boxes in their territories would gain little by advertising
during the nestling period, but males with surplus cavities might still have opportunities
to attract secondary females. An alternative explanation for persistently high song rates
among males with surplus cavities in their territories might be territorial defense against
intruding males trying to usurp nest sites. In a few territories (5 cases in 2002 only),
males did lose control of one of their surplus nest boxes to unmated males. Generally
this phenomenon occurred towards the end of the nestling period of first broods or during
second broods. But since the primary reason for defending surplus cavities is to attract
additional females, defense of surplus cavities would still serve to increase mating
opportunities, even if the song were not specifically directed at females.

Little is known about the specific aspects of male song that females might use in
assessing mates. E.D. Kennedy and J .M. Lueken (pers. comm.) observed that unmated
males sing longer introduction sequences than mated males. I did not measure the
introduction sequence of songs of unmated males but looked for differences between
advertising and non-advertising mated males (i. e., males with surplus versus single cavity
territories). I did not find a difference in the length of the introduction sequence between
mated males with single versus surplus cavities in their territories during the incubation
and nestling periods. However, there was a tendency for males with surplus nest boxes in

their territories prior to the incubation period to sing longer introduction sequences,

40

particularly during the incubation period. Perhaps males with surplus cavities in their
territories prior to the incubation period sang longer introduction sequences earlier in the
nesting cycle and switched to shorter introductions during the late incubation or nestling
periods.

Even though males with surplus cavities in their territories exhibited higher song
rates during the nestling period, there were no differences in provisioning rates between
males with surplus and single nest boxes in their territories. This pattern suggests that
either the costs of mate attraction are low for males with surplus boxes or that the benefits
associated with increased parental effort are low for males with single nest boxes in their
territories. Smith (1995) found that male European starlings with a surplus nest box
added to their territories increased mate attraction effort at the expense of paternal
investment during the incubation period when mate attraction opportunities were high,
but not during the nestling period when mate attraction opportunities were low. It is
unlikely that low mate attraction opportunity accounts for the failure to observe a trade-
off between mate attraction and parental investment during the nestling period in this
study. New females were frequently observed on the study site during late season broods
and were presumably also present during the nestling period of early season broods.
Considering this observation, the fact that polygyny did not occur more often is
surprising (5-10% of males with surplus boxes were polygynous), but one possible
explanation is that the three nest boxes within a territory were Spaced relatively close
together, and resident females may have been able to exclude prospecting females

(Slagsvold and Lifjeld 1994).

41

Males may have faced different, unmeasured costs of increased mate attraction
opportunity, such as increased territorial defense of surplus cavities, or trade-offs might
have occurred earlier in the nestling period. For house wrens, the most critical stage of
paternal care may be during the earliest part of the nestling period when females are
brooding young (Johnson et al. 1992). Male removals later in the nestling period (after
brood day 4) decrease reproductive success only under unfavorable environmental
conditions (Bart and Tomes 1989). Therefore, the trade-off between mate attraction and
paternal care might be most apparent during the first few days after eggs hatch. In this
study, the earliest provisioning observations occurred on brood day 4, and male song was
recorded on brood days 6-8. If paternal care were relatively unimportant later in the
nestling period, it would not be surprising to find similar provisioning rates between
males that were investing in mate attraction and those that were not.

Regardless of cavity availability, male investment in provisioning at early season
nests was lower than at late season nests (see also Dubois and Getty 2003). Virtually all
males provisioned at late season nests, whereas fewer than one-half of males provisioned
for the entire nestling period at early season nests. After nestlings fledge, one or both
parents generally care for the young until they reach independence (Kendeigh 1941).
Males that care for young after fledging are more likely to lose their territory and less
likely to have a second brood (Bart 1990). It is possible that males that desert towards
the end of the nestling period are less likely to end up caring for fledglings. Therefore,
the costs of investing in parental care at early season nests could be high for males of

double-brooded Species, regardless of polygyny potential, if male care towards the end of

42

the nestling period during the first brood decreases the probability of rearing a second
brood.
ACKNOWLEDGMENTS

NSD was supported by a NSF Graduate Research Fellowship, 3 MSU
Distinguished Fellowship, the KBS Research Training Grant funded by NSF, the
departments of Zoology and EEBB (MSU), a Sigma Xi Grants-In-Aid-of-Research
Award, a Lauff Research Award (KBS) and Wallace Award (MSU/Zoology). Field and
technical assistance were provided by E. Dubois, S. Dubois, J. Glazer, T. Grattan,
M. Homish, J. Johnson, C. Noud, T. Poloskey, K. Wharton, and E. Zontek.

E.D. Kennedy provided advice regarding song analysis.

43

LITERATURE CITED

Alworth, T. and I.B.R. Scheiber. 2000. Nest building in House Wrens ( T roglodytes
aedon): a reexamination of male and female roles. Journal of Field Ornithology
71: 409-414.

Bart, J. 1990. Male care, mate switching, and future reproductive success in a double-
brooded passerine. Behavioral Ecology and Sociobiology 26: 307-313.

Bart, J. and A. Tomes. 1989. Importance of monogamous male birds in determining
reproductive success. Evidence for house wrens and a review of male-removal
studies. Behavioral Ecology and Sociobiology 24: 109-116.

Bj elvenmark, J. and E. Forgren. 2003. Effects of mate attraction and male-male
competition on paternal care in a goby. Behaviour 140: 55-69.

Chase, ID. 1980. Cooperative and noncooperative behavior in animals. American
Naturalist 115: 827-857.

Dubois, NS. and T. Getty. 2003. Empty nests do not affect female mate choice or
maternal investment in House Wrens. Condor 105: 382-387.

Gustafsson, L. and T. Part. 1990. Acceleration of senescence in the collared flycatcher
F icedula albicollis by reproductive costs. Nature 347: 279-281.

Johnson, LS. and L.H. Kermott. 1991. The functions of song in male house wrens
(T roglodytes aedon). Behaviour 116: 190-209.

Johnson, L.S., M.S. Merkle, and L.H. Kermott. 1992. Experimental evidence for
importance of male parental care in monogamous House Wrens. Auk 109: 662-
664.

Johnson, L.S., L.H. Kermott, and MR. Lein. 1993. The cost of polygyny in the house
wren T roglodytes aedon. Journal of Animal Ecology 62: 669-682.

Johnson, LS. and WA. Searcy. 1996. Female attraction to male song in house wrens
(T roglodytes aedon). Behaviour 133: 357-366.

Kendeigh, SC. 1941. Territorial and mating behavior of the house wren. Illinois
Biological Monographs 18: 1-120.

Kendeigh, SC. 1952. Parental care and its evolution in birds. Illinois Biological
Monographs 22: 1-356.

Ketterson, E.D., V. Nolan, Jr., M.J. Cawthom, P.G. Parker, and C. Ziegenfus. 1996.
Phenotypic engineering: using hormones to explore the mechanistic and
functional bases of phenotypic variation in nature. Ibis 138: 70-86.

44

Ketterson, E.D., V. Nolan, Jr., L. Wolf, and C. Ziegenfus. 1992. Testosterone and avian
life histories: effects of experimentally elevated testosterone on behavior and

correlates of fitness in the dark-eyed junco (Junco hyemalis). American Naturalist
140: 980-999.

Lazarus, J. 1990. The logic of mate desertion. Animal Behaviour 39: 672-684.

Magrath, M.J.L. and MA. Elgar. 1997. Paternal care declines with increased opportunity
for extra-pair matings in fairy martins. Proceedings of the Royal Society of
London Series B 264: 1731-1736.

Marler, CA. and MC. Moore. 1988. Evolutionary costs of aggression revealed by
testosterone manipulations in free-living male lizards. Behavioral Ecology and
Sociobiology 23: 21-26.

McCabe, RA. 1965. Nest construction by House Wrens. Condor 67: 247-256.

Platt, ME. and MS. F icken. 1987. Organization of singing in House Wrens. Journal of
Field Ornithology 58: 190-197.

Raouf, S.A., P.G. Parker, E.D. Ketterson, V. Nolan, Jr., and C. Ziegenfirs. 1997.
Testosterone affects reproductive success by influencing extra-pair fertilizations

in male dark-eyed juncos (Aves: Junco hyemalis). Proceedings of the Royal
Society of London Series B 264: 1599-1603.

Sinervo, B. and AL. Basolo. 1996. Testing adaptation using phenotypic manipulations.
Pp. 149-185 in Adaptation (Rose, MR. and G.V. Lauder, eds.). Academic Press,
San Diego.

Slagsvold, T. and J .T. Lifjeld. 1994. Polygyny in birds: the role of competition between
females for parental care. American Naturalist 143: 59-94.

Smith, H.G. 1995. Experimental demonstration of a trade-off between mate attraction and
paternal care. Proceedings of the Royal Society of London Series B 260: 45-51.

Sokal, RR. and F]. Rohlf. 1995. Biometry: the principles and practice of statistics in
biological research, third edition. W.H. Freeman and Company, New York.

Soukup, SS. and CF. Thompson. 1997. Social mating system affects the frequency of
extra-pair paternity in house wrens. Animal Behaviour 54: 1089-1105.

Steams, SC. 1989. Trade-offs in life history evolution. Functional Ecology 3: 259-268.

van Noordwijk, A.J. and G. de Jong. 1986. Acquisition and allocation of resources: their
influence on variation in life history tactics. American Naturalist 128: 137-142.

Winkler, D.W. 1987. A general model for parental care. American Naturalist 130: 526-
543.

45

Table 3.1. Proportion of males provisioning nestlings in territories containing surplus

and Single cavities (post-laying treatment) at early and late season nests.“|

 

 

 

 

Early season nests Late season nests
Surplus cavities Single cavities Surplus cavities Single cavities
BTOOd day (11 = 25) (n = 33) (n = 22) (n = 26)
4 0.92 1.00 1.00 0.88
8 0.68 0.67 1.00 0.88
12 0.44 0.45 1.00 0.85

 

a Nests with incomplete data were excluded.

46

Table 3.2. Repeated measures analysis of paternal provisioning rates per nestling on
brood days 4, 8, and 12 for males in territories containing surplus nest boxes and single
nest boxes during the nestling period (post-laying treatment) at early and late season

nests. Only males that provisioned for the entire nestling period were included.a

 

A. MANOVA (profile analysis) of within subject effects. Pillai’s trace values are

 

presented.
Source Num df Den df Value F P > F
Brood day (time) 2 62 0.14 5.2 0.01
Brood day X Season 2 62 0.041 1.3 0.2
Brood day X Treatment 2 62 0.010 0.33 0.7
Brood day X Season X Treatment 2 62 0.016 0.51 0.6

 

B. Repeated measures ANOVA of between-subj ect effects.

 

Source df MS F P > F
Season 1 0.056 0.40 0.5
Treatment 1 0.001 0.01 0.9
Season X Treatment 1 0,016 0.11 0.7
Error 63 0.14

 

C. ANOVAS on each of the contrasts for time differences between successive
observations. Only the brood day effect is shown, others were non-significant.

 

 

Source df MS F P > F
gpgrgriaztaynable: BD8 — BD4 1 0.002 0.04 0.8
Error 63 0.063
Contrast variable: BD12 — BD8
Brood day 1 0.82 7.03 0.01
Error 63 0.12

 

a Data square root transformed. Nests with incomplete data were excluded.

47

Table 3.3. Number of male house wrens retaining the same territory for first and second
broods, switching territories between broods, and losing territories in relation to paternal

provisioning at first brood nests. Only males with successful first brood nests were

 

 

 

 

included.
Reared a second brood
Retained . Lost temtory
. territory Switched No known
First brood temtorres second brood
Provisioned on brood day 12 14 1 12
Did not provision on brood day 12 19a 3 7

 

3 Includes a polygynous male that retained his territory for a secondary female and then

lost it to another male.

48

Table 3.4. Repeated measures analysis of male song rate during the incubation and
nestling periods for males in territories containing surplus nest boxes and single nest

boxes (post-laying treatment) at early season nests.a

 

A. MANOVA (profile analysis) of within subject effects. Pillai’s trace values are

 

presented.
Source Num df Den df Value F P > F
Period (time) 1 46 0.035 1.7 0.2
Period x Treatment 1 46 0.083 4.2 0.05b

 

B. Repeated measures ANOVA of between-subject effects.

 

Source df MS F P > F
Treatment 1 0.22 6.1 0.017
Error 46 0.036

 

a Nests with incomplete data were excluded.
b Period X Treatment effect not significant (P > 0.13) with more conservative analysis

which included returning males only once in the data set.

49

Table 3.5. Repeated measures analysis for the length of the introduction sequence of
male songs during the incubation and nestling periods for male house wrens in territories
containing surplus nest boxes and Single nest boxes during the pro-pairing treatment

(PPT) and post-laying treatment (PLT) periods at early season nests.a

 

A. MANOVA (profile analysis) of within subject effects. Pillai’s trace values are

 

 

 

presented.
Source Num df Den df Value F P > F
Period (time) 1 43 0.015 0.65 0.4
Period X Pre-pairing treatment 1 43 0.067 3.1 0.09
Period X Post-laying treatment 1 43 0.042 1.9 0.18
Period X PPT X PLT 1 43 0.020 0.90 0.3

 

B. Repeated measures ANOVA of between-subject effects.

 

Source df MS F P > F
Pre-pairing treatment 1 0.60 3.1 0.08
Post-laying treatment 1 0.13 0.69 0.4
PPT X PLT 1 0.034 0.17 0.6
Error 43 0.19

 

a Nests with incomplete data were excluded.

50

.Em< on no 3:: we? “no: “we 2E. .38: comma BE 98 .330 5933 .8636 ofi

we cum: we? 2:: o .355 05 we who.» Son E .88 E $5.8 wee: “we: 18 message .353.on .mm BEE

Sena
2 S : 2 a a A e m e m N _

 

 

 

 

srsou pourl J0 roqumN

.Efiwobooam 05 gene xon 05 E “.8865 m_ avg—98¢. 25820 3332 .«o wccflmeoo =5 REES 595— m .3 330:0.“

mm was 35626 63:80 mo £238 oocosaom cowosvobi 5&8 2E. .waom :23 8:0: 2a.: m we Enumobooqm .N.m oBmE

 

 

 

 

 

 

 

 

 

 

m Wm o.m Wm 9N m4 o; md
. . :E :2:th . . 55260::— .
:fififie a was i; ..
- . . -_ . . :r, . , . e
H . . a
. 3
1. NE

 

. :2. a
. :2 .

52

2.5

 

 

 

 

 

a. E E
E : :
‘53 2 — l l
"‘ l l
a : :
8. : : “
'5 2, : :
g) 15 - ll 22 % : :
.2 13 i l
.2 l l
> r I
8 l 3
Q.- l r
2 1 — l i
C“ l l
2 : :
i i meanise
0.5 t ' t ' t
4 8 12

Brood day

Figure 3.3. Average paternal provisioning rates per nestling during 30-minute
observation periods for males in territories containing surplus nest boxes (solid symbols)
and single nest boxes (open symbols) during the nestling period. Early season nests
indicated by circles, late season nests indicated by diamonds. Only males that
provisioned for the entire nestling period were included. Nests with incomplete data
were excluded, therefore samples sizes (indicated) are the same across brood days. There

was a Si gnificant brood day effect (P < 0.01).

53

0.5

 

 

 

 

20
0.4 -
25 0.3 -
E.
B
E
O!)
S 0.2 q 28
U)
0.1 -
mean i se
0 I I
Incubation Nestling
Period Period

Figure 3.4. Song rate index for males in territories containing surplus cavities (solid
symbols) and single cavities (open symbols) during the incubation and nestling periods at
early season nests. The treatment main effect was significant (P < 0.02). The interaction
was marginally significant (P = 0.05). Nests with incomplete data were excluded.

Sample sizes are indicated.

54

 

 

 

 

 

 

1.2
E
‘1’ 15
a .-
“5
0‘ 0.8 -
Gui 13 12 -
s:
.2
g .-
1: ,-
.9.
.S
{.5 7
4:: 0.4 1 I
9:”
O)
b—l ..
meantse
0
Incubation period Nestling period

Figure 3.5. Length of the introduction sequence for songs of male house wrens in
territories containing surplus (solid symbols) and single (open symbols) cavities in their
territories during the incubation and nestling periods. Pre-pairing treatment is indicated
by circles (surplus cavities) and squares (single cavities). Means were not significantly
different. Nests with incomplete data were excluded, therefore samples sizes (indicated)

are the same for the incubation and nestling periods.

55

CHAPTER 4
MATERNAL INVESTMENT BY FEMALE HOUSE WRENS DEPENDS ON
TERRITORY CHARACTERISTICS RELATED TO MALE MATE
ATTRACTION OPPORTUNITY

with Thomas Getty

ABSTRACT

Whether a high quality mate will invest in mate attraction at the expense of
parental care sets up a potential conflict for females of polygynous species. For cavity-
nesting species unable to excavate their own cavities (secondary cavity nesters), a male’s
opportunity to become polygynous is limited by the availability of cavities within his
territory. When male parental care affects reproductive success, then females might
assess male opportunity for polygyny when selecting mates and prefer males with low
polygyny potential. Alternatively, male competition over territories with high polygyny
potential (i. 3., multiple cavities) might result in a positive relationship between male
quality and cavity availability within his territory. If males vary in their ability to provide
benefits to females, either direct or indirect, then females should demonstrate preferences
for males providing benefits and adjust reproductive investment in response the quality of
their mate. We experimentally manipulated cavity availability independently of male
quality and found that first brood females mated to males with multiple cavities in their
territories laid larger clutches than females mated to males with single cavities in their
territories. Cavity availability had no effect on the timing of nest lining (our index of the
timing of female settlement) or on female site fidelity between early and late season

broods.

56

INTRODUCTION

Darwin (1871) identified two components of sexual selection that contribute to
variance in reproductive success among individuals, (1) intrasexual competition,
generally males competing for access to females, and (2) intersexual selection, where
females, often the choosy sex, prefer some males over others. Numerous studies have
shown that females can gain a variety of benefits from mate choice (reviewed in
Andersson 1994). Females might choose mates on the basis of characters that signal
direct benefits such as territory quality (Part and Qvamstrom 1997), paternal care (Hill
1991), and/or indirect benefits such as genetic quality (Moller 1994). When males both
defend territories and provide parental care, male competition over access to high quality
territories may facilitate female choice of high quality mates (Qvarnstr'o’m et al. 2000).

For cavity-nesting species unable to excavate their own cavities (secondary cavity
nesters), cavity availability limits opportunities for polygyny. Experimental studies have
demonstrated that rates of polygyny can increase with the number of cavities available in
a male’s territory (e. g., Johnson and Kermott 1991a, Petit 1991). When the number of
defendable cavities differs between available territories, we might expect competition
among males to lead to higher quality males defending territories with greater numbers of
cavities. If males also provide direct benefits to females, such as parental care, trade-offs
between parental investment and mate acquisition set up a possible conflict between male
and female interests in reproductive investment (Davies 1989, Lifjeld and Slagsvold
1991, Kempenaers 1995, Smith and Sandell 1998).

Many studies have demonstrated costs of polygyny for females, such as reduced

paternal assistance and lower reproductive success (e. g., Litjeld and Slagsvold 1989,

57

Pinxten and Eens 1994, Kempenaers 1995, Sandell et al. 1996). Even prior to the
acquisition of a secondary female, first-mated females may face reduced paternal care
from males attempting to attract additional mates. For example, already-mated male
European starlings (Sturnis vulgaris) with a surplus cavity added to their territory
increased mate attraction effort and contributed less to incubation than did males with
only a single cavity (Smith 1995). The apparent aggressiveness of resident females
towards female intruders in many species with bi-parental care suggests that females may
engage in a form of mate guarding to insure male parental investment (Slagsvold and
Litjeld 1994). In 21 Wyoming population of house wrens (T roglodytes aedon), females
sometimes sang when their mates engaged in courtship activities with additional females
(Johnson and Kermott 1990), perhaps to discourage secondary females from settling.

We investigated reproductive investment decisions in a secondary cavity nester,
the house wren. On our study site in southwest Michigan, house wrens are frequently
double-brooded, and within each brood approximately 5% of males are polygynous. In
other populations, male parental care during the nestling period has been shown to affect
female reproductive success (Johnson et al. 1992), and reduced paternal care contributes
to lower reproductive success of secondary females (Johnson et al. 1993). Male house
wrens frequently advertise for secondary females after pairing (Johnson and Kermott
1991b). Females might therefore be sensitive to male mate acquisition opportunity when
selecting a mate. We tested this hypothesis by manipulating cavity availability in male
house wren territories. We evaluated the effect of increased mate attraction opportunity
on parental investment by both males and females. Specifically, we asked whether

cavity availability affects (1) female mate choice or investment in clutch size and (2)

58

male parental care and mate attraction effort. Here we report our results on female mate
choice and maternal investment.
METHODS
Cavity manipulation

During April through August of 2001 and 2002, we conducted a cavity
manipulation experiment on a population of house wrens at the Kellogg Biological
Station’s Lux Arbor Reserve in Barry County, Michigan (see Chapter 3 for a complete
description of the study site and experimental manipulation). We controlled male access
to multi-cavity territories by manipulating nest boxes within male territories afier male
settlement but prior to pairing. Prior to male settlement, all territories contained three
nest boxes, but males had access to only a single nest box. The remaining nest boxes had
rubber stoppers placed in the entrance holes. After male settlement, but before female
settlement, we randomly assigned territories to single or surplus cavity treatments by
removing rubber stoppers from the nest boxes in surplus cavity territories (the pre-pairing
treatment). Consequently, females chose among males with a single cavity or multiple
cavities in their territories at settlement. This treatment was maintained through the
pairing and egg-laying periods, at which point we re-randomized treatments across
territories (the post-laying treatment, see Chapter 3).

In this paper we focus only on maternal reproductive investment decisions,
addressing mate choice, clutch size, and site fidelity between broods. We used the date of
female nest lining as an index of pairing date and calculated the number of days a male
remained unmated as the time between male settlement and nest lining. The number of

days a male remained unmated was used as an indicator of female mate choice. In cases

59

where we missed the first sign of lining and found a lined nest containing an egg on a
subsequent visit, we used the day between our nest checks as the nest-lining date. We
checked nests every other day to determine clutch size, and used the total number of eggs
laid as our measurement of clutch size, even if clutch size was reduced during the
incubation period. We excluded re-nests that occurred within the early or late season
breeding periods from our analyses. Therefore the data set included a maximum of one
early season nesting attempt and one late season nesting attempt for each individual
female. Nests that failed partway through the nesting period were included in analyses
based on data collected before the failure. Breeding females were captured during the
late incubation periods or nestling periods. In order to assess the relative size of females
settling in single and surplus cavity territories, we collected the following standard
morphological data at the time of capture: length of exposed culrnen, tarsus length, and
unflattened wing chord.
Statistical analyses

When data were approximately normally distributed we used parametric statistical
methods (ANOVAS with type-III sums of squares). We transformed data when necessary
to improve normality of residuals. All parametric analyses were conducted in SAS v.8.1
(SAS Institute Inc. 2000). When interactions were significant, we used the SLICE
command in lsmeans to test for differences between treatments within a brood. Data with
moderate violations of the assumptions of parametric tests (e. g., in cases in which
homogeneity of variance or normality of residuals could not be improved with data
transformations) were also analyzed with non-parametric equivalents. For settlement and

clutch size analyses, error variances between years were homogeneous (Fmax test, Kuehl

6O

1994) and we combined data from both years. When there were significant differences in
dependent variables between years, we standardized data within each year to a standard
normal distribution before combining the data. Standardizing data within each year
simplified our models by eliminating the Year effect and interactions with Year from our
models. For settlement and clutch size analyses, we treated females returning between
years as independent data points. Within a year, approximately 20-30% of adults were
returning breeders. On our study site, fourteen females attempted at least one brood in
both years. Analyses including returning females are presented, but we also ran these
analyses with each female included only once in the data set (randomly selecting which
entry to include). In all cases, these more conservative tests yielded the same results.

We analyzed female size data at early season nests only. We used MANOVAs,
combining years but including each female only once in the analysis (randomly selecting
which entry to use for returning females). We excluded exposed culmen length from the
MANOVA because the sign of the correlation between culmen length and the other
dependent variables differed among groups in some cases, violating key assumptions of
the analysis (Scheiner 1993). Therefore, we analyzed culmen length separately with a
univariate test. Excluding culmen length had no effect on significance of the model. For
all data, means :1: se are reported.

RESULTS
Female settlement

We found no differences in the date of nest lining (our index of the timing of

female settlement) between females settling in territories with three nest boxes or a single

nest box. We considered only the first nesting attempt of the season for each female.

61

The time a male remained unmated, that is the number of days between male settlement
and the date of nest lining, did not differ for males with surplus nest boxes in their
territories and males with a single nest box in their territories (sm'plus = 10.9 :1: 1.06 days
vs. single = 11.6 it 1.05 days, untransformed data; ANOVA: F131 < 0.01, P > 0.9, data
log transformed and standardized within years; Figure 4.1). In addition, there were no
differences in the Julian date of nest lining between surplus cavity and single cavity
territories (ANOVA: F132 = 0.48, P > 0.4, data standardized within years).

Females settling in territories containing surplus cavities were not larger than
females settling in single box territories at early season nests (MAN OVA including tarsus
and wing chord: Pillai’s trace value 0.044, F 2,55 = 1.3, P > 0.2; ANOVA for exposed
culmen: F159 = 1.84, P > 0.17; Table 4.1). We could not age females but could identify
returning breeders banded as adults the previous year (_>_ 2 years old) and first year
breeders banded as nestlings the previous year. We had no information about the age of
unbanded breeders. For analysis, we assumed that unbanded breeders were young birds
and lumped them with first year breeders. We found no difference in the relative age of
females mated to males with surplus nest boxes versus single nest boxes at early season
nests (X2 = 0.14, df= 1, P > 0.6; Table 4.2).

Clutch size

We included double-brooded females in our experiment. Although we re-
randomized treatments across experimental sites during the incubation period, thus giving
each territory an equal probability of being assigned to each treatment during the late
season, the presence of double-brooded females in our data set (included twice per year)

could reduce independence between data points. In fact, clutch sizes of double-brooded

62

females were correlated regardless of the experimental treatment (r = 0.37, P < 0.02,
clutch sizes standardized within each Year x Season combination). We addressed non-
independence of clutch sizes of double-brooded females in two ways. First, we looked
only at early season nests. This maximized our sample sizes for early season nests while
excluding any replicated second brood females. At early season nests, females mated to
males with surplus cavities laid larger clutches than those mated to males with single
cavities (surplus = 6.9 :t 0.12 eggs vs. single = 6.5 :1: 0.13 eggs, untransformed data;
ANOVA: F 1 ,72 = 6.6, P < 0.02, data standardized within years; Wilcoxon two-sample
test: IS = 2.5, P < 0.02; Figure 4.2). Second, in order to include both early and late season
nests in our model, we randomly assigned double-brooded females to either the first or
second brood, thereby including each female only once per year in the data set. In doing
so we decreased the sample sizes in each Treatment x Season combination and thus
reduced the power to detect treatment differences within each brood. There was a
significant Treatment x Season interaction (ANOVA: treatment, F199 < 0.01, P > 0.9;
season, F199 = 7.6, P < 0.01; treatment x season, F139 = 6.1, P < 0.02, data standardized
within years). For early season nests, we observed similar results to our previous
analysis, females mated to males with surplus cavities tended to lay larger clutches than
females mated to males with single cavities (surplus = 6.9 i 0.16 eggs, n = 28 vs. single =
6.5 i 0.17 eggs, n = 26, untransformed data; differences between treatments for early
season nests: P < 0.07, data standardized within years). At late season nests, differences
between treatments were not significant (surplus = 6.1 i 0.19 eggs, n = 19 vs. single =
6.5 d: 0.15 eggs, n = 30, untransformed data; differences between treatments for late

season nests: P > 0.1, data standardized within years).

63

Site fidelity

We looked at whether females mated to males with three box territories were
more likely to stay on the same territory for second broods than females mated to males
with one box territories. We included only successful early season nests because we
suspected that females re-nesting afier a failed attempt might use different decision rules
than females that produced a successful first brood. Females classified as having lefi the
territory after a first brood either produced a second brood on another territory or
disappeared from our study site during the late breeding season. Since the decision to
stay or leave a territory for the second brood occurred after the second experimental
manipulation, we classified females into four treatment groups (Table 4.3). There were
no differences in the proportion of females staying in the same territory for second broods
among treatment groups (X2 = 2.3, df = 3, P > 0.5). In fact, most females attempting
both early and late season nests on the study site remained on the same territory for both
broods (34 out of 43 females).

DISCUSSION

Models of polygyny ofien focus on the decisions of secondary females (i. e., when
a female should choose an already-mated male versus an unmated male). When
polygynous males divide parental care between nests, the decision of the secondary
female can also affect the first-mated female’s reproductive success. Therefore, early
settling females might be able to increase reproductive success if they could assess male
polygyny potential. In the case of secondary cavity nesters, cavity availability provides a

potential indicator of a male’s opportunity to attract additional mates.

For house wrens, paternal provisioning rates do not differ between the nests of
monogamous and primary females (Czapka and Johnson 2000). Polygynous males
preferentially care for primary nests, and secondary females suffer most of the costs of
mate-sharing (Johnson et al. 1993). In our study population, cavity availability does not
appear to affect paternal provisioning rates among males with a single mate (Chapter 3).
Thus, increased opportunities for polygyny might not negatively affect the success of
first-mated females. In fact, our data suggest the opposite: during the early part of the
breeding season, females mated to males with surplus cavities in their territories laid
larger clutches than females mated to males with a single cavity in their territories. Our
experimental manipulation randomized mating opportunity across differences in male
quality. Consequently, females settling in three box territories did so on the basis of
cavity availability, not on some aspect of male quality that might be correlated with
cavity availability under natural conditions, and the observed differences in clutch size
were not in response to male quality or condition per se. We hypothesize two possible
scenarios for increased maternal investment among females mated to males with surplus
cavities in their territories, both of which suggest that females expect to benefit from
mating with males with multiple cavities in their territories.

First, mate choice could result in non-random pairing with respect to cavity
availability. High quality females and/or females in good condition might have greater
access to attractive males or be more choosy in mate selection, thus giving more
attractive males access to higher quality mates (the differential access hypothesis: Burley
1986). For example, in a lekking species, black grouse (T etrao tetrix) females in better

condition paired with dominant males (Rintamaki et al. 1998). If females in better

65

condition or of higher quality are able to produce larger clutches, this could explain the
observed differences in clutch size between females mated to males with surplus versus
single cavities in their territories. Females mated to males with surplus cavities in their
territories were not larger than females mated to males with single cavities, but size alone
might not be a good predictor of female quality and/or condition.

Second, females mated to attractive individuals might contribute greater than
average parental investment to reproduction (the differential-allocation hypothesis:
Burley 1986). Females mated to high quality males have been shown to adjust
reproductive allocation to produce more eggs (Petrie and Williams 1993), lay larger eggs
(Cunningham and Russell 2000), deposit higher levels of testosterone in eggs (Gil et al.
1999), or participate in fewer extra-pair copulations (Burley et al. 1994). In order to
experimentally demonstrate differential allocation, we would have had to manipulate
cavity availability after pairing or have randomly allocated mates (Sheldon 2000). As
well as being logistically difficult in a field study, random allocation of mates would have
prevented us from assessing patterns of female settlement. The second (post-laying)
treatment manipulated cavity availability after pairing, but did so after female investment
in clutch size had been determined. As a result, we cannot assess the relative
contributions of female condition and differential allocation to investment in clutch size,
nor are they mutually exclusive.

Both of these hypotheses suggest that cavity availability might be a male status
symbol that females use in assessing mates. Alternatively, females might choose
territories for the surplus cavities themselves if, for example, the availability of additional

nest sites allowed faster renesting after nest failure. Plissner and Gowaty (1995) found

66

preferences for multi-box territories in Eastern bluebirds (Sialia sialis) before the onset of
breeding activity (mostly males and females visiting together), but no effect on
reproductive success. As a result of our experimental approach, on average, males
holding one box and three box territories should have been equivalent in terms of
condition and/or quality, but we do not know how they responded to cavity availability
after the pre—pairing experimental manipulation. We did not assess courtship intensity
among males but assumed that mate attraction effort by unmated males would be similar
regardless of cavity availability. However, at least in fishes, there is some evidence that
resource attractiveness can affect mate selectivity by males. Male beaugregory
damselfish (Stegastes leucosticus) provided with attractive breeding territories invested
more in courtship and adjusted courtship intensity to female quality (Itzkowitz and Haley
1999). We did observe prospecting females in both single cavity and surplus cavity
territories and have no a priori reason to believe that unmated males with surplus cavity
territories invested more in mate attraction than unmated males with single cavity
territories.
Timing of female settlement

Although female house wrens mated to males with surplus cavities laid larger
clutches than females mated to males with a single cavity during early nests, we did not
observe differences in the timing of female settlement. If females in better condition
were able to acquire “preferred” males or territories (here those with three nest boxes),
we would have expected males with surplus nest boxes in their tenitories to pair earlier
than males with single nest box territories. One possibility is that nest building activities

such as nest lining are a poor index of settlement date (Stutchbury and Robertson 1987,

67

but see Dunn and Harmon 1992). Historically, the primary female contribution to nest
building in house wrens has been thought to be the lining of the nest shortly afier pairing
(Kendeigh 1952, McCabe 1965), but recent evidence suggests that females may
contribute significantly to nest building prior to lining of the nest (Alworth and Scheiber
2000). If females differ in their contribution to nest building, then the correlation
between settlement date and line date might be weak. In addition, if older or better
condition females are able to displace resident females after pairing (e. g., Alworth and
Scheiber 1999) then our measure of the timing of female settlement would not be good
indicator of either female quality or mate choice preferences. We were rarely able to
identify individual females before the start of egg-laying, and we could have missed such
displacements had they occurred.

Females might be prevented from actively displaying preferences among males
during the settlement period if there are high costs associated with continued mate search.
For example, competition among females for access to unmated males might result in
females losing opportunities to pair as first-mated females if they extend their search
effort. Costs of mating as a secondary female can be quite high in terms of reduced male
assistance in rearing young (Johnson et al. 1993). Additionally, females might use cavity
availability as only one of several components in assessing mates. For example, female
sedge warblers (Acrocephalus schoenobaenus) use both song and territory cues in
assessing mates (Buchanan and Catchpole 1997). Our experimental design would have
de-coupled any correlations between cavity availability and male quality that could occur
under natural conditions. In order to examine this possibility, we set up a modified

replicate of the experiment in 2003, in which we assigned territories to one or three box

68

treatments prior to male arrival. Males had access to both one box and three box
territories at the time of settlement, but we found no preferences among males for
territories containing multiple nest boxes, nor did we find a significant difference in date
of nest lining between treatments (unpublished data). Males with surplus cavities in their
territories tended to provision more than males with single cavities (38% vs. 26% of
provisioning trips were by males), but differences were not significant (unpublished
data). Under natural conditions, nest site quality would also vary between territories. In
a nest box study, Johnson and Searcy (1993) found that female house wrens choose mates
at least partly on characteristics of the nest site. On our site, we installed identical nest
boxes in all territories. If females used cavity quality, rather than number of cavities, as
the major criterion in settlement decisions, we would not expect differences in the timing
of female settlement among surplus and single cavity territories containing identical nest
boxes. Post-pairing reproductive investment decisions might still be based on cavity
availability.
Site fidelity between broods and late season nests

Cavity availability had no effect on the likelihood that a female would breed in
the same territory for a second brood. Most double-brooded females retained the same
territory for second broods even though treatments had been randomly re-assigned to
territories prior to initiation of the second brood. Possibly, females did not reassess
territory and/or mate quality between broods, or they used different factors in settlement
and reproductive investment decisions for late season broods. In an Illinois population of
house wrens, mate switching was unrelated to reproductive success of first broods

(Drilling and Thompson 1991 ). Mate switching and territory retention may be more a

69

consequence of which parent has the first opportunity to desert the young and begin a
subsequent breeding attempt.

At late season nests, we found no between-treatment differences in clutch size,
but this could be due to the timing of our second experimental manipulation. We re-
randomized treatments during the incubation period of early broods and carried that
treatment through nest initiation and the laying period of second broods. As a result,
females that maintained the same territory for both broods had an equal probability of
being assigned to either treatment for the second brood. If the differences in clutch size
between treatments during early season nests resulted from differences in female
condition, then our re-randomization of treatments during the laying period decoupled
any potential correlation between treatment and female condition during late season
broods. Double-brooded females laying large first brood clutches were more likely to lay
large second brood clutches regardless of whether they switched treatments for the
second brood. This suggests that female condition contributes, at least in part, to
observed differences in early season clutch sizes between treatments, but does not
exclude the possibility that differential allocation afier pairing is also occurring. Further
experiments are necessary to determine the relative contributions of pre-pairing
(differential access) and post-pairing (differential allocation) decisions to maternal
reproductive investment in this system.

In conclusion, our results show that female house wrens mated to males with
surplus cavities in their territories laid larger early season clutches than females mated to
males with only a single cavity in their territories, suggesting that female house wrens use

cavity availability in male territories to assess male quality. Higher quality males might

70

have greater access to territories containing multiple nest sites; therefore female choice
might be based on indirect benefits. Alternatively, higher quality males able to secure
territories with surplus nest sites under natural conditions might also contribute more
paternal care, thus providing females with direct benefits. Both scenarios could result in
female preferences for males with surplus cavities in their territory and lead to
differences in clutch size between males with surplus and single cavities in their
territories.
ACKNOWLEDGMENTS

NSD was supported by a NSF Graduate Research Fellowship, a MSU
Distinguished Fellowship, the KBS Research Training Grant funded by NSF, the
departments of Zoology and EEBB (MSU), two Sigma Xi Grants-In-Aid-of—Research
Awards, a Lauff Research Award (KBS) and Wallace Award (MSU/Zoology). Field and
technical assistance were provided by E. Dubois, S. Dubois, J. Glazer, T. Grattan,

M. Homish, J. Johnson, C. Noud, T. Poloskey, and E. Zontek.

71

LITERATURE CITED

Alworth, T. and I.B.R. Scheiber. 1999. An incident of female-female aggression in the
House Wren. Wilson Bulletin 111: 130-132.

Alworth, T. and I.B.R. Scheiber. 2000. Nest building in House Wrens (T roglodytes
aedon): a reexamination of male and female roles. Journal of Field Ornithology
71: 409-414.

Andersson, M. 1994. Sexual Selection. Princeton University Press, Princeton, NJ.

Buchanan, KL. and CK. Catchpole. 1997. Female choice in the sedge warbler,
Acrocephalus schoenobaenus: multiple cues from song and territory quality.
Proceedings of the Royal Society of London Series B 264: 521-526.

Burley, N. 1986. Sex-ratio manipulation in color-banded populations of zebra finches.
Evolution 40: l 191 -1206.

Burley, N.T., D.A. Enstrom, and L. Chitwood. 1994. Extra-pair relations in zebra finches:
differential male success results from female tactics. Animal Behaviour 48: 1031-
1041.

Cunningham, E.J.A. and AF. Russell. 2000. Egg investment is influenced by male
attractiveness in the mallard. Nature 404: 74-77.

Czapka, SJ. and LS. Johnson. 2000. Consequences of mate sharing for first-mated
females in a polygynous songbird, the house wren. Wilson Bulletin 112: 72-81.

Darwin, C. 1871. The Descent of Man, and Selection in Relation to Sex. Murray,
London.

Davies, NE. 1989. Sexual conflict and the polygamy threshold. Animal Behaviour 38:
226-234.

Drilling, NE. and CF. Thompson. 1991. Mate switching in multibrooded House Wrens.
Auk 108: 60-70.

Dunn, PO. and SJ. Harmon. 1992. Effects of food abundance and male parental care on
reproductive success and monogamy in Tree Swallows. Auk 109: 488-499.

Gil, D., J. Graves, N. Hazon, and A. Wells. 1999. Male attractiveness and differential
testosterone investment in zebra finch eggs. Science 286: 126-128.

Hill, GE. 1991. Plumage coloration is a sexually selected indicator of male quality.
Nature 350: 337-339.

72

Itzkowitz, M. and M. Haley. 1999. Are males with more attractive resources more
selective in their mate preferences? A test in a polygynous species. Behavioral
Ecology 10: 366-371.

Johnson, LS. and L.H. Kermott. 1990. Structure and context of female song in a North-
temperate population of House Wrens. Journal of Field Ornithology 61: 273-284.

Johnson, LS. and L.H. Kermott. 1991a. Effect of nest-site supplementation on
polygynous behavior in the House Wren (T roglodytes aedon). Condor 93: 784-
787.

Johnson, L.S. and L.H. Kermott. 1991b. The functions of song in male house wrens
(T roglodytes aedon). Behaviour 116: 190-209.

Johnson, L.S., M.S. Merkle, and L.H. Kermott. 1992. Experimental evidence for
importance of male parental care in monogamous House Wrens. Auk 109: 662-
664.

Johnson, L.S., L.H. Kermott, and MR. Lein. 1993. The cost of polygyny in the house
wren T roglodytes aedon. Journal of Animal Ecology 62: 669-682.

Johnson, LS. and WA. Searcy. 1993. Nest site quality, female mate choice, and
polygyny in the house wren (T roglodytes aedon). Ethology 95: 265-277.

Kempenaers, B. 1995. Polygyny in the blue tit: intra- and inter-sexual selection. Animal
Behaviour 49: 1047- 1064.

Kendeigh, SC. 1952. Parental care and its evolution in birds. Illinois Biological
Monographs 22: 1-356.

Kuehl, RC. 1994. Statistical Principles of Research Design and Analysis. Duxbury Press,
Belmont, California.

Litjeld, J.T. and T. Slagsvold. 1989. Allocation of parental investment by polygynous
Pied Flycatcher males. Ornis Fennica 66: 3-14.

Litjeld, J .T. and T. Slagsvold. 1991. Sexual conflict among polygynous pied flycatchers
feeding young. Behavioral Ecology 2: 106-115.

McCabe, RA. 1965. Nest construction by House Wrens. Condor 67: 247-256.

Maller, AP. 1994. Male ornament size as a reliable cue to enhanced offspring viability in
the barn swallow. Proceedings of the National Academy of Sciences of the USA
91: 6929-6932.

Part, T. and A. Qvarnstrom. 1997. Badge size in collared flycatchers predicts outcome of
male competition over territories. Animal Behaviour 54: 893-899.

73

Petit, L]. 1991. Experimentally induced polygyny in a monogamous bird species:
prothonotary warblers and the polygyny threshold. Behavioral Ecology and
Sociobiology 29: 177-187.

Petrie, M. and A. Williams. 1993. Peahens lay more eggs for peacocks with larger trains.
Proceedings of the Royal Society of London Series B 251: 127-131.

Pinxten, R. and M. Eens. 1994. Male feeding of nestlings in the facultatively polygynous
European starling: allocation patterns and effect on female reproductive success.
Behaviour 129: 113-140.

Plissner, J .H. and PA. Gowaty. 1995. Eastern Bluebirds are attracted to two-box sites.
Wilson Bulletin 107: 289-295.

Qvamstrom, A., S.C. Griffith, and L. Gustafsson. 2000. Male-male competition and
parental care in collared flycatchers (F icedula albicollis): an experiment

controlling for differences in territory quality. Proceedings of the Royal Society of
London Series B 267: 2547-2552.

Rintamaki, P.T., A. Lundberg, R.V. Alatalo, and J. Hoglund. 1998. Assortative mating
and female clutch investment in black grouse. Animal Behaviour 56: 1399-1403.

Sandell, M.I., H.G. Smith, and M. Bruum. 1996. Paternal care in the European starling,

Sturnis vulgaris: nestling provisioning. Behavioral Ecology and Sociobiology 39:
301-309.

Scheiner, SM. 1993. MANOVA: multiple response variables and multispecies
interactions. Pp. 94-112 in Design and Analysis of Ecological Experiments
(Scheiner, SM. and J. Gurevitch, eds.). Chapman and Hall, New York.

Sheldon, BC. 2000. Differential allocation: tests, mechanisms and implications. Trends
in Ecology and Evolution 15: 397-402.

Slagsvold, T. and J .T. Lifjeld. 1994. Polygyny in birds: the role of competition between
females for parental care. American Naturalist 143: 59-94.

Smith, H.G. 1995. Experimental demonstration of a trade-off between mate attraction and
paternal care. Proceedings of the Royal Society of London Series B 260: 45-51.

Smith, H.G. and MI. Sandell. 1998. Intersexual competition in a polygynous mating
system. Oikos 83: 484-495.

Stutchbury, B.J. and R]. Robertson. 1987. Do nest building and first egg dates reflect
settlment patterns of females? Condor 89: 5 87-593.

74

Table 4.1. Morphological measurements for females settling in territories containing

surplus and single cavities at the beginning of the breeding season.

 

 

 

Morphological variable Surplus box territory Single box territory
Tarsus length (mm)21 19.6 i 0.11 19.8 :t 0.12
Unflattened wing chord (mm)a 48.8 i 0.31 49.3 d: 0.32
Exposed culmen length (mm)b 12.4 :t 0.12 12.1 :t 0.13
aMANOVA ns
bANOVAns

75

Table 4.2. Numbers of females in “young” and “old” age classes that settled in territories
containing surplus cavities versus single cavities. There were no significant differences

among treatments (X2 = 0.14, df= 1, P > 0.6).

 

Relative female age

 

Young Old
Pre-pairing treatment first year return or new breeder adult return (2 2 years old)

 

Surplus cavity territory 22 17
Single cavity territory 19 12

 

76

Table 4.3. Numbers of females rearing successful first broods that retained the same
territory versus left the territory for the second brood. Females leaving the territory
attempted a second brood elsewhere on the study site or left the study site. There were no

significant differences among treatments (X2 = 2.3, df = 3, P > 0.5).

 

TREATMENT

 

(pre-pairing/post-laying) Retained territory Left territory
Surplus cavities/surplus cavities 7 5
Surplus cavities/single cavity 5 9
Single cavity/single cavity 11 7
Single cavity/surplus cavities 5 5

 

77

0.5

 

 

 

 

 

E 0.25 -
:3
9.
'8 42
v 41
3’ o - ............................... f .............
"d
O.)
N
E
«3 —h.
'8
S
‘0 -0.25 -
-O. 5 mean 1 se
Surplus cavity Single cavity
territory territory

Figure 4.1. Number of days male house wrens with surplus and single cavities in their
territories remained unmated (data log transformed and standardized within years) at the
start of the breeding season. Means were not significantly different (P > 0.9). Sample

sizes are indicated.

78

0.6

 

 

 

 

 

 

0.4 - T
3?
g 0.2 -
.2 __
U
L5
U
'8 0 4 ......................................................
N
'6
23‘
g __
9
m -0.2 -
.35
-0.4 ~
_0 6 mean :1: se
Surplus cavity Single cavity

territory territory

Figure 4.2. The effect of cavity availability on standardized clutch size at early season

nests. Means were significantly different (P < 0.02). Sample sizes are indicated.

79

CHAPTER 5

NESTLING SEX RATIOS VARY WITH A MALE TERRITORY
CHARACTERISTIC IN HOUSE WRENS

with E. Dale Kennedy and Thomas Getty

ABSTRACT

Sex allocation theory predicts that individuals should bias offspring sex ratios
when doing so increases parental fitness. In many cases, females exhibit context-
dependent strategies for biasing offspring sex ratios. For example, females paired with
attractive mates might bias offspring towards males if attractive males have attractive
sons. Females in good condition might bias offspring sex ratio towards sons if males
have higher variance in reproductive success. Or, in cases of sex-specific natal
philopatry, females rrright bias offspring sex ratios towards the sex that disperses in order
to reduce competition. We manipulated the availability of surplus nest sites, a potential
signal of male quality, in male house wren territories. Females mated to males with
surplus nest boxes in their territories produced clutches with higher sex ratios (proportion
males) than females mated to males with single nest boxes in their territories, although
the population offspring sex ratio was not biased. If this facultative adjustment of
offspring sex ratios is adaptive, it suggests that the reproductive value of sons and
daughters differs for females mated to males with surplus nest sites and females mated to
males with single nest sites in their territories.

INTRODUCTION
Fisher (1930) postulated that parents should invest equally in sons and daughters

such that, when the costs of producing male and female offspring were equal, offspring

80

 

sex ratios should be 0.5. Fisher’s principle of equal allocation predicts that the optimal
offspring sex ratio is the same for all females. Trivers and Willard (1973) argued that
selection would favor a conditional strategy of sex ratio allocation when the opportunity
costs of investing in sons and daughters vary with parental condition. For example,
parents reproducing under favorable conditions might increase lifetime reproductive
success by biasing offspring sex ratios towards males when sons in good condition
provide higher inclusive fitness benefits to parents than daughters in good condition. The
Trivers-Willard hypothesis was fiamed in terms of polygynous species, where males have
higher variance in reproductive success than females, but condition-dependent sex
allocation can be applied to any situation in which parents can increase inclusive fitness
by biasing offspring sex ratios (Chamov 1982). For birds, offspring sex ratios have been
shown to vary with parental condition (Bradbury and Blakey 1998), resource availability
(Kilner 1998), and the attractiveness of mates (Sheldon et al. 1999).

House wrens (T roglodytes aedon) are facultatively polygynous obligate cavity
nesters. Offspring sex ratios vary with maternal condition (Whittingham et al. 2002),
hatch order (Albrecht 2000), and mating status (Albrecht and Johnson 2002), and are
biased in the direction predicted by the Trivers-Willard hypothesis. Specifically, these
studies have found that females in poor condition or rearing offspring under unfavorable
conditions (e. g., secondary females receiving reduced paternal assistance or offspring
within the clutch that are at a competitive disadvantage) bias offspring towards females.
In previous work (Chapter 4), we found that, during first broods, female house wrens
mated to males with surplus cavities in their territories laid larger clutches than females

mated to males with single cavities in their territories. We suggested that the observed

81

differences in clutch size might result from female preferences for males able to secure
multi-cavity territories (a potential signal of male quality). Females in better condition
might have greater access to males with surplus cavities in their territories and favor
production of good condition male offspring. Additionally, females mated to attractive
males might bias sex ratios towards males if sons inherit attractiveness or viability from
their fathers. These scenarios are not mutually exclusive, but in both cases we would
expect females mated to males with surplus nest boxes in their territories to produce more
male-biased clutches than females mated to males with single cavities in their territories.
In this study, we tested the hypothesis that female house wrens bias offspring sex ratios in
response to the availability of surplus cavities in their territories.
METHODS
Cavity manipulation
During April through August of 2001 and 2002, we conducted a cavity
manipulation experiment on a population of house wrens at the Kellogg Biological
Station’s Lux Arbor Reserve in Barry County, Michigan (see Chapter 3 for a complete
description of the study site and experimental manipulation). We controlled male access
to multi-cavity territories by manipulating nest boxes within male territories after male
settlement but prior to pairing. Prior to male settlement, all territories contained three
nest boxes, but males had access to only a single nest box. The remaining nest boxes had
rubber stoppers placed in the entrance holes. Afier male settlement, but before female
settlement, we randomly assigned territories to single or surplus cavity treatments by
removing rubber stoppers from the nest boxes in surplus cavity territories (the pre-pairing

treatment). Consequently, females chose among males with a single cavity or multiple

82

cavities in their territories at settlement. This treatment was maintained through the
pairing and egg-laying periods, at which point we re-randomized treatments across
territories (the post-laying treatment, see Chapter 3). We captured breeding females once
during the breeding season and marked individuals with a unique combination of color
bands. Maternal provisioning rates were observed in each territory on brood days 4, 8,
and 12. On brood day 12 we also measured nestling mass, right tarsus length, and right
unflattened wing chord.
Offspring sex identification

We collected blood samples from first-brood nestlings on brood day 12. We used
a sterile 27G needle to puncture the brachial vein and collected 5-40 uL blood samples in
heparinized microcapillary tubes. We then transferred samples to microcentrifuge tubes
containing 1 mL of Queen’s Lysis Buffer (Seutin et al. 1991) and stored samples at 4°C
until analysis. During J une-October 2003, we extracted DNA from blood samples using
the QIAarnp DNA Mini Kit (Qiagen Inc., Valencia, CA). We used the PCR-based
technique described by Fridolfsson and Ellegren (1999) to determine the sex of nestlings,
but modified the annealing temperature specified in the protocol to 46°C or 47°C,
depending on the therrnalcycler we were using. A single set of primers was used to
amplify homologous sections of the CHD-l gene on the avian sex chromosomes. PCR
products were electrophoresed at 90 V on 3% agarose gels for 60-75 minutes and stained
with ethidium bromide. When visualized under UV light, males (ZZ) were identified by
the presence of a single band approximately 650 bases in size, and females (ZW) were
identified by the presence of two bands approximately 650 and 475 bases in size. We

sexed 247 nestlings from 43 first brood nests. We validated the sexing procedure by

83

identifying 12 adults of known sex (6 males and 6 females) using this protocol. All were
sexed correctly. In addition, we recaptured 14 breeding adults in subsequent years that
had been sexed as nestlings. We independently sexed these adults based on behavior and
morphological characteristics in the field and then matched them to the results of our
molecular sex analysis. Thirteen had been sexed correctly. The identity of the fourteenth
could not be determined due to a transcription error in recording keeping.
Statistical analysis

Analyses were conducted in SAS v.8.l (SAS Institute Inc. 2000). To test whether
the experimental treatment affected brood sex ratios, we used generalized linear models
(PROC GENMOD) with a logit link fiinction and binomial errors. We used number of
male nestlings as the dependent variable and the number of nestlings in the brood as the
binomial denominator. We assessed model fit by looking at the residual deviance, which
is distributed asymptotically as X2, and checked standardized residuals for normality.
The data were slightly underdispersed (residual mean deviance = 0.67), which leads to
conservative tests (Wilson and Hardy 2002). Therefore, we also ran the analysis after
rescaling the variance by s (Pearson’s X 2 divided by the residual df, Wilson and Hardy
2002). As both methods produced similar results, we report statistics from the latter
analysis. Offspring sex ratios are reported as proportion males. Means (i se) are derived
from the parameters fitted during statistical analysis. We assessed deviations of
treatment means from 0.5 using replicated Goodness-of-F it tests (Sokal and Rohlf 1995).
In order to calculate the G statistic for the Goodness-of-F it test, we had to exclude one
single-box territory nest that contained no males and three females. The excluded nest

was an incompletely sexed brood (only three nestlings out of seven eggs survived to BD

84

12), and its removal had little effect on the average offspring sex ratio for single box
territories. Nonparametric tests of clutch size were conducted with Wilcoxon two-sample
tests. We sexed offspring from a subset of nests included in the experiment. Therefore
sample sizes vary among analyses.

We transformed data where necessary to improve normality and reduce
heteroscadiscity of data sets analyzed with parametric tests. We used MANOVAs
(PROC GLM) to analyze nestling size data. We compared the average size of male and
female nestlings within a nest in order to avoid pseudoreplication, pairing the data by nest
in order to control for maternal effects. Because several nests contained only one female
or one male, we also analyzed sex differences by comparing sizes for a randomly selected
single male and single female nestling from each nest. When MANOVAs were
significant we used paired t-tests to examine the individual dependent variables. We
excluded one nest from this size analysis that contained only female nestlings.

We analyzed maternal provisioning data (repeated measures) using profile
analysis with MAN OVA. We included primary and monogamous females, but excluded
territories in which another pair of house wrens was able to usurp a ‘surplus’ nest box
(this occurred in five territories in 2002, either during the late nestling period of first
broods or during second broods). Since clutch sizes varied among treatments (Chapter
4), we looked at both total provisioning rates and per-nestling provisioning rates by
females. We were primarily interested in whether any differences in condition between
females settling in single box versus surplus box territories might contribute to female

provisioning effort and nestling condition. Therefore we included only the pre-pairing

85

treatment as the class variable in these analyses. Means for transformed data are back-
transformed into original units.

Females that returned between years were assumed to be independent data points
for analysis. Ten females produced first broods during both years, although of these,
only five returning females were included in the sex ratio analysis and seven in the
provisioning rate analyses. Analyses including returning females are presented, but
analyses were also conducted with each female included only once in the data set
(randomly selecting which entry to include). These more conservative tests yielded
similar results.

RESULTS
Offspring sex ratio

Offspring sex ratio on brood day 12 was significantly higher in territories
containing surplus cavities than territories containing single cavities (surplus vs. single,
sex ratio = 0.57 :t 0.035 vs. 0.44 d: 0.036;F1,41 = 6.2, P < 0.02; Figure 5.1), but neither
offspring sex ratio was significantly different from 0.5 on its own (replicated G-tests:
surplus cavities, Gr: 17.0, df= 21, P > 0.7; single cavities, GT: 10.5, df= 21, P > 0.9).
Combining treatments, average brood sex ratios were close to unity in both years
(0.50 d: 0.033 in 2001 and 0.52 :t 0.041 in 2002). We were unable to sex unhatched eggs
and offspring that did not survive to brood day 12. Sixty percent of nests (26 out of 43)
contained at least one unsexed offspring, either due to unhatched eggs or unsexed
nestlings. We sexed all of the hatched young in 72% of nests. Neither the occurrence of
nests with at least one unhatched egg nor at least one unsexed hatchling differed between

treatments (unhatched eggs: X 2 = 0.18, df= 1, P > 0.6; unsexed hatchlings: X 2 = 2.0,

86

df= 1, P > 0.15). On average 1.1 i 0.26 and 1.0 i: 0.24 offspring (including unhatched
eggs) were unsexed in surplus and single cavity territories respectively. When we limited
the analysis to nests for which we sexed all hatched young, offspring sex ratios still
tended to be higher in territories containing surplus cavities (F130 = 3.8, P < 0.07). Of
the nests included in the sex ratio analysis, mean clutch sizes were higher in territories
with surplus cavities (7.1 :1: 0.18 eggs) than in territories with single cavities (6.6 :t 0.12
eggs) but differences were not significant (Wilcoxon two-sample test: t, = 1.9, P < 0.07).
These nests are a subset of those included in the analyses of clutch size in Chapter 4.
Sex differences in nestling size

Within broods, the average size of male offspring was significantly larger than
average size of female offspring (MANOVA controlling for brood effects: F 3,39 = 7.8,
P < 0.001). Male nestlings had significantly longer tarsi (paired t = 4.9, df = 41,
P < 0.0001) and wings (paired t = 2.3, (if = 41 , P < 0.03) and tended to be heavier (paired
t = 1.9, df = 41 , P < 0.07) than female nestlings. We obtained similar results when we
compared a single randomly selected male and female nestling from each nest
(MANOVA controlling for brood effects: F139 = 3.6, P < 0.03; tarsus: paired t = 3.0,
df= 41, P < 0.01, wing chord: paired t = 1.8, df= 41, P < 0.09, mass: paired t = 2.5,
df = 41, P < 0.02; Table 5.1). But these differences were small (males were less than
3.1% larger than females for all morphological variables). We were unable to sex very
small runts, and they were excluded from this analysis.

Maternal provisioning effort
We compared provisioning rates of females that settled in single nest box versus

surplus nest box territories (excluding the post-laying treatment from our analysis). Total

87

provisioning rates were slightly higher for females that settled in surplus box territories,
but differences were not significant (Table 5.2, Figure 5.2). We found no differences in
female provisioning rates per nestling between females that settled in single nest box
versus surplus nest box territories (Table 5.3, Figure 5.3).
DISCUSSION

As a result of our experimental manipulation, on average, territories in the two
treatments should not have differed in male or territory quality, with the exception of
cavity availability. Differences in female reproductive investment between the two
treatments were likely due to differences in cavity availability. Although we had no
reason to believe a priori that the behavior of unmated males would differ between
treatments, it is possible that males altered their behavior in response to increased cavity
availability, and that female investment might have been a response to male behavior.
For instance, males with surplus cavities added to their territories might have increased
investment in mate attraction in response to a perceived increase in territory quality. As a
result, differences in female reproductive investment could have been an indirect
response to cavity availability via male mate attraction effort, rather than a direct
response to the cavities themselves. Regardless of the specific signal used by females,
we would predict male-biased offspring sex ratios in territories containing surplus nest
sites if, under natural conditions, males of higher phenotypic quality were more likely to
secure multi-cavity territories and/or females in better condition settled in territories
containing surplus cavities.

Under natural conditions, male house wrens compete for cavities and males may

physically challenge one another through chases and physical fights (Kendeigh 1941,

88

Johnson 1998). Polygynous house wrens have greater reproductive success than
monogamous males (Soukup and Thompson 1998). Since opportunities for polygyny are
limited by cavity availability, intense competition among males for access to territories
containing multiple nest sites might lead to a correlation between male quality and cavity
availability. As a result, cavity availability could indicate male attractiveness or quality.
If attractiveness or viability were inherited from fathers, and those characters influence
sons more than daughters, then females mated to attractive or high quality males might
produce male biased clutches (e. g., Ellegren et al. 1996, Svensson and Nillson 1996,
Sheldon et al. 1999).

Female preferences for high quality males might also lead to females in better
condition having greater access to males with surplus nest boxes in their territories, for
example if females in better condition return first to the breeding grounds (e. g., Moller
1991). When maternal condition affects offspring condition and has a greater negative
effect on the future reproductive success of sons than daughters (such as in polygynous
species where males have higher variance in reproductive success), females in good
condition should produce male-biased sex ratios relative to those of females in poor
condition. In a Wisconsin population of house wrens, females in good condition
produced offspring in better condition and had more sons over the breeding season than
did females in poor condition (Whittingham et al. 2002). The observed male bias in
offspring produced by females in good condition was due primarily to male-biased
second broods (Whittingham et al. 2002). We observed differences in offspring sex
ratios during first broods, but we found no differences in the size of females settling in

single box and surplus box territories (Chapter 4) or in female provisioning effort

89

between treatments. Nevertheless, we cannot rule out the possibility that the observed
relationship between cavity availability and offspring sex ratio might be mediated by
differences in female condition between treatments, rather than a direct response to the
availability of surplus nest boxes in male territories.

We found that male nestlings were larger in size than female nestlings, although
differences were small. If one sex is more costly to produce, then females in poor
condition might be restricted to producing the cheaper sex (Myers 1978). While we
could not measure sex-specific provisioning rates, females that settled in single box
territories did not provision nestlings at lower rates than females that settled in surplus
box territories. Other studies have not found evidence for sex-specific provisioning rates
(Albrecht and Johnson 2002). These results do not suggest sex-specific differences in
costs during the nestling period, although costs could manifest themselves during egg
formation if males hatch from larger eggs (Cordero et al. 2000) or from eggs containing
more androgens (Petrie et al. 2001).

Most nests (56%) contained at least one unhatched egg and 28% contained at least
one unsexed young (usually due to mortality before brood day 12), but neither the
frequency of unhatched eggs or unsexed offspring differed between pre-pairing
treatments. Kilner (1998) suggested that food availability might affect the direction of
sex-biased mortality in zebra finches (T aeniopygia guttata). If house wrens are capable
of identifying the sex of nestlings, then selective brood reduction might be possible. In
our study, mortality of hatched young was low (less than 5%). Even when we limited our
analyses to nests from which all hatchlings were sexed, we observed the same patterns in

offspring sex ratio between females mated to males with surplus versus single cavities in

90

their territories. Therefore, it seems unlikely that the observed differences in offspring
sex ratios were due to sex-biased mortality. However, we do not know if sex bias
occurred at laying or whether it might have resulted from differential hatch success.
Since there were no differences in the occurrence of hatch failure between nests in single
cavity and surplus cavity territories, hatch failure would have had to be male biased in
single box territories and female biased in surplus cavity territories in order to account for
the observed differences in offspring sex ratio. If sons are more costly to produce than
daughters, male-biased mortality might be more likely in single box territories if those
females were in poorer condition and incubated less than females mated to males with
surplus cavities. It is more difficult to explain why embryo mortality would be female
biased in surplus cavity territories.

Costs of male and female offspring might differ if sex-specific natal philopatry
resulted in competition between parents and philopatric young for access to resources
(local resource competition: Clark 197 8). When natal philopatry differs between the
sexes, parents might reduce competition with offspring by producing more of the
dispersing sex (Gowaty 1993). Reported rates of natal philopatry are low for house
wrens (generally less than 5%) with males returning more frequently than females
(Drilling and Thompson 1988). We observe similar return rates on our study site, and
approximately 60% of natal returns are males (Dubois, unpublished data). Male-biased
offspring sex ratios in surplus nest box territories could be consistent with local resource
competition if females residing in territories with surplus nest boxes perceived reduced
nest site competition (i. e., there were lots of available nest sites) and invested more in the

philopatric sex than females residing in single box territories. However, low return rates,

91

combined with small differences in return rates between male and female offspring make
it unlikely that local resource competition would have a large impact on biases in
offspring sex ratios.

We have suggested that female preferences for males with territories containing
surplus cavities account for the observed differences in offspring sex ratio, either as a
result of assortative mating among preferred males (or males in preferred territories) and
females in good condition or as a result of differential investment in reproduction by
females mated to preferred males. In order for female condition to account for the
observed bias in offspring sex ratios, male offspring would need to be significantly more
costly to produce than females, or maternal condition would need to have a greater
impact on the future reproductive success of sons than daughters. Alternatively, if sons
of males with high polygyny potential are more likely to be polygynous themselves,
females mated to attractive mates might bias offspring towards males. Research on the
heritability of traits associated with male polygyny potential (or some other indirect
benefit inherited by sons), and whether cavity availability reliably signals male quality in
natural situations, would be needed to confirm this hypothesis.

In summary, we found that female house wrens mated to males with surplus
cavities in their territories produced clutches with higher offspring sex ratios than females
mated to males with single cavities in their territories. The observed sex ratio bias was
likely due to differences in the sex ratio at laying rather than differential mortality during
the nestling period. Our results are consistent with predictions of the Trivers-Willard
(1973) hypothesis based on differences in reproductive value of offspring under different

breeding conditions. Our results may thus have interesting implications for nest box

92

studies, especially those utilizing high densities of nest boxes for polygynous species.
Close spacing of nest boxes probably increases the occurrence of multi-cavity territories
compared to natural cavities and might have consequences on female reproductive
investment, including the sex ratio of offspring produced.

ACKNOWLEDGMENTS

NSD received support from a NSF Graduate Research Fellowship, MSU

Distinguished Fellowship, the KBS Research Training Grant, the departments of Zoology
and EEBB at Michigan State University, two Sigma Xi Grants-In-Aid-of-Research
Awards, and a Lauff Research Award. We thank E. Dubois, S. Dubois, J. Glazer, T.
Grattan, M. Homish, C. Noud, T. Poloskey, and E. Zontek for their long hours in the
field, S. J anota and C. Thompson for advice on blood collection methods, and the Albion
College Biology Department and the Kellogg Biological Station E-lab for generously
allowing us use of their laboratory facilities. The Morris lab in the Department of
Biological Sciences at Ohio University provided helpful comments on a draft of this

chapter.

93

LITERATURE CITED

Albrecht, DJ. 2000. Sex ratio manipulation within broods of house wrens, T roglodytes
aedon. Animal Behaviour 59: 1227-1234.

Albrecht, DJ. and LS. Johnson. 2002. Manipulation of offspring sex ratio by second-
mated female house wrens. Proceedings of the Royal Society of London Series B
269: 461-465.

Bradbury, RB. and J .K. Blakey. 1998. Diet, maternal condition, and offspring sex ratio
in the zebra finch, Peophila guttata. Proceedings of the Royal Society of London
Series B 265: 895-899.

Charnov, EL. 1982. The Theory of Sex Allocation. Princeton University Press,
Princeton.

Clark, AB. 1978. Sex ratio and local resource competition in a prosimian primate.
Science 201: 163-165.

Cordero, P.J., S.C. Griffith, J .M. Aparicio, and D.T. Parkin. 2000. Sexual dimorphism in
house sparrow eggs. Behavioral Ecology and Sociobiology 48: 353-357.

Drilling, NE. and CF. Thompson. 1988. Natal and breeding dispersal in House Wrens
(T roglodytes aedon). Auk 105: 480-491.

Ellegren, H., L. Gustafsson, and BC Sheldon. 1996. Sex ratio adjustment in relation to
paternal attractiveness in a wild bird population. Proceedings of the National
Academy of Sciences of the USA 93: 11723-11728.

Fisher, RA. 1930. The Genetical Theory of Natural Selection. Clarendon, Oxford.

Fridolfsson, A. and H. Ellegren. 1999. A simple and universal method for molecular
sexing of non-ratite birds. Journal of Avian Biology 30: 116-121.

Gowaty, RA. 1993. Differential dispersal, local resource competition, and sex ratio
variation in birds. American Naturalist 141: 263-280.

Johnson, LS. 1998. House Wren ( T roglodytes aedon). In The Birds of North America,
No. 380 (Poole, A. and F. Gill, eds.). The Birds of North America, Inc.,
Philadelphia, PA.

Kendeigh, SC. 1941. Territorial and mating behavior of the house wren. Illinois
Biological Monographs 18: 1-120.

Kilner, R. 1998. Primary and secondary sex ratio manipulation by zebra finches. Animal
Behaviour 56: 155-164.

94

Moller, AP. 1991. Preferred males acquire mates of higher phenotypic quality.
Proceedings of the Royal Society of London Series B 245: 179-182.

Myers, J .H. 1978. Sex ratio adjustment under food stress: maximization of quality or
numbers of offspring? American Naturalist 112: 381-388.

Petrie, M., H. Schwabl, N. Brande-Lavridsen, and T. Burke. 2001. Sex differences in
avian yolk hormone levels. Nature 412: 498.

Seutin, G., B.N. White, and RT. Boag. 1991. Preservation of avian blood and tissue
samples for DNA analyses. Canadian Journal of Zoology 69: 82-90.

Sheldon, B.C., S. Andersson, S.C. Griffith, and J. Omborg. 1999. Ultraviolet colour
variation influences blue tit sex ratios. Nature 402: 874-877.

Sokal, RR. and E]. Rohlf. 1995. Biometry: the Principles and Practice of Statistics in
Biological Research, third edition. W.H. Freeman and Company, New York.

Soukup, SS. and CF. Thompson. 1998. Social mating system and reproductive success
in house wrens. Behavioral Ecology 9: 43-48.

Svensson, E. and J.-A. Nillson. 1996. Mate quality affects offspring sex ratio in blue tits.
Proceedings of the Royal Society of London Series B 263: 357-361.

Trivers, R.L. and DE. Willard. 1973. Natural selection of parental ability to vary the sex
ratio of offspring. Science 179: 90-92.

Whittingham, L.A., S.M. Valkenaar, N.E. Poirer, and PO. Dunn. 2002. Maternal
condition and nestling sex ratio in House Wrens. Auk 119: 125-131.

Wilson, K. and I.C.W. Hardy. 2002. Statistical analysis of sex ratios: an introduction. Pp.
48-92 in Sex Ratios: Concept and Research Methods (Hardy, I.C.W., ed.).
Cambridge University Press, Cambridge.

95

Table 5.1. Size of male and female offspring on brood day 12.3

 

MORPHOLOGICAL VARIABLE Male offspring Female offspring

 

Mass (g) 10.5:1:0.10 10.2i0.12
Tarsus (mm) 19.5 is 0.08 19.2 i 0.09
Wing (mm) 34.5 d: 0.30 33.9 :t 0.38

 

3 Analysis comparing a randomly selected male and female nestling from each nest.

MANOVA significant at the P < 0.03 level. See text for alternate method of analysis.

96

Table 5.2. Repeated measures analysis of total maternal provisioning rates on brood days
4, 8, and 12 for females that settled in territories containing surplus nest boxes and single

nest boxes (pre-pairin g treatment) at early season nests.“

 

A. MANOVA (profile analysis) of within subject effects. Pillai’s trace values are

 

 

presented.
Source Num df Den df Value F P > F
Brood day (time) 2 54 0.67 54.1 0.0001
Brood day >< Treatment 2 54 0.01 0.27 0.7

 

B. Repeated measures ANOVA of between-subject effects.

 

 

Source df MS F P > F
Treatment 1 2.5 1 .7 0. 19
Error 55 1 .5

 

C. ANOVAS on each of the contrasts for time differences between successive
observations. Only the brood day effect is shown, others were non-significant.

 

 

Source df MS F P > F b
Contrast variable: BD8 — BD4
Brood day 1 56.3 69.8 0.0001
Error 55 0.8
Contrast variable: BD12 — BD8
Brood day 1 5.3 4.6 0.04
Error 55 1.1

 

a Data square root transformed. Nests with incomplete data were excluded.

b Bonferroni adjusted at = 0.025.

97

Table 5.3. Repeated measures analysis of maternal provisioning rates per nestling on
brood days 4, 8, and 12 for females that settled in territories containing surplus nest boxes

and single nest boxes (pre-pairing treatment) at early season nests.a

 

A. MANOVA (profile analysis) of within subject effects. Pillai’s trace values are

 

 

 

 

 

presented.

Source Num df Den df Value F P > F
Brood day (time) 2 54 0.67 55.1 0.0001
Brood day X Treatment 2 54 0.03 1.0 0.3
B. Repeated measures ANOVA of between-subject effects.

Source df MS F P > F
Treatment 1 0.03 0.48 0.4
Error 55 0.06

 

C. AN OVAs on each of the contrasts for time differences between successive
observations. Only the brood day effect is shown, others were non-significant.

 

 

Source df MS F P > F°
Contrast variable: BD8 — BD4
Brood day 1 2.3 78.1 0.0001
Error 55 0.03
Contrast variable: BD12 — BD8
Brood day 1 0.15 3.7 0.06
Error 55 0.04

 

a Data reciprocal square root transformed. Nests with incomplete data were excluded.

b Bonferroni adjusted at = 0.025.

98

 

0.75

0.5 « ..........................................

Offspring sex ratio

 

mean i se

 

 

0.25
Surplus cavity Single cavity
territory territory

Figure 5.1. Average offspring sex ratios (proportion males) of clutches produced by
females mated to males with surplus and single cavities in their territories. Means were
significantly different at the P < 0.02 level. Sample sizes are indicated above the

symbols.

99

 

H
N
I

 

 

 

 

,8.
E
E"
.9. 8 7
'E
a
Q.
2
t 4 _
L1.
0 I I I mean 1 se
4 8 12
Brood day

Figure 5.2. Average total maternal nestling provisioning rates during 30-minute
observation periods for females that settled in territories containing surplus nest boxes
(solid symbols) and single nest boxes (open symbols) through the laying period. Only

the time effect was significant (P < 0.0001). Sample sizes are indicated.

100

 

3.5

 

 

 

 

 

3.
on
.E “
a
8- 2.5
,8.
E
on
c:
’5‘
:a 2‘
>
E.
2
a 15 g

1 meanise
4 8 12

Brood day

Figure 5.3. Average maternal provisioning rates per nestling during 30-minute
observation periods for females that settled in territories containing surplus nest boxes
(solid symbols) and single nest boxes (open symbols) through the laying period. Only the

time effect was significant (P < 0.0001). Sample sizes are indicated.

101