io
r
.
og
, B
r
m
nt
u
g
of young fl edged. Simulations suggest that selection most strongly favors a reduction in nest predation when breeding
seasons are short and predation rates are low (temperate characteristics). Conversely, selection favors shorter renesting
intervals when breeding seasons are long and nest predation rates are high (tropical characteristics). Reducing already low
Lati
brat
(Lac
196
1984, Yomtov et al. 1994, Brawn et al. 1995, Martin 2004,
Jetz et al. 2008, Lima 2009). Latitudinal trends include longer
breeding seasons, and consequently more breeding attempts
per season, and smaller clutches or litters towards lower lati-
tudes (Kulesza 1990). Th ese trends are not restricted to birds:
many reptiles and mammals also breed repeatedly each year in
tropical regions while similar species are typically single-
brooded in temperate areas (Conaway et al. 1974, Ricklefs



activity may not always infl uence predation risk (Roper and
Goldstein 1997). Because nest predation tends to be greater
in tropical compared to temperate regions (Ricklefs 1969,
Skutch 1985), nest-site and habitat selection should be
important for tropical birds. Some experimental studies have
supported this view by demonstrating variation in nest pre-
dation between diff erent nest sites (Gibbs 1991, but see
Filliater et al. 1994, Roper 2000, 2003). Th e idea that the
eda-
97),
ests
t al.
iors
06).and Bloom 1977, Dunham et al. 1988, Eisenberg 1989).
Th e impact of predation on reproductive success in birds
highlights adaptations to reduce nest predation, including
nest-site and habitat selection (Ricklefs 1970, 1988, Martin
and Roper 1988). Since nest predation is usually greater in
tropical birds (Ricklefs 1969, Skutch 1985, Martin 1995,
greater activity associated with larger broods attracts pr
tors has received little support (Roper and Goldstein 19
while some evidence suggests that greater activity at n
during feeding can increase predation risk (Martin e
2000). Also, parent birds may adjust their nesting behav
as a response to predation risk (Fontaine and Martin 20but
tho
such
(Ma
imp
predrates provides a proportionately greater increase in annual nesting success than does the same reduction when nest preda-
tion rates are higher. In some tropical species, individuals increase reproductive success not by avoiding predation in sub-
sequent nesting attempts, which is largely beyond their control, but rather by reducing renesting intervals. We suggest that
the emphasis on nest predation avoidance has biased our perspectives for alternative hypotheses of how birds should
respond to nest predation and the consequences of those alternatives for life history theory. Similarly to the need to control
for phylogenetics in examining life history strategies, future studies must also control for diff erences in breeding season
lengths and renesting intervals to better understand the infl uence of nest predation on avian life histories.
tudinal diff erences in life histories among terrestrial verte-
es remain poorly understood in spite of decades of research
k 1947, Skutch 1949, 1985, Ricklefs 1968a, 1968b,
9, 1970, 1977, Stearns 1977, Perrins 1977, Slagsvold
and Whelan 1999, Roper 2000). Predation risk may also
cause reduction of activity at the nest, even when that reduc-
tion bears a cost to brood size and reproductive rate (Skutch
1949, Conway and Martin 2000, Martin et al. 2000). But,Avoid nest predation when predat
lessons: testing the tropical–tempe
James J. Roper, Kimberly A. Sullivan and Robert E
J. J. Roper (jjroper@gmail.com), Univ. Federal do Paraná, Graduate Pr
BR–81531-990 Curitiba, PR, Brasil. – K. A. Sullivan, Dept of Biology
Ricklefs, Dept of Biology, Univ. of Missouri at St Louis, 8001 Natural B
Nest predation is the most important cause of nest failure in
and life histories suggest that nest predation has been infl ue
should favor reduction of nest predation, yet evidence is eq
combined eff ects of variation in nest predation rates, breedin see Oniki 1979, Robinson et al. 2000), nest predation is
ught to have driven trends in life history evolution in
a way as to explain, at least partially, latitudinal trends
rtin 1996). However, debate continues on the general
ortance of nest site selection and its infl uence on
ation risk (Holway 1991, Filliater et al. 1994, Schmidt Oikos 119: 719–729, 2010
doi: 10.1111/j.1600-0706.2009.18047.x
© 2009 Th e Authors. Journal compilation © 2010 Oikos
Subject Editor: Richard Stevens. Accepted 24 November 2009
n rates are low, and other
ate nest predation paradigm
Ricklefs
ram in Ecology and Conservation, Centro Politécnico, CP 19034,
NR 313, Utah State Univ., Logan, UT 84322, USA. – R. E.
idge Road, St Louis, MO 63121-4499, USA.
ost birds and latitudinal diff erences in nest predation rates
ial in life history evolution. All else equal, natural selection
ivocal. We used Monte Carlo simulations to examine the
season length and renesting intervals on the annual number 719
Th e infl uence of nest predation on annual reproductive
success depends on additional factors, particularly breeding
season length and renesting intervals following either nest
success or failure, that contribute to number of reproductive
attempts each year (Ricklefs and Bloom 1977). Temperate
regions usually have short breeding seasons during which

only one or few breeding attempts are possible (Wyndham
1986). Tropical breeding seasons are often longer and pro-
vide opportunities for many nesting attempts (Table 1).
Sub-tropical breeding seasons are very poorly known, yet
often seem to be as short as temperate breeding seasons (Auer
et al. 2007) and how these birds fi t in the temperate–tropical
debate remains to be seen. Th e entire breeding season must
be considered, because low nest success does not necessarily
imply low annual reproduction (Th ompson et al. 2001,
Roper 2005). Modeling additional factors that infl uence
annual reproductive success has been applied to a temperate
bird species (Podolsky et. al 2007) but this approach has not
been applied to the infl uence of latitudinal diff erences in
reproductive success on breeding adaptations.
Here, we used Monte Carlo-based numerical simulations
to analyze the consequences for annual reproductive success
of variation in breeding season length, renesting intervals
and predation rates for tropical and temperate passerine
birds. Specifi cally, we wished to test the hypothesis that as
daily nest mortality rates increase, natural selection should
increasingly favor adaptations to reduce those rates.
Methods
The model
Breeding season length, nest cycle length, renesting interval
lengths, and nest survival rate determine annual reproduction.
Here, we assume a time-dependent nest mortality model,
due to predation and accident, and exclude starvation and
partial predation from our models. Also, although clutch size
varies among species and between regions at high and low
latitude, we assume that clutch size is constant within a pop-
ulation and otherwise unrelated to the other nesting
parameters used in the model. With these assumptions,
annual nesting success (the number of nesting attempts that
fl edge at least one off spring) is the product of the number of
breeding attempts per year and the individual nest success.
We assume that nest survival rate is constant over the nest
cycle (Roper and Goldstein 1997, Sullivan et al. 1999). Accord-
ingly, nest success (S) is the daily survival rate (p) taken to the
power of the number of days a nest is at risk (nest cycle length,
T), that is, S  pT. Breeding season length, duration of each
nesting attempt, and renesting intervals determine the annual
number of breeding attempts (Ricklefs and Bloom 1977).
We used Monte Carlo methods to simulate annual repro-
duction in birds based on these nesting parameters. Simula-
tions provide an estimate of the variation in outcomes due to
the stochastic nature of nest loss. Th e simulations in this
study were iterated through combinations of nest survival
rates (p), renesting intervals (after-success, r
s
, and after-fail-
ure, r
f
) and breeding season lengths (B). Th e duration of an
individual nesting attempt (nest cycle length) is the number
of days between its initiation and termination (either through
fl edging or predation, maximum at fl edging). For each day
of the nesting attempt, a uniform random variate (URV) was
generated (0  URV  1). Th e nest was considered to have
survived a given simulated day when the URV was less than
p; otherwise the nest failed. Iterations through individual
nesting attempts were continued until nests failed or until 25
days were reached (the nest cycle length in this analysis). 720A single nest cycle length was used for all simulations to sim-
plify the analyses and their interpretation.
Following each nesting attempt, a renesting interval was
added, the length of which depended on whether the nest
failed (after-failure interval) or succeeded (after-success
interval). After-success intervals were longer because they
included a period of post-fl edging parental care. Th e length
of each complete nest cycle was added to that of previous
cycles until the sum equaled the breeding season length
minus the time required for another successful attempt.
Th is process was repeated 30 times at all combinations of
breeding season lengths, renesting intervals, and nest sur-
vival rates. Th irty individual nesting seasons for each com-
bination of variables provide a suffi cient sample for
comparisons of nesting success in fi eld studies and we
wished to simulate those values likely to be encountered in
fi eld studies. Values used in the simulations were obtained
from studies of breeding in tropical and temperate passer-
ine birds (tropical: Skutch 1950, 1985, Snow 1964, Dyrcz
1983, Young 1994, data presented herein; temperate: Rick-
lefs 1968b, Zimmerman 1984, Skutch 1985, Weathers and
Sullivan 1989, Gawlik and Bildstein 1990, Morton et al.
1993, Norment 1993, Table 1).
Th e sequence of simulations proceeds from a more gen-
eral perspective, in which values used in the model vary
widely (refl ecting the general view of latitudinal variation
and its consequences), to more specifi c, realistic scenarios.
Th e fi nal simulation is the most specifi c, based on pub-
lished breeding data for three species of passerine bird (two
temperate and one tropical). At higher latitudes, avian
breeding seasons are typically shorter, nest survival is usu-
ally higher, and birds tend to renest more rapidly following
nest success. Th is combination will be referred to as the
‘temperate scenario’. Conversely, long breeding seasons,
lower nest survival, and longer renesting intervals comprise
the ‘tropical scenario’.
Simulation 1. Broad variation in nest survival rates
In this simulation, breeding seasons were 90 and 250 days long
(temperate and tropical scenarios respectively; Snow and Snow
1963, Snow 1964, Ricklefs and Bloom 1977, Table 1). Two
ranges of renesting intervals were used. Th e renesting intervals
were either the same in temperate and tropical regions or the
same proportion of the breeding season length for both. When
intervals were the same in both regions, we used a range of
intervals representative of temperate birds (4–22 days after suc-
cess, and 3–12 days after failure, Table 1). When intervals were
the same proportions of the breeding seasons, the temperate
zone range of 4–22 days was increased to 11–61 days for trop-
ical birds. Conversely, 1–8 day intervals span the same propor-
tion of 90 days (for temperate birds) as 4–22 day intervals of
250 days (for tropical birds, cf. Table 2). Lastly, the broad range
of nest survival rates spanned values reported across latitudes
(0.915–0.995 d–1 in 0.02 d–1 increments; Skutch 1950, 1985,
Ricklefs 1969).
Simulation 2. Variation in breeding season lengths
In the second simulation, we examined variation in breeding
season length. Temperate breeding seasons were 60 and 90
days, and tropical breeding seasons were 167 and 250 days (the
short season was arbitrarily 2/3 the length of the long season).
In this simulation, nest survival rates for the temperate scenario

Temperate 93–126c 1–20/3–11 Ricklefs 1968b
51, 120 3.5–7.5, 3–16 Scott et al. 1987
~60–135 Mason 1985e
105–120 Gawlik and Bildstein 1990
~120 4d Weathers and Sullivan 1989
~90 Zimmerman 1984
~105 Payne 1989
Hötker 1988
Fitzpatrick and Woolfenden 1989f
Summary 51–150 1
8

8
su
. dSwere 0.975–0.995 d–1 (in 0.005 d–1 increments, average
0.985 d–1) and for the tropical scenario, 0.91–0.93 d–1 (also
in 0.005 d–1 increments, average 0.92 d–1; Roper and Gold-
stein 1997). Finally, renesting intervals were the same in
both simulations (4–22 days after success, 3–12 days after
failure, Table 3).
Simulation 3. Renesting interval variation
Renesting intervals in the temperate scenario were short (1–8
days after nesting success and 1–4 days after nest failure) and
long (4–22 days after nest success and 3–12 days after nest
failure). Some temperate birds renest almost immediately
Tropical 198–235c
150–365
180–300
~155
120-year round
90–201
225–273
Summary 150–365
aValues were estimated if not explicitly stated in the references. bAfter
cFrom Table 1, without arctic birds with a breeding season of 40 days
near Buenos Aires, Argentina. fColonial breeding jays.after nest failure or chick fl edging (Scott et al. 1987, Weathers
perate birds, and long intervals were 11–61 days following
success and 8–33 days following failure (Table 3). Th e short
and long intervals were the same proportions of the temper-
ate and tropical breeding season lengths as in Table 2.
Simulation 4. Narrow variation in nest survival rate
Nest survival rates were varied with two options. First, for
large rate variation, the diff erence between the maximum
Table 2. Values used in simulations to generate Fig. 1, results in Table 4
Simulation DSRa (increment) Breeding seasonb
Temperate 0.915–0.995
(0.02)c
90
Tropical 0.915–0.995
(0.02)c
250
aDaily survival rates (1-daily predation rates) used in the Monte Carlo s
breeding season lengths at both latitudes and the values were chosen tnest survival rate and the minimum nest survival rate was
0.04 d–1, and second, the diff erence in rates for the low
variation was 0.01 d–1. Th us, temperate nest survival rates
ranged from 0.9650–0.99975 d–1 and from 0.9825–0.9925
d–1 (for high and low variation, respectively, both averaging
0.9875 d–1). Tropical nest survival rates ranged from 0.90–
0.94 d–1 and 0.915–0.925 d–1 (high and low variation respec-
tively, average  0.92 d–1). In the larger temperate range, the
greatest nest survival rate could not be the same increment
larger than the next, because it would have exceeded 1.0 and
so 0.99975 d–1 was used to be less than, but near, 1.0. Breed-
ing season length was the same as in the preceding simula-
–20/3–11
–59/6–24 Ricklefs 1968b
Snow 1964
Young 1994
Lawton and Lawton 1985
Isler and Isler 1987
Wikelski et al. 2003
30/5–72 Roper 2005
–59/4.5–72
ccess/after failure, only one value indicates after nest failure.
hortest renesting interval after a successful nest. eSouthern latitudes tion, and renesting intervals were the same proportion of the
Study species
Th e red-faced warbler Cardellina rubrifrons (Martin and
Barber 1995) and the yellow-eyed junco Junco phaeonotus
(Weathers and Sullivan 1989) provided model temperate
species for detailed simulations. Th e western slaty antshrike
Th amnophilus atrinucha provided a tropical comparison. Th e
breeding data of the western slaty antshrike are reported here
(see also Roper and Goldstein 1997, Roper 2005).
and to illustrate the protocol.
Renesting interval
Post-successb Post-failureb
4, 10, 16, 22
1, 4, 6, 8
3, 6, 9, 12
1, 2, 3, 4
11, 28, 44, 61
4, 10, 16, 22
8, 16, 25, 33
3, 6, 9 12
imulations. The renesting intervals are the same proportions of the
o approximate the data in Table 1. bBreeding season lengths, post-success and post-failure renesting intervals are all in days. cThis is a very broad range of predation rates that encompasses both tropical and
temperate values.and Sullivan 1989). In the tropical scenario, short renesting
intervals were the same as the long renesting intervals of tem-
breeding season for both simulations (Table 3).35–150
~90Table 1. Breeding seasons and renesting intervals (in days) from selected studies of temperate and tropical passerine birds that provided
values used in the simulations.a
Region Breeding Renestingb Reference721

Breeding season lengths, renesting intervals, and nest sur-
vival rates were either calculated or estimated from these stud-
ies. Red-faced warbler data did not include the standard
deviation of the nest survival rate, so we used the value for the
yellow-eyed junco. Red-faced warbler breeding seasons were
approximately 67 days and they have only one successful nest
each year, but they may renest after early failure (Martin and
season (Weathers and Sullivan 1989). Renesting after failure
was also assumed to be brief (1–7 days). Nest survival rate
for juncos in 1986 was 0.975 (SE  0.013, Table 5).
Th e western slaty antshrike Th amnophilus atrinucha has a
breeding season length of 225–270 d. Th e nesting cycle lasts
21–26 days, with 12–14 d incubation period and a 9–12
d nestling stage. Renesting intervals varied from 48–100 d
Simulation (increment) Breeding seasonb Post-success renestingb Post-failure renestingb
Temperate2 0.975 – 0.9975 (0.005) 60 – 90 4, 10, 16, 22 3, 6, 9 12
Tropical2 0.91 – 0.93 (0.005) 167 – 250 4, 10, 16, 22 3, 6, 9, 12
Temperate 0.975 – 0.9975 (0.005) 90 4, 10, 16, 22
1, 4, 6, 8
3, 6, 9, 12
1, 2, 3, 4
Tropical 0.91 – 0.93 (0.005) 250 11, 28, 44, 61
4, 10, 16, 22
8, 16, 25, 33
3, 6, 9, 12
Temperate3 0.9675 – 0.99975 (0.01)
0.9825 – 0.9925 (0.0025)
90 4, 10, 16, 22 3, 6, 9 12
Tropical3 0.90 – 0.94 (0.01)
0.915 – 0.925 (0.0025)
250 4, 23, 42, 61 3, 13, 23, 33
aSurvival rates used in the Monte Carlo simulations. bBreeding season lengths, Post-success and post-failure renesting intervals are all in days.
2, 3These simulations are illustrated in Fig. 2 and 3 respectively.
lmore than once successfully was possible only when the fi rst
nesting attempt was successful. Renesting intervals after nest
failure were not reported, so we assumed them to be relatively
brief (1–7 days, Scott et al. 1987). Th e nest survival rate for red-
faced warblers (0.971) was calculated from the reported mean
nesting success (43.5%) and nesting cycle length (28 days).
Yellow-eyed juncos have a maximum breeding season of
80 days (Weathers and Sullivan 1989, Sullivan unpubl.).
Juncos may renest immediately following independence of
young, and so renesting intervals used in the simulation were
short (2–11 days). In contrast to warblers, however, jun-
cos may nest successfully up to three times in any breeding
Table 4. Variance explained by the regression models and each variab
results are illustrated in Fig. 1–3.
Simulation Treatmen
Large range in nest survival, short
and long breeding seasons
temperate short
temperate long
tropical short
tropical long722
tropical long
Renesting interval length
temperate small
temperate large
tropical small
tropical large
Nest survival rate
temperate small
temperate large
tropical small
tropical large
aSurvival, success and failure are daily nest survival, renesting intervals adetails of the western slaty antshrike reproductive season, see
Supplementary material Appendix 1.
Statistical analysis
Multiple regression analysis was used to calculate the relative
potential strength of selection on each of the breeding-season
variables (Neter et al. 1985, Phillips and Arnold 1989). Th is
analysis illustrates how annual reproduction (dependent
variable) may be infl uenced by renesting intervals, nest sur-
vival rate, and length of breeding season (independent vari-
ables). Th us, in this analysis, the dependent variable is
e in each model. Partial r2 values less than 0.10 are not shown. These
t Model R2
Partial r2 a
Survival Success Failure
0.92 0.87
0.89 0.74 0.12
0.97 0.89
0.93 0.73 0.140.82 0.13 0.20 0.5
0.86 0.22 0.63
0.90 0.10 0.80
0.79 0.30 0.14 0.34
0.75 0.17 0.20 0.39
0.87 0.83
0.88 0.25 0.63
0.81 0.18 0.59
0.89 0.41 0.14 0.35
fter success and failure respectively (values in boldface are the partial
r2 values that explain the most variation in annual reproductive success in each model).Breeding season length
temperate short 0.74 0.18 0.56
temperate long 0.89 0.12 0.76
tropical short 0.79 0.29 0.44Barber 1995). To refl ect these details, intervals for renesting
after success (14–20 days) were long and chosen so that nesting
after success, and 4–73 d after failure. Nest survival rate was
0.914 (nest predation rate  0.086 d–1, SE  0.009). For Table 3. Values used in simulations to generate Fig. 2 and 3 and results in Table 4. Cells in bold text indicate those that provided the variation
in the simulation while the other variables maintained the same variation about values that were equal in both, or typical of each latitude.
Daily survival ratesa

in each model that was uniquely explained by each indepen-
dent variable. Simulations and statistical analyses were car-
ried out with SAS (ver. 6.11, 1995) and JMP (ver. 7.01).
Results
Broad variation in nest survival rate
In simulations with large variation in nest survival rates, nest
survival rate itself was the most infl uential variable for annual
Tropical short renesting Temperate short renesting
Yr
–
1
30
p
ai
rs
–
1
Post failure
o
eTropical long renesting
Su
cc
es
sf
ul
n
es
ts

Post success Post success
Post failure
Temperate long renesting
Figure 1. Th ree dimensional representation of the results of the simu-
lation. In this simulation, daily nest survival rates (DSR) are equal
and variable at both latitudes, while renesting intervals vary within
realistic limits at each latitude (Table 2). Th e planes are at constant
DSRs at varying renesting intervals (after successful and failed nests).
Th e Z axis (successful nests year ) is the sum of the number of suc-
cessful nests of 30 pairs of birds as calculated in the simulation. Th is
simulation represents the general current view of how birds should
respond to nest predation. Note that the slopes of the planes as well
as the distances between them are greater in the tropical (left) panels
and thus, one would expect selection to favor those birds that reduce
predation rates. On the other hand, in the temperate latitudes,
changes in DSR as well as changes in renesting intervals are less infl u-reproduction (explaining 73–88% of the variance, Table 4,
Fig. 1). Renesting intervals also explained a relatively large part
of the variation in annual reproduction in both temperate and
tropical scenarios. Th e rank importance (in terms of number
of successful nests per breeding season) of the independent
variables was also similar in temperate and tropical scenarios.
Variation in breeding season lengths
Simulations using short breeding seasons explained less of
the variance in annual reproduction than did simulations
with long breeding seasons. For temperate birds, the model
R2 increased from 0.743 (short breeding season) to 0.893
(long breeding season), and for tropical scenarios it increased
from 0.791 to 0.822, respectively (Table 4, Fig. 2).
For temperate birds, the renesting interval after success
infl uenced annual reproduction most, and more so when the
breeding season was longer (explaining 56–76% of the vari-
ance). Variation in renesting interval after predation did not
contribute to variation in annual reproduction for either
long or short seasons (Fig. 1, 2, Table 4).
In the tropical scenario, the renesting interval after fail-
ure was the most important variable for annual reproduc-
tion (explaining 44–50% of the variance). Variation in nest
survival rate only weakly infl uenced annual reproduction,
especially during short breeding seasons, when it explained
 10% of the variation (Table 4, Fig. 2).
Renesting interval variation
In the temperate scenario, renesting after success explained
most of the variation in annual reproduction when the
variation in renesting intervals was large and was still impor-
tant when renesting intervals were less variable. Th e impor-
tance of nest survival rate was similar regardless of variation
in renesting intervals. And, as before, renesting after preda-
tion was unimportant (Table 4).
In the tropical scenario, annual reproduction was most
lers and yellow-eyed juncos (Fig. 2).
Renesting intervals after
Breeding seasonb Successb Failureb
250 48, 65, 82, 100 4, 27, 50, 73
67 14, 16, 18, 20 1, 3, 5, 7
80 2, 5, 8 11 1, 3, 5, 7
f each was that calculated for each species. bBreeding season lengths,
r the western slaty antshrike are described herein, and for the red-faced
-eyed junco.e (From Sullivan unpubl. data). fThe mid-point of the daily
d by the standard error as estimated in the program MARK.measured as the number of successful nests produced by 30
pairs in each breeding season. Coeffi cients of partial determi-
nation (r2) were used to show the proportion of the variance
Table 5. Values used in simulations for slaty antshrikes, red-faced warb
Species Survival ratesa (increment) Nesting period
SLASc 0.897 – 0.932 (0.0087)f 24
RFWAd 0.945 – 0.997 (0.013)f 28
YEJUe 0.949 – 0.999 (0.013)f 25
aPredation rates used in the Monte Carlo simulations and the mid-point
post-success and post-failure renesting intervals are all in days. cValues fo
warbler.d Were calculated from Martin and Barber (1995), for the yellow
survival rates was the actual value for each species, and was incrementinfl uenced by renesting after nest failure (predation) regard-
reproduction (explaining 30 and 34% of the variance respec-
tively). Th e relative importance of renesting after success was
less in the tropical than in the temperate scenario (Table 4).
Narrow variation in nest survival rate
For temperate birds, renesting after success was most impor-
gardless of the variation in tant for annual reproduction, reential on the number of successful nests.less of the range of variation in renesting intervals. When
renesting intervals varied little, nest survival rate and renest-
ing after failure were similar in their importance for annual
Th e slope of the plane indicates the infl uence of the variation in ren-
esting interval over the particular interval of interest, while the diff er-
ence between planes indicates the importance of that change in DSR.
–1723

Tropical wide DSR
ce
ss
fu
l n
es
ts
Y
r–1
3
0
pa
ir
Temperate wide DSR
Post failureSu
cc
es
sf
ul
n
es
ts
Y
r–1
3
0
pa
irs
–
1
Tropical short season Temperate short season
Tropical long season
Post success Post success
Post failure
Post failure
Temperate long season
Figure 2. See Fig. 1, Table 3. Here, with general values for renesting
intervals at long and short breeding season lengths, we see some-
what similar latitudinal responses to DSR and renesting intervals.
However, note that in the upper right panel (temperate, short
breeding seasons) there is only time for one successful nest per pair,
regardless of renesting intervals – hence the short fl at planes – and
so change in DSR provides a greater increase in nesting success.
Similarly, in the lower right panel, at long breeding seasons, very
short renesting intervals can increase annual reproduction – hence 724
tion in nest survival rate was low, renesting after failure
contributed the most to annual reproduction and renesting
after success contributed the least (explaining 59% and ~3%
of the variance, Table 4). When variation in nest survival rate
was larger, nest survival rate and post-predation renesting
intervals contributed nearly equally to reproduction (explain-
ing 40% and 35% of the variance, respectively). Renesting
after success was less important at both levels of variation in
nest success (at 18% and 14% respectively, Table 4, Fig. 3).
Simulation based on data
Th e relevant data used in the simulations, obtained from the
literature for the red-faced warbler and yellow-eyed junco,
and as described in supplemental information for the west-
ern slaty antshrike, are summarized in Table 5. For both tem-
perate species, variation in nest predation contributed most
to annual reproduction (explaining 81% for the warbler, and
79% for the junco) and the variance explained by the
the sudden steep slope. For the tropical scenarios, however, the dif-Renesting intervals (days)
Post failure
Post success Post success
Figure 3. See Fig. 1, Table 3. Th e top two panels have low variation
and the bottom two have wide variation in DSR. Note that the
planes are virtually identical for tropical birds when variation in
DSR is small. Also, when variation in DSR is large, the diff erence
between long and short renesting intervals results in a greater
change in nesting success than does the diff erence between DSRs.
Low variation in DSR in temperate birds also suggests that renest-
ing intervals are important, but at wide variation in DSR a change
in DSR aff ords a greater increase in nest success than does a change
in renesting intervals, but at relatively short renesting intervals, a
change in DSR results in much greater reproductive success.nest survival rates, when these rates were typical of temperate
latitudes. However, with greater variation in nest survival
rates, the proportion of the variance explained by renesting
interval decreased (from 83–63%) and that explained by nest
survival rate increased (from 3–25%, Table 4). Renesting after
predation did not contribute to annual reproduction, regard-
less of the variation in nest survival rate. (Table 4, Fig. 3).
For tropical ranges of daily nest survival rates, when varia-
ference in successful nests between the maximum and minimum
renesting intervals is greater than that for diff erent DSRs (vertical
diff erences between planes) and so changes in renesting intervals are
more important than change in DSR. See text for details. regression (84% and 87% respectively, Table 6, Fig. 4). For
the tropical western slaty antshrike, nest predation accounted
for only 12% of the variance explained by the model, while
renesting interval after failure accounted for 62% of the vari-
ance (the model explained 83%, Table 6, Fig. 4).
Discussion
Simulations, from general to specifi cSu
cTropical narrow DSR
s–
1
Tropical narrow DSRSimulation results suggest the surprising conclusion that low
nest predation rates are associated with stronger selection
pressure to reduce those rates even further. Th is arises from
the fact that a given change in nest mortality rate has a greater
proportional eff ect on lower mortality rates than on higher
Table 6. Variance explained by regressions, and the proportion
explained by each variable.
Simulation (Fig. 5) Model R2
Partial r2a
Predation Success Failure
Western slaty antshrike 0.826 0.122 0.082 0.622
Red-faced warbler 0.836 0.812 0.011 0.013
Yellow-eyed junco 0.870 0.787 0.063 0.019
aPartial r2 values indicate the proportion of the variance of annual
nesting success that is explained by each independent variable. Pre-
dation, success and failure indicate the variance associated with
nest predation rate, renesting intervals after success and after failure
respectively. See Table 5 for the simulation protocol.

os
c v
pic
d-
on
e
d
, t
es
zo
lo
enmortality rates. Th is simple insight might help to explain why
several studies fi nd relationships between nest predation and
life history in temperate birds while similar relationships are
absent in the few tropical studies (Martin 2004, Roper 2005,
Martin et al. 2006). Th is brings up the question of what vari-
ation in predation rate, breeding season length or renesting
intervals is actually available to the birds in the wild?
n
es
ts
fail
P
Figure 4. Graphical representation of the fi tness components of realisti
one tropical and two temperate species of birds. Left panel is the tro
yellow-eyed junco J. phaeonotus and the right panel is the temperate re
cessful nests per breeding season for 30 pairs of birds in the simulati
success and after failure) and represents a constant daily nest survival rat
renesting values that result in a larger increase in successful nests than
the increase in successful nests due to a change in DSR. For the junco
the change in after success renesting interval suddenly allows two succ
from changing renesting intervals and all planes are more or less hori
larger than that in the junco or antshrike. Th us, the tropical bird with
renesting intervals more than increasing DSR while the warbler only bTh e fi rst simulation considered the large range of nest
became clearer when more typical ranges in nest survival
rates were used (Fig. 2, 3). Time is obviously important and
so birds should renest rapidly in both latitudes. However,
temperate birds, with low predation rates, may seldom have
the chance to renest. Once a nest is successful, they may
spend the remainder of the season in post-fl edging care, after
which there is little time for another attempt. In the tropics,
most nests are unsuccessful and so birds should renest quickly
after failure (Fig. 2). But, since a time constraint may often
be less important, tropical birds may renest after both unsuc-
cessful and successful nests. Nest survival rate is important
(but less so than in the previous simulation), probably
because it determines how many renesting attempts are
required for successful breeding: higher predation rates will
result in a greater number of nest attempts per successful
nest. Th e importance of renesting intervals will depend on their variance and that of the variance in renesting intervals
(Fig. 2, 3). While temperate birds may renest more quickly
after failure than tropical birds (Table 1), the time it takes for
renesting is a larger proportion of the total breeding season
length than that for tropical birds. Th us, the relative strength
of selection will depend on the combination of these vari-
ables and their variance. When temperate birds do have time
t success
alues of renesting intervals and a range of daily nest survival rates for
al western slaty antshrike T. atrinucha, middle panel the temperate
faced warbler C. rubrifrons. Each plane indicates the number of suc-
. Th e plane is the consequence of the two renesting intervals (after
(DSR). Note for the tropical antshrike, the slope of the plane includes
oes the diff erence at any point between two planes, which would be
he planes are nearly horizontal except for the stepwise change where
sful nests per year per pair instead of one. Th e warbler never benefi ts
ntal, while the diff erence between planes (change in DSR) is much
ng breeding season and high predation rates benefi ts from reducing
efi ts from increasing DSR. See text and Table 5 for details.and manage to carry out additional nesting attempts, the
survival rates found across latitudes. If we assume large varia-
tion in survival rate across latitudes, then we bias our expec-
tations towards reduction in nest predation in both temperate
and tropical regions (Fig. 1). If natural variation in nest sur-
vival rate were this large, then birds that could reduce preda-
tion would always be favored. However, in nature such a
large variation in nest survival rates within populations has
not been demonstrated and perhaps is inherently diffi cult to
demonstrate.
Latitudinal diff erences in response to nest predation
benefi ts for annual reproduction may be very large (Nagy
and Holmes 2005) and emphasizes that small reductions in
low predation rates may allow these repeated attempts.
Th ese fi rst three simulations were intended to become
more realistic at each step. If variation in nest survival rate is
low, variation in renesting intervals becomes more impor-
tant. Th is sequence of simulations illustrates that to predict
the importance of any one variable associated with reproduc-
tion, one must understand the available variation in all the
variables together.725
Red-faced warblers, yellow-eyed juncos
and slaty antshrikes
Striking latitudinal diff erences become clear with data-based
simulations for particular species. Patterns for the two tem-
perate species are quite similar, even though breeding season
lengths and renesting intervals diff er (Table 5, Fig. 4). For
both species, variation in nest survival rate has the largest
infl uence on annual reproduction as measured by the regres-
sion analyses (Table 6). In the fi gures, this translates to the
diff erence between the planes being larger than the diff er-
ence within a plane over the range of the two renesting inter-
val axes (Fig. 4). Note that renesting intervals for the red-faced
warbler are unimportant (planes are parallel) because they
rarely successfully nest twice (Martin and Barber 1995). Still,

how these factors may interact (Fig. 5). Th e fi rst (top panel)
represents how varying nest predation probability infl uences
reproductive success in tropical and temperate birds, while
the bottom panel shows the same for renesting intervals.
Over combinations of breeding season length and predation
rates typical or possible at each latitude, a unit change in
nesting success will have a large eff ect on nesting success in
temperature birds and a much smaller eff ect in tropical birds
(Fig. 5A). Conversely, a unit change in renesting interval
after failure will have a large infl uence on nesting success
in tropical, but not temperate, birds (Fig. 5B). Th erefore,
to properly interpret possible adaptive bases of life history
traits across latitude, we must have information about breed-
ing season length and nesting success as critical components
83% of the red-faced warbler pairs may have a successful A
al
Figure 5. A graphical model illustrating the trends suggested here.
(A) Shows that changing predation rates, when all other factors are
considered at the appropriate latitude, will increase annual repro-
ductive success more for temperate than for tropical birds. (B) Simi-
larly, shows that changing renesting intervals (especially after failure),
when all other factors are considered at the appropriate latitude, will
increase annual reproduction more for tropical birds. Th e two verti-
cal arrows indicate the change in annual reproduction for tropical
(solid line, left) and temperate (dashed line, right) birds given the
unit change in the abscissa implication is that we must have infor-
mation for nesting success, breeding season length and renesting
intervals to predict how birds might increase reproductive success at
any given location.nest by the third attempt (at 44% overall nesting success).
But, that will only be possible if the fi rst nest attempts fail
rapidly and renesting is rapid.
Renesting intervals were relatively unimportant (planes
mostly fl at) for the yellow-eyed junco except for the quan-
tum jump where the combinations of breeding season length
and renesting intervals suddenly allow three rather than two
successful nests per year (Fig. 4). Indeed, juncos may success-
fully nest up to three times in a season (Weathers and Sullivan
1989, Sullivan et al. 1999) as suggested in Fig. 4. Also, the
longer breeding season, which allows juncos to nest more
than once in any year, shows that by increasing nest success
(through reducing predation) a large infl uence on annual
reproduction follows. For these birds, 88% of pairs will have
successfully fl edged young by the third nesting attempt (at
51% overall nesting success). In the black-throated blue war-
bler, birds that double-brooded initiated their fi rst brood
only a few days prior to those that did not, however, the
most important factor determining double brooding was
food availability (Nagy and Holmes 2005). Th us, perhaps
food, not time, is the constraint for repeated brooding in
temperature latitudes.
Nest success in the western slaty antshrike is very low
(0.914 d–1) compared with both temperate species (red-faced
warblers, 0.971 d–1, and yellow-eyed juncos, 0.974 d–1). Vari-
ation in the renesting interval after failure contributed most
to variation in nest success, which generally came after a
series of failed attempts. Because of their low nest survival
rates, only 31% of antshrike pairs are successful by their third
nesting attempt (compared to 83–88% for the temperate
species), and only by the sixth attempt have 50% of the pop-
ulation successfully fl edged young. Th us, selection favors
tropical birds that renest more rapidly or more often, while
selection favors temperate birds that reduce nest predation
risk. It should be noted that while predation was very high on
the western slaty antshrike, the rate of production of young
during the year exceeded that of mortality of adults (Roper
2005). In all three species, we see that understanding the rate
of production of young must consider the entire season rather
than the nest survival rate alone (Th ompson et al. 2001).
Whether selection acts on life history traits in a popu-
lation depends on heritable variation in the trait of inter-
est. Th e stronger selection gradient in temperate birds for
changes in nest mortality rate can produce evolutionary force
if individuals vary in behaviors that infl uence nest success,
including nest site selection and nest defense behaviors. Pos-
sibly, temperate regions, where nest predation rates are more
often low, have greater intra-specifi c variation in nest site
selection, with a greater consequent infl uence on nest preda-
tion. Or, behaviors at and near nests may be more important
for predation in temperate birds. Both of these may be the
case (Martin and Roper 1988, Fontaine and Martin 2006,
Chalfoun and Martin 2009). For tropical birds, in contrast,
nest predation might not vary in a predictable manner among
a wide range of potential nest sites (Robinson et al. 2000,
Roper 2000, 2003). Given that nest predation is uniformly
high, the best option for increasing annual nest success
might be to increase the number of attempts by reducing the
interval after failure. We present a simple model to suggest 726 re
pr
od
uc
tio
n
Predation rateB
An
nu

of the environmental context. For example, the life histo-
ries of tropical species having short breeding seasons, at high
altitude or where rainfall is seasonal, should resemble those
of temperate regions in most respects owing to the strong
infl uence of breeding season on selection regimes. A simi-
lar change in breeding season dynamics due to elevational
variation has been demonstrated in a temperate bird (Bears
et al. 2009, Tieleman 2009) and we predict that similar
elevational and seasonal variation will be found in tropical
birds under similar conditions. Also, perhaps small clutches
and high predation risk favors parents that make few trips
to the nest (Fontaine and Martin 2006). Th ese fewer trips
by the parents may provide the additional benefi t of reduc-
ing their own energy costs of reproduction which will then
permit more rapid renesting after the current nest fails
(Roper 2005). Clearly, breeding parameters modeled here
do not encompass the ranges possible. However, by using
this simple simulation, we show that we must include breed-
ing parameters if we wish to compare species over latitudinal
ranges to understand how predation may infl uence life his-
tory evolution.
Finally, the perspective and simulations presented here
respond to the call for alternative hypotheses for interpreting
avian life histories (Martin 2004, Lima 2009). For example,
smaller clutch size in the tropics might not be a response to
low food availability (Lack 1947) nor to reduce nest preda-
tion rates (Skutch 1949, 1985, Roper and Goldstein 1997).
Instead, small clutches may reduce reproductive costs for any
attempt to help insure future nesting attempts. Future stud-
ies of latitudinal trends in life history evolution will require
detailed breeding season information to examine the impor-
tance of variation in any of the parameters infl uencing repro-
ductive success. Simulations of the kind reported here can
quantify the strength of selection resulting from variation in
the breeding parameters. Here we have shown that low nest
predation rates, as in temperate latitudes, may be associated
with strong selection favoring individuals with lower than
average predation risk. In contrast, when predation rates are
high, as in tropical birds, long nesting seasons and low nest
survival rates place a premium on short renesting intervals
rather than predation avoidance.
Acknowledgements – Th anks to the Smithsonian Tropical
Research Institute for all the support that made this work
possible. Th anks to Stan Rand, Peter Wrege, Steve Emlen,
Cathy Robb and Janna Ellingson for support while in
Panamá. Th anks to Paul Sievert, Graham Watkins, Dina
Fonseca and Frank Gill for discussions and constructive crit-
icisms. Th anks to the Berryman Inst. of Utah State Univ. for
their postdoctoral support. Special thanks to Amanda Bakian
for her kind help with the simulations.
References
Auer, S. K. et al. 2007. Breeding biology of passerines in a sub-
tropical montane forest in northwestern Argentina. – Condor
109: 321–333.
Bears et al. 2009. Breeding in high–elevation habitat results in shift
to slower life-history strategy within a single species. - J. Anim.
Ecol. 78: 365-375.Brawn, J. D. et al. 1995. Demography of birds in a neotropical forest:
eff ects of allometry, taxonomy and ecology. – Ecology 76: 41–51.
Chalfoun, A. D. and Martin, T. E. 2009. Habitat structure mediates
predation risk for sedentary prey: experimental tests of alterna-
tive hypotheses. – J. Anim. Ecol. 78: 497–503.
Conaway, C. H. et al. 1974. Geographic variation in litter size and onset
of breeding in cottontails. – J. Wildlife Manage. 38: 473–481
Conway, C. J. and Martin, T. E. 2000. Evolution of passerine incu-
bation behavior: infl uence of food, temperature and nest preda-
tion. – Evolution 54: 670–685.
Dunham, A. E. et al. 1988. Life history patterns in squamate
reptiles. – In: Gans, C. and Huey, R. B. (eds), Biology of the
Reptilia. Vol. 16. Alan R. Liss Inc., p. 659.
Dyrcz, A. 1983. Breeding ecology of the clay-coloured robin Turdus
grayi in lowland Panamá. – Ibis 125: 287–304.
Eisenberg, J. F. 1989. Mammals of the neotropics, Vol I. – Univ. of
Chicago Press.
Filliater, T. S. et al. 1994. Predation on northern cardinal nests:
does choice of nest-site matter. – Condor 96: 761–768.
Fitzpatrick, J. W. and Woolfenden, G. E. 1989. 13. Florida scrub
jay. – In: Newton, I. (ed.), Lifetime reproduction in birds.
Academic Press, pp. 201–218.
Fontaine, J. J. and Martin, T. E. 2006. Parent birds assess nest pre-
dation risk and adjust their reproductive strategies. – Ecol. Lett.
9: 428–434.
Gawlik, D. E. and Bildstein, K. L. 1990. Reproductive success and
nesting habitat of loggerhead shrikes in northcentral South
Carolina. – Wilson Bull. 102: 37–48.
Gibbs, J. P. 1991. Avian nest predation in tropical wet forest: an
experimental study. – Oikos 60: 155–161.
Holway, D. A. 1991. Nest-site selection and the importance of nest
concealment in the black-throated blue warbler. – Condor 93:
575–581.
Hötker, H. 1988. Lifetime reproductive output of male and female
pipits Anthus pratensis. – J. Anim. Ecol. 57: 109–118.
Isler, M. I. and Isler, P. R. 1987. Th e tanagers: natural history, dis-
tribution, and identifi cation. – Smithsonian Inst.Press.
Jetz, W. et al. 2008. Th e worldwide variation in avian clutch size
across species and space. –PLoS Biol. 6: e303. doi:10.1371/
journal.pbio.0060303.
Kulesza, G. 1990. An analysis of clutch-size in New World passe-
rine birds. – Ibis 132: 407–422.
Lack, D. 1947. Th e signifi cance of clutch size. – Ibis 89: 302–352.
Lawton, M. F. and Lawton, R. O. 1985. Th e breeding biology of the
brown jay in Monteverde, Costa Rica. – Condor 87: 192–204.
Lima, S. L. 2009. Predators and the breeding bird: behavioral and
reproductive fl exibility under the risk of predation. – Biol. Rev.
84: 485–513.
Martin, T. E. 1995. Avian life history evolution in relation to nest
sites, nest predation and food. – Ecol. Monogr. 65: 101–127.
Martin, T. E. 1996. Life history evolution in tropical and south
temperate birds: what do we really know? – J. Avian Biol. 27:
263–272.
Martin, T. E. 2004. Avian life-history evolution has an eminent
past: does it have a bright future? – Auk 121: 289-301.
Martin, T. E. and Roper, J. J. 1988. Nest predation and nest-site
selection of a western population of the hermit thrush. –
Condor 90: 51–57.
Martin, T. E. and Barber, P. M. 1995. Red faced warbler (Cardellina
rubrifrons). – In: Poole, A. and Gill, F. (eds), Th e birds of North
America, No. 152. Acad. Nat. Sci., Philadelphia, and
Am. Ornithol. Union, Washington, DC.
Martin, T. E. et al. 2000. Nest predation increases with parental
activity: separating nest-site and parental activity eff ects. – Proc.
R. Soc. Lond. B 267: 2287–2293.
Martin, T. E. et al. 2006. Life history and ecological correlates of
geographic variation in egg and clutch mass among passerine
species. – Evolution 60: 390–398.727

Mason, P. 1985. Th e nesting biology of some passerines of Buenos
Aires, Argentina. – In: Buckley, P. A. et al. (eds), Neotropical
ornithology. Ornithol. Monogr. Vol. 36. Allen Press.
Morton, M. et al. 1993. Nest predation in the mountain white-
crowned sparrow. – Condor 95: 72–82.
Nagy, L. R. and Holmes, R. T. 2005. To double-brood or not? Indi-
vidual variation in the reproductive eff ort in black-throated
blue warblers (Dendroica caerulescens). – Auk 122: 902–914.
Neter et al. 1985. Applied linear statistical models (2nd ed.). –
Richard D. Irwin, Inc.,
Norment, C. J. 1993. Nest-site characteristics and nest predation in
Harris‘ sparrows and white-crowned sparrows in the Northwest
Territories, Canada. – Auk 110: 769–777.
Oniki, Y. 1979. Is nesting success of birds low in the tropics? – Bio-
tropica 11: 60–69.
Payne, R. B. 1989. 10. Indigo bunting. – In: Newton, I. (ed.),
Lifetime reproduction in birds. Academic Press, pp. 153–172.
Perrins, C. M. 1977. Th e role of predation in the evolution of
clutch size. – In: Stonehouse, B. and Perrins, C. (eds), Evolutio-
nary ecology. MacMillan Press, pp. 181–191.
Schmidt, K. A. and Whelan, C. J. 1999. Nest placement and mor-
tality: is nest predation a random event in space and time? –
Condor 101: 916–920.
Scott, D. M. et al. 1987. Relaying interval after nest failure in
gray catbirds and northern cardinals. – Wilson Bull. 99:
708–712.
Skutch, A. F. 1949. Do tropical birds rear as many young as they
can nourish? – Ibis 91: 430–455.
Skutch, A. F. 1950. Th e nesting seasons of central American
birds in relation to climate and food supply. – Ibis 92:
185–222.
Skutch, A. F. 1985. Clutch size, nesting success, and predation on
nests of neotropical birds, reviewed. – In: Buckley, P. A. (eds),
Neotropical ornithology. Ornithol. Monogr. Vol. 36. Allen
Press, pp. 575–594.
Slagsvold, T. 1984. Clutch size variation of birds in relation to nest
predation: on the cost of reproduction. – J. Anim. Ecol. 53:
945–953.Phillips, P. C. and Arnold, S. J. 1989. Visualizing multivariate selec-
tion. – Evolution 43: 1209–1222.
Podolsky, A. L. et al. 2007. Modeling population growth of the
ovenbird (Seivrus aurocapilla) in the southern Appalachians. –
Auk 124: 1359–1372.
Ricklefs, R. E. 1968a. On the limitation of brood size in passerine
birds by the ability of adults to nourish their young. – Proc.
Natl. Acad. Sci. USA 61: 847–851.
Ricklefs, R. E. 1968b. Th e nesting cycle of songbirds in tropical and
temperate regions. – Living Bird 8: 165–175.
Ricklefs, R. E. 1969. An analysis of nesting mortality in birds. –
Smithsonian Contrib. Zool. 9: 1–48.
Ricklefs, R. E. 1970. Clutch size in birds: outcome of opposing
predator and prey adaptations. – Science 168: 599–600.
Ricklefs, R. E. and Bloom, G. 1977. Components of avian breed-
ing productivity. – Auk 94: 86–96.
Robinson, W. D. et al. 2000. Nesting success of understory forest
birds in central Panama. – J. Avian Biol. 31: 151–164.
Roper, J. J. 2000. Experimental analysis of nest sites and nest pre-
dation for a neotropical bird: stuck between a rock and a hard
place. – Ararajuba 8: 85–91.
Roper, J. J. 2003. Nest sites infl uence predation diff erently at natu-
ral and experimental nests. – Ornithol. Neotrop. 14: 1–14.
Roper, J. J. 2005. Try and try again: nest predation favors persis-
tence in a neotropical bird. – Ornithol. Neotrop. 16: 253–262.
Roper, J. J. and Goldstein, R. R. 1997. A test of the skutch hypo-
thesis: does activity at nests increase nest predation risk? – J.
Avian Biol. 28: 111–116.728Snow, D. W. 1964. Breeding seasons and annual cycles of Trinidad
land-birds. – Zoologica 49: 1–35.
Stearns, S. C. 1977. Th e evolution of life history traits: a critique of
the theory and a review of the data. – Annu. Rev. Ecol. Syst. 8:
145–171.
Sullivan, K. A. et al. 1999. Individual variation in renesting
intervals, dairly energy expenditure and reproductive success in
passerines. – In: Adams, N. J. and Slotow R. H. (eds), Proc.
22nd Int. Ornithol. Congr., Durban, South Africa, pp.
401–411.
Th ompson, B. C. et al. 2001. Nest success is not an adequate com-
parative estimate of avian reproduction. – J. Field. Ornithol.
72: 527–536.
Tieleman, B. I. 2009. High and low, fast or slow: the complemen-
tary contributions of altitude and latitude to understand life-
history variation. – J. Anim. Ecol. 78: 293–295.
Weathers, W. W. and Sullivan, K. A. 1989. Juvenile foraging profi -
ciency, parental eff ort, and avian reproductive success. – Ecol.
Monogr. 59: 223–246.
Wikelski, M. et al. 2003. Reproductive seasonality of seven Neotro-
pical passerine species. – Condor 105: 683–695.
Wyndham, E. 1986. Length of birds’ breeding seasons. – Am. Nat.
128: 155–164.
Yomtov, Y. et al. 1994. Clutch size in passerines of southern South
America. – Condor 96: 170–177.
Young, B. E. 1994. Th e eff ects of food, nest predation and weather
on the timing of breeding in tropical house wrens. – Condor
96: 341–353.
Zimmerman, J. L. 1984. Nest predation and its relationship to
habitat and nest density in dickcissels. – Condor 86: 68–72.

Appendix 1 Western slaty antshrike repro-
duction.
Methods
Th e breeding season for the western slaty antshrike Th amno-
philus atrinucha was studied in the fi eld on Pipeline Road in
Soberania National Park, in central Panama (1 January – 12
December, 1993). Antshrikes weigh approximately 24 g, are
common, permanently monogamous, territorial, and nest in
and total annual reproduction was determined for 30 pairs.
Clutch size is invariably two and both members of the pair
build nests, incubate eggs and nestlings and feed nestlings
and fl edglings (Oniki 1975, Stiles and Skutch 1989). Th e
nesting cycle lasts 21–26 days and consists of a 12–14 d
incubation period and a 9–12 d nestling stage (Roper 2005).
Nesting cycle (dates of construction, egg laying, hatching,
Breeding-season length
On 1 January 1993, the fi rst nest was found under construc-
tion and near 1 October the last nest failed. No breeding
attempts occurred for the remainder of 1993 (until 12 Decem-
ber). From 1 January – 1 October is 270 days. Breeding-season
length, when calculated to include variation in the number of
nests initiated each month, following Ricklefs and Bloom
(1977) was 225 days. In all months in which breeding
occurred, 16–80% of the pairs had active nests.
ing interval ranged from 48 to at least 157 days (average 
91 d). Th e 157 d interval represents seven pairs that fl edged
young and did not renest for the remainder of the year, during
which time the young remained in their natal territory. Drop-
ping these seven nests, the maximum renesting interval after
success was 86 days. Th e interval between the end of nest con-
struction and laying the fi rst egg was 1–90 days (Table 5).
.
ge
t w
a
va
e program Mark (2009).
Results
Nest predation and nesting success
Of 107 nesting attempts, 89 were lost to predation. Daily
survival was 91.4% (nest predation rate  0.086 d–1, stan-
dard error  0.009). Seventeen nests were successful, of
which young from 14 survived until independence. Pairs
averaged 5.1 nesting attempts. Th e modal number of nesting
attempts was four, with a range of 2–12 attempts pair–1 year–1
(Table A1).
Table A1. Intervals in days for western slaty antshrike nesting in 1993
Interval after n Avera
Successa 14 91
Successb 7 71
Failurec 69 23
Constructiond 55 7.1
Attempts/paire 21 5.1
aFrom the date of fl edging until the pair initiated the next nest or that nes
still with the parents, and they had not nested again at the end of observ
date of predation until the next attempt is initiated or found. dThe inter
was laid in the nest. eNumber of nest attempts of all pairs with completReferences
Howell, S. N. G. and Webb, S. 1995. A guide to the birds of Mexico
and northern central America. – Oxford Univ. Press.
Oniki, Y. 1975. Th e behavior and ecology of slaty antshrikes
(Th amnophilus punctatus) on Barro Colorado Island, Panamá
Canal Zone. – Anais Acad. Bras. Ciência 47: 471–515.
Ricklefs, R. E. and G. Bloom. 1977. Components of avian bree-
ding productivity. – Auk 94: 86–96.
Ridgley, R. S. and G. Tudor. 1994. Th e birds of South America.
Vol. II. – Univ. of Texas Press.
Roper, J. J. 2005. Try and try again: nest predation favors persis-
tence in a neotropical bird. – Ornithol. Neotrop. 16: 253–262.
Stiles, F. G. and Skutch, A. F. 1989. A guide to the birds of Costa
Rica. – Cornell Univ. Press.
Minimum Maximum
48 157
48 100
4.5 72
1 90
2 12
as found. Includes seven pairs that were open-ended: the young were
tions in 1993. bExcludes those seven observations. cFrom the estimated
l between apparent end of nest construction and the date the fi rst egg
data. Mode was 4 attempts and 13 pairs attempted  4 nests.fl edging) and the number of nesting attempts for each pair
were recorded. Daily nest survival was estimated using the the forest understory (Oniki 1975, Roper 2005). Th ey occur
from Guatemala and Belize to western Colombia (Ridgley
and Tudor 1994, Howell and Webb 1995). All breeding
attempts were recorded for 21 pairs of color-banded birds,
Renesting-intervals
Birds initiated new nests from 4.5–72 days (average  23 d)
after failed nesting attempts. After successful nests the renest-729