.+.
a. a... s.

. “.3...
u

by:
‘Is
b

( rlun. . '
[43.1.93 #1. a
a: a»

H: x: .3
- "j
.5

. ....»|.. In)
:I {.12.}...
0......

.
6 . I.
:.5: 4.

5...... z.

x
z. 2 7!: .
$13.55

1.7..
. z...

1;: l
:9... u I)?
I ‘~.y )I.

 

 

.fikaflfi 1 . . :..5 is.
. .. . . . . .14.: . ! «.fl... thvrwvhm...c.,!nuuWHr4Ltu

 

 

 

THESIS

~_:~) \,‘
s '\ '

MICHIGAN STATE UNIVERSITY LIBRARIES

I I l WWIll/WIIWHIHIH ”Hill/ll

3 1293 01420 2810

ll

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

The Effects of Female Age on
Reproduction, Parental Care and
Growth of Young in Tree Swallows

(Tachycineta bicolor)

presented by

Patrick E. Lederle

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in ZOOIOQN

 

Wa’m

Major professor

 

Date 22 MIL fiyj

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE II RETURN BOX to remove thin chockout from your record.
TO AVOID FINES Mum on or More data duo.

l DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Intuition
W

ans-m

EFFECTS OF FEMALE AGE ON REPRODUCTION,
PARENTAL CARE AND GROWTH OF YOUNG IN TREE
SWALLOWS (TACHYCINETA BICOLOR)

BY

Patrick E. Lederle

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

DOCTOR OF PHILOSOPHY

Department of Zoology

1995

ABSTRACT

EFFECTS OF FEMALE AGE ON REPRODUCTION,
PARENTAL CARE AND GROWTH OF YOUNG IN TREE
SWALLOWS (TA CHYCINETA BICOLOR)

BY

Patrick E. Lederle

It has been argued that clutch size in altricial birds is limited by the parent’s
abilities to adequately feed nestlings. In many species of birds, younger females
breeding for the first time produce smaller clutches, and foraging deficiencies or
inexperience have been thought to be the determining factors resulting in smaller

clutches produced by younger females.

One-year Old (SY) females initiated nesting later, laid smaller clutches,
produced smaller eggs, and had a lower likelihood of hatching eggs, compared to older
(ASY) females. I compared patterns of parental care and the resulting growth of young
at four treatment groups characterized by age of female and brood size, to test the
hypothesis that SY females are constrained from producing larger clutches by
deficiencies in their abilities to care for young. Treatment groups were ASY nests with
four young, ASY with six, SY with four and SY with six. During 1990, nestlings at

SY nests with six young exhibited the poorest growth of any of the treatment groups

and this corresponded to the lowest rates of per young visitation rates by parents.
Mortality rates were also highest for this treatment group. In contrast, during 1993,
nestlings at ASY nests with six young received the lowest rate of per young visitation
and growth measures were much more uniform between treatment groups. Results from
1990 appear to support the hypothesis that SY females may be constrained from
producing larger clutches due to inadequate young rearing capabilities, yet the results
from 1993 do not. During both years, comparison of parental visitation rates and
measures of growth showed that within treatment groups, higher levels of visitation did
not always translate into higher quality growth. This is likely due to differences in
quality of parental care provided as opposed to visitation rates which only assess
quantity. Analyses of growth variables and load sizes of food delivered to young which
show significant effects due to the individual nest suggest that the performance of
individual pairs of adults is more important in determining reproductive success than

age of the female or brood size per se.

To Kate

iv

ACKNOWLEDGMENTS

I could not have undertaken and completed this study without the help and sup-
port from many individuals. In agreeing to serve as my Major Professor, Don Beaver
took on a dual role since he was also my boss on the ELF Project. Because of his will-
ingness to serve in this capacity, he played a major role in my professional develop-
ment as Research Coordinator for MSU-TVG, as well as contributing to the overall
success of my graduate program. I am grateful for all of the freedom and support he

provided throughout my involvement with both projects.

I am grateful to my research committee members Dick Hill, Steve Stephenson
and Don Straney, who were more than patient over the years. Dick Hill was especially

encouraging and insightful during several low periods in my studies.

The actual field work that was so prevalent in this study would never have been
completed without the able and willing assistance of many MSU-TVG employees. I
especially benefited from the direct efforts of Eileen Eliason (1990), Mark Nelson
(1991), Pat “Newt” Van Daele (1992), Joe Lipar (1993), and Dave Gauthier (1994).

Mark Nelson, the quintessential field worker, was the moving force behind the devel-

Opment of the blind system used to collect boluses. Jerry Burke suggested using the

“carny” method of modular design.

Many others from TVG also provided assistance: Darrin Bauer, Bryan Cherry,
Keith Davisson, Tim Klaes, Tom Knoedler, Andy Mullard, Chris Murray, John Nie-
woonder, Grigoris Papakostas, Tom Ryder, Rollin Sachs, Hilda Sexauer, Peter Smith,

Terry Trier, Brad White, Bill Wise, John Yunger and Mary Zaloga.

Ned Walker and Rich Merritt provided advice on collecting insects and devel-
opment of Tabanid traps, and their lab for drying and weighing insects. Karen Strickler
also provided a balance. Bryan Pijanowski shared with me his techniques for bolus

collections.

I benefited greatly from long discussions with Dan O’Brien on the art and sci-
ence of statistics. My analyses and interpretations were greatly strengthened due to his
advise and insights. Of course, any shortcomings are my own. Terry Trier also shared
insight on the repeated measures analysis of variance techniques, software troubleshoot-

ing, and the use of SYSTAT.

Jim Hammill of the Michigan Department of Natural Resources helped insure
that Panola Plains was developed into the research site that provided the Tree Swallows
for study. John Force taught us how to effectively map the research plots and laid out

buffer zones which also insured that the research area would be protected.

Financial support was provided by Sigma Xi—The Scientific Research Society,

The George and Martha Wallace Endowed Scholarship of the Department of Zoology,

vi

Graduate student funds from the Department of Zoology, and the Illinois Institute of
Technology Research Institute ELF Ecological Monitoring Program Subcontract num-

ber D06205-93-006 administered by D.L. Beaver.

I am especially grateful to my wife, Kate, who provided assistance with logis-
tics, database management and graphics during all phases of the study. Most impor-
tantly, Kate provided endless encouragement and support which helped me put my
graduate studies in proper perspective. As a true co-investigator, while I was at Panola
Plains considering the vagaries of parental care in Tree Swallows, she was conducting
practical long-term studies by providing parental care to our children, Daryl and

Eileen.

vii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................... xi
LIST OF FIGURES ........................................................................... xiv
INTRODUCTION ................................................................................. 1
The Effects Of Female Age on Breeding Performance ............................... 3
TREE SWALLOW NATURAL HISTORY .......................................... 10
METHODS ......................................................................................... 14
Study Sites ............................................................................................... 14

Egg Weights and Hatching Times ............................................................ 16

Adult Weights and Measures ................................................................... 18
Growth Of Young ...................................................................................... 20
Parental Care Measures .......................................................................... 21
Establishment Of Treatment Nests ........................................................... 25
Weight Of Food Delivered tO Young ......................................................... 26
Ambient Monitoring .................................................................................. 29
Statistical Analyses .......................................................................... . ....... 30
RESULTS .......................................................................................... 31

viii

AGE-RELATED FECUNDITY ........................................................................ 31

Percent SY Females In Population .......................................................... 31
Nest Initiation And Egg Laying ................................................................. 32
Clutch Size ............................................................................................... 35
Clutch Size Repeatability ......................................................................... 39
Egg Weights ............................................................................................. 42
Egg Weight Repeatability ......................................................................... 47
Likelihood TO Hatch ................................................................................. 47
Nesting Success Based on Exposure ...................................................... 50
PARENTAL CARE ......................................................................................... 53
Nest Visitation Rates ................................................................................ 53
Weight Of Food Delivered to Young ......................................................... 64
GROWTH OF YOUNG .................................................................................. 67
Growth Curves ......................................................................................... 67
Weight: Fitted Growth Constants ............................................................. 70
Weight: Maximum Values Attained .......................................................... 72
Wan Length: Fitted Growth Constants .................................................... 75
Wan Length: Maximum Values Attained ................................................. 77

RELATIONSHIPS BETWEEN PARENTAL CARE AND GROWTH OF

YOUNG ................................................................................................. . ........ 79
Mortality Rates of Treatment Group Young ............................................. 88
WEIGHT CHANGES IN PARENTS FEEDING YOUNG ................................ 90

ix

DISCUSSION .................................................................................... 92

AGE CLASS DIFFERENCES: PRE-HATCHING ........................................... 92
Clutch Size and Egg Weights ................................................................... 92
Hatching Success, Egg and Nest Failure ................................................. 98

THE ENERGETICS OF EGG PRODUCTION AND INCUBATION .............. 101

AGE CLASS DIFFERENCES: POST-HATCHING ....................................... 103
Male versus Female Contributions ......................................................... 103
Overall Parental Care ............................................................................. 106
Growth Of Young .................................................................................... 108
Relationship Between Parental Care and Growth of Young .................. 110
Mortality Of Treatment Group Young ...................................................... 115
DO Adults Undergo Increased Stress While Feeding Young? ................ 116

CONCLUSIONS .............................................................................. 118
Suggestions for Future Research .......................................................... 124
LITERATURE CITED ...................................................................... 128

LIST OF TABLES

Table 1. Percentage of breeding females in the SY and ASY age classes,

1987-1994. ................................................................................ 31
Table 2. Comparisons of yearly nest initiation dates between SY and ASY nests

using Kruskal-Wallis tests. ............................................................. 34
Table 3. Analysis of variance on nest initiation date. ....................................... 35
Table 4. Analysis of variance on clutch size. ................................................. 36

Table 5. Analysis of covariance on clutch size, using nest initiation date as the
covariate. .................................................................................. 37

Table 6. Distribution of clutch sizes, 1987-1994. Clutches of two, three, and
seven were collapsed into adjacent categories. Female age classes were

compared using x2 tests. ................................................................ 39
Table 7. Analysis of variance on individual egg weights. .................................. 43
Table 8. Analysis of variance on nest mean egg weights. .................................. 44

Table 9. Nested analyses of variance on egg weights by year. Egg weights are

nested within the nest (i. e. the individual female) ................................... 46
Table 10. Likelihood of eggs to hatch, 1987-1994. Yearly comparisons between

SY and ASY females were made using x2 tests. .................................... 49
Table 11. Probability of egg mortality at SY and ASY nests, 1987-1994, based

on the Mayfield exposure method. See text for further explanation. ............. 51
Table 12. Probability of nest failure at SY and ASY nests, 1987-1994, based on

the Mayfield exposure method. See text for further explanation. .......... . ....... 52
Table 13. Repeated measures analysis of variance on total visits/hour. .................. 59
Table 14. Repeated measures analysis of variance on female visits/hour ................. 59

xi

Table 15. Repeated measures analysis of variance on male visits/hour. .................. 60

Table 16. Repeated measures analysis of variance on total, female and male
visits/hour at nests with four young only. ............................................ 61

Table 17. Repeated measures analysis of variance on total, female and male
visits/hour, for nests with six young only. ........................................... 63

Table 18. Analysis of variance on total visits/hour. Data from day 13 posthatch
only. ....................................................................................... 64

Table 19. Linear regression analyses on log transformed bolus weights. Independent
variable was days posthatch. Probability values indicate slopes significantly
different from zero. ...................................................................... 65

Table 20. Nested analysis of variance on log transformed bolus weights. Covariate
was days posthatch. ...................................................................... 67

Table 21. Analysis Of variance on weight growth constants, R2 = 0.035. ............... 71
Table 22. Nested analysis of variance on weight growth constants, R2 = 0.747 ........ 72

Table 23. Analysis of variance on maximum weights attained by nestlings,

R2 = 0.283 ................................................................................ 74
Table 24. Nested analysis of variance on maximum weights attained by nestlings,

R2 = 0.569 ................................................................................ 74
Table 25. Analysis of variance on wing growth constants, R2 = 0.365. ................. 75
Table 26. Nested analysis of variance on wing growth constants, R2 = 0.722. ......... 76

Table 27. Analysis of variance on maximum wing length attained by nestlings.
Data were transformed by squaring and values for SS and MS are presented
x 10'5 for clarity. R2 = 0.297. ......................................................... 78

Table 28. Nested analysis of variance on maximum wing length attained by nestlings.
Data were transformed by squaring and values for SS and MS are presented
x 10'5 for clarity. R2 = 0.729. ......................................................... 79

Table 29. Multiple linear regression of nest mean growth variables: Weight growth
constants and maximum weights. ...................................................... 85

Table 30. Multiple linear regression of nest mean growth variables: Wing growth
constants and maximum wing values. ................................................. 86

xii

Table 31. Multiple linear regression of day 13 posthatch mean growth index
variables: Brood weights and nest mean nestling weights. ......................... 87

Table 32. Likelihood of mortality of treatment group young, 1990 and 1993. .......... 89

Table 33. Linear regressions of adult weights as a function of age of young during
nestling rearing. All data are from 1990 .............................................. 9O

xiii

LIST OF FIGURES

Figure 1. Mean nest initiation dates (iS.E.) for SY and ASY females, 1987-1994. ...33

Figure 2. Mean clutch size (iS.E.) for SY and ASY females, 1987-1994. .............. 36
Figure 3. Distribution of clutch sizes. Clutch sizes of two, three and seven are
collapsed into adjacent categories and all years are pooled. ....................... 38
Figure 4. Mean egg weights (iS.E.) for SY and ASY females, 1990-1994.
Individual eggs are the unit of measure. .............................................. 42
Figure 5. Nest mean egg weights (iS.E.) for SY and ASY females. ..................... 45

Figure 6. Mean total visits/hour (iS.E.) during 1990 at SY and ASY treatment
groups ...................................................................................... 55

Figure 7. Mean total visits/hour (iS.E.) during 1990 at treatment groups with
four or six young. ........................................................................ 56

Figure 8. Mean total visits/hour (iS.E.) during 1993 at SY and ASY treatment
groups ...................................................................................... 57

Figure 9. Mean total visits/hour (iS.E.) during 1993 at treatment groups with
four or six young. ........................................................................ 58

Figure 10. Regression lines representing log of bolus weights increasing over time.
All slopes were homogeneous and significantly different from zero. ............ 66

Figure 11. Increase in nestling weight over time during 1990 for all four treatment
groups ...................................................................................... 68

Figure 12. Increase in nestling wing length over time during 1990 for all four
treatment groups .......................................................................... 68

Figure 13. Increase in nestling weight over time during 1993 for all four treatment
groups ...................................................................................... 69

xiv

Figure 14. Increase in nestling wing length over time during 1993 for all four
treatment groups .......................................................................... 69

Figure 15. Mean weight growth constants (iS.E.) during 1990 and 1993 for all
four treatment groups .................................................................... 70

Figure 16. Mean maximum weight (iS.E.) attained by nestlings during 1990 and
1993 for all four treatment groups. .................................................... 73

Figure 17. Mean wing length growth constants (iS.E.) during 1990 and 1993
for all four treatment groups. .......................................................... 75

Figure 18. Mean maximum wing length (:S.E.) attained by nestlings during 1990
and 1993 for all four treatment groups ................................................ 77

Figure 19. Total visits/young/hour during 1990 and 1993 at all four treatment
groups ...................................................................................... 81

XV

INTRODUCTION

Clutch size—the number of eggs laid by a female during one breeding attempt—
is an easily measured variable of reproductive effort and determines the upper limit on
reproductive output for a particular breeding attempt. David Lack, a pioneer in the
development of clutch size theory in birds, hypothesized that the clutch size which
evolved was the one that maximized the number of young which fledged (Lack 1947,
1954, 1966). He argued that the main factor in altricial birds which limited clutch size
and the resulting number of fledged young, was the ability of the parent(s) to feed the
young. Lack’s hypothesis is appealing intuitively and has gained wide acceptance, this
acceptance based mostly on brood enlargement studies that have shown that the most
common clutch size is also the most productive in terms of producing young which
fledge (e.g. Alatalo and Lundberg 1989, Crossner 1977, Lack 1954, Perrins 1965), yet
few studies have been able to support his theory unequivocally (Bryant 1975, DeSteven
1980, Hussell 1972, Klomp 1970, Perrins and Moss 1975, von Haartrnan 1971). Some
studies, in fact, have concluded that females are not optimizing clutch or brood size in
relation to their abilities to feed young (Nur 1984, 1986). Nur showed that Blue Tits
could successfully rear broods larger than the average clutch size (some much larger),

and other studies have shown similar results (Bryant 1975, DeSteven 1980, HOgsted

1

2
1980, Loman 1980, Raskaft 1985, Slagsvold 1982, among others). It is thought that
larger clutch sizes which result in larger broods could have detrimental effects on both
adults and their offspring. Adults attempting to keep up with the demands of artificially
expanded broods have been shown to lose greater amounts of weight than those from
control broods (Askenmo 1979, DeSteven 1980, Hussell 1972). Weight loss of adults
during the feeding of young has been used as an indication that parents are stressed
from the rigors of providing parental care (e.g. Hussell 1972, Nur 1984, Smith et at.
1988, yet see Freed 1981 or Norberg 1981 for alternative explanations), and young
from enlarged broods often fledge at lower weights which can contribute to lower
survival rates (Hochachka and Smith 1991, Magrath 1991, Perrins 1965, Smith et al.

1989).

Even though Iack’s hypothesis has been both supported and refuted by various
studies, it has provided ornithologists a powerful theoretical model from which to base
years of productive research (Klomp 1970), and clutch size theory today is applied to
many other taxonomic groups as well, such as invertebrates and even plants (Godfray,
et at. 1991). Lack’s hypothesis is easily understood when dealing with semelparous
organisms where maximizing fitness for one breeding attempt (the only one in the
individual’s entire life) would be equivalent to lifetime fitness. The hypothesis becomes
more difficult to apply in iteroparous organisms (those with repeated breeding attempts)
where lifetime fitness is summed over all breeding attempts, and there are many fac-

tors, both intrinsic and extrinsic, which can limit or prevent breeding opportunities.

3
Because of this, Lack’s hypothesis has been modified over the years, mainly due to the
recognition that there are factors related to clutch size which can affect the fitness of
young following fledging, as well as potentially affecting future breeding Opportunities
of adults. It has become more appropriate, yet many times not possible, to use a closer
approximation of parental fitness than the number of young fledging, such as the num-

ber of Offspring that survive to breed or the number of grandoffspring produced.

The Effects of Female Age on Breeding Performance

Females breeding for the first time have been observed in many species of ver-
tebrates to produce smaller clutch sizes, and differ in other aspects of reproductive
biology as well, when compared to older individuals (Clutton-Brock 1984, Salthe 1969,
Tinkle and Ballinger 1972). This relationship has been observed most notably in birds
(reviewed by Stether 1990). In general, younger individuals or first-time breeders
initiate nesting later, lay smaller clutches, produce smaller eggs, show a lower likeli-
hood of hatching and fledging success, and produce fewer fledglings (DeSteven 1978,
F inney and Cooke 1978, Hannon and Smith 1984, Harvey et al. 1979, Middleton
1979, Perrins and Moss 1974, Ross 1980, Saether 1990, Stutchbury and Robertson
1988). There is also evidence that young which fledge from younger females breeding
for the first time are of a smaller average size (Crawford 1977) and may be less likely
to survive to breed (Lessells and Krebs 1989). The significance of the relationship

between age and clutch size is underscored by F inney and Cooke (1978) who have

4
shown that although young female Snow Geese are less productive than older females,
it is almost entirely due to initial differences in clutch size. Although clutch size differ-
ences between age classes of females are quite common, these differences are not uni-
versal (e.g. Bédard and LaPointe 1985, Erikstad et al. 1985, Hannon and Smith 1984,

Leinonen 1973).

Two of the hypotheses explaining the differences in clutch size between first-
time breeders and older individuals are the so—called restraint and constraint hypotheses
(Curio 1983, Desrochers 1992). The restraint hypothesis argues that it is adaptive for
first-time breeders to limit their reproductive output, whereas the constraint hypothesis
assumes that younger breeders are deficient in some aspects of their abilities to produce
offspring. These two hypotheses are not mutually exclusive (Curio 1983, Desrochers

1992, Wooller et al. 1990).

The restraint hypothesis predicts that natural selection would favor decreased
reproductive output in younger, first-time breeders if it increased the probability of
breeding again when they were older and more experienced (Charnov and Krebs 1974,
Curio 1983, Stearns 1976, Williams 1966). At that time they would presumably be
more likely to raise a larger brood due to increased age and experience (Finney and
Cooke 1978, Hamann and Cooke 1987, Perrins and Moss 1974). Increased brood sizes
have been shown to delay post-breeding molt (Slagsvold and Litjeld 1989), delay
breeding (Lessells 1986, Reskaft 1985) and decrease reproductive output in subsequent

seasons (Gustafsson and Part 1990, Gustafsson and Sutherland 1988, Reskaft 1985), or

5
may lead to reduced adult survival (Askenmo 1979, Nur 1984, Reid 1987, yet see
Alatalo and Lundberg 1989, or Alerstam and HOgstedt 1984). Implicit in this argument
is that older individuals would increase efforts to produce Offspring as their reproduc—
tive value declined with age (N01 and Smith 1987, Reid 1988, Williams 1966) produc-
ing larger clutches and thus potentially fledging a greater number of young. Testing the
restraint hypothesis requires a demonstration that a skill necessary for successful
breeding is available to an individual yet is not utilized. The hypothesis is difficult to
test because it is necessary to follow individuals through time and compare lifetime
reproductive success of individuals that delayed breeding to individuals that bred at the
first opportunity. Due to the problems involved with obtaining longitudinal data, par-
ticularly in short-lived species which migrate or do not show high site fidelity, com-

plete data sets have been nearly impossible to obtain.

Testing the constraint hypothesis requires demonstrating that some of the skill or
mechanisms necessary for successful breeding are lacking or deficient in younger
breeders. The constraint hypothesis is supported circumstantially by many studies,
many focusing on foraging abilities, which show that younger birds do less well than
older individuals (reviews in Burger 1990, Marchetti and Price 1989, Wunderle 1991).
It is postulated that younger birds breeding for the first time may be less able to pro-
vide food for themselves and produce smaller clutches as a result ( Aldrich and Ravel-
ing 1983, Lack 1968, Perrins 1970, Perrins and Moss 1974). While some argue that

deficiencies in providing parental care to the brood resulting from inadequate foraging

6
skills limit clutch size in younger birds (Ainley and Schlatter 1972, Bryant 1975, Jones
1987b and 1987c, Perrins and Moss 1974), others contend that limitations acting on the
female during egg formation and incubation limit clutch size (Hussell and Quinney
1987, Nur 1984, Yom-Tov and Hilborn 1981). Food supplementation studies ( Des—
rochers 1992, Kallander 1974) have shown experimentally that young breeders start
breeding earlier than normal when provisioned with additional food, and Desrochers
(1992) demonstrated that it was because young females were less successful at finding
food. Desrochers (1992) further concluded that although supplemental food did allow
first-time breeders to increase the length of their breeding season compared to older
individuals by allowing earlier nest initiation, they were still deficient in other aspects
of reproductive performance as evidenced by lower annual reproductive output. In
addition, there may be physiological or developmental constraints which contribute to
observed differences in clutch sizes between age classes. Westin 1989 (cited in
Enoksson 1993) showed that the ovaries and follicles of one-year Old Willow Tits
(Parus montanus) were smaller than those from older females, and smaller clutches and

eggs potentially result from these size differences.

The mechanisms causing smaller clutch size and subsequent reduced reproduc-
tive output by females breeding for the first time are poorly understood. As Ricklefs
(1977) points out, aspects of reproductive effort manifest themselves differently de-
pending upon which of the many factors (e. g. age, territory quality, food supply, pre-

dation pressures) contribute to reproductive success or failure in any one breeding

7
attempt. Many potentially important yet subtle differences in behavior or physiology
are difficult or impossible to measure in the context of field studies. As a result, most
researchers have been able to measure only fecundity or mortality as a function of

initial clutch size, and more subtle measures, often out of necessity, have been ignored.

Tree Swallows (T achycineta bicolor) are unique among North American passer-
ines because young adult females have a distinct subadult plumage (Rohwer et al.
1980) which makes it possible to assign females to age classes based on plumage char-
acteristics (Hussell 1983a, USFWS and CWS 1991). Females in their second calendar
year of life (SY—second year) can be distinguished from females that are Older (ASY—
after second year), and these characteristics represent an unusual Opportunity to investi-
gate questions concerning age-related reproductive performance. In many species of
birds, female ages can be determined with confidence only if the individual was marked

prior to fledging and then returned to breed in subsequent years.

It would seem that in a short-lived passerine, selection would favor early aged
breeding, if at all possible. In Tree Swallows, which have an average expected lifespan
of only 2.7 years (Butler 1988), each breeding season represents a significant portion of
an individuals’ potential reproductive output, so the benefits of early breeding are very
high (Perrins and Moss 1974, Studd and Robertson 1985, Wittenberger 1979). In
addition the effects of adult survival on average fitness will have less impact on short-

lived species (many passerines) when compared to longer-lived species (such as many

8
seabirds), further suggesting that breeding by younger females would be favored if at

all possible (Charnov and Krebs 1974).

Tree Swallow SY females have been shown generally to initiate nesting later,
lay smaller clutches with lower egg weights, and produce smaller broods and fledge
fewer young than ASY females (e. g. DeSteven 1978, Stutchbury and Robertson 1988).
Previous studies on Tree Swallows have focused on relating observed differences in
clutch size between SY and ASY females to nesting chronology, growth of young,
fiedging success and survival of adults (DeSteven 1978, 1980, Lombardo 1991, Stutch-
bury and Robertson 1988). The objective of this research was to investigate patterns
and processes of reproduction and parental care in Tree Swallows as related to the age
of the female parent in an attempt to clarify the mechanisms underlying the observation
that SY females produce smaller clutch sizes. More specifically, this study attempts to
answer the question of whether or not female Tree Swallows breeding for the first time
are constrained from producing larger clutches by their abilities to provide parental care
to the young they produce. This study is unique in that it attempts to test if the ability
to provide parental care in Tree Swallows is an important factor in determining clutch
size for this species. Because of the general observation of breeding deficiencies in
first-time breeders (Sather 1990) and, specifically, smaller clutch sizes in first-time
breeders, coupled with Lack’s postulated relationship between clutch size determination
and parents’ abilities to provide food for the young, the following working hypotheses

and predictions were formulated:

1.

9
Tree Swallow SY females have been shown to lay smaller clutches and produce
lighter eggs than ASY females. If improvement is age-related, both clutch size
and egg weights will increase when SY individuals breed again upon reaching

the ASY age class.

Growth of young is influenced by the quantity and/or quality of parental care
and those nests with the highest level of parental care will produce young with
the best growth. Specifically, the growth of young at SY nests will be poorer,

because parents at SY nests will provide lower levels of parental care.

. Experimentally enlarged brood sizes will accentuate any deficiencies on the part

of parents at SY nests by increasing the level of difficulty associated with rear-
ing young. This will be reflected in poorer growth of young and smaller maxi-
mum size attained prior to fledging. In addition, parents at ASY nests will be

better able to provide increased levels of parental care to larger broods.

The rigors of providing parental care are reflected in the condition Of the par-
ents as well as the young. Experimentally enlarged brood sizes will accentuate
any potential deficiencies on the part of the parents at SY nests and this will be
reflected in greater weight loss by parents at SY nests during the time of feeding

young.

TREE SWALLOW NATURAL HISTORY

Tree Swallows are obligate secondary cavity-nesters which breed across much
of North America. The southern limit of their breeding range extends across the central
United States and the northern limit is the tree line. Populations are distributed based
on the availability of nesting sites, and in some areas Tree Swallow populations are
limited by the availability of artificial nestboxes (Austin and Low 1932, Chapman

1935, Holroyd 1975, Kuerzi 1941, Low 1933).

Tree Swallows are aerial insectivores, yet sometimes consume plant material,
especially berries, fruits and seeds during periods of poor weather when insects are
unavailable (Chapman 1955, Turner and Rose 1989). Adults generally feed on medium
sized insects (4-6 mm long) and prey items taken closely correspond to insects which
are available within one or two km of the nesting area (Kuerzi 1941, Quinney and
Ankney 1985). As such, the diet varies by location, and site specific insect abundance

has been shown to influence clutch size (Hussell and Quinney 1987).

Migrants arrive at breeding sites approximately one month prior to nest build-
ing. Adults have been shown to return to the same general breeding area as previous

years, with reports of adults using the same nest cavity as previous years (Turner and

10

11
Rose 1989, personal observation). There are also anecdotal reports of females being
less likely to return to the same area to nest if they have been unsuccessful (Cohen
1981, Turner and Rose 1989). Juveniles, whose return rates are generally very low, are
reported to return to their natal area to breed (Butler 1988, Houston and Houston
1987). Adults are not territorial in the classic sense, yet defend an area of approxi-
mately 15 m radius around the nest (Robertson and Gibbs 1982) by chasing intruders
and excluding others from the nest by perching in the entrance hole. The female col—
lects most of the material for the nest which she builds over the course of several days
to two weeks. The male collects feathers for lining the nest, and feathers have been
demonstrated to provide a positive influence on the growth of young by providing

additional insulation (Winkler 1992).

There are reports of populations of non-breeding Tree Swallows (termed
“ floaters”) which do not breed because of severe competition for nesting sites. It has
been estimated in a well studied Ontario population that approximately 25 % of all
females were floaters, and 50-80% of these floaters were members of the SY female
age class (Stutchbury and Robertson 1985, 1987a). The unique and distinguishable
plumage of SY females has been hypothesized to suppress aggression from pairs estab-
lished at nest sites by signaling to the resident female that the intruder is a subordinate
and at the same time signaling to the resident male that the intruder is a female

(Stutchbury and Robertson 1987b).

12

Females are monogamous and single brooded, although there have been cases of
polygyny reported in very high prey abundance areas (e. g. Quinney 1983) and there
have also been a few reports of two broods being produced in one season (Hussell
1983b). Eggs are laid generally at daily intervals, yet sometimes days are skipped,
particularly during poor weather. Incubation lasts for approximately 14 days and only
the female sits on the eggs, although there have been anecdotal reports of males incu-
bating (Kuerzi 1941). Incubation often begins with the laying of the penultimate egg,
and this leads to asynchronous hatching with the last egg hatching one day later than
the rest of the clutch (Zach 1982). Brooding of the young is undertaken primarily by
the female and effective thermoregulation of the brood in an enclosed nestbox occurs at
approximately four days (Dunn 1979). Both females and males feed the young, bring-
ing approximately equal amounts of food (Quinney 1986). Young fledge from the nest
at approximately 18-21 days of age, depending upon the weather, and generally do not
return to the nest following fledging (Turner and Rose 1989). There is little evidence
that any post-fledging parental care occurs, yet Wheelwright et al.(1991) mention
unpublished data showing that adults guard and feed young for several days following

fledging.

Nesting success for Tree Swallows is highly variable from year to year due to
the unpredictable insect food supply. Butler (1988) reports an average from several
studies that 76% of eggs which are laid produce young which fledge. The population

considered in this study showed a much lower success rate, with only approximately

13
45 % of eggs laid resulting in fledged young over eight years (Beaver et al. 1994). Cold
and wet weather can have a severe impact on fledging rates by causing mortality of
young in the nest. Chapman (1955) reported mortality rates of young ranging from 6-
44%, while Beaver et al. (1994) report a 95 % loss in one year, 75 % in another and
55 % in a third. All three of these years were characterized by prolonged episodes of
cold and wet weather when insect food was virtually unavailable (personal observa-

tion).

Following the breeding season, flocks congregate near bodies of open water and
then migrate to the southern United States, Central America and the Caribbean. Some-
times the rare individual moves as far south as the western coast of Argentina. Adults
undergo a complete molt following breeding from approximately mid-July through
October. This time period of molt sometimes conflicts with late breeders still feeding
young (Hussell 1983b), as well as coinciding with the migration period (Stutchbury and

Rohwer 1990).

Although individual Tree Swallows have been known to live for up to eleven
years (Hussell 1982), on average they live for only 2.7 years (Butler 1988) which is
typical for small passerines (Bulmer and Perrins 1973). Mortality rates are about 60%
for adults and may be higher for SY adults (Butler 1988, Houston and Houston 1987,

Lombardo 1986).

METHODS

Study Sites

This study was conducted during 1989-1994 on state-owned land at Panola
Plains (T42N R32W, Section 10), seven miles south of Crystal Falls, in the Upper
Peninsula of Michigan. Additional data were collected during 1993 at Tachycineta
Meadows (T42N R31W, Section 3), 12 miles east of Crystal Falls. Some additional
clutch size and fecundity data from Panola Plains collected prior to 1989 are included
for comparison. Both study areas were located in large openings which had been main-
tained by the Michigan Department of Natural Resources using fire and herbicide
treatments during the 19605 and 703 in an attempt to provide habitat for Sharp-tailed
Grouse. Each area has had resident populations of Tree Swallows since 1984 when
grids of nestboxes were deployed in order to establish the areas as control plots for
studies on potential effects of electromagnetic fields from the Navy’s Project ELF (see
Beaver et al. 1994). Panola Plains is situated on a glacial outwash plain with soils
dominated by loamy sands (V ilas-Karlin complex, USDA Soil Conservation Service
1992). Predominant vegetative cover consists of low woody shrubs: specifically, sweet-
fern (Comptonia peregrina) and blueberry (Vaccinium spp.), interspersed with areas of

bracken (Pteridium aquilinum), and various species of grasses, forbs and sedges. Aspen

14

15
clones (Populus tremuloides) and scattered groups of Amelanchier spp. , Salix spp. ,
Crataegus spp. and Prunus virginiana which have grown up since cessation of burning
and other habitat management treatments in the late 19703, were interspersed with the
nestboxes. Tachycineta Meadows is also situated on a glacial outwash plain of very fine
sandy loam soils (Oconto series, USDA Soil Conservation Service 1992), yet the pre-
dominant vegetative cover is very uniform and consists of several species of grass, with
only very sparsely distributed clumps of Salix spp. and Amelanchier spp. and a few
scattered jack pines (Pinus banksiana). The small aspen clones which were present on
the site when the initial nestboxes were deployed, were removed by roller-chopping in
1987 by the Michigan Department of Natural Resources, Wildlife Division. Nestboxes
constructed of rough-sawn white cedar (9.4 cm W x 15 cm D x 22.5 cm H, inside
dimensions), were placed on cedar posts approximately 1.5 m high and the posts were
wrapped with high-density polyethylene to discourage climbing predators (Lederle et
al. 1985). The front panel on the boxes was latched and hinged to allow easy access to
the nest contents. Nestboxes were placed approximately 30 m from their nearest neigh-
bor. This spacing has been found to be a preferred inter-nest distance for this obligate
hole-nesting species (Muldal et al. 1985, Robertson and Rendell 1990). Originally,
there were 75 nestboxes at Panola Plains and this number was increased to 100 in 1988,
125 in 1989, and 165 in 1990 and for the remainder of the study. Tachycineta Mead-

ows originally had 75 boxes and this number was increased to 100 for the 1993 field

16
season. Tree Swallow occupancy rates on both sites have averaged greater than 80%

since 1987.

Egg Weights and Hatching Times

Nestboxes were checked daily during the nesting period in order to assess any
nesting activity which occurred. From these daily checks, generally made between
0630 and 1200 CDT, basic measures of reproductive activity were recorded: date of
first egg laid, subsequent egg laying, total numbers of eggs laid (clutch size), numbers
of eggs hatching, numbers of young fledging, and any mortality of young. Eggs at
Panola Plains during 1990-1994 were marked lightly with a sequence number using a
pencil on the first day they were observed, and weighed with a 5 g capacity Pesola®
spring scale readable to the nearest 0.05 g. Weights were taken with the spring scale
hanging inside a specially designed weighing box which ameliorated the effects of wind
and rain on measurements. Spring scales were calibrated periodically (every 2 or 3
days, more frequently in wet weather) using a set of standard weights (Fisher Scien-
tific) through the full range of the scale, with particular attention paid to the range of
weights represented by the eggs being weighed. Occasionally an egg was missed on the
first day it was laid due simply to not observing it in the nest or other logistical prob-
lems, so these eggs were weighed on the day following laying. First and second laid
eggs were easily distinguished from one another by the size of the air space located in

the blunt end of the egg; the first laid egg had the larger air space (personal observa-

17
tion). Eggs lose weight over time and Wiggins (1990) adjusted weights of Tree Swal-
low eggs missed on the first day by adding 0.02 g to the measurement on day two. I
did not make this adjustment since the precision of the spring scales (measurement was
possible to the nearest 0.05 g) used in this study was of a greater increment then the

correction made by Wiggins.

During 1990-1993 egg weights were obtained by myself. During 1994 most of
the weights were collected by a field assistant. In order to test for observer bias, 40
eggs from Tachycineta Meadows were weighed by both myself and the field assistant.
Results showed that we measured weights slightly differently, yet, overall, these differ-
ences averaged only 0.008 g/egg less for the field assistant (less than 1/2 of 1% of the
mean). A one-way analysis of variance detected no differences between observers (F =
0.039, P = 0.844) and the observer factor explained far less than 1% of the variability
in egg weight (R2 = 0.001). As such, no corrections were made for the 1994 egg

weight data.

When nests neared the end of incubation and hatching was imminent (13-15
days after the last egg was laid), nests were visited more frequently (four times daily in
1989, three times daily 1990-1993, twice daily at Tachycineta Meadows in 1993) in
order to establish hatching times. The first time a young was observed following
hatching, it was marked by nail clipping and weighed with a Pesola® spring scale to the
nearest 0.1 g. In addition, the condition of down was noted: wet or dry, matted or

fluffy (Quinney et al. 1986). These distinctions allowed estimation of hatching times to

18
within two hours. Because eggs were numbered and nests were checked frequently, the
egg from which a specific young hatched could be determined approximately 15-50%

of the time.

Adult Weights and Measures

Adults were captured, generally while feeding young, using a simple nestbox
trap (Magnusson 1984), or were handled in the course of daily box checks. Adults were
handed using aluminum United States Fish and Wildlife Service hands if they were not
already banded. Authorization for handling, color-marking and banding adults and
young was secured from the proper state and federal agencies prior to the study through
D.L. Beaver, Department of Zoology, Michigan State University (Master Permit
#0966, my subpermit #0966-L). In addition, the Michigan State University Animal
Care Committee approved the techniques and treatments used in this study, under the
auspices of D.L. Beaver et al. (1994) ELF Project studies. Attempts were made each
year to band as many adults as possible. As many young as possible were also banded,
usually on day 16 following hatching. Adults were sexed by the presence of a distinct
brood patch in females, or cloacal protuberance or lack of a brood patch in males.
Females were aged using the criteria of Hussell (1983a). Females handled for the first
time could most often be placed into one of two age categories by scoring the percent-
age of blue plumage on the dorsal surface of the body. The two categories (US Fish

and Wildlife Service and Canadian Wildlife Service 1991) were: 1) SY, or second

19
year, which indicated that the female hatched during the previous calendar year breed-
ing season and was actually approximately one year old, and 2) ASY, or after second
year, which indicates that the female was at least two years old. These female age
categories (SY and ASY) will be used throughout the remainder of the text. SY females
were characterized by 0-50% coverage of dorsal blue plumage, and ASY females
showed greater than 90% coverage. Females with 50-90% coverage are rare, yet are of
uncertain age unless banded as nestlings, so were excluded from the study. Males do
not show any plumage differences following their post-juvenile molt (Stutchbury and
Rohwer 1990) and could only be aged positively if they were initially banded as young
and then returned in later years. A minimum known age could be determined if a male
was handed as an adult and returned in later years. Adults were handled once to de-
termine age and sex and again only if necessary as part of the experimental treatment or
design. Upon capture, adults were weighed to the nearest 0.1 g using a 50 g capacity
Pesola" spring scale. Weights were taken by placing the adult head first into a coin
envelope (#4, 11.5 X 7.5 cm) which held the bird without struggle and allowed easy
release. The right wing chord was measured in a flattened position to the nearest mm
using a stainless steel ruler fitted with a stop which allowed anchoring the wrist. During
the 1990 field season I attempted to capture and weigh all treatment group adults (see
below) a minimum of three times during the course of the nestling period in order to
assess changes in body weight during the course of feeding young. Typically, weights
were recorded on day 1 (day of hatching was day 0) or day 2 (rarely on days 0 or 3),

on days 6, 7 or 8, and approximately day 14. There were many reasons why some

20
adults were not measured exactly on the day desired: weather, logistical problems, or
difficulties with capture. Some adults were much easier (or more difficult) to capture
than others. Although highly desirable, logistical problems prevented standardization of
the time of day that adults were weighed. Wing chords were also remeasured upon
recapture to assess repeatability within and across observers. Attempts were also made
during 1991 to obtain multiple measures on all treatment group adults. I found that
episodes of inclement weather caused adults to be in poor physical condition and the
added invasiveness of capture was causing some nest abandonment, even though cap-
tures took place primarily during fine weather. Because of the problems in 1991, se-
quential adult weights and measures were not attempted during 1992-1994. Adults were
color marked when necessary for behavioral study using “magic markers” (El Marko
brand) for positive identification using video cameras or observing through binoculars.
These colors faded after approximately seven days, so it was desirable to recapture
treatment birds for remarking. This presented another opportunity for taking additional

weights and measures, if desired.

Growth of Young

Young in the nest that had been originally marked by nail clipping which were
included in the growth study were color banded (A.C. Hughes Ltd. , Oxford, England)
when they reached approximately 7 g. Growth of young at each of the nests under.

observation was assessed by visiting the nest every other day and taking measures of

21
weight to the nearest 0.1 g using a 10 g capacity Pesola‘” spring scale when the young
were small and later a 50 g capacity spring scale also read to the nearest 0.1 g. Length
of tarsus (tibiometatarsus), ulna, and wing (from elbow to tip of longest feather, or to
tip of the fleshy part prior to feather eruption) were measured to the nearest 0.01 mm
using a digital caliper (Multitoyo Digimatic). Data were entered on N EC PC8201A or
PC8300 portable computers as they were collected in the field. Error trapping pre-
vented entering of values out of the possible range of measurement. Young were only
measured through day 16 (day of hatching was considered day 0) since nest disturbance
after that time has been shown to increase the likelihood of premature fledging of
young (DeSteven 1980, Kuerzi 1941, Paynter 1954, personal observation). Following
day 16, nests were checked only briefly to determine if fledging had occurred, and

during these checks the young were not handled unless absolutely necessary.

Parental Care Measures

Levels of parental care were quantified by observing the frequency of visits by
adults to selected nests. The assumption was made that each visit constituted some
aspect of parental care: brooding or feeding young, nest sanitation, nest defense or
guarding, or more subtle behaviors. Nests were observed on days 1, 5, 9, 13 and 17
following hatching. These observation periods represent distinct phases of the nesting
period. During day 1 most activity involved brooding of the young by females, some—

times for longer than 60 minutes. Although males visit just as frequently on day 1, the

22
duration of their visits is very short and probably does not involve brooding the young,
since they are often observed perching with their head outside of the entrance hole.
Days 5, 9 and 13 represent a period of very rapid growth of the young and large num-
bers of visits/h generally occur on these days, particularly day 13. Maximum nestling
weight has generally been reached by day 17, and the young often are entering a phase

of weight recession at this time prior to fledging (Zach and Mayoh 1982).

During 1990, observations were conducted using video cameras. Once treatment
nests were established on the day of hatching, video equipment (Canon VCZOOA cam—
eras with Canon VR30A VHS format recorders or Canon E61 8 mm camcorders)
placed approximately 3 m from the nestbox entrance on a tripod, recorded visits to the
nest by color-marked adults. A timer (Micronta model 63-5012 weather-resistant stop-
watch) attached to the nestbox using velcro was used to indicate the time of clay and
allowed computation of the duration of each event. Observations were conducted at
various times during the day depending upon weather, logistics and other observations
scheduled on the same day. Nests observed with the VHS cameras were recorded for
the duration of a 360 minute tape, those observed with an 8 mm camera were watched
for the duration of a 120 minute tape whereupon the tape was changed and another 120
minute tape was recorded. An attempt was made to view each treatment nest alternately
with each camera type throughout the study. In order to lessen the disturbance caused
by setting up and taking down the camera equipment, a false camera setup was used on

each treatment nest and remained in place from the day of hatching through fledging.

23
The only time the false setup was not in place was during actual observations. Each
false setup mimicked the actual setup and consisted of a tripod constructed of 2 cm
diameter electrical conduit painted in a camouflage pattern like the actual tripods used.
A block of wood shaped like a camera was mounted atop the false tripod and was
covered by a camouflage cloth bag of the same pattern used over the actual cameras. A
small block of wood painted black served as a false clock and was attached to the nest-
box using velcro. The adults adjusted to the presence of the equipment very rapidly,
sometimes entering the box within five minutes after setup. During 1990, data on
parental care were obtained by viewing the video tapes and recording information from
each event on NBC PC8201A or PC8300 portable computers. Variables recorded for
each event included: sex of parent making the visit, the duration of time spent in the
nestbox, and whether or not a fecal sac was removed from the nest. The number of
visits was standardized to a per hour basis to take into account differences in time under
observation for each nest. Time under observation was computed for each tape as the
time from the start of the first event recorded to the end of the last complete event on

the tape.

During 1993, events were recorded manually by observing the focal nest with
binoculars for a sampling time of one h during days 1, 5, 9, 13 and 17. This sampling
technique allowed an increase in the number of nests observed from 16 in 1990 to 29 in
1993 (four in each treatment group in 1990 and a minimum of six in each group in

1993). In addition, because of the increase in the number of nests observed and the use

24
of two plots in 1993 in an attempt to increase sample size, it was not possible to use the
video recording technique. Nests were observed with 8.5 X or 10x binoculars from a
distance of 50-150 m. The distance to the nestbox differed for each nest under observa-
tion due to terrain, distance to adjacent boxes (which was maximized) and availability
of cover in which to locate the observation post. When setting up for observation, care
was taken to note any alarm calls given by adults and the observation post was moved
if alarm calls from adjacent nests persisted. Binoculars were held in a tripod system
which allowed the observer to sit comfortably in a lawn chair and look through the
binoculars without moving. Observers used personal communicators (Maxon 49—SX)
which allowed constant contact between all members of the research team on the plot.
This allowed perfect coordination of observations and other research tasks being con-
ducted on the plot and eliminated any possible interference. In addition, any questions
which arose could be dealt with in an immediate fashion rather than waiting for a later

time.

Observations in 1993 were of one h duration and the number of visits by each
adult were recorded. Although fecal sac removals could often be observed and they
were recorded, manual observations did not have the benefit of instant replay, so I
could not be certain of fecal sac removals and these data were not included in any

analyses.

25

Establishment of Treatment Nests

Four treatment groups were established for comparison: 1) ASY nests with four
or, 2) six young, and 3) SY nests with four or, 4) six young. Female ages were deter-
mined during the course of daily box checks by viewing the adults as they exited the
nestbox, by observing through binoculars or by direct handling if past day ten of incu-
bation. Handling of adults during egg laying and early incubation was avoided since
this can lead to abandonment of the nest (Burtt and Tuttle 1983, Cohen 1985, Lom-
bardo 1989, personal observation). An experienced observer could make the distinction
between ASY and SY females most of the time, and final determinations were made
during handling to color mark the adults prior to any observations. Several nests were
dropped from the study at this point because the adults fell into the unknoWn age cate-
gory (Hussell 1983a). Treatment nests were established on the day of hatching. Deci-
sions as to which nests were to be included in the treatment groups were based on age
of the female and date of hatching. Treatments were established as rapidly as possible,
the limiting factor being the maximum number of nests which could be observed on
any one day, or the total number of young which could be measured for growth by one
observer on any one day. Establishment of treatments took five days in 1990 and eight
days in 1993. Hatching times were determined by visiting the nests frequently (see
above) and if necessary, the number of young in the nest was manipulated so all young
hatched within approximately eight hours of one another. Manipulations took place on

the day of hatching and were used if it was necessary to increase the number of young

26
in the nest, decrease the number of young in the nest, or to even out cases of hatching
asynchrony which is a common occurrence in Tree Swallows (Clark and Wilson 1981,
Zach 1982, personal observation). If it was necessary to move young into a treatment
nest, care was taken to choose a nestling that was hatched at approximately the same
time as the rest of the young in the treatment nest and to assure that the weights of the
young were approximately the same. These precautions helped eliminate any weight
hierarchies which could confound results (Zach 1982). Manipulations took place at
approximately the same time that nests were checked for hatching, so the manipulations
did not significantly increase the amount of disturbance which was already occurring. It
does not appear that Tree Swallow adults can distinguish between their own and fos-
tered young (Beaver et al. 1994), and it is well established that the major influence on
growth is the nest (i.e. parental effect) in which the young are raised, rather than the
nest of origin (Beaver et al. 1994, Pettifor et al. 1988, Quinney et al. 1986, Ricklefs

and Peters 1981).

Weight of Food Delivered to Young

During 1989, 1990 and 1991 the weight of food delivered to the young at se-
lected nests was determined using the “ligature method” (Johnson et al. 1980). This
involved placing around the young’s neck a “ligature” or “collar” which constricted the
esophagus and prevented swallowing. Once fed by an adult, the food bolus could be

removed from the mouth and esophagus with forceps and stored in 70% ethanol for

27
further analyses. Collars were made of paper coated wire commonly referred to as
“twist ties”, and stayed on the young for a maximum bout time of 75 minutes. Gener-
ally, bouts were much shorter, as longer times often resulted in disgorged boluses
which tended to collect nest debris which contaminated the sample and potentially
biased final bolus weights. All young in the nest were collared during a bout. Collaring

took place generally at 5-14 days posthatch.

During 1989, from 30 June to 23 July, young were collared at 15 nests varying
from three to five young per nest. Nine nests had ASY females, and six nests had SY
females. From these nests 244 boluses were collected, many of which were collected
and weighed as a group associated with one bout. Because of this, means of groups
were used (n = 94 groups). During 1990, from 26 June to 13 July, young. were col-
lared at 21 nests with three to five young per nest. Fourteen nests had ASY females,
and seven nests had SY females. From these nests 239 boluses were collected, and
because some were collected as groups, an n of 177 was used in the analyses. No at—
tempt was made to differentiate boluses delivered by the male or female parent, so

during 1989 and 1990 only nest types categorized by female age could be distinguished.

Protocols were much different during 1991 when boluses were collected from
six nests (three ASY and three SY) between 30 June and 18 July. Five nests contained
four young and one of the SY nests contained only three young. Adults were captured
prior to any bolus collection bouts and color marked for positive identification. When

young were between one and four days posthatch, nestboxes were modified to open

28
additionally from the back and a blind was attached to the back of the nestbox. Blinds
were constructed of lightweight wooden frames covered with cardboard sheets painted
in a camouflage pattern. Each blind consisted of four panels screwed together making
up the four sides and a similarly constructed roof. During deployment, modification of
the nestbox, and assembling and attaching the blind to the nestbox took approximately
20 minutes. Tree Swallow adults readily accepted the blinds, some returning to feed
young almost immediately after deployment was complete. Because the blind was
attached to the nestbox, an observer inside the blind could open the nestbox and quickly
remove boluses from young after an adult had left the nest. One observer was located
in the blind to collect boluses, another observer was situated 75-150 m away from the
nestbox hidden in vegetation, and it was this observer who determined the sex of the
adult entering or leaving the nestbox based on color markings. Communication between
observers was achieved using the radios described earlier. In this way it could be de-
termined with confidence from which adult the food delivered had come from. Sixty-
nine female and 43 male boluses were collected. The sex of the delivering adult could

not be determined with confidence for an additional 24 boluses.

Samples were removed from ethanol and placed on pre-dried and pre-weighed
filter paper (1989) or polystyrene weighing dishes (1990 and 1991). Samples were
dried to a constant weight in an drying cabinet with the temperature maintained at 40°C
for a minimum of 48 h (N. Walker, Dept. of Entomology, Michigan State University,

personal communication). Dried samples (bolus plus weighing dish) were weighed to

29
the nearest 0.0001 g using a Sartorius H51 or Sartorius 1207 MP2 electronic scale, and

the final dried weight of the bolus was determined by subtraction.

Ambient Monitoring

Ambient temperatures were recorded at 9 minute intervals at both Panola Plains
and Tachycineta Meadows beginning prior to egg laying and ending when all fledging
had occurred on the plot. Temperatures were recorded using an On-site Weather Log-
ger (O.W.L., EME Systems, Berkeley, CA) and an NBC PC8201A portable computer
powered by a 12 v gel-cell battery. Data were collected by transferring files directly
from the NEC to a portable disc drive (Purple Computer Products or Tandy). The
temperature probes (microcircuit type, EME Systems, Berkeley, CA) were attached
directly underneath a nestbox approximawa 1.5 m above ground level insuring that
they were situated in the shade at any time of the day (Christian and Tracey 1985).
Prior to deployment, probes were calibrated using thermometers whose calibrations are
traceable to the National Bureau of Standards. For this study, two temperature probes
were used, one situated at a low elevation portion of the plot, the other at a high eleva-
tion portion of the plot. High and low elevations differed by approximately 5 m. Any
temperature data presented is derived from an average of these two probes. Once ap-
pended to a database, data were scanned using error detection routines, corrected for

probe calibrations, and hourly means were calculated.

30

Statistical Analyses

Prior to statistical analyses data sets were tested for normality using a combina-
tion of graphical and statistical techniques. Probability and stem/leaf plots were used to
view the distribution of data, and statistical assessments for normality included the
procedures outlined by D’Agostino et al. (1990) or Lilliefors’ test (Lilliefors 1967).
Variance heterogeneity was tested using either Bartlett’s test or the Fmax test. Bartlett’s
test is overly sensitive to departures from normality (Sokal and Rohlf 1981) so the Fmax
test was used in cases where this was a potential problem. If possible, data were trans—
formed in an attempt to eliminate any problems with normality or variance heterogene—
ity. Means are presented with standard errors, and figures represent untransformed data
unless otherwise noted. All statistical procedures were carried out using SYSTAT®
(Wilkinson 1992). All tests were two-way unless otherwise noted as warranted by the
type of hypothesis tested and CL = 0.05 was used as the standard to determine whether
departures from the null hypothesis were significant. When using nested analysis of
variance techniques, denominator mean squares used in the calculation of F ratios were
determined using the methods outlined in Zar (1984, pp. 470-476). Other techniques

are explained in the text where first used.

RESULTS

AGE-RELATED FECUNDITY

Percent SY Females In Population

The percentages of nesting SY females in the population under study at Panola
Plains has ranged from 11.3 % of the known age females in 1993 to 31.5 % in 1994

(Table 1). These values are slightly lower when all of the females (including those of

Table 1. Percentage of breeding females in the SY and ASY age classes, 1987-1994.

 

 

Total % SY of
Total % SY Known Age Known Age

Year # ASY # SY # Unknown Females Females Females Females
1994 74 34 21 129 26.4 108 31.5
1993 94 12 6 112 10.7 106 11.3
1992 94 24 10 128 18.8 118 20.3
1991 92 17 27 136 12.5 109 15.6
1990 95 23 12 130 17.7 118 19.5
1989 89 19 6 114 16.7 108 17.6
1988 49 14 16 79 17.7 63 ‘ 22.2
1987 43 13 3 59 22.0 56 23.2

 

31

32
unknown age) in the population are included. Lombardo (1986) reported that 48.9% of
all nesting attempts (laying at least one egg) were made by SY females, and that the
proportion of SY’s in the population remained stable from year-to-year. Stutchbury and
Robertson (1985) reported that SY females accounted for 22.8% of breeding attempts.
This value is likely inflated because SY females made up a high proportion of the
floating population that was encouraged to nest by the provision of extra nestboxes later
in the season. Interestingly, the lowest percentage of SY females in the Panola Plains
population was recorded during 1993, following a geographically widespread weather-
related mortality event during 1992. Mortality of nestlings was nearly 100% in nests at
Panola Plains and surrounding populations of Tree Swallows (Beaver et al. 1994).
Presumably, recruitment into the 1993 breeding population was low and resulted in
small numbers of SY females. In contrast, the highest percentage of SY females
breeding occurred during the following year in 1994. This may reflect increased
breeding opportunities for SY females hatched in 1993 due to a potential decrease in

ASY population levels associated with the lack of recruitment from the 1992 cohort.

Nest Initiation And Egg Laying

Egg laying started later, on average, for SY females during all years, except for
1987 (Figure 1). Data were corrected for yearly variation in nest initiation by counting
from the day the first egg was observed for that year rather than counting from a fixed

date (i. e. day 1 = day first egg was encountered in the population for that year). The

33

 

1a
1 O ASY

16 A ° 3"
.l

14 4 \
_ /\

/

INITIATION DATE

 

”i // ~¥ \ /I\I
1

a I
.. fl

 

 

8788899091929394

Figure 1. Mean nest initiation dates (iS.E.) for SY and

ASY females, 1987-1994.

distribution of laying dates was
skewed toward low-numbered
dates and could not be normal-
ized by transformations, and
variances were heterogeneous
as well, so a Kruskal-Wallis
one-way analysis of variance
was used to compare the two
female age groups. Computed
on a yearly basis (Table 2),

Kruskal-Wallis tests showed

significant differences between SY and ASY females in initiation dates for all years (all

P < 0.039), except 1987. With data pooled across years there was a significant differ-

ence between the two age classes (at2 = 46.522, df = 1, P < 0.001).

Although I could not meet all of the strict assumptions of the test, analysis of

variance (Table 3) revealed a significant effect of female age on initiation date, with

SY’s beginning egg laying later in the season (F = 51.150, P < 0.001). Year-to-year

fluctuations in the mean initiation dates (corrected for year) were also significant (F =

4.725, P < 0.001), and a significant female age x year interaction was detected (F =

2.792, P = 0.007) due to data from 1987 where the SY mean date was earlier than the

ASY date (Figure 1).

34

Table 2. Comparisons of yearly nest initiation dates between SY and ASY nests using Kruskal-

 

 

 

 

 

 

 

 

 

Wallis tests.

Female Rank
Year Age # Nests Sum 76 P
1994 ASY 74 3605.5

SY 34 280.5 8.132 0.004
1993 ASY 92 4549.5

SY 11 806.5 6.361 0.012
1 992 ASY 94 5286.0

SY 24 1735.0 4.259 0.039
1991 ASY 91 4615.0

SY 17 1271.0 8.529 0.003
1990 ASY 94 4946.0

SY 23 1957.0 17.226 < 0.001
1989 ASY 86 4095.5

SY 17 1260.5 11.346 0.001
1988 ASY 49 1410.0

SY 14 606.0 6.928 0.008
1 987 ASY 43 1252.5

SY 13 343.5 0.277 0.599

 

35
Table 3. Analysis of variance on nest initiation date.

 

 

 

Source SS DF MS F P
Female Age 1014.929 1 1014.929 51.150 <0.001
Year 656.344 7 93.763 4.725 <0.001
AgeXYear 387.741 7 55.392 2.792 0.007
Error 15079963 760 19.842

Clutch Size

Birds lay eggs in discrete numbers, so strictly speaking clutch size is not a con-
tinuous variable. Clutch size has been treated historically in the literature as a continu-
ous variable, with mean clutch sizes being reported and analyses being conducted on
these values. Even though all of the strict assumptions of the analysis of variance could
not be met, due to historical precedence and for purposes of comparison, I treated
clutch size in the same manner. I also analyzed these data using the more appropriate

distribution-free tests as well.

Average clutch sizes were smaller for SY females throughout the period of ob-
servation (1987-1994, Figure 2). Analysis of variance revealed a strong effect of fe-
male age on clutch size (Table 4, F = 28.924, P < 0.001), and although there was a
significant year effect (F = 3.591, P = 0.001), the relationship between SY and ASY

clutches remained stable across years (female age x year interaction, F = 0.735, P =

36

0.642). Overall, clutch sizes for

 

 

 

 

 

6.0
I O ASY SY females averaged 0.4
, 0 SY
1- 5.5 ; eggs/nest less than those observed
to .
w 1
E 1 in ASY females (Tukey’s HSD
8 50 4
8 post-hoc test, P < 0.001).
=1:
4.5 a
1 Because Tree Swallows
1
4.0 ‘ f I I I I I I I (Stutchbury and Robertson 1988)
97 88 89909192 93 94 .
YEAR as well as other spec1es (e.g.

Erikstad et al. 1985, Hussell
Figure 2. Mean clutch size (iS.E.) for SY and ASY
females, 1987-1994. 1972, Murphy 1986, Perrins
1965), have been observed to lay
smaller clutches as the season progresses (yet see Conrad and Robertson 1992), and the
fact that SY females initiate laying significantly later in the season, on average, than
ASY females, the relationship between female age and clutch size may be due, in part,

to the time of the season when clutches are initiated. For both SY and ASY females,

clutch size decreased significantly with date over all years of the study (SY females,

Table 4. Analysis of variance on clutch size.

 

 

Source SS DF MS F F

Female Age 16.751 1 16.751 28.924 <0.001
Year 14.558 7 2.080 3.591 0.001
Age XYear 2.982 7 0.426 0.735 0.642

Error 440.139 760 0.579

 

37
Table 5. Analysis of covariance on clutch size, using nest initiation date as the covariate.

 

 

Source SS DF MS F P
Female Age 5.419 1 5.419 10.339 0.001
Year 11.393 7 1.628 3.105 0.003
Date 42.326 1 42.326 80.756 <0.001
AgeXYear 2.736 7 0.391 0.746 0.633
Error 397.813 759 0.524

 

slope = -0.048, R2 = 0.144, n = 119, P < 0.001; ASY females, slope = 0.055,
R2: 0.093, n = 550, P < 0.001). Controlling for the date of nest initiation over all
years in an analysis of covariance (slopes were homogeneous), the effects of female age
and year on clutch size both persist strongly (Table 5). Although clutch size is affected
significantly by the age of the female, the year, and the date of initiation of egg laying,
these factors only explain 16.8 % of the variation in clutch size in this population. Other
factors, such as location within the plot (Perrins 1965) and habitat (HOgstedt 1980,

Krebs 1970) have also been shown to affect clutch sizes.

The distribution of clutch sizes (1987-1994) ranged from two to seven, which is
typical, although clutch sizes of one and eight eggs have been reported (Turner and
Rose 1989). For analysis, clutch sizes were collapsed into adjacent categories of four
eggs for clutches of two and three, and six for clutches of seven (Figure 3), because of

the low numbers of clutches in some years with two, three, or seven eggs (overall, 41

38
of 776 clutches or 5.3 %). Only ASY females laid clutches of seven eggs (n = 18),

and, overall, clutches of two and three were rare (25, n = 5; 3s, n = 18).

A significant lack of independence in clutch size distribution was observed be-
tween SY and ASY females during 1990, 1991 and 1992 (all P < 0.040, Table 6).
Heterogeneity x2 testing (Zar

1984, p. 67) revealed that years

 

 

 

 

 

 

 

 

70 were homogeneous and thus
. CZ] ASY
so _ sv
. could be pooled for further
50 _.
.12 . analysis (heterogeneity x2 =
a 40 4
E 1 9.04, df = 6, P > 0.1). Pooling
O 30 _
°\° 20 1 across years, the resulting 3 x 2
1
1o _ table shows a significant lack of
0 4 6 independence between SY and
CLUTCH SIZE ASY females with regard to

clutch size distribution (112 =

Figure 3. Distribution of clutch sizes. Clutch sizes of two, 35157 df = 1 p < 0001) and
three and seven are collapsed into adjacent ’ ’

cate cries and all ears are ooled'. . . . . .
g y p this is ev1dent in Figure 3. Al-

though clutch sizes of five were
the most common for both SY and ASY age classes throughout the study, clutch sizes
of four were more prevalent for SY females and clutch size of six were more prevalent

for ASY females.

39

Table 6. Distribution of clutch sizes, 1987-1994. Clutches of two, three, and seven were col-
lapsed into adjacent categories. Female age classes were compared using 38 tests.

 

Frequency of clutch size

 

 

 

 

 

 

 

 

 

Year Age 4 5 6 x2 P

1994 ASY 5 49 20 4.934 0.085
SY 7 21 6

1993 ASY 10 47 35 1 .825 0.402
SY 7 21 6

1992 ASY 16 43 35 6.419 0.040
SY 8 13 3

1991 ASY 15 56 20 12.811 0.002
SY 9 8 0

1 990 ASY 16 48 30 8.477 0.014
SY 8 14 1

1989 ASY 3 47 36 3.486 0.175
SY 2 1 1 4

1988 ASY 2 27 20 3.068 0.216
SY 2 9 3

1987 ASY 5 26 12 3.177 0.204
SY 4 5 4

 

Clutch Size Repeatability

Repeatability is the ratio of the between-individual variance and the total pheno-
typic variance in a measure (Falconer 1981, Lessells and Boag 1987), and‘is described

by the intraclass correlation coefficient. High repeatability implies little or no change

40
by individuals over time, whereas low repeatability indicates a greater degree of change
by individuals between measurement periods. For individual females aging from SY to
ASY (n = 44, pooled over years), repeatability of clutch size was low and non-
significant (r = 0.095, F43,“ = 1.209, P = 0.266). This indicates that there is little
correlation between clutch sizes within an individual female during her SY and ASY
years. For observations of ASY females encountered two years in a row (n = 94,
pooled over years), repeatability of clutch size was higher, and the intraclass correla-
tion coefficient was significant (r = 0.176, F9334 = 1.435, P = 0.041). This indicates
a higher correlation between clutch sizes produced by individual females between two
ASY years than between SY and ASY years, or in other words, less change between
years. Repeatabilities calculated for the ASY/ASY dataset were nearly twice as large as
the SY/ASY dataset, yet overall, these ASY/ASY values are slightly lower than what
has been reported for repeatability of clutch size in other species (r = 0.23 in Song
Sparrows, r = 0.248 in Lesser Snow Geese, both reported in Lessells and Boag
(1987), r = 0.51 in Great Tits, Perrins and Jones (1975), r = 0.28 also in Great Tits,

van Noordwijk et al. (1980)).

Analysis of mean clutch sizes from the repeatability dataset, comparing SY fe-
males moving into the ASY age class (n = 44 individuals, pooled over all years),
shows that females increased their clutch size with age, on average, 0.182 eggs/nest.
This value represents approximately half of the average 0.4 eggs/nest difference ob-

served between the SY and ASY age classes in the entire data set (see Figure 2), yet

41
mean clutch sizes in this data subset were not different between age classes (SY =
4.773i0.122 eggs/nest, ASY = 4.955;t0.121 eggs/nest, paired t-test, t = 1.159, P =
0.253). For ASY females encountered two years in a row, no change in clutch size was
detected (ASYl = 512810.077 eggs/nest, ASY2 = 5.138:l:0.079 eggs/nest, paired t-

test, t = -O.105, P = 0.916).

Because the data in my study were collected over eight years and there was a
significant effect due to year when comparing clutch size between female age classes
(Table 4), it is desirable to control for this effect (a portion of the environmental vari-
ance) when calculating repeatabilities. This was done by using standard normal deviates
of clutch size (Hochachka 1992, Perdeck and Cavé 1992) which were calculated using
the equation:

INDIVIDUAL CLUTCH SIZE - YEAR IVE/1N CLUTCH SIZE

CL UTCH SIZE =
4’” YEAR CLUTCH SIZE STANDARD DEVIA TION

 

9

where yearly means and standard deviations used in the equation correspond to respec-
tive female age groups (SY or ASY) within each year. Following these corrections,
repeatability for the SY/ASY group was still very low (r = 0.044, F43,“ = 1.093, P =
0.385), whereas the ASY/ASY value was greater and statistically significant (r =

0.209, F9334 = 1.529, P = (1.021)

Even though the mean clutch sizes were not significantly different from one
another as individual females aged from SY to ASY, the trend suggests that SY females

are more likely to increase clutch size as they age (low r) compared to older females

42
showing a higher degree of repeatability and very little, if any, change in mean clutch
size. This could be explained by several factors, including: 1) females show improve-
ment with age and are better able to produce larger clutches as ASY’s one year later, or
2) there is a higher probability of mortality on those females which lay smaller clutches
when they are SY compared to SY females that produce larger clutches (N01 and Smith

1987).

Egg Weights

Egg weights were smaller for SY females throughout the study, except in 1993
(Figure 4, Table 7, F = 35.059, P < 0.001). Weights also showed effects due to the
year of the sample (F = 19.638,

P < 0.001) as well as a signifi-

 

 

 

 

 

2.0 cant female age x year interac-
, o ASY
19: O sv tion (F = 2.691, P = 0.030)
’63 , which was due to 1993 data
V 1.8 9
1— .
5 where mean egg weights were
E 1.7 J
1 nearly the same for SY and ASY
1'6 1 females.
1-5 i . . i . The effect of individual
90 91 92 93 94
YEAR females was not included in this
Figure 4. Mean egg weights (iS.E.) for SY and ASY analysis as intraclass correlations

females, 1990-1994. Individual eggs are the unit
of measure.

43
during every year showed that the variation among females was greater than the varia-
tion within an individual female’s clutch of eggs ( all r < 0.72, all P > 0.13, see
Winkler 1993). As such, the sample n for the preceding analyses is the number of

individual eggs in each female age class.

There is some controversy, however, concerning the use of eggs from the same
nest as independent measures, even given the results of the intraclass correlations
(Jover, et al. 1993). Therefore, two additional approaches were taken, one using mean
egg weight for each nest as the unit of measure (n = 546 nests, rather than n = 2789

eggs), and the second using a nested analysis of variance.

Table 7. Analysis of variance on individual egg weights.

 

 

Source SS DF MS F P

Female Age 0.837 1 0.837 35.059 <0.001
Year 1.876 4 0.469 19.638 <0.001
AgeXYear 0.257 4 0.064 2.691 0.030

Error 66.364 2779 0.024

 

44

Table 8. Analysis of variance on nest mean egg weights.

 

 

Source SS DF MS F P
Female Age 0.151 1 0.151 7.863 0.005
Year 0.385 4 0.096 5.007 0.001
AgeXYear 0.046 4 0.012 0.603 0.661
Error 10.292 536 0.019

 

Analysis of variance on mean egg weights (Table 8) shows a significant effect
due to female age (F = 7.863, P = 0.005), and year (F = 5.007, P = 0.001), with no
interaction detected. Mean egg weights at SY nests were significantly smaller than
means at ASY nests (Tukey’s HSD post hoc test, P = 0.005). These results are gen-
erally the same as when individual eggs are used, with some notable exceptions. Using
the nest as the unit of measure, standard errors are much larger due to the large drop in
sample size; they now overlap in three of the five years of study (Figure 5). Although
SY females produce, on average, lower weight eggs in four of five years, t—tests com-
puted by year show significant differences between mean egg weight at SY and ASY
nests only during 1992 (t = 2.343, P = 0.021). In addition, the female age X year

interaction term now becomes nonsignificant.

45

The second approach, using nested analyses of variance (Table 9), showed a

significant effect of the nest (i. e. presumably attributes of the individual female) during

each year of the study (all P < 0.001). Because the factor of the nest accounted for a

much greater proportion of the variability in egg weights compared to the factor of

female age, significant effects of female age noted earlier were not significant in the

 

 

 

 

 

2.0
l O ASY
: 0 SY
1.9—
g .
Ii ..
g1.8':
2 1.7-«
fl 1
2 ..
1
1.64
1
1
15 f l I j I
90 91 92 93 94
YEAR

Figure 5. Nest mean egg weights (iS.E.) for SY and
ASY females.

nested models, except during 1992
(P = 0.028). Depending upon the
year, between 74 and 81% of the
variability in egg weight was
attributable to the female that laid

the eggs.

The relationship between
nest mean egg weights and clutch
size for SY nests was found to be
weak and nonsignificant using
regression analysis (R2 < 0.01, n

= 103, P = 0.963). For ASY

nests the relationship was also weak, yet the slope was positive and significant (Nest

Mean = 1.740 + 0.016(Clutch Size)), indicating a trend of increasing egg size with

increasing clutch size (R2 = 0.009, n = 443, P = 0.048).

46
The relationship between mean egg weight and date of first egg laid (corrected
for year) was non-significant for both ASY females (R2 < 0.01, n = 443, P = 0.683)
and SY females (R2 = 0.008, n = 103, P = 0.939). Because of the lack of relation-
ship between mean egg weight and clutch size or date of first egg, neither of these two
measures proved useful as a covariate in an attempt to explain more of the variability in

nest mean egg weights.

Table 9. Nested analyses of variance on egg weights by year. Egg weights are nested within the
nest (i. e. the individual female).

 

 

 

 

 

 

Year Source SS DF MS F P

1990 Female Age 0.097 1 0.097 1.031 0.312
Nest 10.466 111 0.094 14.542 <0.001
Error 2.983 460 0.006

1991 Female Age 0.195 1 0.195 1.797 0.183
Nest 12.582 116 0.108 17.800 <0.001
Error 2.821 463 0.006

1992 Female Age 0.557 1 0.557 4.954 0.028
Nest 12.153 108 0.113 17.997 <0.001
Error 2.864 458 0.006

1993 Female Age 0.000 1 0.000 0.000 0.986
Nest 8.731 99 0.088 17.538 <0.001
Error 2.177 433 0.005

1994 Female Age 0.266 1 0.266 3.073 0.083
Nest 8.822 102 0.086 13.428 <0.001
Error 2.763 429 0.006

 

47

Egg Weight Repeatability

Egg weight repeatability of individual SY females moving into the ASY age
class (n = 21 individuals, pooled over years) was high (r = 0.737, F2031 = 6.907, P
< 0.001), which indicates that SY female clutch mean egg weights remain stable with
age. Although mean egg weights did increase slightly within individual females aging
from SY (1.775 i0.033g/egg) to ASY (1.806;t0.027g/egg), these differences were not
significant (paired t-test, t = -1.477, P = 0.155). Likewise, individual ASY females
encountered over two consecutive years (n = 73, pooled over years) also showed high
repeatability of mean egg weight (r = 0.830, F72,” = 11.847, P < 0.001). Although
mean egg weights increased slightly over years (ASY, = l.817;t0.017g/egg, ASY2 =
l.830:l:0.018g/egg), again there were no differences in mean egg weight (paired t-test,
t = -1.261, P = 0.211). Other studies have also shown that egg weights are highly
repeatable fromyear to year in the same individual (Ojanen et al. 1979, Smith et al.
1993). When corrected for yearly variation in mean egg weights (see section on clutch
size repeatability), SY/ASY repeatability was slightly lower (r = 0.696), and
ASY/ASY repeatability increased (r = 0.891). Again, there were no significant differ-

ences between groups using paired t-tests.

Likelihood To Hatch

Comparing the likelihood of eggs to hatch between female age classes from all

of the eggs laid, it was found that significantly greater proportions of eggs hatched

48
from ASY nests during four of eight years (1989-1992); from SY nests in 1993; and no

differences between age classes during 1987, 1988 and 1994 (Table 10).

The likelihood to hatch was not homogeneous over years (heterogeneity x2 test-
ing, x2 = 46.293, df = 7, P < 0.001). A test of mutual independence was also re-
jected (x2 = 234.84, df = 22, P < 0.001) which indicates that some combination of
year and female age is influencing the likelihood to hatch. Three tests of partial inde-
pendence were all rejected (all P < 0.001) which further indicates that likelihood to
hatch is conditionally dependent upon the factors of year and female age. Because these
tests of partial independence were rejected, the most reasonable approach is to consider
these data on a yearly basis (Table 10). The significantly lower likelihood of hatching
of eggs in SY nests in four of eight years could be due to several reasons, including
higher rates of infertility, deficits in incubation behavior, higher rates of predation, or a

greater likelihood of nest abandonment.

One of the most striking results is the high likelihood of batching of eggs for SY
females during 1993, which is in sharp contrast to all other years. Population levels of
SY females were lowest during 1993 compared to all other years. Although only
speculative, if there was differential mortality on young in 1992 resulting in SY indi-
viduals nesting in 1993 that were of inherently higher quality in terms of breeding

capabilities, this could account for some of the difference observed in hatch rates.

49

Table 10. Likelihood of eggs to hatch, 1987-1994. Yeariy comparisons between SY and ASY
females were made using X2 tests.

 

 

 

 

 

 

 

 

 

Female # Eggs # # not %
Year Age Hatch Hatch Hatch X2 / P
1994 ASY 382 336 46 88.0 x2 = 2.318
SY 169 156 13 92.3 P= 0.128
1993 ASY 475 379 96 79.8 x2 = 9.678
SY 58 56 2 96.6 P = 0.002
1992 ASY 458 356 102 77.7 x1 = 10.289
SY 114 72 42 63.2 P = 0.001
1991 ASY 456 371 85 81.3 x2 = 4.859
SY 74 52 22 70.3 P = 0.027
1990 ASY 485 404 81 83.3 x2 = 19.568
SY 103 66 37 64.1 P < 0.001
1989 ASY 460 338 122 73.5 x2 = 6.611
SY 96 58 38 60.4 P = 0.010
1988 ASY 258 227 31 88.0 x2 = 0.305
SY 69 59 10 85.5 P = 0.581
1987 ASY 201 137 64 68.2 x2 = 0.580
SY 53 39 14 73.6 P = 0.446

 

50

Nesting Success Based on Exposure

Using the Mayfield exposure method (Mayfield 1961, 1975), the probabilities
of mortality of individual eggs and the incidence of nest failure during egg laying and
incubation were compared between SY and ASY females. This method was not used to
address mortality of nestlings or nest failure during the time young were in the nest
because many of the nests used (nearly all of the SY nests) were manipulated by adding

or subtracting young, potentially resulting in a biased sample.

Briefly, exposure is calculated by summing the number of eggdays (for individ-
ual eggs) or nestdays (for entire nests) over the time that the nest was under observa-
tion. For example, a nest with five eggs which was under observation for 13 days
would represent 13 nestdays or 85 eggdays (5 eggs X 13 days). These exposure values
can then be summed within female age classes and probabilities of egg disappearance or
nest failure over the time under observation can then be calculated. Comparisons be-
tween age classes of females were made on a yearly basis using a x2 test (Mayfield
1961, 1975) and, additionally, Z statistics based on maximum likelihood estimates

(Hensler and Nichols 1981).

51

Significantly higher probabilities of losing eggs during egg laying and incuba-

tion were recorded for SY females during four of eight years (Table 11, both methods

of calculation, all P < 0.03). These same four years also showed a significantly lower

likelihood of egg hatchability for eggs from SY females (Table 10). During 1993, the

probability for egg mortality was significantly lower for SY females calculated by both

methods (both P < 0.004). This result is noteworthy because 1993 was the same year

Table 11. Probability of egg mortality at SY and ASY nests, 1987-1994, based on the Mayfield

exposure method. See text for further explanation.

 

 

 

 

 

 

 

 

 

# Prob. of

Female Total Eggs Egg

Year Age Days Lost Loss 2 P x2 P

1994 ASY 6125 45 0.007 0.342 0.367 0.121 0.728
SY 2736 22 0.008

1993 ASY 8512 100 0.012 5.427 <0.001 8.310 0.004
SY 1026 2 0.001

1992 ASY 8299 108 0.013 3.338 <0.001 17.019 <0.001
SY 1839 48 0.026

1991 ASY 7400 93 0.012 1 .929 0.027 5.255 0.022
SY 1255 26 0.021

1990 ASY 8705 96 0.011 3.306 <0.001 16.654 <0.001
SY 2047 46 0.022

1989 ASY 8294 127 0.015 3.204 <0.001 15.440 <0.001
SY 1681 49 0.029

1988 ASY 4373 38 0.009 0.194 0.425 0.039 0.843
SY 1073 10 0.009

1987 ASY 3756 97 0.025 1.173 0.121 1.549 0.213
SY 1 164 38 0.033

 

52

when egg weights were not different between SY and ASY females (Figure 4) and SY

females showed a significantly greater likelihood to hatch eggs as well (Table 10).

Comparing the probability of nest failure (Table 12) between SY and ASY fe-

males, showed that during 1989 and 1992, SY females had a significantly higher rate of

nest failure during egg laying and incubation than did ASY females calculated by both

methods. During 1993, Z statistics indicated a significantly higher rate of nest mortality

Table 12. Probability of nest failure at SY and ASY nests, 1987-1994, based on the Mayfield

exposure method. See text for further explanation.

 

 

 

 

 

 

 

 

 

, # Prob.

Female Total Nests of

Year Age Days Lost Loss 2 P x2 P

1994 ASY 1500 0 0.000 1.000 0.159 2.405 0.121
SY 624 1 0.001

1993 ASY 1737 6 0.003 2.454 0.007 0.870 0.351
SY 251 0 0.000

1992 ASY 1464 9 0.006 2.277 0.011 14.279 <0.001
SY 253 8 0.032

1991 ASY 1670 11 0.007 0.785 0.215 0.396 0.529
SY 289 1 0.003

1990 ASY 1058 12 0.011 0.180 0.429 0.031 0.860
SY 484 5 0.010

1989 ASY 1076 15 0.014 1.715 0.043 5.322 0.021
SY 219 8 0.037

1988 ASY 935 3 0.003 0.186 0.425 0.039 0.843
SY 248 1 0.004

1987 ASY 707 9 0.013 0.264 0.397 0.065 0.799
SY 279 3 0.01 1

 

53
for ASY females (P = 0.007), yet the x2 test did not (P = 0.351). The P values in
Tables 11 and 12 indicate clearly that the x2 test is the more conservative of the two
methods. All other years showed no differences in probability of nest failure between

the two female age groups.

These results are in agreement with hatching likelihoods presented earlier and
further indicate that SY females have a higher likelihood to lose eggs during the laying
and incubation period, yet the incidence of total nest failure was approximately the
same for both age classes of females—only during 1989 and 1992 did SY females show

a significantly higher rate of nest failure.

PARENTAL CARE

Nest Visitation Rates

Whereas fecundity variables were measured at as many nests as possible, meas-
ures of parental care and growth (later section of Results) were obtained from a subset
of nests, many of which were manipulated in order to standardize brood sizes to either
four or six young, or to eliminate problems with hatching asynchrony (see Methods).

These nests are referred to throughout the text as treatment nests.

Although I could not determine the exact age of the males at many treatment
nests (unless males were banded as young, which was rare given the very low return

rates in this population of less than 1%), I could still test to see if the behavior or effort

54
of the male at any individual nest was different from that of the known-age female, or
was influenced by the behavior or effort of the known-age female. During 1990, four
of 14 nests showed lack of independence between frequencies of male and female visits
pooled over the course of the entire observation period, when compared to an expected
50/50 ratio of visits. At two of these nests the male visited more frequently, whereas
the female visited more frequently at the other two nests. During 1993, none of the 22
nests showed any differences between frequencies of male and females visits over the
course of the observation period. The relative proportion of visits made to the nest by
males and females was found to be significantly different within the treatment group of
ASY nests with four young, where females made a greater number of visits during 24
of 38 observation periods (Wilcoxon test, Z = -2.975, P = 0.003). For the other three
treatment groups, no such differences were found (all Z > 0.047, all P > 0.387). The
adult pairs at SY nests showed a high degree of symmetry with approximately 50% of
the observation periods having a greater number of female visits and the other half
having a greater number of male visits. Within all four treatment groups, male and
female frequencies of visits to the nest were shown to be significantly correlated with
one another (Pearson correlations, all rp > 0.57, all P < 0.001), so although there
were some differences at individual nests, overall, male and female visits were posi-
tively and significantly associated. For further analyses of parental care measures by

treatment groups, all nests within a treatment group were pooled.

55

Mean total visits/h (summed female plus male visits), used as an index of paren-

tal care, is presented graphically for various treatment group combinations in Figures

50

 

 

 

a:

D

O

E ASY
‘0 0

1-

5 V 4YNG
S O 6YNG
i 40—

'—

O

..—

 

 

 

 

S Y
0 7 F l j T
1 5 9 13 1 7
DAYS PO STHATCH

Figure 6. Mean total visits/hour (:tSE.) during 1990 at
SY and ASY treatment groups.

6-9. During 1990 at ASY nests
(top half Figure 6), the number of
visits/h was greater at nests with
six young when compared to nests
with only four young. During
1990 at SY nests (bottom half of
Figure 6), the number of visits/h
was approximately the same for
nests with six young compared to
nests with only four young. This
shows that adults at SY nests did
not increase total visits/h with
increased brood size as is evident
at ASY nests, suggesting a lower
level. of parental care on the part
of the adults at SY nests. Viewing
the same data, but now comparing

nests with only four young (top

half Figure 7), no differences in total visits/h are shown between ASY and SY nests.

However, at nests with six young (bottom half Figure 7), a greater number of visits/h

56

was recorded at ASY nests when

 

O ASY 1990 compared to SY nests, again
0 SY

suggesting a relatively lower level
of parental care on the part of the

parents at SY nests with six

 

 

 

 

 

 

g young.
0 J
E 4 Yng
'9 0 During 1993, no differ-
E O ASY
> 40 g ences are evident between ASY
2‘
'—
O nests with four or si 0 n to
,_ 30 . x y u g( p
1 half Figure 8), whereas at SY
20 1
nests (bottom half Figure 8), those
10 ~
nests with broods of six oun
1 6 Yng y g
0 I I r 1 1 showed higher total visits/h. This
1 5 9 13 17
DAYS POSTHATCH is in contrast to results observed in

Figure 7. Mean total visits/hour (:tSE.) during 1990 at 1990 when ViSitS/h at SY nests
treatment groups with four or six young.
with six young were shown to be

less than visits/h at ASY nests. Viewing the same data from a different perspective,
total visits/h were approximately equal at ASY and SY nests with four young (top half
Figure 9). At nests with six young, however, total visits/h were greater for SY nests

when compared to ASY nests (bottom half Figure 9).

57

 

50
V 4 YNG

<> 6YNG
40 —

1 993

 

TOTAL VISITS I HOUR

 

 

 

 

1 5 9 13

l

17

DAYS POSTHATCH

Figure 8. Mean total visits/hour (:tS.E.) during 1993 at

SY and ASY treatment groups.

In order to test statistically
for effects due to female age or
brood size, repeated measures
analysis of variance was used, the
repeated measure being visits/h on
days 1, 5, 9, 13 and 17 posthatch.
Visits/h were represented by total
visits/h (female plus male), female
visits/h, or male visits/h. Because
observations were missing from
the data sets during both 1990 and
1993 due to inclement weather or
logistical problems, some values
were estimated. If only one of five
days was missing, the missing
observation period value was

estimated based on the mean for

that treatment group on that day. If more than one day was missing, the nest was ex-

cluded from analyses. For example, if an ASY nest with four young was missing day

five of observation, this value was estimated from the other ASY nests with four young

which were also observed on day five. Fourteen nests from 1990 and 15 nests from

58

 

so
0 ASY 1993
0 SY

4o —

so ~

 

TOTAL VISITS / HOUR

 

 

 

 

6 Yng
O 1 i T l l
1 5 9 1 3 1 7
DAYS POSTHATCH

Figure 9. Mean total visits/hour (iS.E.) during 1993 at
treatment groups with four or six young.

1993 were used in these analyses.
Data were square root transformed
in order meet the assumption of

normality.

Results of the repeated
measures analysis of variance on
total visits/h during 1990 and
1993 using the grouping factors of
female age, brood size and year
(Table 13) shows a significant
effect due to brood size (F =
18.714, P < 0.001), yet no sig-
nificant effects due to female age,
year, or any of the interaction

terms.

Considering only female

visits/h (Table 14), a significant effect due to brood size (F = 11.003, P = 0.003) and

a significant three-way female age x year X brood size interaction (F = 4.418, P =

0.048) were detected. For male visits/h (Table 15), only brood size was significant (F

= 13.109, P = 0.002).

59

Table 13. Repeated measures analysis of variance on total visits/hour.

 

 

 

 

 

Source SS DF MS F P
Female Age (A) 0.177 1 0.177 0.269 0.609
Year (Y) 0.779 1 0.779 1.186 0.289
Brood Size (B) 12.300 1 12.300 18.714 <0.001
AXY 2.478 1 2.478 3.770 0.066
AXB 0.010 1 0.010 0.015 0.904
YXB 0.000 1 0.000 0.000 0.991
AXYXB 2.436 1 2.436 3.707 0.068
Error 13.803 21 0.657

Table 14. Repeated measures analysis of variance on female visits/hour.
Source SS DF MS F P
Female Age (A) 0.008 1 0.008 0.024 0.877
Year (Y) 0.410 1 0.410 1 .253 0.276
Brood Size (B) 3.598 1 3.598 11.003 0.003
AXY 0.776 1 0.776 2.374 0.138
AXB 0.164 1 0.164 0.502 0.486 .
YXB 0.173 1 0.173 0.530 0.475
AXYXB 1.445 1 1.445 4.418 0.048
Error 6.867 21 0.327

 

60

Table 15. Repeated measures analysis of variance on male visits/hour.

 

 

Source SS DF MS F P
Female Age (A) 0.192 1 0.192 0.262 0.614
Year (Y) 0.324 1 0.324 0.442 0.513
Brood Size (8) 9.594 1 9.594 13.109 0.002
AXY 1.969 1 1.969 2.691 0.116
AxB 0.371 1 0.371 0.507 0.484
YXB 0.211 1 0.211 0.288 0.597
AXYXB 1.153 1 1.153 1.575 0.223
Error 15.369 21 0.732

 

Although comparison of Figures 6—9 seems to suggest that female age (i. e. nest
type characterized by the age of the female) and year may be important factors, their
influence was too small to be detected due to the high variability in the frequency of
visits over time. Brood size (four versus six young) was the most important factor
influencing the number of visits made to the nest by parents in this design. This was
expected, since if equal provisioning of individual young is the norm, then visits/h
would be higher at nests with six young. In order to control for this factor, each brood

size was considered separately in the following analyses.

Considering only nests with four young (Table 16), repeated measures analysis
of variance on the dependent variables of total, female or male visits/h, shows that no
significant effects due to female age or year were detected in any of the models. This
suggests that across both years (1990 and 1993) and both treatment groups (SY and

ASY nest types), the level of parental care, as indexed by the number of visits made to

61

Table 16. Repeated measures analysis of variance on total, female and male visits/hour at nests
with four young only.

 

Total visits/hour

 

 

 

 

 

 

 

 

Source SS DF MS F P
Female Age (A) 0.123 1 0.123 0.150 0.707
Year (Y) 0.363 1 0.363 0.443 0.522
AXY 0.000 1 0.000 0.000 0.992
Error 7.381 9 0.820

Female visits/hour

Source SS DF MS F P
Female Age (A) 0.046 1 0.046 0.167 0.692
Year (Y) 0.023 1 0.023 0.084 0.779
AXY 0.047 1 0.047 0.172 0.688
Error 2.457 9 0.273

Male visits/hour

Source SS DF MS F P
Female Age (A) 0.501 1 0.501 0.586 0.464
Year (Y) 0.483 1 0.483 0.565 0.471
AXY 0.050 1 0.050 0.058 0.815
Error 7.688 9 0.854

 

the nest, is uniform at nests with four young. These results are graphically represented

in the top halves of Figures 7 and 9.

1 Considering only nests with six young (Table 17), a different picture emerges.
No significant effects due to female age or year were detected, yet a significant female

age x year interaction was detected for total visits/h (F = 10.147, P = 0.008), female

62
visits/h (F = 6.522, P = 0.025), and male visits/h (F = 5.296, P = 0.040). These
results from treatment nests with six young indicate a significantly different response on
the part of parents at SY vs. ASY nests between the two years of observation. This is
evident in the bottom halves of Figures 7 and 9 where total visits/h at nests with six
young is highest at ASY nests during 1990, yet highest at SY nests in 1993. During
1990, at nests with six young, total visits/h was significantly higher at ASY nests (one-
way repeated measures analysis of variance, F = 7.657, P = 0.033). During 1993.,
however, even though graphically visits/h were greater at SY nests, no statistically

significant differences were detected (F = 4.125, P = 0.089).

Another approach is to assess parental care activity on a daily basis. Although
this analysis could be conducted on all five days of observation, day 13 was chosen
because it represents the time of maximum number of visits made to most nests and the
assumption was made that any differences between treatment groups would be accentu-
ated at this time when the parents were presumably working the hardest. Analysis of
variance (Table 18) on the dependent variable of total visits/h detected a significant
effect due to the number of young in the nest (F = 5.369, P = 0.029), yet no other
factors or interaction terms were significant. Post-hoc comparisons showed that overall,
rates of visitation by adults at ASY nests with six young did not differ on day 13 com-
pared to adults at SY nests with six young (P > 0.9). The same result was shown at

nests with four young (P > 0.9). Similar results of significant effects of brood size ‘

63

Table 17. Repeated measures analysis of variance on total, female and male visits/hour, for
nests with six young only.

 

Total visits/hour

 

 

 

 

 

 

 

 

Source SS DF MS F P
Female Age (A) 0.057 1 0.057 0.107 0.749
Year (Y) 0.422 1 0.422 0.788 0.392
AXY 5.430 1 5.430 10.147 0.008
Error 6.421 12 0.535

Female visits/hour

Source SS DF MS F P
Female Age (A) 0.135 1 0.135 0.367 0.556
Year (Y) 0.616 1 0.616 1.677 0.220
AXY 2.397 1 2.397 6.522 0.025
Error 4.410 12 0.368

Male visits/hour

Source SS DF MS F P
Female Age (A) 0.016 1 0.016 0.025 0.876
Year (Y) 0.007 1 0.007 0.010 0.920
AXY 3.390 1 3.390 5.296 0.040
Error 7.680 12 0.640

 

were recorded for female visits/h, yet for male visits/h no significant main effects were

detected.

In summary, the analyses show that brood size most strongly influenced the
number of visits made to the nest by parents, which was expected given the fact that
only two brood sizes were used in the experimental design. At nests with four young,

parents maintained a fairly even rate of visitation across treatment groups and years. At

64
nests with six young, significantly higher rates of visitation were recorded at ASY nests
in 1990 (P = 0.033), yet during 1993 higher rates were observed at SY nests although

the comparison was marginally non-significant (P = 0.089).

Table 18. Analysis of variance on total visits/hour. Data from day 13 posthatch only.

 

 

Source SS DF MS F P
Female Age (A) 1.794 1 1.794 0.025 0.876
Year (Y) 80.166 1 80.166 1.111 0.302
Brood Size (B) 387.421 1 387.421 5.369 0.029
AXY 223.258 1 223.258 3.094 0.091
AXB 34.084 1 34.084 0.472 0.498
YXB 47.140 1 47.140 0.653 0.427
AXYXB 160.045 1 160.045 2.218 0.149
Error 1803.811 25 72.152

 

Weight of Food Delivered to Young

All bolus collection data were log transformed to meet the assumption of nor-
mality. Regression analyses on data collected from all three years (1989-1991) showed
that bolus weights increased significantly with the age of the young (Table 19). In
addition, slopes of the regression equations were found to be homogeneous, with 1989

having the largest boluses and 1991 the smallest (Figure 10). Nested analysis of covari-

65
ance (Table 20) revealed a significant effect of year (F = 6.476, P = 0.004) and nest
(F = 1.785, P = 0.005) after controlling for age of the young (F = 28.450, P <
0.001), yet no significant effect due to nest type (i.e. SY or ASY nest). Together these
factors accounted for 31.5 % of the variability in bolus weight delivered. Because data
collection protocols were slightly different between years (see Methods), I also ana-
lyzed years separately and found no effect due to nest type (SY or ASY nest) after

controlling for age of young, a significant covariate.

Table 19. Linear regression analyses on log transformed bolus weights. Independent variable
was days posthatch. Probability values indicate slopes significantly different from zero.

 

 

Year n Slope Ft2 P

1989 87 0.091 0.221 <0.001
1990 176 0.064 0.090 <0.001
1991 136 0.057 0.128 <0.001

 

During 1991, using the blind observational system developed for the purpose
(see Methods), I was able to distinguish between male and female boluses at six nests,
three of which were SY nests and three ASY nests. Mean bolus weight, pooled over
sexes, nests, and days of observation, was lower at SY nests (0.025:t0.002g), when
compared to ASY nests (0.02810.001g). Although one-way analysis of variance de-
tected significant differences in bolus weight (F = 6.259, P = 0.014), the nest type

(SY vs. ASY nest) explained only 4.5 % of the variability in bolus weight. However,

66

 

 

 

 

is»:
E ///5
g -3 a 1989/x”
E ’,,/’ 1990/x»
m /// ,//”: ””””””
3 r’/ ,,,,,,,,,,, 1991
_, ,,,,,,,, ,,
O *,,,—/
(D _4 .1
(D
o
...l

'5 I I I . I I I I I

6 8 1o 12 14
DAYS POSTHATCH

Figure 10. Regression lines representing log of bolus
weights increasing over time. All slopes were
homogeneous and significantly different from
zero.

after controlling for age of the
young using analysis of covari-
ance, the nest type factor became

nonsignificant.

Comparing male bolus
weights only, no differences
between males at SY nests were
detected when tested against males
at ASY nests (F = 0.748, P =
0.392), again controlling for age
of the young. Although mean

female bolus weight was larger for

ASY females (0.02610.001g) than SY females (0.022:t0.002g), they were not signifi-

cantly different (F = 0.744, P = 0.392) after controlling for age of the young.

Whereas female bolus weight increased significantly over time with the age of the

young (n = 69, Log Bolus Weight = -4.375 + 0.067(Days), P < 0.001), male bolus

weight increases over time were smaller and not significant (n = 43, Log Bolus Weight

= -3.783 + 0.022(DayS), P = 0.363)-

67

Table 20. Nested analysis of variance on log transformed bolus weights. Covariate was days

 

 

 

posthatch.
Source SS DF MS F P
Female Age 0.064 1 0.064 0.177 0.715
Year 3.735 2 1 .867 6.476 0.004
AgexYear 0.720 2 0.360 1 .248 0.299
Nest 10.380 36 0.288 1 .785 0.005
Days Posthatch 4.597 1 4.597 28.450 <0.001
Error 57.518 356 0.162
GROWTH OF YOUNG

Growth Curves

Graphical representations of the mean measures of weight and wing lengths in—
creasing over time for the four treatment groups during 1990 are presented in Figures
11 and 12 and values for 1993 are shown in Figures 13 and 14. For clarity, the figures
are shown without standard errors, which are very small. Each mean is representedhby

12-36 young depending upon the treatment group and year. During 1990, the topmost

68

 

 

  

 
 
   

 

 

25
1 v ASY wl4Yng
g 20; O ASYw/6Yng
o 1 v SYw/4Yng .
g « 9 SYw/6Yng /O/""
2 151 07f
u. 1
O
'— -1
I 10—4
Q .1
5Q
1990
0 1 T I If I I I r
0 2 4 6 8 1O 12 14 16
DAYS POSTHATCH

Figure 11. Increase in nestling weight over time during

1990 for all four treatment groups.

 

 

  

 
  
 

 

 

3° .
3 v ASY w/4Yng
70 —
: <> ASYw/6Yng
E 60—: V SYw/4 Yng )
g i O SYw/6Yng //Q
I 50 J
5 ‘ <>//Q/
z /
LIJ 4° "
.1
(.9 .2
Z 30 :
§ 20%
10 _— 1990
0 i I T I I I I I I
2 4 6 8 10 12 14 16
DAYS POSTHATCH

Figure 12. Increase in nestling wing length over time

during 1990 for all four treatment groups.

curve (open triangles) for both
weight (Figure 11) and wing
(Figure 12) measures represents
ASY nests with four young,
whereas the bottom curve (filled
diamond) in each 1990 figure
represents SY nests with six
young. It appears as though nes-
tlings during 1990 in ASY nests
with four young are growing most
rapidly and reach a higher final
measure than do the nestlings in
SY nests with six young which do
more poorly. The other two
treatment groups, ASY nests with
six young and SY nests with four

young, show intermediate values.

During 1993, for the vari-
able of weight (Figure 13), the
topmost curve represents SY nests

with four young and the lower

WEIGHT OF YOUNG (9)

Figure 13. Increase in nestling weight over time during
1993 for all four treatment groups.

WING LENGTH (mm)

Figure 14. Increase in nestling wing length overtime

69

 

 

 
   

 

 

 

25
. v ASY wl4Yng
‘ /T‘V
« O ASYw/6Yng
20— V SYw/4Yng
O SYw/6Yng
15?
105
1
5‘ 1993
0 I I I I I I I I I
0246810121416
DAYS POSTHATCH

 

ASY wl4 Yng
ASY w/6 Yng
SY wl4 Yng
SY w/6 Yng

OdO<l

1993

  
  

 

 

T I I I I I

j
2 4 6 8 10 12 14
DAYS POSTHATCH

16

 

during 1993 for all four treatment groups.

curve ASY nests with six young.
For the variable wing length
(Figure 14), however, there is
very little visual difference be-
tween the four treatment groups.
Detailed analyses on nestling

weights and wing lengths follow.

Individual nestling growth
measures were fit to growth mod-
els (Ricklefs 1967, 1983) in a
non-linear curve fitting routine in
SYSTAT" (Wilkinson 1992).
Highest overall R2 values were
obtained by using the logistic
model for weight and the expo-
nential model for wing length.
The same procedures have been
used for Tree Swallows by Zach
and Mayoh (1982) and Beaver et

al. (1994).

70
Variables used in the following analyses include the weight growth constant
which is estimated by the logistic model and the wing growth constant which is esti-
mated by the exponential model. Also considered were maximum weights and maxi-
mum wing lengths, both of which have been shown to be less variable than fitted pa-

rameters from the non-linear growth models (Zach 1988).

Weight: Fitted Growth Constants

Mean weight growth constants (Figure 15) were slightly higher for 1993 meas-
ures in three of the four treatment groups compared to 1990 values, yet standard errors
are large and overlapped in all combinations of treatment groups and years. Analysis of
variance revealed no significant effects due to female age, year or brood size (Table

21), and there were no significant

interactions, yet this model only

 

 

 

 

0.54
g 052 A ' 199° explains a small amount of the
5 ' A 1993
6 05° ‘ variability in the weight growth
2 0.48 _
I; constant (R2 = 0.035). Using a
Q 0.46 4
or
B 0.44 4 é nested analysis of variance model
I
9 0.42 4 - -
I; (Table 22) Wthh Includes the
0.40 4
0 as factors of female age, year, brood
' 4 6 4 6
l-ASY—l l—SY—l size and nest (nested within all
TREATMENT GROUP

three of the first factors), plus

Figure 15. Mean weight growth constants (iS.E.) during
1990 and 1993 for all four treatment groups.

71
Table 21. Analysis of variance on weight growth constants, R2 = 0.035.

 

 

Source SS DF MS F P
Female Age (A) 0.032 1 0.032 3.506 0.063
Year (Y) 0.010 1 0.010 1.075 0.301
Brood Size (8) 0.000 1 0.000 0.005 0.944
AXY 0.000 1 0.000 0.000 0.995
AXB 0.001 1 0.001 0.121 0.728
YXB 0.007 1 0.007 0.803 0.371
AXYXB 0.001 1 0.001 0.090 0.764
Error 1.446 160 0.009

 

interaction terms, shows significant effects due to the nest (F = 13.272, P < 0.001) as
well as female age (F = 349.384, P = 0.034), with no significant interactions. Mean
growth constants were lower at nests with SY females compared to nests with ASY
females (see Figure 15). This nested model explains a large amount of the variability in
the dependent variable of weight growth constant (R2 = 0.781). In other words, much
of the variability shown in the growth constant values for weight are due to the individ-

ual nest from which they were obtained.

Although growth constants were normally distributed, some outliers were evi-
dent in stem/leaf plots. With three outliers removed, the female age factor in a nested
model becomes non-significant (F = 4.072, P = 0.293). All three outliers represented
very high growth constant values from two ASY nests with six young during 1993.

These three nestlings all showed rapid growth for several days, quickly reaching as-

72
ymptotic weights which were very low, indicating poor overall growth. Generally, high
growth constants translate into high quality growth, yet in these cases the opposite was

true .

Table 22. Nested analysis of variance on weight growth constants, R2 = 0.747.

 

 

Source 88 DF MS F P
Female Age (A) 0.044 1 0.044 349.384 0.034
Year (Y) 0.019 1 0.019 0.497 0.487
Brood Size (B) 0.000 1 0.000 0.001 0.979
Nest 1.067 28 0.038 13.272 <0.001
AXY 0.000 1 0.000 0.003 0.955
AXB 0.004 1 0.004 21.740 0.658
YXB 0.014 1 0.014 0.357 0.555
AXYXB 0.000 1 0.000 0.004 0.948
Error 0.379 132 0.003

 

Weight: Maximum Values Attained

Weight asymptotes were also calculated from the logistic models, yet the results
were very similar to the maximum values of weight which are reported here. The
maximum values were also slightly less variable, a result also reported by Zach (1988)

in Tree Swallows.

73

Mean maximum values of

 

 

 

 

26
I 1990 . . . .
25 _ A 1993 weight attamed during the period
E 24 - .1. of measurement show higher
a I g
E 23 1 values for three of the four treat—
: 22 1
g 21 _ § § [I ment groups during 1993 (Figure
3 2o 6 + 16). Results of an analysis of
19 T variance model including the
13
4 6 4 6
l—ASY--I I--SY—l factors of female age, year, brood
TREATMENT GROUP

size, and interaction terms (T able

Figure 16. Mean maximum weight (iS.E.) attained by 23) showed significant effects due
nestlings during 1990 and 1993 for all four

t t 1 .
rea men groups to year (F = 13,106, P < 0.001)

and brood size (F = 31.308, P <
0.001). The female age x year interaction term was also significant. Using a nested
analysis of variance model (Table 24) which further included the factor nest (nested
within all three of the first factors), plus interaction terms, shows significant effects of
year (F = 5.099, P = 0.032), brood size (F = 410.304, P = 0.031) and nest (F =
3.136, P < 0.001). The female age X year interaction term was also significant. This
interaction is due to mean maximum values at ASY nests being approximately equal
between 1990 and 1993, whereas mean maximum values were higher for SY nests

during 1993 than 1990 (see Figure 16). The nested model also explains a greater

amount of variability (R2 = 0.569) than the unnested model (R2 = 0.283). In addition,

74

results changed very little with the exclusion of two outliers.

Table 23. Analysis of variance on maximum weights attained by nestlings, R2 = 0.283.

 

 

Source SS DF MS F P
Female Age (A) 3.955 1 3.955 0.893 0.346
Year (Y) 58.058 1 58.058 13.106 <0.001
Brood Size (B) 138.697 1 138.697 31 .308 <0.001
AXY 49.067 1 49.067 11.198 0.001
AXB 6.441 1 6.441 1.454 0.230
YXB 0.056 1 0.056 0.013 0.911
AXYXB 10.767 1 10.767 2.430 0.121
Error 708.805 160 4.430

 

Table 24. Nested analysis of variance on maximum weights attained by nestlings, R2 = 0.569.

 

 

Source SS DF MS F P
Female Age (A) 3.305 1 3.3305 0.071 0.834
Year (Y) 51.564 1 51.564 5.099 0.032
Brood Size (B) 137.337 1 137.337 410.304 0.031
Nest 283.129 28 10.122 3.136 <0.001
AXY 46.565 1 46.565 4.605 0.041
AXB 7.774 1 7.774 0.690 0.559
YXB 0.335 1 0.335 0.033 0.857
AXYXB 11.274 1 11.274 1.115 0.300
Error 425.676 132 3.225

 

WING GROWTH CONSTANT

75

 

 

 

 

0.140
3 I 1990
. A
0.135 1 1993
0.130 { 6 %
0.125 5 I
0.120 j i
0.115 —: I
1
0.110
4 6 4 6
l—ASY—l l—-SY--l
TREATMENT GROUP

Figure 17. Mean wing length growth constants (iS.E.)

during 1990 and 1993 for all four treatment
groups.

Wing Length: Fitted Growth

Constants

Mean values for wing
growth constants during the period
of measurement show higher
values for three of the four treat-
ment groups during 1990 (Figure
17). Results of an analysis of
variance model including the
factors of female age, year, brood

size, and appropriate interaction

terms (Table 25) showed significant effects due to female age (F = 23.110, P <

0.001) and brood size (F = 32.793, P < 0.001). Three of the four interaction terms

Table 25. Analysis of variance 0n wing growth constants, R2 = 0.365.

 

 

Source SS DF MS F P
Female Age (A) 0.001 1 0.001 23.110 <0.001
Year (Y) 0.000 1 0.000 0.033 0.857
Brood Size (B) 0.001 1 0.001 32.793 <0.001
AXY 0.000 1 0.000 3.967 0.048
AXB 0.000 1 0.000 0.479 0.490 _
YXB 0.001 1 0.001 17.483 <0.001
AXYXB 0.000 1 0.000 9.061 0.003
Error 0.007 160 0.000

 

76

were also significant.

Using a nested analysis of variance model (Table 26) which includes the factors
of female age, year, brood size and nest (nested within all three of the first factors),
plus interaction terms, shows a significant effect due to nest (F = 6.052, P < 0.001)

and a significant year x brood size interaction. This interaction is due to mean values

Table 26. Nested analysis of variance on wing growth constants, R2 = 0.722.

 

 

Source SS DF MS F P
Female Age (A) 0.001 1 0.001 6.676 0.235
Year (Y) 0.000 1 0.000 0.058 0.811
Brood Size (B) 0.001 1 0.001 1.586 0.427
Nest 0.004 28 0.000 6.052 <0.001
AXY 0.000 1 0.000 1.231 0.227
AXB 0.000 1 0.000 0.126 0.783
YXB 0.001 1 0.001 6.478 0.017
AXYXB 0.000 1 0.000 2.849 0.103
Error 0.003 132 0.000

 

at nests with four young being greater in 1993, whereas means were greater during

1990 at nests with six young (see Figure 17). This nested model explains greater than
70% of the variability in the wing growth constant values (R2 = 0.722); the unnested
model explains less (R2 = 0.365). Although several outliers were detected by viewing

the distribution of the wing growth constant data, all of the individual nestlings

77
exhibited slow but steady growth throughout the observation period. None was re-

moved for further analyses.

Wing Length: Maximum Values Attained

Growth measures were only recorded through 16 days posthatch and wing
feathers are still growing at that time. As such, the maximum wing length attained
during this time period represents an index of potential wing size at the time of fledg-
ing, which occurs within several days, rather than an absolute measure of maximum

wing length.

Wing maximum values were squared to render the data set normal. Mean
maximum values attained for wing

measurements were greater during

 

 

 

 

80
‘ ' 199° 1990 at ASY nests and greater
75 fl A 1993
E 3 f during 1993 at SY nests (Figure
3 70 :1
E j i 18). Whereas marked differences
2 65 : § § § [3%
g are shown between treatment
3 60 j
2 i * groups during 1990, no differ-
55 —
ences were shown during 1993.
50 I
4 6 4 6
l--ASY—| I—-SY-—|
TREATMENT GROUP

Figure 18. Mean maximum wing length (iS.E.) attained
by nestlings during 1990 and 1993 for all four
treatment groups.

78
An analysis of variance model including the factors of female age, year, brood
size, and appropriate interaction terms (Table 27), showed significant effects due to

female age (F = 33.915, P < 0.001), brood size (F = 5.793, P = 0.017), and one

Table 27. Analysis of variance on maximum wing length attained by nestlin 5. Data were trans-
formed by squaring and values for SS and MS are presented x 10‘ for clarity. R2 =

 

 

0.297.

Source SS DF MS F P
Female Age (A) 1079.517 1 1079.517 32.503 <0.001
Year (Y) 0.101 1 0.101 0.003 0.956
Brood Size (B) 196.072 1 196.072 5.904 0.016
AXY 1181.930 1 1181.930 35.587 <0.001
AXB 2.566 1 2.566 0.077 0.781
YXB 171.915 1 171.915 5.176 0.024
AXYxB 19.944 1 19.944 0.600 0.440
Error 5314.043 160 33.213

 

interaction term. Using a nested analysis of variance model (Table 28) which further
included the factor of nest (nested within all three of the first factors), plus interaction
terms, shows a significant effect due to nest (F = 7.501, P < 0.001) and a significant
female age x year interaction. This interaction is due to higher overall ASY nest
means during 1990, whereas overall SY nest means were higher during 1993 (see
Figure 18). This nested model explains a much greater proportion of the variability in
maximum wing values (R2 = 0.729) when compared to the unnested model (R2 =

0.297). No outliers were detected in the maximum wing length data set.

79
It is apparent from the results of the nested models presented above that the nest
effect is one of the most important factors influencing growth of young in this system.

While the factors of female age, year and brood size help explain some of the variabil-

Table 28. Nested analysis of variance on maximum wing length attained by nestlings. Data were
transformed by squaring and values for SS and MS are presented x 10'5 for clarity. R2 =

 

 

0.729.

Source SS DF MS F P
Female Age (A) 1074.965 1 1074.965 0.905 0.516
Year (Y) 0.820 1 0.820 0.007 0.932
Brood Size (B) 220.581 1 220.581 1.422 0.444
Nest 3141.505 28 112.197 6.817 <0.001
AXY 1188.281 1 1188.281 10.591 0.003
AXB 1.359 1 1.359 0.065 0.841
YXB 155.157 1 155.157 1.383 0.250
AXYXB 20.854 1 20.854 0.186 0.670
Error 2172.539 132 16.459

 

ity in the dependent variables and are sometimes significant in the non-nested models,

values of R2 nearly double with the addition of the nested factor to the models.

RELATIONSHIPS BETWEEN PARENTAL CARE AND GROWTH OF
YOUNG

The prediction that young at SY nests, and especially at those nests with six.

young, would exhibit lower rates of growth or overall poorer growth and corresponding

80
lower levels of parental care was assessed using graphical as well as direct and indirect

statistical approaches.

Weight of nestlings prior to fledging has been shown to be an important factor
influencing survival of young during the post-fledging period (Hochachka and Smith
1991, Magrath 1991, Perrins 1965). Condition of young at fledging is likely even more
critical in Tree Swallows given that fact that there is little evidence of post-fledging
parental care (Turner and Rose 1989, yet see comments by Wheelwright et al. 1991).
Because of the relationship between overall size and subsequent survival, it is important
to determine the strength of correspondence between the levels of parental care

delivered and measures of growth of young.

During 1990, mean total visits/young/h over the five days of observation (top
half Figure 19) showed very similar results at three of four treatment groups, the ex-
ception being at SY nests with six young, where young were visited at a much lower
rate, on days 9, 13 and 17. This lower rate of visitation (i. e. feeding) amounted to
approximately one to two less visits per young per hour at these nests compared to all

other treatment groups.

81

 

 

TOTAL VISITS / YOUNG l HOUR

 

 

 

 

j
1 5 9 13 1 7
DAYS POSTHATCH

Figure 19. Total visits/young/hour during 1990 and 1993
at all four treatment groups.

Mean weight growth
constants (Figure 15), however,
showed high variability and no
significant differences between
any of the four treatment groups
were recorded during 1990 (post-
hoc tests, all pairs comparisons
with Bonferroni adjustments, all
P > 0.8). Mean maximum
weight values for weight during
1990 (Figure 16) were signifi-
cantly higher for young at ASY
nests with broods of four com-
pared to the other three treatment
groups (all P < 0.05), yet the
number of visits/young/h made

by adults at this treatment group

appeared to be no different than from the number made to ASY nests with six young or

to SY nests with four young. Mean values for wing growth constants (Figure 17) were

significantly larger during 1990 at SY nests with four young when compared to ASY

nests with four young (P = 0.007) or ASY nests with six young (P = 0.003), yet these

differences are not reflected in visits/young/h which are approximately the same for

82
these three treatment groups. Mean values of maximum wing length attained (Figure
18) were significantly smaller at SY nests with six young when compared to ASY nests
with four young (P < 0.001) or ASY nests with six young (P < 0.001), and this
relationship corresponds to the values for visits/young/h. In contrast, maximum wing
lengths did not differ between SY nests with four or six young (P = 0.743), yet vis-
its/young/h at SY nests with four young were high compared to SY nests with six

young.

During 1993, mean total visits/young/h over the five days of observation
(bottom half Figure 19) showed very similar results at three of four treatment groups,
this time the exception being at ASY nests with six young, where young were visited at
lower rates on days 9, 13 and 17 compared to the other three treatment groups. This
lower rate of visitation (i.e. feeding) amounted to approximately one to two less visits
per young per hour at these nests compared to all others. Mean weight growth constants
(Figure 15) during 1993, which were highly variable, showed no significant differences
between the treatment groups (post-hoe tests, all pairs comparisons with Bonferroni
adjustments, all P > 0.9), even though visitation rates were much lower at ASY nests
with six young. Mean maximum weights (Figure 16) were significantly lower for ASY
nests with six young when compared to ASY nests with four young (P = 0.005) or SY
nests with four young (P < 0.001), yet not when compared to SY nests with six young
(P > 0.9). For wing length measures in 1993, calculated growth constants (Figure 17)

were significantly lower for ASY nests with six young compared to the other three

83
treatment groups (all P < 0.005). This corresponds very well to visitation rates which
were lowest for ASY nests with six young and approximately equal for the other three
treatment groups (Figure 19). Maximum wing length attained in 1993 shows no differ-
ences between any of the four treatment groups (all P > 0.9, Figure 18), even though
visitation rates were lowest at ASY nests with six young, and essentially the same at the

other three treatment groups.

There are difficulties inherent in comparing the relationships between parental
age and growth statistically due to the non-independent nature of some of the measure-
ments involved. For example, parental care assessed as visits/h during five observation
periods is not independent over time since day 5 activity is likely influenced by day 1
activity and so on. Although some have argued that these measures are independent
(e.g. Lombardo 1991) and have treated them as such, this is not the case, as days 1, 5,
9, 13 and 17 are clearly repeated measures on the same subject. In addition, growth is
described using an index for many of the above analyses (i. e. growth constants calcu-
lated from the logistic equations), and it is difficult to relate an index, which is a single

measure, to parental care factors measured over time.

In an attempt to address the issue of non-independent measures, I computed an
index of parental care for each nest by summing visits/h over the five observation
periods. Missing data were dealt with in the same manner as described earlier in the
parental care section for the repeated measures analysis of variance. Fourteen nests,

from 1990 and 15 nests from 1993 were used in the following analyses. I also com-

84
puted a new response variable of growth for each nest based on the average for all
nestlings. For example, weight growth constants for all nestlings in a nest were aver-
aged to produce a nest average growth constant which was used in the following analy-
ses. Using multiple linear regression, the dependent or response variables of nest mean
growth constants and nest mean maximum values attained (for both weight and wing
length) were tested for strength of relationships to the independent or explanatory vari-
ables, female age, year, brood size, and indices of parental care (delineated by total,
female or male visits/h). Three models were run for each growth measure and included
all of the explanatory variables above plus one of the parental care indices (total, fe-

male, or male visits/h).

Weight growth constants were not significantly related to any of the explanatory
variables used in the regression models (Table 29) which explained only approximately
10% of the variability in the calculated weight growth constant (R2 range = 0.082-
0.111). For maximum weight attained (Table 29), the factors of year and number of

young were significant in all three models (R2 range = 0.480 to 0.487).

Wing growth constants (Table 30) were significantly related to age of female
and number of young for the models using total visits/h or female visits/h (R2 = 0.352
and 0.431, respectively), and to female age for the model using male visits/h (R2 =
0.390). Maximum wing values attained (Table 30) were significantly related to female
age and number of young for the model including total visits/h (R2 = 0.453), to female

age only for the model including female visits/h (R2 = 0.385), and to female age,

85

Table 29. Multiple linear regression of nest mean growth variables: Weight growth constants and
maximum weights.

 

Dependent variable: Nest Mean Weight Growth Constant

 

 

 

r?2 = 0.086 R2 = 0.062 R2 = 0.111
Model P Model P Model P
Female Age 0.168 Female Age 0.180 Female Age 0.147
Year 0.715 Year 0.798 Year 0.718
Brood Size 0.628 Brood Size 0.973 Brood Size 0.451

Total Visits/h 0.644 Female Visits/h 0.746 Male Visits/h 0.352

 

Dependent variable: Nest Mean Weight Maximum Values Attained

 

 

 

R2 = 0.487 R2 = 0.480 R2 = 0.465
Model P Model P Model P
Female Age 0.325 Female Age 0.351 Female Age 0.322
Year 0.036 Year 0.041 Year 0.042
Brood Size 0.002 Brood Size 0.001 Brood Size 0.001

Total Visits/h 0.459 Female Visits/h 0.620 Male Visits/h 0.492

 

number of young and male visits/h for the model which incorporated male visits/h as

the index of parental care (R2 = 0.478).

These multiple regression analyses show that although female age, year and
brood size often contribute significantly to explaining the variability in the growth
indices, the parental care factors (total, female or male visits/h) are only significant in
one case (male visits/h for wing maximum values). These results suggest that the num-
ber of visits made to the nest is less important to the overall growth of the young when

placed in context with the other explanatory factors.

86

Table 30. Multiple linear regression of nest mean growth variables: VVIng growth constants and
maximum wing values.

 

Dependent variable: Nest Mean Wing Growth Constant

 

 

 

R2 = 0.352 R2 = 0.431 R2 = 0.390
Model P Model P Model P
Female Age 0.021 Female Age 0.015 Female Age 0.014
Year 0.932 Year 0.713 Year 0.986
Brood Size 0.047 Brood Size 0.003 Brood Size 0.155

Total Visits/h 0.916 Female Visits/h 0.079 Male Visits/h 0.232

 

Dependent variable: Nest Mean Wing Maximum Values Attained

 

 

 

R2 = 0.453 R2 = 0.365 R2 = 0.476
Model P Model P Model P
Female Age 0.001 Female Age 0.003 Female Age 0.001
Year 0.446 Year 0.535 Year 0.508
Brood Size 0.016 Brood Size 0.083 Brood Size 0.010

Total Visits/h 0.072 Female Visits/h 0.485 Male Visits/h 0.038

 

Another approach avoiding the problems with non-independence of measure-
ments is to test the strength of the relationship between the condition of young in the
nest (brood weight or mean young weights) and parental care measures (visits/h: total,
female or male) by day of observation. Day 13 represents the time when, overall,
adults are exerting maximum levels of effort as evidenced by the highest of visits/h
being made to the nest. Day 13 is also the time when young are reaching their maxi-
mum weights and will soon undergo a weight recession prior to fledging. Because of

these factors, day 13 is likely the best of any of the five days of observation with which

87

to assess whether or not there are differences in abilities in providing parental care.

Because growth of young was not measured on the same days that behavioral observa-

tions were obtained, weights of young on days 1, 5, 9, 13 and 17 were interpolated

from data points surrounding them. For example, day 13 data were calculated as the

midpoint between days 12 and 14 growth measures.

Multiple linear regression analyses of brood weights (weight of all nestlings in a

nest summed) on day 13 (Table 31) show a significant effect of the number of young in

the nest for all three models (R2 range = 0.486-0.507), which is expected given the fact

Table 31. Multiple linear regression of day 13 posthatch mean growth index variables: Brood
weights and nest mean nestling weights.

 

Dependent variable: Day 13 Brood Weights

 

 

 

 

 

 

 

R2 = 0.466 R2 = 0.501 R2 = 0.507
Model P Model P Model P
Female Age 0.214 Female Age 0.172 Female Age 0.163
Year 0.644 Year 0.854 Year 0.633
Brood Size <0.001 Brood Size <0.001 Brood Size <0.001
Total Visits/h 0.819 Female Visits/h 0.347 Male Visits/h 0.270
Dependent variable: Day 13 Nest Mean Nestling Weights

R2 = 0.260 R2 = 0.289 R2 = 0.260
Model P Model P Model P
Female Age 0.212 Female Age 0.163 Female Age 0.171
Year 0.791 Year 0.996 Year 0.763
Brood Size 0.025 Brood Size 0.031 Brood Size 0.010
Total Visits/h 0.989 Female Visits/h 0.291 Male Visits/h 0.383

 

88
that only two brood sizes were included in the experimental design. The other factors
were not significant. For mean nestling weights on day 13 (Table 31), there was also a
significant effect of the number of young in the nest for all three models, yet less of the
variability in the dependent variable was explained (R2 range = 0.260-0.289). Mean
nestling weights on day 13 for nests with four young were significantly higher than
those for nests with six young (t-test, t = 2.920, P = 0.006), suggesting that young in
smaller broods are receiving, on average, more food per trip to the nest since the paren-

tal care factor was not significantly different.

Mortality Rates of Treatment Group Young

During 1990, mortality of young occurred only in the SY with six young treat-
ment group (Table 32). In addition, each of the four nests in this treatment group lost
young: two young each from two nests, and one young each from two nests. Because
there was lack of mortality in three of the four treatment groups (i.e. cells with zeros),

I did not analyze these data using multidimensional contingency tables.

During 1993, mortality was spread among all four treatment groups and ranged
from 8.3 to 42.9% (Table 32). A test of mutual independence was rejected (x2 =
11.44, df = 4, P = 0.022), so three tests of partial independence were conducted.
Two of the three tests were rejected (both P < 0.014). The test not rejected (x2 =
2.03, df = 3, P = 0.567), was that female age was independent of brood size and the

likelihood to fledge. Because it was not rejected, it was permissible to test the hy-

89
pothesis that likelihood to fledge was independent of brood size (Zar 1984, pg. 77).
This hypothesis was rejected (x2 = 10.208, df = 1, P < 0.001), which indicates that
overall fledge rate was dependent upon the number of young in the nest. During 1993,
mortality rates were higher at nests with six young, regardless of the age of the female

(Table 32).

Table 32. Likelihood of mortality of treatment group young, 1990 and 1993.

 

 

 

Treatment # # not % %
Year Group Fledge Fledge Fledge Mortality
1990 ASY/4 12 0 100.0 0.0
ASY/6 24 0 1 00.0 0.0
SY/4 12 0 100.0 0.0
SY/6 18 6 75.0 25.0
1993 ASY/4 23 5 82.1 17.9
ASY/6 24 1 8 57.1 42.9
SY/4 22 2 91 .7 8.3
SY/6 38 16 70.4 29.6

 

Mortality rates reported here are from all treatment group nests which were es-
tablished. Unfortunately, not all of these nests were used for growth and parental care
measures because of early failure. Because of this, complete comparisons of parental

care and potential impacts on mortality rates were not possible.

90

WEIGHT CHANGES IN PARENTS FEEDING YOUNG

The relationship between adult weights (both male and female, total n = 102
measurements on adults from 16 nests) and age of the young during 1990 showed that,
in general, adult weights decreased through time (Table 33, all data pooled). Linear
regression showed that the slope for pooled data was marginally significantly different
from zero (P = 0.050), yet the fit was poor (R2: 0.038). Although adult weights
decline from the time following hatching through the middle portion of the nestling

period, weights appear to increase at the end of the nestling period.

Analyzed by treatment group (Table 33), only adults at ASY nests with four
young show a slope significantly different from zero (P = 0.017) that corresponds to a
decline in weight over time. Adults at ASY nests lost more weight over time than did
adults at SY nests (slightly steeper slopes), and adults at SY nests with four young

actually increased in weight slightly over time (slope = 0.016), although not

Table 33. Linear regressions of adult weights as a function of age of young during nestling rear-
ing. All data are from 1990.

 

 

 

Adult Weights

Treatment Group n R2 Slope P
ASY With 4 Young 21 0.266 -0.104 0.017
ASY With 6 Young 27 0.087 -0.077 0.136 .
SY With 4 Young 26 0.005 0.016 0.720
SY With 6 Young 28 0.013 -0.033 0.570

Pooled Groups 102 0.038 -0.050 0.050

 

91
significantly. Both of these observations are contrary to predictions that adults at SY
would be stressed to a greater degree from the rigors of feeding young and as a result
would lose more weight. In addition, during 1990, adults at ASY nests with six young
showed much higher levels of parental care measures, yet, on average, lost less weight

than parents at ASY nests with four young.

All regressions for treatment groups, with the exception of ASY nests with four
young (R2 = 0.266, P = 0.017), show a poor and non-significant fit to the linear
model (all R2 < 0.087, all P > 0.13), suggesting both a high degree of variability
between individual adults and perhaps that a non—linear relationship would be more
appropriate. Some individuals within the same treatment group gained weight, while
others lost weight, suggesting some individuals are better able to maintain or even
increase energy reserves while at the same time feeding young. This finding is
interesting in light of the significance of the nested growth models presented earlier
which suggested that much of the variability in growth can be attributed to the

individual nest.

DISCUSSION

AGE CLASS DIFFERENCES: PRE-HATCHING

Clutch Size and Egg Weights

Clutch sizes were significantly smaller (Figure 2, Table 4) and were initiated
significantly later (Figure 1, Table 3) for SY females throughout the study. Although
there were some years when mean clutch size values were similar for the two age
classes and standard error bars overlapped (e. g. 1987), and all of the assumptions of
the test could not be met, analysis of covariance shows a strong effect of female age,
even after controlling for a significant effect of initiation date (see also Perrins 1965,
Stutchbury and Robertson 1988). Similar relationships between female age and clutch
size in Tree Swallows have been reported previously (DeSteven 1978, Stutchbury and
Robertson 1988). The significant initiation date finding is important because it has been
shown in a number of species, with all else being equal, later nesters produce fewer
young which survive (e. g. Reese and Kadlec 1985, Perrins 1970). Because the number
of young banded at Panola Plains which return to breed is very low ( < 1%), it has not

been possible to test for differential young survival rates.

Both inexperience and foraging deficiencies on the part of SY females have been

thought to be important factors resulting in smaller clutches and lower reproductive
92

93
output (Lack 1968, Perrins 1970). Since food abundance impacts both clutch size (e. g.
Hussell and Quinney 1987), and growth of young (6. g. Quinney er al. 1986) in Tree
Swallows, Stutchbury and Robertson (1988) reasoned that if clutch size differences
between SY and ASY females were due to foraging deficits, then these differences
would likely be lessened in a food rich environment. This is a compelling and testable
hypothesis, yet they provided no evidence for this. Desrochers (1992) found that clutch
size in European Blackbirds overall was increased when he provided supplemental
food, yet the magnitude of differences between SY and ASY females persisted. DeSte-
ven’s (1978) study was conducted at Long Point, Ontario where prey abundance, in the
years following her study (measured as insect biomass indices, Hussell and Quinney
1987) was found to be much higher than at Panola Plains (Hussell et al. 1990). The
range of differences between age classes reported by DeSteven (197 8) was greater (0.8
and 0.9 eggs/nest, respectively, in a two year study) than reported in my study (0.186
eggs/nest in 1987 to 0.647 eggs/nest in 1991), yet she also reported overall larger
clutch sizes for both female age classes as well. Even though there were potentially
large differences in prey abundance between DeSteven’s study and this study, the dif-
ference in magnitude of clutch size between SY and ASY females is in the opposite
direction of that predicted by Stutchbury and Robertson (1988). Although prey abun-
dance has been documented to have a significant proximate influence on clutch size
(Hussell and Quinney 1987, Jarvinen and Viisanen 1982, 1984, Perrins and McCleery

1989), it does not appear to affect the relationship between SY and ASY females.

94

In this study individual egg weights were also shown to be smaller for SY fe-
males, and yearly differences ranged from 0.007g/egg in 1993 to 0.094g/egg in 1992.
The range of differences was similar when nest mean egg weight was considered. Egg
weights in SY Tree Swallows were also shown to be significantly smaller by DeSteven
(1978) with a similar range of differences. Wiggins (1990) reported marginally non-
significant differences in mean egg weight between the two age classes, yet his samples
sizes were much smaller than what is reported here (71 = 16 SY female nests, 121 ASY
female nests, over three years). In this study, depending upon the year, nested analysis
of variance showed that between 74 and 81% of the variability in egg weights was
attributable to the nest (i. 6. individual female) where the eggs were laid. Factors con-
tributing to the importance of the nest include primarily attributes of the individual
female, and possibly also nestbox location and time of season. Because of the signifi-
cant effect due to the individual female in the nested analysis of variance, female age
was a significant factor only during 1992 (Table 9, P = 0.028). Wiggins (1990) also
found a high degree of among-clutch variability in Tree Swallows, which amounted to
79% of the total variation in egg weights. The large proportion of the variability in egg
weights and measurements attributable to individual females is also reported in other
species. Ojanen er al. (1981), in four species of passerines, also showed that approxi-
mately 50-70% of the variation in egg dimensions (weights were not reported) was
attributable to the female from which they came (see also Jarvinen and vaisanen 1983,

Jover er al. 1993).

95
Differences in egg weights could be due to genotypic differences of females, or
phenotypic adjustments due to factors such as seasonality, weather, food abundance or
laying sequence (O’Connor 1979). Although it appears that the age of the female has
considerable influence on the weight of eggs, this effect is difficult to discern using
nested models where an individual female factor is included. There appear to be inher—
ent differences in the abilities of individual females to produce larger eggs, and these

differences have an age-related component as well.

Ricklefs (1974) argued that if a female is stressed energetically, the quality and
size of eggs she produces are not altered appreciably, but only the number, whereas
others have shown that poor feeding conditions at the time of egg formation can result
in lower quality or smaller eggs (Martin 1987, Ojanen et al. 1981, Perrins 1979). Egg
weights have been shown to be correlated with the overall physical condition of the
female (Jarvinen and Vaisanen 1984, Murphy 1980, Murphy 1986), and the size and
quality of eggs have been shown to increase in times of high food abundance (Bryant
1978a, Ewald and Rohwer 1982, Hogstedt 1981b, Jarvinen and Véiisanen 1984). Egg
weights in my study varied from year to year (Figure 4), and presumably these differ-
ences could be due to yearly changes in food availability. The magnitude of differences
between SY and ASY egg weights has also varied across years, yet whether or not this

yearly variability is due to age-related responses to food abundance is unknown.

In this study, repeatability of clutch size was lower in SY females advancing

into the ASY age class when compared to ASY females encountered over two subse-

96
quent years. This suggests there is at least some improvement with age for SY females,
and the improvement is in the direction predicted, yet comparing mean clutch sizes for
this subset of females showed they were not significantly different. In addition, the
improvement recorded (approximately 0.18 eggs/ nest) was of a lower magnitude than
the differences between age classes in an average year (approximately 0.4 eggs/nest).
Similar results were reported by Smith (1993) in Marsh Tits (Parus palustris), where
significant differences of 0.37 eggs/nest for age class comparisons were larger than the
non-significant 0.14 eggs/nest for individual females aging from SY to ASY. The
majority of studies investigating age-related reproduction have been cross sectional
rather than longitudinal, and although studies of long-lived species have shown in-
creases in clutch size with age (6. g. Coulson 1966, Hamann and Cooke 1987), rela-
tively few studies in short-lived species have shown actual improvements in individuals
as they age. Changes in mean clutch size with age can be the result of an improvement
of performance in individuals, or it can be a result of differential mortality on poorer
SY performers, thus shifting the mean values for the cohort upward (the selection
hypothesis of N01 and Smith 1987). H0gstedt (1981a) found no increase in clutch size
as individuals aged in a small sample of Magpies (Pica pica). Similar results were
reported by van Noordwijk et al. (1981) for a large sample of Great Tits and by N01
and Smith (1987) in Song Sparrows (Melospiza melodia). Desrochers and Magrath
(1993), studying European Blackbirds (T urdus merula), did report an increase in clutch
size of 0.4 egg/nest in individual females aging from one to two years, and this being

in a species with a smaller clutch size than Tree Swallows (range 1-5 eggs, no mean

97
value presented). Desrochers and Magrath (1993) concluded that this increase in clutch
size was not due to a reduction in the proportion of poor performers, but improvements
in individuals over time—they found no correlation between fecundity in SY females

and their probability to survive to the next year.

In contrast to clutch size repeatability, egg weight repeatability was much higher
for both groups considered in the analyses (SY aging to ASY, and ASY individuals
encountered over two years). In other words, eggs produced by the same female re-
mained generally stable in weight from year to year, regardless of age class. Significant
intraclass correlations indicate that the variability associated with year-to-year changes
in the same individual is low compared to the variability between individual females.
Even though repeatability of egg weights was high, the values were lower for SY fe-
males, suggesting that the potential for change as SY females aged was greater than for
ASY females encountered over two years. This suggests there is at least some im-
provement with age for SY females, and the improvement is in the direction predicted,
yet comparing mean egg weights for this subset of females showed they were not sig-
nificantly different. The improvement recorded (approximately 0.031g/egg) was of a
lower magnitude than the differences between age classes in an average year

(approximately 0.041g/egg).

98

Hatching Success, Egg and Nest Failure

In this study, SY females showed a greater likelihood of egg loss over time than
ASY females. This is evidenced by both a lower rate of hatching (Table 10) and higher
mortality rate of eggs in the nest (Table 11). Lower egg hatching rates by SY females
in my study are contrary to DeSteven’s (1978) findings where the hatching rate of eggs
from SY female Tree Swallows was slightly higher than eggs from ASY females.
DeSteven reported only two years of data, and although mean hatching rates were
higher for SY females, they were significantly higher only in one of two years when
analyzed separately. An overall lower rate of hatching was also observed in SY females
by Brown (1978) in Purple Martins (Prague subis), as well as other species (Aldrich
and Raveling 1983, Boekelheide and Ainley I989, Roskaft et al. 1983). Many studies,
however, show no differences in hatch rate due to age (Hannon and Smith 1984, N01
and Smith 1987, Perrins and Moss 1974, Rockwell et al. 1993, Sather 1990). The
lower rates of hatching by SY females reported in this study may be due to poorer
quality incubation behavior, higher rates of infertility, or a higher rate of developmen-
tal abnormalities resulting in embryo death. A lack of constancy in egg temperature
maintained during incubation can lead to developmental abnormalities and influence
hatching success (Drent 1975). Aldrich and Raveling (1983) showed that younger
Canada Geese were less attentive incubators and this was correlated with lower hatch-
ing success. Older females started incubation in better body condition than younger

females and thus were able to spend less time foraging which resulted in more time

99
sitting on the eggs. In my study, SY females produced smaller eggs, and hatching
success is also generally thought to be related to egg size (Martin 1987). Rofstad and
Sandvik (1985) reported that smaller eggs in the Hooded Crow (Corvus corona) were
less likely to hatch. Jarvinen and vaisanen (1983) showed that hatching success in the
Pied Flycatcher (F icedula hypoleuca) at the northern end of the species’ range was
related to egg size and that larger eggs hatched with greater frequency. They suggested
that larger eggs could better withstand cooling conditions when the female left the nest

to forage because of their larger thermal mass. In my study, SY females only produced

 

approximately 90% of ASY clutch biomass. It is not clear whether or not differences of
this magnitude would produce differences in thermal mass great enough to influence

foraging dynamics and the time spent off the nest during incubation.

The higher rate of egg mortality observed could be due to inherently higher pre-
dation rates on eggs in SY nests, which may be related to nestbox location. Nest site
location has been shown to influence several breeding variables in other species, such
as nest initiation date, clutch size and nest failure rate (Hogstedt 1980, Middleton 1979,
Nilsson 1984, Perrins and Moss 1975), and younger females have been shown to have
a higher likelihood of settling in marginal habitats in Black-billed Magpies (Pica pica)
(Reese and Kadlec 1985). It appears that the dispersion of SY nests at Panola Plains
was not random, but was clumped, and SY females were more likely to nest in areas of
the plot characterized by higher density of shrubs, or on the edge of the nestbox arrays

which are often adjacent to forest edge. Tree Swallows prefer open habitats, and, if

100
possible, avoid brushy areas and forest edge (Rendell and Robertson 1990). These
brushier areas of the plot are ideal House Wren (T roglodytes aedon) habitat, and inter-
ference competition by wrens (Finch 1990, Rendell and Robertson 1990) for nesting
sites may partially explain the higher egg losses in SY females. House Wrens will enter
nestboxes of other species and poke holes in eggs, remove eggs entirely from the nest-
box, and in some cases, kill young (Belles-Isles and Picman 1986, Kendeigh 1941).
Although I did note wren predation when it was obvious, I do not have detailed enough
information to test this interference hypothesis. Wren interference often results in nest—
box takeover (personal observation), and total nest failure was significantly higher for
SY females during two of eight years of the study (Table 12) compared to ASY fe-
males. These same two years (1989, 1992) were also characterized by episodes of cold
and wet weather during the nestling phase which had a significant impact on young in
the nest (Beaver er al. 1994). During these periods of inclement weather, adults aban-
don the nest, leaving the young to die from exposure. Late nesters which were still
incubating are also impacted and may also abandon nests. Since SY females initiate
nests significantly later, this may partially explain their higher rate of nest failure in

those years.

Another factor correlated with increased likelihood of nest abandonment during
incubation reported in the literature is egg size. Smith et al. (1993) found that higher

rates of nest abandonment in European Starlings occurred at nests with smaller eggs.

101
These results are only correlational, yet are interesting in view of my results where SY

females produced smaller eggs and also showed higher rates of nest failure.

THE ENERGETICS OF EGG PRODUCTION AND INCUBATION

In terms of clutch biomass, SY females, over the entire study, produced on av-
erage 0.912 g/nest less than ASY females, or approximately 90% of ASY output.
Walsberg (1983) calculated that for five small passerines, the energetic content of the
clutch is only approximately 2% of the female’s total energy expenditure during a
single reproductive event. So, although differences in clutch size and egg weight be-
tween SY and ASY female Tree Swallows are statistically significant, they appear
initially to be very small in light of a female’s total energy output for the breeding
attempt. However, energy for egg formation is not the only expenditure needed for
successful egg production and hatching, as there are many other related costs. Although
the male helps defend the nestbox and provides the nest lining consisting of feathers he
has collected, nest building, egg laying and incubation behaviors in Tree Swallows are
exclusively female (Turner and Rose 1989). Female Tree Swallows, like most passer—
ines, are “daily surplus” users of energy during egg formation (Perrins and Birkhead
1983), that is, they increase intake of daily energy in order to produce eggs. There are
no reports of male Tree Swallows feeding their mates during this time, so energetically
speaking, the female is entirely responsible for the formation of eggs. She must take in

enough energy for maintenance, plus additional energy for egg formation and produc-

102
tion of one egg per day for two to seven days, depending on the size of her clutch. In
addition, there is the cost in energy prior to egg formation required for gonadal devel-
opment. Once the clutch is complete, the female must then keep the eggs warm for
approximately 14 days during the incubation period. The incubation period is likely the
least energy demanding phase of breeding (King 1974, Walsberg 1983, Williams
1988), yet the male Tree Swallow does not feed the female during this time, so she
must leave the nest to forage in order to maintain her energy balance. Even in species
where males feed the incubating female, weight loss by females during incubation has
been reported (van Balen 1973). Rewarming of eggs upon returning from foraging in
single sex incubators (e. g. Tree Swallows) is energetically expensive, as it has been
shown that metabolic rate increases up to three times when cold eggs are rewarmed in
the European Starling (Biebach 1979). Westerterp and Bryant (1984) also found in a
comparative study of aerial insectivores that incubation is more energetically expensive

in a single-sex incubator compared to species which share incubation duties.

Although egg formation and incubation can be energetically expensive when
viewed in a broader context than simply laying and then sitting on eggs, in terms of
age-related effects, there is little information available. However, if SY females were
indeed deficient in foraging abilities, or, perhaps more importantly, deficient in the
overall management of time and energy, one would expect them also to be less effec-
tive incubators and this could result in fewer eggs hatched or a higher likelihood of nest

abandonment.

103

AGE CLASS DIFFERENCES: POST-HATCHING

Male versus Female Contributions

In species that exhibit biparental care, it is thought that male parental care con-
tributions to the young are made in inverse proportion to the abilities of females to
successfully fledge young without male assistance (Clutton-Brock 1991, Emlen and
Oring 1977, Orians 1969b), and that geographical areas with greater resource avail-
ability will be characterized by lower levels of parental care by males. Dunn and Rob-
ertson (1992) reported that male parental care in Tree Swallows was more important to
overall nesting success in areas of comparatively lower food abundance, which sup-
ported this view, yet how this is related to SY females’ abilities to provide parental care
and the assistance they receive from male partners is unknown. In areas of high insect
abundance, Tree Swallows, which are normally monogamous, may exhibit a low inci-
dence of polygyny (5-8% of males, Dunn and Harmon 1992, Dunn and Robertson
1992, Quinney 1983). Because of this, some females may go without any male assis-
tance in feeding young, since males rarely contribute at secondary nests (Dunn and
Harmon 1992). Available insect biomass was much lower at Panola Plains (D. Hussell,
personal communication, Hussell er al. 1990) when compared to the Ontario and Al-
berta sites used by Dunn and Robertson (1992). In my population of Tree Swallows,
few differences were shown in the frequency of visits to individual nests when males
and females were compared, and within treatment groups, the frequencies of male and

female visits were significantly correlated. The importance of biparental care in the

104
Panola Plains population is emphasized by the observation that in all cases where a nest
appeared to have only one adult parent actively feeding young, the nest failed. There
are reports of single parents successfully fledging young in areas characterized by high
insect abundance (Dunn and Harmon 1992, Quinney 1986). If SY females were less
able to raise young because of deficiencies in their abilities to provide parental care, it
is expected there would be differences in the proportion of male and female contribu-
tions at SY and ASY nests, yet this was not observed in this study. Equal provisioning
could be the case if SY females were more likely to mate with SY males that were also
deficient in breeding capabilities, yet I do not have any evidence for assortative mating
according to age. Most often, I was not able to determine the age of the male other
than a minimum known age, yet there was no difference in the likelihood that an SY
female would mate with an unbanded male (potentially an SY male) or a previously
banded male (a certain ASY male, unless banded as a young in the nest) when com-
pared to ASY females (n = 74 pairs, x2 = 1.66, df = l, P > 0.1). While some stud-
ies have shown assortative mating with respect to age (Bryant 1979, Crawford 1980,
Lessells and Krebs 1989, Smith 1993), others have not (Harvey er al. 1985, Perrins
and McCleery 1985). Furthermore, Quinney (1986) and Leffelaar and Roberston
(1986) also found that male and female visits by Tree Swallows to the nest were ap-
proximately equal, whereas Lombardo (1991) did find some significant differences,
although all of these studies measured and analyzed parental care data somewhat differ-
ently. In the present study, of the four nests in 1990 which showed significant differ-

ences in the proportion of male and female visits, two were SY nests with six young

105
(both of which had significantly higher levels of female visits, which incidentally is
contrary to predictions), one was an SY nest with four young (higher male visits), and

one was an ASY nest with four young (higher male visits).

Enoksson (1993) reported in European nuthatches (Sitta europaea) that the
male’s age had no effect on any of the reproductive variables measured in his study.
Other studies have found that the male’s age was a significant factor contributing to the
fledging of young, yet had no effect on the clutch size produced by the female. Harmon
and Smith (1984), for example, showed in Willow Ptarmigan (Lagopus lagopus) that
pairs of two adults (versus some combination including a subadult) fledged the most
young, yet male age did not influence clutch size or laying date of the female. Perrins
and McCleery (1985), in Great Tits, also found that the age of the male significantly
affected fledge rate, yet had no influence on female clutch size (see also Reese and
Kadlec 1985). Female European Blackbirds did not adjust clutch sizes based on the age
of the male, even though older males provided higher quality parental care, as it was
shown that they were more proficient foragers than younger males (Desrochers 1992,
Desrochers and Magrath 1993). It may be the case in many species that females are not
able to assess the quality of their mates prior to the time when males and females begin
sharing of parental care duties by feeding young (6. g. Desrochers and Magrath 1993,

Slagsvold and Lifjeld 1988, 1990), unless they had previously mated.

106

Overall Parental Care

Williams (1988) found a positive association in Tree Swallows between the
number of visits made to the nest and metabolic rates measured using the doubly la-
beled water method. Greater rates of visitation resulted in higher rates of energy ex-
pended, and he suggested that nest visitation rates were a reasonable measure of paren-
tal effort (see also Bryant and Tatner 1990). Bryant and Westerterp (1983) found no
significant differences in average daily metabolic rates that could be attributable to
female age during the time of feeding nestlings. Even though no differences were
found, the amount of energy collected for delivery to young in the nest per unit of
energy expended could be vastly different depending on potential differences in forag-
ing efficiency (Bryant 1982). Because of potential differences in load sizes of food
delivered to young which vary according to year, age of young, individual nest, and to
a lesser degree, the sex of the adult parent (see below), inferences drawn from the
number of visits made to the nest as they relate to the overall growth of the young will
be much more conclusive if these factors can be taken into consideration (Royama

1966).

Graphically, there appear to be differences in parental care associated with
treatment groups and years (Figures 6 through 9), yet repeated measures analysis of
variance on total visits/h did not detect any significant effects due to female age (nest
type) or year. The response variables female visits/h and male visits/h gave the same

results. Only brood size had significant effects on nest visitation rates in all three of

107
these models. At nests with four young, 6 consistent pattern of total, male and female
visits/h was detected, with no effects due to age of the female or year, so it appears that
the level of parental care is very uniform at these nests, including across years which
may differ in extrinsic factors such as food abundance and weather conditions. The
pattern at nests with six young, however, was much different, with a significant female
age x year interaction for total, male and female visits/h models. This indicates a
variable response between the two years under study for nests with SY vs. ASY fe-
males and this is apparent in Figures 6 and 8. The prediction that adults at SY nests
would provide lower levels of parental care appears to be partially supported during
1990 at nests with six young when adults at SY nests provided parental care at only
approximately four-young levels. During 1993, however, the prediction is not sup-
ported, and the response is actually in the opposite direction of that predicted; adults at
SY nests provided a higher level of parental care when compared to ASY nests which

were now visiting at approximately four-young levels.

The uniform response in terms of visits/h at nests with four young (with no year
effect) indicates that parents generally do not differ in their abilities to provide care at
the levels necessary for this brood size. With brood sizes of six young, however, the
variability in response between treatment groups increases, indicating that the increased
burden of feeding greater numbers of young may be taxing the abilities of some adult

pairs to provide adequate parental care. Interestingly, however, this apparent inade-'

 

108
quacy of response switches from SY nests with six young in 1990 to ASY nests with

six young in 1993.

Growth of Young

Graphically, the growth curves in 1990 appear to indicate effects due to both
age of female and number of young in the nest, and the observed responses are in the
direction predicted (Figures 11 and 12). As predicted, ASY nests with four young show
the best growth over the observation period (for both weight gain and wing length
increase), whereas SY nests with six young show the poorest. The young at the two
remaining treatment groups (ASY nests with six young and SY nests with four young)
show intermediate values, and appear to grow at approximately the same rates. These
results are similar to those of DeSteven (1978), who showed that fledging weights of
young were negatively affected by brood enlargement at SY nests, but not at ASY
nests. During 1993 (Figures 13 and 14), however, the same relationship does not hold,
particularly for wing growth where little, if any, differences are shown between treat-
ment groups. For the variable of weight there is some separation between treatment
groups, yet the separation is slight and relationships between treatment groups (best to

poorest) are different than those observed during 1990.

The non-nested analysis of variance models on growth variables reveal some
significant main effects (with the exception of weight growth constants), yet the inter-

pretation of these results is severely confounded by significant interaction terms (Sokal

109
and Rohlf 1981) which are also frequently present in the models. For example, results
of analysis of variance on wing growth constants (Table 25) show that three of the four
interaction terms are significant. Even though there is a significant effect due to female
age, the female age factor also interacts with year alone, and year and brood size in
combination. These significant interactions suggest that although there are statistically
significant differences between treatment groups for the growth variables measured,

these differences are not consistent across levels of factors used in the models. Signifi—

 

cant second order interactions make interpretation particularly difficult, yet do indicate

highly variable growth patterns across years and nest types.

Nested analysis of variance reveals that growth of young in all four treatment
groups was much more heavily influenced by the factor of nest than by female age,
year, or number of young in the nest, and this was true of all the growth variables
reported, especially the fitted growth constants for weight, where only 3.5 % of the
variability could be attributable to factors other than the nest. Growth of individual
young has been shown to be highly dependent upon the nest in which they are raised
(DeSteven 1980, Quinney et al. 1986). This study is in agreement since variability in
the growth patterns observed was much more strongly influenced by the individual nest
than treatment groups characterized by the age of the female and brood size. Because
the factor of the nest is significant for all of the dependent growth variables tested, this
result suggests that individual adult pairs feeding the young may be performing quite

differently when compared to one another. Regardless of the age of the female or year,

110
adult pairs are likely delivering food at different rates, or in different quantities at the
same rate of visitation (Royama 1966), and highly variable growth patterns result.
O’Connor (1975) concluded that the young of aerial insectivores exhibited highly vari-
able growth patterns due to the unpredictable patterns of food availability is determined
primarily by prevailing weather conditions (see also Bryant 1978a, Jones 1987c, Turner
1983). The unpredictability of food resources results in the number of visits made to
the nest and the amount of food delivered to the young to be highly variable (Bryant
1978a). Studies on other aerial insectivores have drawn similar conclusions; growth of
young is closely associated with the amount of food available to the adults (Bryant
1975, 1978b, Quinney er al. 1986). Depending upon the time frame, some nests can be
impacted by inclement weather and the associated decrease in insect abundance much
more than other nests, and this can be reflected in the overall growth of young. With
these relationships in mind, the data presented here suggest that the observed patterns
of growth potentially result from differences in pairs’ abilities to capitalize on fluctuat-

ing resources.

Relationship Between Parental Care and Growth of Young

Ricklefs (1977) makes the point that, evolutionarily speaking, parental care,
feeding of the young, and growth of young are interrelated and optimized as a unit.
Because the growth of young is a direct result of the quantity and/or quality of parental

care, the relationship between these factors is the key to understanding whether or not

 

111
the ability to provide parental care is a possible determinant of clutch size in younger

females.

During 1990, adults at SY nests with six young made visits to the nest at a rate
approximately 80% of the other three treatment groups (Figure 19), and this is re-
flected in the poorer growth of young in this treatment group (Figure 11). The 1990
growth data and parental care data, when considered as separate results, support the
hypothesis that adults at SY nests provided a lower level of parental care and that
young grew more poorly as a result. Viewed within the context of all treatment groups,
however, there is a lack of correspondence between parental care and growth measures;
high rates of visitation did not always translate into better overall levels of growth.
Even though SY nests with six young showed the poorest growth and lowest per young
visitation rates, the lack of correspondence between parental care and growth in the
other treatment groups weakens the conclusions. Support for the major hypothesis that
SY females are constrained by their inability to provide parental care would be much

stronger for the 1990 data if all treatment groups responded as predicted.

During 1993, the lowest levels of visits/young/h were measured at ASY nests
with six young and performance was approximately 80% of that observed for means
from the other three treatment groups (Figure 19). Growth curves also showed that
ASY nests with six young had the poorest weight gain as well (Figure 13), yet the

curves were not clearly separated, and this is particularly evident for wing growth

1 12
(Figure 14). This is in contrast to 1990 data when SY nests with six young clearly

showed the lowest levels of parent visitation and the poorest growth curve.

Correspondence between visitation rates and growth measures during 1993 do
not support the prediction that adults at SY nests would provide lower quality parental
care resulting in poorer growth of young. The correspondence between visitation rates
and growth measures within treatment groups is better during 1993 than that observed
during 1990, yet still there is a marked lack of consistency; high rates of visitation did
not always translate into better overall levels of growth. This is underscored by the
results of multiple regression analyses on growth variables which showed that indices
of parental care rarely explained significant proportions of the variability, whereas the
other independent variables of nest type (SY or ASY), year and brood size often did.
This suggests that factors other than the number of visits per se made to the nest are

more important contributors to the growth patterns of young observed in this study.

Even though there were generally no differences in proportions of male and fe-
male visits, the quality of the visits themselves may be highly variable. Across two
years of observations (1990 and 1993), the proportion of male and female visits to the
nest did not differ to any appreciable degree. However, when boluses were collected
during 1991 (unfortunately, sample sizes were small and only collected during late or
renesting), I found that males delivered significantly larger boluses to the young (after
controlling for age of young using analysis of covariance), which essentially can be

translated into higher quality parental care for the same quantity of visits to the nest.

113
The measures of bolus weights for males did not change over time (slopes did not differ
significantly from zero), yet bolus weights did increase for females. Females increased
load sizes as the young aged, yet the males did not. This could be attributable to a
tendency of males not to increase parental investment as the young aged, similar to that
which was observed in Great Tits (Slagsvold and Lifjeld 1990). Jones (1988) also
reported that males were more likely to engage in self feeding during times of low prey
abundance, whereas females were more likely to feed young and lost more weight as a
result. Boluses were also collected during 1989 and 1990, and when the years are
analyzed together, it was found that mean bolus weight (once again taking age of young
into account using analysis of covariance) was significantly different between years.
This is not likely attributable to different collection times since these were all late
season nests, and all years showed significant increases in bolus weight as the young
aged. More likely the cause is differences in prey abundance between years, and this
can have a profound effect on how hard the adults have to work in order to keep up
with the demands of the young in the nest. During times of low prey abundance adults
most likely have to work harder to maintain the number of visits/h and to maintain the
amount delivered per bolus. Another factor contributing to these yearly differences is
the fact that during 1990 and, in particular 1989, some boluses were collected as
groups which effectively reduces the variance and thus makes yearly differences more

likely.

114

I did not measure load size delivered by males and females at the treatment
nests that I was observing the number of visits at during 1990 and 1993, primarily
because of the invasiveness of the ligature technique used to determine load sizes
(Johnson et al. 1980). Results of other studies on aerial insectivores have revealed
highly variable relationships between load sizes delivered and age of young and sex of
adults (e.g. Jones 1987b, 1988, Martins and Wright 1993). Nests from late in the
season in this study showed clearly that bolus size increased with the age of the young,
and this trend was significant in all three years. Others have also shown a positive
relationship between bolus size and age of the young (e. g. Johnson and Best 1982,
Knapton 1984, Walsh 1978), yet Turner (1983) found that bolus size did not vary with
age of the young between 7 and 18 days (Swallows, Hirando rustica, and Sand Mar-
tins, Riparia riparia), and similar findings were reported by Jones (1987b, also in H.
rustica). Significant differences were found in 1991 (not measured in the other two
years) between male and female load size after controlling for the age of young using
analysis of covariance. Females delivered loads approximately 80% of male size. These
differences could be due to body size differences as male Tree Swallows are signifi-
cantly larger in both weight and wing length. These findings are contrary to the study
by Jones (1987b) who found, with a much larger sample size, that males, which were
also significantly larger in body size, delivered significantly smaller load sizes. Another
important finding was that of significant differences in average load size between years,
suggesting that either prey abundance or frequency of prey sizes taken by adults dif-

fered between years. In addition, much of the variability in bolus weight from all three

115
years was due to day-to-day changes in load size (significant days posthatch covariate)
and this variability obscured any differences due to nest type (SY or ASY nest). It
appears that the lack of correspondence between the number of visits to the nest and the
growth of young may be partially due to differences in load sizes delivered on each

visit. A much more detailed study would be needed to assess this fully.

Mortality of Treatment Group Young

Overall mortality rates across treatment groups during 1993 were 27.7% while
during 1990 were only 8.3 %. In addition, mortality of young during 1993 was spread
across all treatment groups, whereas during 1990 only SY nests with six young were
impacted. However, even though mortality was recorded in all treatment groups during
1993, nests with six young were much more severely impacted, regardless of the age of
the female. While nests with six young suffered 35.4% mortality of young, nests with
four young suffered only 13.5 % mortality. Turner (1983) also found that larger broods
were more severely impacted by inclement weather, primarily due to larger disparities
in weight of young within a brood, apparently putting some young at greater risk dur-
ing times of food deprivation. Inclement weather was a major factor on mortality rates
during 1993, since most mortality occurred following cold and wet weather on 16 and
19 June. As mentioned earlier, some of the nests that were included in the mortality
values were not included in the parental care analyses simply because not enough ob-

servational data was collected on some of the nests.

116

D0 Adults Undergo Increased Stress While Feeding Young?

Although sample size was small and restricted to only one year, it does not ap-
pear that adults in this study lost weight over time to any appreciable degree during the
time of feeding young. Collectively, Tree Swallow adult weights appear to decline
during the early portion of the nestling phase and then increase toward the end, yet the
data are highly variable. Similar results were reported for Swifts (Apus opus) (Martins
and Wright 1993) and European Starlings (Ricklefs and Hussell 1984). Jones (1987c)
also found that the greatest amount of weight loss occurred prior to and during the
early phase of feeding young, weights then leveled off through the remainder of the
nestling period. Jones (1988) also found that adults were heavier during times of high
prey abundance, and large changes in weight could occur in a matter of only several
hours (see also Bryant 1979). This seems to be a general rule in aerial insectivores, and
caution must be used when comparing weight dynamics of adults over time unless a
specific time frame is used and can be somehow controlled for in each individual. This
was not possible in my study. Some individuals may be impacted to a greater degree
than others by events such as inclement weather, and their respective weight changes
may reflect this. The data set presented here suggests that changes in weight are related
to the age of the young, yet the data set is too small to adequately test the hypothesis
that adults at SY nests, and particularly those with six young, would lose a greater
amount of weight. When compared to other treatment groups, only adults at ASY nests
with four young showed a linear decline in weight that was significantly different from

zero, which is contrary to both the predictions in this study, and the results of Martins

117
and Wright (1993), who showed that with increased brood sizes adults lost a greater
amount of weight. More detailed comparisons of weight changes over time in relation

to female age would require a much larger sample size than what is reported in this

study.

CONCLUSIONS

Prior to the hatching of eggs, when adults begin a vastly different phase of
providing parental care, it is clear that SY females are deficient in several aspects of
reproductive biology when compared to older ASY females. Clutch sizes and egg
weights were smaller for SY females, which were shown to produce approximately
90% of ASY clutch biomass. Rates of egg loss and nest failure were also higher for SY
females, and hatch rates were lower, so compared to ASY females, potential reproduc-
tive output is already lower for SY females prior to the time that nestlings need to be
cared for. These results contribute to the large body of evidence which circumstantially
supports the constraint hypothesis. This type of evidence, however, does not provide
further insight into the mechanisms of clutch size determination and any potential influ-
ence of female age. Because there are many factors which influence reproductive suc-
cess (including age), it is extremely difficult to separate the ultimate causes of clutch

size determination from proximate constraints (Slagsvold and Lifjeld 1988).

In some populations, a large number of SY females never have the opportunity
to breed (Brown 1969, Smith 1978, Stutchbury and Robertson 1985). Due to the fact
that so many do not breed, the real differences between SY and ASY females are po-
tentially much larger than those reported here and elsewhere. It seems clear that it

would be advantageous to breed if at all possible. For example, Gustafsson and Part

118

119
(1990) report that lifetime reproductive success in the short-lived Collared Flycatcher
(Ficedula albicollis) is higher for those individuals that breed in their first year of life
compared to those that wait. In Tree Swallows, if SY females do not breed, it is esti-
mated that a female must produce an average-sized brood for the following three years
in order to replace herself, yet the average life span is only 2.7 years (Butler 1988).
Given these results, clearly there are advantages to breeding as SY’s, even at a reduced
rate of production. These advantages to early breeding would be offset only if inordi-
nately high mortality rates resulted following breeding as an SY, or if the possibility of
producing surviving young was extremely low compared to the population mean. The
importance of clutch-size-dependent adult mortality has been debated in the literature
(6. g. Alerstam and Hogstedt 1984), yet the impact in relation to female age has not
been investigated directly. There are hints that SY females may have higher mortality
rates than older females (see Lombardo 1986), yet there have been no comparisons of
mortality rates of breeding versus non-breeding SY females. This would be extremely
difficult to test given the fact that individuals who do not breed are rarely captured and

banded: a necessary event in order to follow an individual over time.

The potential for improvement of clutch size and egg weight with age was
somewhat greater for SY females (i. 6. lower repeatability), yet the magnitude of im-
provement was less than the yearly differences observed between age classes. Because

of the high mortality rates of adults (> 50%), it is difficult to determine whether 'or not

120
improvement is truly age-related or is due to differential mortality on the part of poorer

performing SY females.

Following hatching, the entire age-related aspect of the study changes since
males are now involved to a much greater degree by providing direct parental care to
the young in the form of food deliveries. In general, efforts by male parents in this
study did not differ significantly from that of the female. In most cases the numbers of

visits made to the nest by females and males were highly correlated.

The data from 1990 seem to support the hypothesis that adults at SY nests were
unable to provide adequate levels of parental care at nests with six young, when com-
pared to ASY nests, and this was reflected in the poorer growth of those young.
Moreover, the only mortality of treatment group young during 1990 occurred at SY
nests with six young. In addition, the mortality that did occur took place in all four of
the SY nests with six young. Two nests lost two young each, and the other two nests
lost one young each. These results also support the predictions that adults at SY nests
would provide lower quality or quantity of parental care. It appears that in 1990 adults
at ASY nests were better able to cope with the rigors of providing care to a larger
brood of six young. During 1993, results were much different than those observed
during 1990. The largest discrepancy was that SY nests with six young did very well
during 1993—better than ASY nests with six young, which did the most poorly of all
the treatment groups. As such, the data from 1993 do not support the hypothesis that

SY nests would fare more poorly, and, in fact, the opposite is true. Results from 1993

121
are distinct in other aspects as well. Although clutch sizes were significantly smaller for
SY females, egg weights were not, as means were nearly identical in 1993. Hatching
success was also significantly higher for SY females when compared to ASY females;
1993 being the only year where this occurred. Numbers of nesting SY females during
1993 were also the lowest reported in the study. Nest failure during 1992 approached
100% and was geographically widespread (Beaver et al. 1994), and this presumably
resulted in very low recruitment into the 1993 breeding population, hence the low
numbers of SY females. It may be that the characteristics that facilitated survival of
fledging females from the 1992 cohort also contributed to the unusual reproductive
variables measured in 1993. In other words, the females that survived to breed as SY’s
in 1993 may have possessed attributes which contributed to enhanced reproductive

SUCCCSS .

During both 1990 and 1993, the correspondence within treatment groups be-
tween levels of parental care and growth of young observed was not one-to-one. That
is, higher rates of visitation did not always translate into better growth. This could be
attributable to differences in bolus weight delivered per visit, underscored by the fact
that a significant portion of the variation in bolus weight can be accounted for by the
nest itself. In other words, lower nest visitation rates could potentially provide more
than adequate parental care, if weight delivered/visit was large enough. It appears that

the size of the bolus delivered, as well as the frequency of visits made to the nest, is a

122
highly variable and highly individualized phenomenon, and these relationships, in turn,

contribute to the high degree of variability in the growth of the young.

Even though there are numerous references of foraging deficiencies in juveniles
and younger adults (Burger 1980 and 1990, Dunn 1972, Groves 1978, Orians 1969a,
Recher and Recher 1969, Wunderle 1991), to my knowledge there is no evidence in the
literature concerning such age-related foraging deficiencies in aerial insectivores or
specifically Tree Swallows. In this study, no differences were found in bolus weight
delivered to young between SY and ASY nests, yet the trend (although not statistically
significant) was for SY females to deliver smaller loads than ASY females. Interest-
ingly, in addition to SY females only producing 90% of ASY clutch biomass, mean
bolus size for SY females was approximately 85 % of ASY bolus size. Desrochers
(1992) argued that low foraging success may be a general constraint causing the age
effects seen in so many avian species. This may also be the case in Tree Swallows,
judging from the trend for SY females to deliver smaller loads to young in the nest.
Foraging skills can improve over time (Desrochers 1992), yet if females were deficient
foragers during the time of egg formation, laying and incubation, it seems unlikely that
improvements within individuals of the scope needed to adequately care for young
could occur over the course of only approximately three weeks time while feeding
young. Marchetti and Price (1989) review age-related differences in foraging efficiency
and one of the points they raise is experiential constraints. For example, some foraging

techniques require a great deal of learning which involves interaction with the appro-

123
priate environmental stimuli. Whereas many fledglings undergo a period of “training”
by adults, particularly in precocial species (6. g. Hannon and Smith 1984), there is little
evidence for post-fledging care in Tree Swallows, and fledglings are completely on
their own upon leaving the nest (yet see comments by Wheelwright et al. 1991). Even
if SY females can adequately feed themselves because of the nearly year-long experi-
ence of foraging on their own, the environmental stimulus represented by young in the
nest is unique, as well as very demanding, in terms of time and energy allocation.
Experiential constraints could explain some of the apparent deficiencies observed (for
example in the 1990 growth and parental care data), and may also help explain the

ultimate reasons for clutch size differences between SY and ASY females.

The argument that smaller clutch sizes in SY females are a result of inexperi-
ence and foraging deficiencies and that these factors contribute to an overall inability to
provide adequately for themselves or their young is pervasive in the literature (e.g.
Lack 1968, Perrins 1970). One of the reasons this hypothesis is popular is due to its
intuitive appeal; however, there is little but circumstantial evidence to support it. In this
study, nestlings of SY females sometimes exhibited poorer growth, lower levels of
parental care, and sometimes higher levels of mortality. In addition, there is some
evidence that SY females deliver smaller sized boluses to nestlings. Overall, however,
the evidence is inconsistent and sometimes contrary, and as such, is not strong enough
to contend that SY females are constrained from producing larger clutches by their

inability to feed nestlings.

124

The evidence reported in this study from growth analyses and patterns of paren-
tal care does not support Lack’s contention that clutch size is determined by the par-
ent’s collective abilities to feed young, yet does not exclude the possibility of female
control over clutch size based on her own abilities (e. g. Slagsvold and Lifjeld 1990). In
species like Tree Swallows where both parents feed nestlings, but only the female
builds the nest and produces the eggs, clutch size differences are likely determined by
factors prior to laying or during incubation, as opposed to factors operating during the

nestling period.

Suggestions for Future Research

As is the case with many field investigations, this study raises many new ques-
tions which it would be fruitful to pursue further. The question of whether or not Tree
Swallow SY females are deficient foragers has not been addressed adequately. Al-
though Williams (1988) suggests that the number of trips is a good reflection of paren-
tal effort, the combination of visits and amount delivered is an approach which would
provide the most useful information (as originally suggested by Royama 1966). A
rigorous sampling scheme using the ligature method and the blinds developed in this
study would provide useful information in this regard. This was attempted during 1992
following the preliminary study of 1991 reported above, yet inclement weather pre-

vented the gathering of any useful data. All of the nests set up for study failed.

125

This design would allow determination of whether or not SY females and males
paired with SY females are indeed deficient foragers and whether or not partners com-
pensate for each other’s deficiencies. Another option would be to use an accurate bal-
ance system located under the nest to determine load sizes, as well as parental weight
dynamics (e.g. Jones 1987a, b, c, 1988). In either case, a design that uses fixed brood
sizes as was done in this study would be useful to assess potential differences associated
with brood size as well. The difficulty with this scenario in Tree Swallows is that the
percentage of SY females in the population does not lend itself well to a study of this

nature and sample sizes could be restricted.

Interestingly, the bolus weight changes measured in this study over time were
reflected in females but not in the males. It would be very interesting to see if this
changes with a larger sample size or if the relationship remains the same. These sexual
differences indicate that although the number of visits may not be different between the
sexes, as evidenced by this and other studies, the amount of food delivered may be

considerably different, particularly later in the nestling stage when young are largest.

There is a need to assess mortality rates of the age classes of females more
closely to see if it is possible to test the selection hypothesis of N01 and Smith (1987).
Some suggest higher rates of mortality in the SY age class, yet this needs to be tested
further. This has not been possible in the present study because so many of the SY

female nests were manipulated. This would require a long-term study, and longitudinal

126
studies of this nature have been attempted on short-lived passerines, yet definitive

results have been difficult to obtain.

Desrochers and Magrath (1993) make a strong point that ignoring the potential
age effect of mates (as was done, for the most part, in this study) could lead to spurious
conclusions, especially if the effects of the sexes are in opposite directions. Future
work should address this in species where there is biparental care. Because Tree Swal-
lows are single-sex incubators, only females would exhibit egg-formation, egg-laying
and incubation constraints, yet because both sexes feed the young in the nest, potential
constraints impacting nestling rearing could influence both sexes. In other words, there
may well be conflicting selection pressures acting upon each of the sexes. In addition,
much of the available evidence in my study points to factors acting outside of the nest-

ling rearing phase which are constraints to larger clutches in SY females.

Stutchbury and Robertson (1988) suggest that age-related differences would be
less in a food rich environment. For example, Dunn (1972) reported differences be-
tween young and older birds in foraging abilities, yet these differences disappeared
when prey abundance was artificially increased as a result of human activities. Al-
though comparisons between my study and DeSteven’s (1978) study have suggested
otherwise, this could be tested at sites which differed in prey abundance, similar to
Hussell and Quinney’s (1987) work on clutch size and prey abundance. They compared
clutch size output at two sites which differed primarily in the levels of insect biomass
available, yet also reported that few, if any, SY females nested at the high prey abun-

dance site. This potential problem could be eliminated by establishing nestboxes on

127
sites after relative prey abundance was measured in previous years. Newly established
nestbox plots have been shown to attract larger numbers of SY females. Because of egg
size differences observed in this study, it would be interesting to see if these differences
were also correlated with gonad size as was suggested by Westin (1989, cited in
Enoksson 1993). This would necessarily be an invasive study, yet might provide one of
the more definitive pieces of information concerning the possibility of physiological or

morphological constraints on egg size.

LITERATURE CITED

Ainley, DC. and RP. Schlatter. 1972. Chick raising ability in Adelie Penguins. Auk
89:559-566.

Alatalo, RV. and A. Lundberg. 1989. Clutch size of the Pied Flycatcher—an
experiment. Ornis Fennica 66: 15-23.

Aldrich, T.W. and DC. Raveling. 1983. Effects of experience and body weight on
incubation behavior of Canada Geese. Auk 100:670-679.

Alerstam, T. and G. Hogstedt. 1984. How important is clutch size dependent adult
mortality? 0ikos 43:253-254.

Askenmo, C. 1979. Reproductive effort and return rate of male Pied Flycatchers.
American Naturalist 113:748-753.

Austin, D.L. and Low, S.H. 1932. Notes on the breeding of the Tree Swallow. Bird-
Banding 3:39-44.

Beaver, D.L., R.W. Hill and SD. Hill. 1994. Extremely Low Frequency
Communications system Ecological Monitoring Program: Small Mammals and
Nesting Birds. Final Report. (Submitted to Illinois Institute of Technology
Research Institute and the US. Navy).

Bédard, J. and G. LaPointe. 1985. Influence of parental age and season on Savannah
Sparrow reproductive success. Condor 87:106-110.

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

Biebach, H. 1979. Energetik des Brfitens biem Star (Sturnus vulgaris). Journal filr
Ornithologie 120:121-138.

Boekelheide, KL and DO. Ainley. 1989. Age, resource availability, and breeding
effort in Brandt’s Cormorant. Auk 106:389-401.

Brown, CR. 1978. Clutch size and reproductive success of adult and subadult Purple
Martins. Southwestern Naturalist 23:597-604.

128

 

 

129

Brown, J .L. 1969. Territorial behavior and population regulation in birds: A review
and re-evaluation. Wilson Bulletin 81:293-329.

Bryant, D.M. 1975. Breeding biology of House Martins Delichon urbica in relation to
aerial insect abundance. Ibis 117:180-216.

. 1978a. Environmental influences on growth an survival of nestling House
Martins Delichon urbica. Ibis 120:271-293.

. 1978b. Establishment of weight hierarchies in the broods of House Martins
Delichon urbica. Ibis 120: 16-26.

. 1979. Reproductive costs in the House Martin (Delichon urbica). Journal of
Animal Ecology 48:655-675

Bryant, D.M. and P. Tatner. 1990. The cost of brood provisioning: Effects of brood
size and food supply. Proceedings of the International Ornithological Congress
19:364-379.

Bryant, D.M. and KR. Westerterp. 1982. Evidence for individual differences in
foraging efficiency amongst breeding birds: A study of House Martins Delichon
urbica using the doubly labeled water technique. Ibis 124:187-192.

Bryant, D.M. and KR. Westerterp. 1983. Time and energy limits to brood size in
House Martins (Delichon urbica). Journal of Animal Ecology 52:905-925.

Bulmer, MG. and CM. Perrins. 1973. Mortality in the Great Tit Parus major. Ibis
115:277-281.

Burger, J. 1980. Age differences in foraging Black-necked Stilts in Texas. Auk 97 :633-
636.

. 1990. Effects of age on foraging in birds. Proceedings of the International
Ornithological Congress 19: 1 127-1 140.

Burtt, EM. and RM. Tuttle. 1983. Effect of timing of banding on reproductive
success of Tree Swallows. Journal of Field Ornithology 54:319-323.

Butler, R.W. 1988. Population dynamics and migration routes of Tree Swallows,
Tachycineta bicolor, in North America. Journal of Field Ornithology 59:395-
402. ,

Chapman, LB. 1935. Studies of a Tree Swallow colony. Bird-Banding 6:45-57.

. 1955. Studies of a Tree Swallow colony. Bird-Banding 26:45-70.

 

130

Charnov, EL. and LR. Krebs. 1974. On clutch-size and fitness. Ibis 116:217-219.

Christian, K.A. and CR. Tracey. 1985. Measuring air temperatures in field studies.
Journal of Thermal Biology 10:55-56.

Clark, AB. and D.S. Wilson. 1981. Avian breeding adaptations: Hatching
asynchrony, brood reduction and nest failure. Quarterly Review of Biology
56:253-277.

Clutton-Brock, T.H. 1984. Reproductive effort and terminal investment in iteroparous
animals. American Naturalist 123:212-219.

. 1991. The Evolution of Parental Care. Princeton University Press,
Princeton, NJ.

Cohen, RR. 1981. Dispersal in adult female Tree Swallows (Iridapracne bicolor) in
Colorado. Journal of the Colorado- Wyoming Academy of Science. 13:62.

. 1985. Capturing male breeding Tree Swallows with feathers. North
American Bird Bander 10:18-19.

Conrad, K.F. and R]. Robertson. 1992. Intraseasonal effects of clutch manipulation
on parental provisioning and residual reproductive value of Eastern Phoebes
(Sayomis phoebe). Oecalogia 89:356-364.

Coulson, LC. 1966. The influence of pair-bond and age on the breeding biology of the
Kittiwake Gull Rissa tridactyla. Journal of Animal Ecology 35:269-279.

Crawford, RD. 1977. Breeding biology of year-old and older female Red-winged and
Yellow-headed Blackbirds. Wilson Bulletin 89:73-80.

. 1980. Effects of age on reproduction in American Coots. Journal of Wildlife
Management 44: 183-189.

Crossner, K.A. 1977. Natural selection and clutch size in the European Starling.
Ecology 58:885-892.

Curio, E. 1983. Why do young birds reproduce less well? Ibis 125:400-404.

D’Agostino, R.B., A. Belanger, and RB. D’Agostino, Jr. 1990. A suggestion for
using powerful and informative tests of normality. American Statistician -
44:316-321.

Desrochers, A. 1992. Age-related differences in reproduction by European Blackbirds:
Restraint or constraint? Ecology 73: 1 128-1 131.

131

Desrochers, A. and RD. Magrath. 1993. Age-specific fecundity in European
Blackbirds (T urdus merula): Individual and population trends. Auk 110:255-
263.

DeSteven, D. 1978. The influence of age on the breeding biology of the Tree Swallow
Iridoprocne bicolar. Ibis 120:516-523.

. 1980. Clutch size, breeding success, and parental survival in the Tree
Swallow (Iridapracne bicolor). Evolution 34:278-291.

Drent, R.H. 1975. Incubation. Pp. 333-420 in Farner, D.S. and J.R. King, eds. Avian
Biology vol. 5. Academic Press, New York.

Dunn, EH. 1979. Age of effective homeothermy in nestling Tree Swallows according
to brood size. Wilson Bulletin 91:455-457.

Dunn, E.K. 1972. Effect of age on the fishing ability of Sandwich Terns Stema
sandvicensis. Ibis 114:360-366.

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

Dunn, PC. and R]. Robertson. 1992. Geographical variation in the importance of
male parental care and mating systems in Tree Swallows. Behavioral Ecology
3:291-299.

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

Enoksson, B. 1993. Effects of female age on reproductive success in European
Nuthatches breeding in natural cavities. Auk 110:215-221.

Erikstad, K.E., H.C. Pedersen and J .B.Steen. 1985 . Clutch size and egg size variation
in the Willow Grouse Lagopus lagopus lagopus. Omis Scandinavica 16:88-94.

Ewald P.W. and S. Rohwer. 1982. Effects of supplemental feeding on timing of
breeding, clutch size and polygyny in Red-winged Blackbirds Agelaius
phoeniceus. Journal of Animal Ecology 51:429-450.

Falconer, D.S. 1981. Introduction to Quantitative Genetics. Longman Scientific and
Technical, Essex, England. -

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

 

 

132

Finney, G. and F. Cooke. 1978. Reproductive habits in the Snow Goose: The influence
of female age. Condor 80: 147-158.

Freed, LA. 1981. Loss of mass in breeding wrens: stress or adaptation? Ecology
62:1179-1186.

Godfray, H.C.J., L. Partridge, and RH. Harvey. 1991. Clutch size. Annual Review of
Ecology and Systematics 22:409-429.

Groves, S. 1978. Age-related differences in Ruddy Turnstone foraging and aggressive
behavior. Auk 95:95-103.

Gustafsson, L. and T. Part. 1990. Acceleration of senescence in the Collared
Flycatcher Ficedula albicallis by reproductive costs. Nature 347:279-281.

Gustafsson, L. and WJ. Sutherland. 1988. The costs of reproduction in the Collared
Flycatcher Ficedula albicallis. Nature 347 :279-281.

Hamann, J. and F. Cooke. 1987. Age effects on clutch size and laying date of
individual female Lesser Snow Geese Anser caerulescens. Ibis 129:527-532.

Harmon, SJ. and J .N.M. Smith. 1984. Factors influencing age-related reproductive
success in Willow Ptarmigan. Auk 101:848-854.

Harvey, P.H., P.J. Greenwood, C.M. Perrins and A.R. Martin. 1979. Breeding
success of Great Tits (Parus major) in relation to age of male and female parent
Ibis 121:216-219.

Harvey, P.H., M.J. Stenning and B. Campbell. 1985. Individual variation in seasonal
breeding success of Pied F lycatchers (Ficedula hypoleuca). Journal of Animal
Ecology 54:391-398.

Hensler, G.L. and JD. Nichols. 1981. The Mayfield method of estimating nesting
success: A model, estimators and simulation of results. Wilson Bulletin 93:42-
53.

Hochachka, W. 1992. How much should reproduction cost? Behavioral Ecology 3:42-
52.

Hochachka, W. and J .N .M. Smith. 1991. Determinants and consequences of nestling
condition in Song Sparrows. Journal of Animal Ecology 60:995-1008.

Hogstedt, G. 1980. Evolution of clutch size in birds: Adaptive variation in relation to
territory quality. Science 210: 1148-1150.

133

. 1981a. Should there be a positive or negative correlation between survival
of adults in a bird population and their clutch size? American Naturalist
118:568-571.

. 1981b. Effect of additional food on reproductive success in the Magpie
(Pica pica). Journal of Animal Ecology 50:219-229.

Holroyd, G.L. 1975. Nest site availability as a factor limiting population size of
swallows. Canadian Field-Naturalist 89:60-64.

Houston, M.I. and GS. Houston. 1987. Tree Swallow banding near Saskatoon,
Saskatchewan. North American Bird Bander 12: 103-108.

Hussell, D.J.T. 1972. Factors affecting clutch size in Arctic passerines. Ecological
Monographs 42:317-364.

. 1982. Longevity and fecundity records in the Tree Swallow. North
American Bird Bander 7: 154.

. 1983a. Age and plumage color in female Tree Swallows. Journal of Field
Ornithology 54:312-318.

. 1983b. Tree Swallow pairs raise two broods in a season. Wilson Bulletin
95 :470-47 1 .

Hussell, D.J.T., and T.E. Quinney. 1987. Food abundance and clutch size of Tree
Swallows (T achycineta bicolor). Ibis 129:243-258.

Hussell, D.J.T., T.E. Quinney, P.O. Dunn, D.L. Beaver, P.E. Lederle, E.H. Burtt,
Jr., S.R. Derrickson, E.M. Landre, S.L. Leathery, L.A. Wakelyn, N.T.
Wheelwright, D.A. Wiggins. 1990. Geographic variation in food abundance
and clutch size of Tree Swallows T achycineta bicolor. Poster presentation.
International Ornithological Congress, XX, Christchurch, New Zealand.

Jarvinen, A. and R.A. Vaisanen. 1982. Effects of exceptionally favorable weather on
the breeding of the Pied Flycatcher Ficedula hypoleuca in Finnish Lapland. Ibis
124:196-198.

Jarvinen, A. and R.A. Vaisanen. 1983. Egg size and related reproductive traits in a
southern passerine Ficedula hypoleuca breeding in an extreme northern
environment. Ornis Scandinavica 14:253-262. '

Jarvinen, A. and R.A. Viiisanen. 1984. Reproduction of Pied Flycatchers (Ficedula
hypoleuca) in good and bad breeding seasons in a northern marginal area. Auk
101:439-450.

134

Johnson, EL, L.B. Best and P.A. Heagy. 1980. Food sampling biases associated with
the “ligature method”. Condor 82:186-192.

Johnson, EL and LB. Best. 1982. Factors affecting feeding and brooding of Gray
Catbird nestlings. Auk 99: 148-156.

Jones, G. 19876. Time and energy constraints during incubation in free-living swallows
(Hirundo rustica): An experimental study using precision electronic balances.
Journal of Animal Ecology 56:229-245.

. 1987b. Parental foraging ecology and feeding behaviour during nestling
rearing in the swallow. Ardea 75:169-174.

. 1987c. Parent-offspring resource allocation in Swallows during nestling
rearing: An experimental study. Ardea 75:145—168.

. 1988. Concurrent demands of parent and offspring Swallows Hirundo
rustica in a variable feeding environment. Ornis Scandinavica 19:145-152

Jover, L., X. Ruiz and M. Gonzalez-Martin. 1993. Significance of intraclutch egg size
variation in the Purple Heron. Ornis Scandinavica 24: 127-134.

Kallander, H. 1974. Advancement of laying of Great Tits by the provisioning of food.
Ibis 116:365-367.

Kendeigh, SC. 1941. Territorial and mating behavior of the House Wren. Illinois
Biological Monographs 18:1-20.

King, J.R. 1974. Seasonal allocation of time and energy resources in birds. Pp. 4-70,
in Paynter, R.A., ed. Publication of the Nuttall Ornithological Club, No. 5,
Cambridge, MA.

Klomp, H. 1970. The determination of clutch size in birds. Ardea 58:1-124.

Knapton, R.W. 1984. Parental feeding of nestling Nashville Warblers: The effects of
food type, brood-size, nestling age, and time of day. Wilson Bulletin 96:594-
602.

Krebs, J.R. 1970. Regulation of numbers in the Great Tit (Aves: Passeriformes).
Journal of Zoology 162:317-333.

Kuerzi, R.G. 1941. Life history studies of the Tree Swallow. Proceedings of the
Linnaean Society of New York 52: 1-52.

Lack, D. 1947. The significance of clutch size. Ibis 89:309-352, 90:25-45.

135

. 1954. Natural Regulation of Animal Numbers. Clarendon, Oxford.
. 1966. Population Studies of Birds. Oxford University Press, Oxford.

. 1968. Ecological Adaptations for Breeding in Birds. Metheun and Co. ,
Ltd., London.

Lederle, P.E., B.C. Pijanowski and D.L. Beaver. 1985 . Predation of Tree Swallows
by the least chipmunk. Jack-Pine Warbler 63: 135.

Leffelaar, D. and RJ. Robertson. 1986. Equality of feeding roles and the maintenance
of monogamy in Tree Swallows. Behavioral Ecology and Sociobiology 18: 199-
206.

Leinonen, M. 1973. Comparisons of breeding biology of year-old and older females of
the White Wagtail (Matacilla alba) in central Finland. Ornis Fennica 50:126-
133.

Lessells, OM. 1986. Brood size in Canada Geese: A manipulation experiment. Journal
af Animal Ecology 55 :669-689.

Lessells, GM. and RT. Boag. 1987. Unrepeatable repeatabilities: A common mistake.
Auk 104:116-121.

Lessells, GM. and J .R. Krebs. 1989. Age and breeding performance of European Bee-
eaters. Auk 106:375-382.

Lilliefors, H.W. 1967. On the Kolmogorov-Smirnov test for normality with mean and
variance unknown. Journal of the American Statistical Association 64:399-402.

Loman, J. 1980. Brood size optimization and adaptation among Hooded Crows Corvus
carone. Ibis 122:494-500.

Lombardo, M .P. 1986. Yearling-biased female mortality in the Tree Swallow. Condor
88:520-521.

. 1989. More on the effect of the timing of handing on female Tree Swallow
nest site tenacity. Journal of Field Ornithology 60:68-72.

. 1991. Sexual differences in parental effort during the nestling period in Tree
Swallows. Auk 108:393-404.

Low, S.H. 1933. Further notes on the nesting of Tree Swallows. Bird-Banding 6:76-
87.

Magnusson, A. 1984. Ny fangstmetod for holkfaglar. Vdr Fagelva’rld 4:318.

136

Magrath, RD. 1991. Nestling weight and juvenile survival in the Blackbird, T urdus
merula. Journal of Animal Ecology 60:335-35 1 .

Marchetti, K. and T. Price. 1989. Differences in the foraging of juvenile and adult
birds: The importance of developmental constraints. Biological Review 64:51-
70.

Martin, T.E. 1987. Food as a limit on breeding birds: A life-history perspective.
Annual Review of Ecology and Systematics 18:453-487.

Martins, T.L.F. and J. Wright. 1993. Cost of reproduction and allocation of food
between parent and young in the Swift (Apus apus). Behavioural Ecology 4:213-
223.

Mayfield, H. 1961. Nesting success calculated from exposure. Wilson Bulletin 73:255-
261.

. 1975. Suggestions for calculating nesting success. Wilson Bulletin 87 :456-
465.

Middleton, A.L. 1979. Influence of age and habitat on reproduction by the American
Goldfinch. Ecology 60:418-432.

Muldal, A., H.L. Gibbs and RJ. Robertson. 1985. Preferred nest spacing of an
obligate cavity-nesting bird, the Tree Swallow. Condor 87:356-363.

Murphy, EC. 1980. Body size of House Sparrows: Reproduction and survival
correlates. Proceedings of the International Ornithological Congress 17:1155-
1161.

Murphy, M.T. 1986. Body size and condition, timing of breeding, and aspects of egg
production in Eastern Kingbirds. Auk 103:465-476.

Nilsson, S.G. 1984. The evolution of nest-site selection among hole-nesting birds: The
importance of nest predation and competition. Ornis Scandinavica 15:167-175.

Nol, E. and J .N .M. Smith. 1987. Effects of age and breeding experience on seasonal
reproductive success in the Song Sparrow. Journal of Animal Ecology 56:301-
313.

Norberg, R.A. 1981. Temporary weight decrease in breeding birds may result in more
fledged young. American Naturalist 118:838-850.

Nur, N. 1984. The consequences of brood size for breeding Blue Tits. II: Nestling
weights, offspring survival and optimal brood size. Journal of Animal Ecology
53:497-517.

137

. 1986. Is clutch size variation in the Blue Tit (Parus caeruleus) adaptive?
Journal of Animal Ecology 55:983-999.

O’Connor, R.J. 1975. Initial size and subsequent growth in passerine nestlings. Bird-
Banding 46:329-340.

. 1979. Egg weights and brood reduction in the European Swift (Apus apus).
Condor 81: 133—145.

Ojanen, M., M. Orell and R.A. Viiisanen. 1979. Role of heredity in egg size variation
in the Great Tit Parus major and the Pied Flycatcher F icedula hypoleuca. Ornis
Scandinavica 10:22-28.

Ojanen, M., M. Orell and R.A. Viisiinen. 1981. Egg size variation within passerine

clutches: Effects of ambient temperature and laying sequence. Ornis Fennica
58:93-108.

Orians, G.H. 1969a. Age and hunting success in the Brown Pelican (Pelecanus
occidentalis). Animal Behavior 17:316-319.

. 1969b. On the evolution of mating systems in birds and mammals.
American Naturalist 103:589-603.

Paynter, R.A. 1954. Interrelations between clutch-size, brood-size, prefledging
survival, and weight in Kent Island Tree Swallows. Bird-Banding 25:35-58 and
25: 102-1 10.

Perdeck, AC. and A.J. Cavé. 1992. Laying date in the coot: Effects of age and mate
choice. Journal of Animal Ecology 61:13-19.

Perrins, C.M. 1965. Population fluctuations and clutch-size in the Great Tit, Parus
major L. Journal of Animal Ecology 34:601-647.

. 1970. The timing of birds’ breeding seasons. Ibis 112:242-255.
. 1979. British Tits. Collins, London.
Perrins, C.M. and TR. Birkhead. 1983. Avian Ecology. Blackie, London.

Perrins, C.M. and P.J. Jones. 1975 . The inheritance of clutch size in the Great Tit
(Parus major L.). Condor 76:225-229.

Perrins, C.M. and D. Moss. 1974. Survival of young Great Tits in relation to age of
female parent. Ibis 116:220-224.

138

Perrins, C.M. and D. Moss. 1975. Reproductive rates in the Great Tit. Journal of
Animal Ecology 44:695-706.

Perrins, C.M. and R.H. McCleery. 1985. The effect of age and pair bond on the
breeding success of Great Tits Parus major. Ibis 127:306-315.

Perrins, C.M. and R.H. McCleery. 1989. laying dates and clutch size in the Great Tit.
Wilson Bulletin 101:236-253.

Pettifor, R.A., C.M. Perrins and R.H. McCleery. 1988. Individual optimization of
clutch size in Great Tits. Nature 336:160-162.

Quinney, T.E. 1983. Tree Swallows cross a polygyny threshold. Auk 100:750-754.

. 1986. Male and female parental care in Tree Swallows. Wilson Bulletin
98:147-150.

Quinney, T.E. and CD. Ankney. 1985. Prey size selection by Tree Swallows. Auk
102:245-250.

Quinney, T.E., D.J.T. Hussell and CD. Ankney. 1986. Sources of variation in growth
of Tree Swallows. Auk 103:389—400.

Reese, K.P. and J.A. Kadlec. 1985. Influence of high density and parental age on the
habitat selection and reproduction of Black-billed Magpies. Condor 87:96-105.

Reid, W.V. 1987. The cost of reproduction in the Glaucous-winged Gull. Oecalogia
74:458-467.

. 1988. Age-specific patterns of reproduction in the Glaucous-winged Gull:
Increased effort with age? Ecology 69: 1454-1465.

Recher, H.F. and J .A. Recher. 1969. Comparative foraging efficiency of adult and
immature Little Blue Herons (Florida caerulea). Animal Behavior 17 :320-322.

Rendell, W.B. and R.J. Robertson. 1990. Influence of forest edge on nest-site selection
in Tree Swallows. Wilson Bulletin 102:634-644.

Ricklefs, RE. 1967. A graphical method of fitting equations to growth curves. Ecology
48:978-983.

. 1974. Energetics of reproduction in birds. Pp. 152-192 in Paynter, R.A.,
ed. Publication of the Nuttall Ornithological Club, No. 15., Cambridge, MA.

. 1977. On the evolution of reproductive strategies in birds: Reproductive
effort. American Naturalist 111:453-478.

139

. 1983. Avian postnatal development. Pp. 1-83 in Farrier, D.S. and J.R.
King, eds. Avian Biology, vol. 7, Academic Press, New York.

Ricklefs, RE. and D.J.T. Hussell. 1984. Changes in adult mass associated with the
nesting cycle in the European Starling. Ornis Scandinavica 15:155-161.

Ricklefs, RE. and S. Peters. 1981. Parental components of variance in growth rate and
body size of nestling European Starlings (Stumus vulgaris) in eastern
Pennsylvania. Auk 98:39-48.

Robertson, R.J. and H.L. Gibbs. 1982. Superterritoriality in Tree Swallows: A re-
examination. Condor 84:313-316.

Robertson, R.J. and W.B. Rendell. 1990. A comparison of the breeding ecology of a
secondary cavity nesting bird, the Tree Swallow (T achycineta bicolor), in nest
boxes and natural cavities. Canadian Journal of Zoology 68: 1046-1052.

Rockwell, E.G. Cooch, C.B. Thompson and F. Cooke. 1993. Age and reproductive
success in female Lesser Snow Geese: Experience, senescence and the cost of
philopatry. Journal of Animal Ecology 62:323-333.

Rofstad, G. and J. Sandvik. 1985 . Variation in egg size of the Hooded Crow Corvus
corone cornix. Ornis Scandinavica 16:38-44.

Rohwer, S., S.D. Fretwell and D.M. Niles. 1980. Delayed maturation in passerine

plumages and the deceptive acquisition of resources. American Naturalist
1 15:400-437.

Roskaft, E. 1985. The effect of enlarged brood size on the future reproductive potential
of the Rock. Journal of Animal Ecology 54:255-260.

Raskaft, E., Y. Espmark and T. Jarvi. 1983. Reproductive effort and breeding success
in relation to age by the Rook Corvus frugilegus. Ornis Scandinavica 14:169-
174.

Ross, HA. 1980. The reproductive rates of yearling and older Ipswich Sparrows,
Passerculus sandwichensis princeps. Canadian Journal of Zoology 5 8: 1557-
1563.

Royama, T. 1966. Factors governing feeding rates, food requirements and brood size
of nestling Great Tits Parus major. Ibis 108:313-347. .

Saether, B.-E. 1990. Age-specific variation in reproductive performance in birds.
Current Ornithology 7 :25 1-283.

140

Salthe, SN. 1969. Reproductive modes and the number and sizes of ova in urodeles.
American Midland Naturalist 81 :467-490.

Slagsvold, T. 1982. Clutch size, nest size, and hatching asynchrony in birds:
experiments with the F ieldfare (T urdus pilaris). Ecology 63: 1389-1399.

Slagsvold, T. and J.T. Lifjeld. 1988. Ultimate adjustment of clutch size to parental
feeding capacity in a passerine bird. Ecology 69:1918-1922.

Slagsvold, T. and J .T. Lifjeld. 1989. Hatching asynchrony in birds: The hypothesis of
sexual conflict over parental investment. American Naturalist 134:239-253.

Slagsvold, T. and J .T. Lifjeld. 1990. Influence of male and female quality on clutch
size in tits (Parus spp.). Ecology 71: 1258-1266.

Smith, H.G. 1993. Parental age and reproduction in the Marsh Tit Parus palustris. Ibis
135: 196-201.

Smith, H.G., H. Kallander, K. Fontell and M. Ljungstrom. 1988. Feeding frequency
and parental divisions of labour in the double brooded Great Tit Parus major.
Behavioural Ecology and Sociobiology 22:447—453.

Smith, H.G., H. Kallander and J.-A. Nilsson. 1989. The trade-off between offspring
number and quality in the Great Tit Parus major. Journal of Animal Ecology
58:383-401.

Smith, H.G., U. Ottosson and T. Ohlsson. 1993. Interclutch variation in egg mass
among Starlings Stumus vulgaris reflects female body condition. Ornis
Scandinavica 24:311-316.

Smith, S.M. 1978. The “underworld” in a territorial sparrow: Adaptive strategy for
floaters. American Naturalist 112:571-582.

Sokal, R.R. and El. Rohlf. 1981. Biometry. W.H. Freeman and Company, New
York.

Stearns, SC. 1976. Life-history tactics: A review of the ideas. Quarterly Review of
Biology 51:3-47.

Studd, M.V. and R.J. Robertson. 1985. Life span, competition, and delayed plumage
maturation in male passerines: The breeding threshold hypothesis. American
Naturalist 126:101-115.

Stutchbury, BJ. and R.J. Robertson. 1985. Floating population of female Tree
Swallows. Auk 102:651-654.

141

Stutchbury, B.J. and R.J. Robertson. 1987a. Behavioral tactics of subadult female
floaters in the Tree Swallow. Behavioral Ecology and Sociobiology 20:413-419.

Stutchbury, B.J. and R.J. Robertson. 1987b. Signaling subordinate and female status:
two hypotheses for the adaptive significance of subadult plumage in Tree
Swallows. Auk 104:717-723.

Stutchbury, B.J. and R.J. Robertson. 1988. Within-season and age-related patterns of
reproductive performance in female Tree Swallows (T achycineta bicolor).
Canadian Journal of Zoology 66:827-834.

Stutchbury, B.J. and S. Rohwer. 1990. Molt patterns in the Tree Swallow (T achycineta
bicolor). Canadian Journal of Zoology 68: 1468-1472.

Tinkle, D.W. and R.E. Ballinger. 1972. Sceloporous undulatus: A study of the
intraspecific comparative demography of a lizard. Ecology 53:570-584.

Turner, A. 1983. Time and energy constraints on the brood size of Swallows, Hirundo
rustica, and Sand Martins, Riparia riparia. Oecalogia 59:331-338.

Turner, A. and C. Rose. 1989. Swallows and Martins, An Identification Guide and
Handbook. Houghton Mifflin Company, Boston.

United States Department of Agriculture, Soil Conservation Service. 1992. Iran
County, Michigan Survey. Crystal Falls, Michigan.

United States Fish and Wildlife Service and Canadian Wildlife Service. 1991. Bird
Banding Manual. Patuxent, Maryland.

van Balen, J .H. 1973. A comparative study of the breeding ecology of the Great Tit
Parus major in different habitats. Ardea 61: 1-93.

van Noordwijk, A.J., J.H. van Balen and W. Scharloo. 1980. Heritability of
ecologically important traits in the Great Tit. Ardea 68:193-203.

van Noordwijk, A.J., J .H. van Balen and W. Scharloo. 1981. Genetic and
environmental variation in clutch size of the Great Tit. Netherlands Journal of
Zoology 31:342-372.

von Haartman, L. 1971. Population dynamics. Pp. 391-459 in Farner, D.S. and J.R.
King, eds. Avian Biology, vol. 1, Academic Press, New York. ‘

Walsberg, GE. 1983. Avian ecological energetics. Pp. 161-220 in Farner, D.S. and
J.R. King, eds. Avian Biology, vol. 7, Academic Press, New York.

Walsh, H. 1978. Food of nestling Purple Martins. VVllsan Bulletin 90:248-260.

142

Westin, J. 1989. Endocrine studies on the Willow Tit Parus montanus, with special
emphasis on the female. Ph.D. thesis, Gothenburg University, Gothenburg,
Sweden.

Westerterp, K. and D.M. Bryant. 1984. Energetics of free existence in swallows and
martins (Hirundinidae) during breeding: A comparative study using doubly
labeled water. Oecalogia 62:376-381.

Wheelwright, N.T., J. Leary and C. Fitzgerald. 1991. The cost of reproduction in Tree
Swallows (T achycineta bicolor). Canadian Journal of Zoology 69:2540-2547.

Wiggins, DA. 1990. Sources of variation in egg mass of Tree Swallows T achycineta
bicolor. Ornis Scandinavica 21: 157-160.

Williams, G.C. 1966. Natural selection, the cost of reproduction and a refinement of
Lack’s principle. American Naturalist 100:687-690.

Williams, LB. 1988. Field metabolism of Tree Swallows during the breeding season.
Auk 105:706-714.

Wilkinson, L. 1992. SYSTATfor Windows: The System for Statistics. SYSTAT, Inc.,
Evanston, Illinois.

Winkler, D.W. 1993. The use and importance of feathers as nest lining in Tree
Swallows (T achycineta bicolor). Auk 110:29-36.

Wittenberger, J.F. 1979. A model for delayed reproduction in iteroparous animals.
American Naturalist 114:439-446.

Wooller, R.D., J .8. Bradley, 1.]. Skira and D.L. Serventy. 1990. Reproductive
success of Short-tailed Shearwaters Pufiinus tenuirostris in relation to their age
and breeding success. Journal of Animal Ecology 59: 161-170.

Wunderle, J .M., Jr. 1991. Age-specific foraging proficiency in birds. Current
Ornithology 8:273-324.

Yom-Tov, Y. and R. Hilborn. 1981. Energetic constraints on clutch size and time of
breeding in temperate zone birds. Oecalogia 48:234-243.

Zach, R. 1982. Hatching asynchrony, egg size, growth and fledging in Tree Swallows.
Auk 99:695-700 '

Zach, R. 1988. Growth-curve analysis: A critical reevaluation. Auk 105:208-210.

Zach, R. and K.M. Mayoh. 1982. Weight and feather growth of nestling Tree
Swallows. Canadian Journal of Zoology 60: 1080-1090.

 

143

Zar, J.H. 1984. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, NJ.

HI

HIGRN STRTE U

C NIV. LIBRRRIES
llIllIIIIIIIIIIIIIIIIIIllIIIIIIIIIIIIIIIIIIII
31293014202810