Evidence of the trade-off between starvation and predation risks in ducks.

Abstract

The theory of trade-off between starvation and predation risks predicts a decrease in body mass in order to improve flight performance when facing high predation risk. To date, this trade-off has mainly been validated in passerines, birds that store limited body reserves for short-term use. In the largest avian species in which the trade-off has been investigated (the mallard, Anas platyrhynchos), the slope of the relationship between mass and flight performance was steeper in proportion to lean body mass than in passerines. In order to verify whether the same case can be applied to other birds with large body reserves, we analyzed the response to this trade-off in two other duck species, the common teal (Anas crecca) and the tufted duck (Aythya fuligula). Predation risk was simulated by disturbing birds. Ducks within disturbed groups were compared to non-disturbed control birds. In disturbed groups, both species showed a much greater decrease in food intake and body mass during the period of simulated high risk than those observed in the control group. This loss of body mass allows reaching a more favourable wing loading and increases power for flight, hence enhancing flight performances and reducing predation risk. Moreover, body mass loss and power margin gain in both species were higher than in passerines, as observed in mallards. Our results suggest that the starvation-predation risk trade-off is one of the major life history traits underlying body mass adjustments, and these findings can be generalized to all birds facing predation. Additionally, the response magnitude seems to be influenced by the strategy of body reserve management.

Full text

Page 1

Evidence of the Trade-Off between Starvation and
Predation Risks in Ducks
Ce ´dric Zimmer1,2*, Mathieu Boos3, Nicolas Poulin1,2, Andrew Gosler4, Odile Petit1,2, Jean-Patrice Robin1,2
1Universite ´ de Strasbourg, IPHC, Strasbourg, France, 2CNRS, UMR 7178, Strasbourg, France, 3Research Agency in Applied Ecology, Naturaconst@, Wilshausen, France,
4Edward Grey Institute of Field Ornithology, Oxford University, Oxford, England
Abstract
The theory of trade-off between starvation and predation risks predicts a decrease in body mass in order to improve flight
performance when facing high predation risk. To date, this trade-off has mainly been validated in passerines, birds that store
limited body reserves for short-term use. In the largest avian species in which the trade-off has been investigated (the
mallard, Anas platyrhynchos), the slope of the relationship between mass and flight performance was steeper in proportion
to lean body mass than in passerines. In order to verify whether the same case can be applied to other birds with large body
reserves, we analyzed the response to this trade-off in two other duck species, the common teal (Anas crecca) and the tufted
duck (Aythya fuligula). Predation risk was simulated by disturbing birds. Ducks within disturbed groups were compared to
non-disturbed control birds. In disturbed groups, both species showed a much greater decrease in food intake and body
mass during the period of simulated high risk than those observed in the control group. This loss of body mass allows
reaching a more favourable wing loading and increases power for flight, hence enhancing flight performances and reducing
predation risk. Moreover, body mass loss and power margin gain in both species were higher than in passerines, as
observed in mallards. Our results suggest that the starvation-predation risk trade-off is one of the major life history traits
underlying body mass adjustments, and these findings can be generalized to all birds facing predation. Additionally, the
response magnitude seems to be influenced by the strategy of body reserve management.
Citation: Zimmer C, Boos M, Poulin N, Gosler A, Petit O, et al. (2011) Evidence of the Trade-Off between Starvation and Predation Risks in Ducks. PLoS ONE 6(7):
e22352. doi:10.1371/journal.pone.0022352
Editor: Sean A. Rands, University of Bristol, United Kingdom
Received December 10, 2010; Accepted June 23, 2011; Published July 18, 2011
Copyright: ? 2011 Zimmer et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by grants from ‘‘Re ´gion Alsace’’, National Fund for Biological Research on Game and Wildlife Species and CNRS. The funders
had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: cedric.zimmer@c-strasbourg.fr
Introduction
The competition between two or more processes for the
allocation of limited resources generally results in a trade-off that
underlies different life-history traits [1]. One important trade-off
occurs for species acquiring food while avoiding predation [2,3].
Animals have to build up or maintain body fuel reserves which are
an important buffer against starvation, especially when harsh
winter weather conditions involve unpredictable food availability
and energy requirements [4]. However, it is surprising to see that
birds often maintain their level of body reserves below the
maximum threshold [5]. Assuming that maintaining a high level of
body fuel, i.e. a high body mass, also incurs a significant cost in
terms of enhanced mortality risk due to predation vulnerability
[5], the amount of body reserves that a bird carries has generally
been viewed as a trade-off between the risk of starvation and the
risk of predation [6,7]. Body mass adjustment is considered to be
the consequence of this trade-off.
In this context, the mass-dependent predation risk theory
predicts that if the probability of an individual being caught by a
predator depends on its body mass, its weight should be
maintained at an appropriate level to balance predation risk
against the risk of starvation [6,8–10]. Such body mass adjustment
has the advantage of improving flight performance and reducing
the associated metabolic demands. It also results in a lower
investment in foraging time and less exposure to predation
[7,10,11]. A high body mass is correlated with high wing loading
and a greater cost of flight. These two factors could impair flight
performance, particularly during take-off, due to a smaller angle of
ascent and a lower speed [10,12–18]. Conversely, birds have to
maintain a level of body reserves which is high enough to limit the
risk of starvation [6,10]. It has generally been assumed that this
strategy would lead animals to carry greater body reserves when
starvation risk is high and vice versa [6,13]. Nevertheless,
empirical data on the starvation-predation risk trade-off that
illustrates a decrease in body mass when individuals are under
higher predation risks mainly originate from studies on small
passerine birds [2,19–26]. Furthermore, experimental studies have
demonstrated that when predation risk was increased or when
predator attacks were simulated by chasing the birds, food
consumption decreased in order to adjust body mass [19,21,23].
This body mass adjustment improves take-off performance
because the available power for flying increases when body mass
declines [27] and this ultimately maximizes survival.
To our knowledge, apart from the afore-mentioned studies of
passerines the only other species studied in relation to the
starvation-predation risk trade-off are the redshank (Tringa totanus)
[28], a larger species, the mallard (Anas platyrhynchos) [29] and one
non-bird species, the harbour porpoise (Phocoena phocoena) [2]. In
the two last species, it has been shown that body mass or body
PLoS ONE | www.plosone.org 1July 2011 | Volume 6 | Issue 7 | e22352

Page 2

reserves were linked to predation risk. In mallards, the relative
body mass decrease was twice as high as in passerines [29] and it
was hypothesized that this was due to a difference in body mass
and the amount of body reserves of each species: whereas
passerines build up body reserves during the day and use them
during the following night for energetic purposes [4,30] mallards
store more body reserves than required immediately in order to
cope with possible future periods of cold spells [31–33].
Furthermore, large birds have higher body reserves and a lower
metabolism per unit body mass than small species which have a
higher surface/volume ratio [34]. Thus, large birds can sustain
greater body mass variations than small ones, even in proportion
to lean mass, without dramatically increasing their starvation risk.
Moreover, this is consistent with the idea that greater body mass
loss allows a greater power margin gain in large birds than in
passerines. The power margin is defined as the ratio between
power available and power required for flight. It therefore appears
that the magnitude of the response to increased predation risk
depends on species size, with a higher body mass loss in large birds
than in small ones due to the difference in the amount of body
reserves stored [29].
The present study was carried out on one small and one
medium-sized duck species, the common teal (Anas crecca) and the
tufted duck (Aythya fuligula) respectively. Predation risk was
artificially increased in order to confirm that the starvation-
predation risk trade-off applied in large birds. These two species
were chosen as their exposure to predation is similar to that of
other duck species sharing the same habitat (see [29]). Moreover,
teals and mallards have similar body reserve dynamics throughout
their biological cycle [35], and while diving tufted ducks show a
similar body weight variation to dabbling ducks (i.e. mallards and
teals) during winter [36], in terms of size they are intermediate
between teals and mallards. They also differ from passerines in
both size and body fuel storage strategy, with mass variations that
are low on a daily basis but are seasonally high [4,30,35,36]. We
predicted that the extent of body mass loss in these two duck
species should be greater than in passerines: although they need to
improve their escape performance, the response should be
approximately the same as that observed in the mallard, because
the three species have the same relative amount of body reserves
[37,38].
Methods
Ethics Statement
This work was performed with governmental authorizations
delivered by the Pre ´fecture du Bas-Rhin (Strasbourg, France) to
conduct experiments on ducks numbers 67-99 and 67–285, and
was approved by the Direction De ´partementale des Services
Ve ´te ´rinaires du Bas-Rhin (Strasbourg, France). The experiment
complied with the ‘‘Principles of Animal Care’’ publication
No. 86–23, revised 1985 of the National Institute of Health, and
with current legislation (L87–848) on animal experimentation in
France. After the study, ducks were released in the field under the
control of the ‘‘Office National de la Chasse et de la Faune
Sauvage’’ and with the authorization of the ‘‘Direction De ´parte-
mentale de l’Agriculture et de la Fore ˆt du Bas-Rhin’’.
Animals and experimental conditions
The study was conducted on 42 common teals (21 females and
21 males) from the Fauna Leroy rearing centre (Westvleteren/
Belgium) and 28 tufted ducks (14 females and 14 males) from the
‘‘Les Canards de Mormal’’ rearing centre (Jolimetz/France).
Groups of 14 individuals (7 males and 7 females) were constituted
in both species: three groups in teals and two groups in tufted
ducks. Only two groups could be studied in the latter species. This
was due to a limited supply of individuals, which was insufficient
for three groups to be created within the same season. Each group
of 14 individuals was maintained in an outdoor tunnel aviary of
100 m2(206562.5 m). Each aviary contained a 4 m2pool
(0.60 m depth) containing clear running water which was
positioned at the same location in each tunnel. Birds were
subjected to natural photoperiod and ambient temperature. A
species-specific balanced commercial diet (Standard duck food
7751, Sanders Corporation; Teurlings premium duck food) was
provided ad libitum. The food was provided in feeders placed on
262 m tarpaulins to account for food spillage. The aviaries were
located close to the laboratory, and were protected against
predators within an electric enclosure and visually separated by
opaque barriers. A two-month period of acclimation to the aviaries
was applied for both species (September-October 2007 for teal,
September-October 2008 for tufted ducks).
Experimental procedure
Disturbance.
and one group of tufted ducks were disturbed over a one-week
period. These disturbances were carried out three times at
intervals of approximately 1.5 months (Table 1). Birds in group
1 (G1), teals only, were disturbed twice daily for 15 minutes
between 08:00 and 11:00. In both species, birds in the group 2
(G2) were disturbed four times daily for 15 minutes during the
same period of time. In each species, birds in the control group
(CG) were not disturbed. During disturbance sessions, each aviary
was monitored throughout the night via a night-view camera to
ensure that the ducks were not disturbed by any other external
factors.
The disturbance was created by steering a radio-controlled car
(E-Zilla FWD Hot-boddiesTM) towards the ducks at high speed
until they took off. This was the most efficient way to induce
simultaneous take-off flights for all birds in the group i.e. all birds
During the winter period, two groups of teal
Table 1. Date (mm/dd/yy) of the beginning and the end of the three disturbance sessions for each group in teals and tufted
ducks.
session 1session 2session 3
beginning endbeginning endbeginning end
teals G111/14/0711/20/07 01/04/0801/10/0802/13/0802/19/08
G211/21/07 11/27/0701/11/0801/17/0802/20/0802/26/08
tufted ducksG211/26/0812/02/0801/10/0901/16/0902/18/0902/24/09
doi:10.1371/journal.pone.0022352.t001
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org2July 2011 | Volume 6 | Issue 7 | e22352

Page 3

flew from one end of the tunnel aviary to the other, hence leading
to a response similar to that induced by a real predator.
Furthermore, ducks had no previous experience of this type of
stressor, therefore precluding any previous learning mechanism
[29,39]. No ducks were hurt by the car during these experiments.
During disturbance phases, two experimenters (C.Z, M.B.) were
near the aviaries to control the radio-controlled car and to record
the number of individuals taking off.
Weighing and wing loading.
disturbed groups were caught with a net and were weighed (61 g)
in a nearby room the day before the beginning of each
disturbance, on the fourth day and on the last day, immediately
after the end of the last disturbance. Control birds were also
caught and weighed at the same frequency as disturbed groups.
The birds of each group were released together in their respective
aviaries after weighing.
The wing area of the birds was determined from the outline of
the stretched two wings, drawn onto paper. After the images were
digitised, the wing area of each wing was measured using Sigma-
Scan software (version 5.0). Total wing area was obtained by
adding together the wing areas of the left and right wing of each
individual. Wing loading (g.cm-2) was determined by dividing body
mass by total wing area.
In the two species, ducks in
Power Margin
Power margin was calculated from the power available (Pa) and
the power requirement (Pr) for flight in all disturbed individuals
before and after disturbance sessions. Equations for calculating Pa
and Pr were derived from Norberg [40]. The equations are
Pa=21.946m(2/3)and Pr=6.3336m(7/6), where m denotes the
body mass in kg and the power is in watts. The power margin (PM)
is the ratio of the power available divided by the power required
for flight (Pa/Pr). The PM gain is the difference between the PM
values recorded before and after the disturbance.
Food intake
Daily food intake determination began one week before each
disturbance session and ended one week after its completion. Each
day at 18:00 the food remaining from the preceding 24 h was
removed and food spilled on the tarpaulin was collected. One kg of
standard duck fresh food (teal) and 0.8 kg of standard duck fresh
food plus 0.8 kg of premium duck food (tufted duck) was then given
to birds. The food given to the birds and removed from feeders
was dried in an oven for 24 h at +40uC before being weighed to
avoid errors due to changes in water content.
Statistical analysis
Two-way repeated measures ANOVAs were used to test for
differences in the number of individual daily flights between
sessions and groups in teal, and also between sessions and sexes in
tufted duck. We used general linear mixed models (GLMM) to
examine the effects of disturbance on body mass and wing loading
Table 2. Statistics values and P values for all simple fixed
effects of each model and for useful interactions between
fixed effects and multiple comparisons in teals.
ParameterTest valueP value
flight number sessionF2,41=4.3 0.04
groupF1,41=163.7
,0.001
session*group F2,41=13.8
,0.001
G2 session1 vs G2 session2t12=20.50.99
G2 session1 vs G2 session3t12=24.30.005
G2 session2 vs G2 session3t12=23.80.01
body masssessionF2,46.5=74.5
,0.0001
groupF2,23.8=4.00.033
sex F1,35.7=0.44 0.51
stateF1,34.5=538.5 ,0.0001
group*stateF2,34.5=41.2
,0.0001
session*group*stateF4,51.8=25.6
,0.0001
CG initial vs CG finalt34.4=6.4
,0.0001
G1 initial vs G1 finalt34.4=214.9 ,0.0001
G2 initial vs G2 finalt34.9=18.9
,0.0001
wing loadingsessionF2,45.7=55.7
,0.0001
groupF2,26.1=4.1 0.028
sexF1,39.4=0.3 0.56
stateF1,33.1=459.8 ,0.0001
session*group*stateF4,51.4=20.4
,0.0001
CG initial vs G1 initialt30.6=1.30.81
CG initial vs G2 initialt31.1=1.2 0.84
G1 initial vs G2 initialt31.9=0.0 1
CG final vs G1 finalt24.2=3.40.02
CG final vs G2 finalt25.1=4.30.002
G1 final vs G2 finalt25.3=21.1 0.89
G1 final session1 vs G2 final
session 1
t31.4=0.91
G1 final session2 vs G2 final
session 2
t34.5=21.00.99
G1 final session3 vs G2 final
session 3
t25.2=21.2 0.99
body mass loss sessionF2,34.8=89.2
,0.0001
groupF2,19.9=57.7
,0.0001
sexF1,23.8=0.90.36
session*groupF4,40.9=22.7
,0.0001
session1 vs session2t35.4=26.8
,0.0001
session1 vs session3t35.4=212.9 ,0.0001
session2 vs session3t36=28.0
,0.0001
CG vs G1t17.6=8.6
,0.0001
CG vs G2t17.5=8.5
,0.0001
G1 vs G2t19=2.60.049
power margin
gain
sessionF2,22.9=66.7
,0.0001
groupF1,20.5=7.10.015
sexF1,20.5=0.00.99
session*groupF2,22.9=3.63 0.043
session1 vs session2t23.4=9.7
,0.0001
session1 vs session3t23.6=11.7
,0.0001
session2 vs session3t24=4.70.0003
ParameterTest valueP value
G1 session1 vs G2 session1 t24=23.3 0.04
G1 session2 vs G2 session2t23.1=22.10.35
G1 session3 vs G2 session3t21.5=20.9 0.95
doi:10.1371/journal.pone.0022352.t002
Table 2. Cont.
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org3 July 2011 | Volume 6 | Issue 7 | e22352

Page 4

changes. Session, group, sex and state (before disturbance and
after disturbance) were included as fixed factors. Session and state
were also specified as repeated factors. Individuals were defined as
a random factor to account for inter-individual variability. To
examine the effect of disturbance on body mass loss, a GLMM was
performed with session, group and sex as fixed factors. Session was
specified as a repeated factor and individuals as random factor.
For each disturbed group, we used a GLMM to test the difference
in daily body mass loss between the first four days and the last
three days of disturbance with session, sex and state (daily body
mass loss during the first fourth days and daily body mass loss
during the last third days). Session and state were used as repeated
factors and individuals as a random factor. In teals, a GLMM was
fitted to compare PM gain between the two disturbed groups.
Session, group and sex were specified as fixed factors, session was
also stated as a repeated factor and individuals as a random factor.
In tufted ducks, the same model without the factor group was run
to study PM gain in the disturbed group. For each GLMM,
Tukey-Kramer multiple comparison adjustment was applied to
obtain corrected p-values. Proc Mixed in SAS 9.1.3 (SAS Institute
Corporation) was used to fit GLMM.
Daily food intake measurements have been investigated as time
series. For each group and each session the autocorrelation
functions (ACF) were calculated for all lags between 210 and 10.
These functions give the correlation between the signal and itself
with an increasing lag (i.e. correlation between the response at
time t and the response at time t +l lag). This allowed us to
investigate periodicity in the time series. In order to assess if a
series is drawn at random, we used the Ljung-Box test. This is a
portemanteau test which null hypothesis is ‘‘data are random’’ (the
name portemanteau test refers to a test that is made for each lag).
The tests were performed for each group and each session at each
lag. Time series analyses were conducting using R 2.9.2.
Probability levels ,0.05 were considered as significant. Mean
values are reported 6 S.E.
Results
In both species, escape flights, body mass, wing loading and
body mass loss did not differ between sexes (Table 2, 3).
Teal
On average, G2 birds performed 1.5-fold more escape flights
than those of G1 (Figure 1a). Individual daily flights did not differ
between sessions in G1birds (P.0.07) but did in those from G2,
with more flights during session 3 (Table 2, Figure 1a).
Body mass, wing loading and body mass loss differed over time
between birds in the different groups (Table 2).
During disturbance, body mass loss was significant in birds in all
three groups and differed significantly between the three sessions,
with a greater loss in the first session (Table 2, Figure 1a). Birds in
both treatment groups showed body mass loss at least two times
higher than that seen in control birds (Table 4), and body mass loss
was higher in G2 than in G1 (Table 2). Birds in disturbed groups
showed a mean daily body mass loss that was not linear over the
sessions, it was higher during the first four days of disturbance
(G1=28.0460.62 g.day21, G2=210.3860.83 g.day21) than
duringthethree last days
G2=23.5560.74 g.day21) (Table 2).
Although initial wing loading did not differ between groups,
final wing loading was lower in disturbed birds (G1 and G2) than
control birds at the end of the disturbance sessions (Table 2, 4). For
all sessions, final wing loading was similar in both disturbed groups
(Table 2, 4).
(G1=22.7760.45 g.day21,
Body mass loss was not related to final wing loading (R2=0.09,
F2,11=1.54, P=0.27), but the greater the initial wing loading, the
greater the body mass loss was seen to be (R2=0.88, F2,11=37.36,
P,0.0001, Figure 2a).
Power margin gain was higher during the first session (0.7660.03)
then decreased over the following sessions (session 2: 0.4360.03 and
session3:0.3160.03).Powermargingain wasgreaterinG2thanG1
birds, although the difference was only significant for the first session
(G1: 0.6560.04; G2: 0.8760.05; Table 2).
Table 3. Statistics values and P values for all simple fixed
effects of each model and for useful interactions between
fixed effects and multiple comparisons in tufted ducks.
Parameter Test valueP value
flight number sessionF2,41=118.1
,0.001
sexF1,41=3.8 0.10
session*sexF2,41=2.10.17
session1 vs session2t23=211.5
,0.0001
session1 vs session3t23=214.6
,0.0001
session2 vs session3t23=23.10.006
body masssessionF2,28.5=47.2
,0.0001
groupF1,26.5=28.2
,0.0001
sexF1,26.5=5.3 0.03
stateF1,23.3=136.3 ,0.0001
session*group*stateF2,30.2=30.20.0009
wing loadingsessionF2,27.5=47.0
,0.0001
groupF1,28.5=18.30.0002
sexF1,28.5=0.3 0.60
stateF1,22.1=115.8 ,0.0001
session*group*stateF2,28.6=8.90.001
GC initial vs G2 initialt40.5=1.3 0.55
GC final vs G2 finalt27.3=6.8
,0.0001
G2 final session1 vs G2 final session2 t31.1=1.20.98
G2 final session1 vs G2 final session3 t39.4=2.6 0.30
G2 final session2 vs G2 final session3 t29.8=1.7 0.86
body mass
loss
sessionF2,23=2.4 0.11
groupF1,15.7=41.7
,0.0001
sex 1.341.70.28
session*groupF2,23=12.00.0003
CG session1 vs CG session2t24=1.8 0.51
CG session1 vs CG session3t24=2.00.37
CG session2 vs CG session3t24=20.40.99
G2 session1 vs G2 session2t24=23.00.07
G2 session1 vs G2 session3t24=25.10.0005
G2 session2 vs G2 session3t24=20.90.94
power margin
gain
sessionF2,11=4.00.048
sexF1,12=0.010.92
session*sexF2,11=1.50.26
session1 vs session2t12=2.00.16
session1 vs session3t12=3.00.03
session2 vs session3t12=0.30.97
doi:10.1371/journal.pone.0022352.t003
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org4July 2011 | Volume 6 | Issue 7 | e22352

Page 5

The food intake of control birds was random, that is to say not
dependent on previous food consumption (Figure 3). The food
intake for birds in the disturbed groups was dependent on their
food consumption over the 2 previous days. These results were
confirmed by Ljung-box test (see figure S1). When comparing the
food intake of a particular day with that recorded the same day of
the previous or following week, correlations were stronger in G2
than in G1 birds (Figure 3). All correlations were negative,
meaning that when the food intake was high one day it would be
low 7 days later, and vice versa. This means that food intake
decreased during the disturbance and returned to normal values
following the disturbance; this decrease was greater in G2 than in
G1 birds (Figure 4a).
Tufted duck
The number of daily escape flights in the disturbed birds
increased throughout the three sessions (Table 3, Figure 1b). Body
mass, wing loading changes and body mass loss were different over
time between disturbed and control birds (Table 3).
Body mass significantly decreased during the disturbance
session in both groups, but this decrease was three times greater
in disturbed birds than in control birds (Table 3, 4). Body mass loss
was greater in the first than in the third session in disturbed birds,
whereas it did not differ between sessions in control birds (Table 3,
Figure 1b). Mean daily body mass loss in G2 birds was greater
during the first four days of disturbance (225.4861.24 g.day21)
than during the last three days (2.0661.74 g.day21) (Table 3).
Figure 1. Escape flights and body mass variations. Mean number (6 SE) of daily escape flights for the three sessions in disturbed groups and
average body mass variations (g) for the three disturbance sessions in control and in disturbed groups in teals (a) and tufted ducks (b). In each group,
black bars correspond to the first session, white bars to the second session and grey bars to the third session. Letters indicate significant difference
between groups. * indicate differences between sessions in the different groups.
doi:10.1371/journal.pone.0022352.g001
Table 4. Mean (6SE) body mass loss (g), initial and final wing loading (g.cm22) for each group and mean (6SE) power margin gain
and relative power margin gain for the disturbed groups in the two species for all sessions.
body mass loss (g) initial wing loading (g.cm22)final wing loading (g.cm22) power margin gainpower margin gain (%)
G145.062.40.9060.02 0.7860.020.4360.03 7.360.4
tealsG257.364.10.9060.02 0.7560.02 0.5760.049.660.7
CG19.361.7 0.9460.03 0.8960.03//
tufted
ducks
G2117.468.81.4360.021.1960.010.3960.039.660.7
CG 36.063.71.4860.01 1.4160.01//
doi:10.1371/journal.pone.0022352.t004
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org5July 2011 | Volume 6 | Issue 7 | e22352

Page 6

Initial wing loading did not differ between groups (Table 3, 4).
Final wing loading at the end of disturbance in G2 birds did not
differ between sessions and was significantly lower than in control
birds (Table 3, 4).
Although, there was no relationship between final wing loading
and body mass loss (R2=0.75, F2,5=8.75, P=0.06), we noted that
the greater the initial wing loading was, the greater the body mass
loss was seen to be (R2=0.93, F2,5=34.80, P=0.008, Figure 2b).
Disturbance led to a mean power margin gain of 9.660.7 % in
G2. The power margin gain differed between sessions, values
being higher in the first (0.4760.04) and lower in the third
(0.3560.04) (Table 3).
Food intake depended on what had been eaten the previous day
(Figure 5). In the disturbed group, food intake was dependent on
that of the 2 previous days. These results were confirmed by
Ljung-box test (see figure S2). The comparison of the food intake
of a specific day with that recorded the same day of the previous or
following week revealed negative correlations, indicating that if the
food intake was high one day, it would be low 7 days later and vice
versa. This means that food intake decreased during the
disturbance and returned to its normal value after the disturbance
ended (Figure 4b).
Discussion
We show in this experimental study that teal and tufted ducks
responded to an increased predation risk by reducing food
consumption and body mass. This strategy is consistent with an
improved capacity to escape, achieved by the reduction of wing
loading. This study also confirms that the magnitude of the
response to the starvation-predation risk trade-off depends on
body reserve management strategy.
In both species, the relative body mass decrease was at least
twice as high in the disturbed groups as in the control one. We
suggest that this loss probably resulted from mass-dependent costs
associated with disturbance. This result is in accordance with
findings obtained in mallards and in passerines exposed to an
increased predation risk [15,16,20–25,29]. Such body mass
decrease should be advantageous since it reduces energetic
maintenance costs, foraging time and wing loading. In natural
conditions, reducing the impact of these factors would allow
reduced exposure to predators and enhanced escape capabilities
[8,9,11,18]. Our results therefore support the starvation-predation
risk trade-off theory which predicts that an increase in predation
pressure leads to a decrease in body mass [6,9,10]. Similarly,
Nebel and Ydenberg [41] have shown that wing loading
differences of non-breeding waterbirds are linked to changing
predation risk.
Teal and tufted ducks in control groups lost a significant amount
of weight, probably as a direct consequence of handling stress
[3,23,42,43]. Moreover, body mass loss was also higher in
disturbed groups of both species during the first session than
during the following ones. This could be related to a stress
response associated with the novelty of the disturbance situation.
During the first session, stress response associated with the mass-
dependent costs of the disturbance may result in a greater body
mass loss. However, this is unlikely to be the correct explanation,
since corticosterone levels were the same in both disturbed and
control groups, and were not seen to be higher during the first
session in either species [Zimmer unpublished data]. Another
possibility is that both species started the first session with relatively
high fat stores, meaning that mass-dependent costs of disturbance
flights were greater during this first session.
In tufted ducks, body mass loss was at its highest during the first
session, during which the lowest number of daily escape flights was
also recorded. Interestingly, the converse was observed during the
third session (Figure 1b). A similar result was found in teal from the
G2 disturbed group (Figure 1a). Escape flights require at least
three times more energy expenditure than sustained flapping
flights [44,45]. Thus, if body mass loss is directly related to an
increase in energy expenditure in response to the number of
escape flights, one would expect a greater decrease in body mass to
result from a larger number of flights. Our results do not support
this prediction, therefore underlining the fact that the extent of
body mass loss in disturbed groups for these two species is not
consistent with an increase in the energy expenditure associated
with escape flights.
Figure 2. Body mass loss and initial wing loading. Relationship between body mass loss (g) and initial wing loading (g.cm22) for the three
disturbance sessions in disturbed groups in teals (a) and tufted ducks (b) (values plotted are means). In teals, individuals in G1 and G2 are indicated by
open and closed circles, respectively. The relationship is best described by y=85.31612.86+21.92, R2=0.88, F2,11=37.36, p,0.0001 in teals and by
y=58.6862.77241.65, R2=0.93, F2,5=34.73, p=0.008 in tufted ducks. The results indicate that body mass loss was higher when the initial wing
loading was high.
doi:10.1371/journal.pone.0022352.g002
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org6 July 2011 | Volume 6 | Issue 7 | e22352

Page 7

Although food was provided ad libitum, food intake decreased
during the week of disturbance in the treatment groups of both
species (Figure 4). The food intake of disturbed groups was not
random, but depended at least on food consumption of the
previous day. Therefore, teals and tufted ducks did not increase
their energy intake to compensate for the increased energy
expenditure incurred by take-off flights. Both species are mainly
nocturnal foragers whose diurnal foraging represents only
approximately 10% of their total daily time budget [46–49].
Since the maximum disturbance in the study lasted for a
maximum of 1 hour, (i.e. less than 12% of the daytime) we
assume that ducks were not constrained by their available foraging
time. The decrease in food intake does not fit the interrupted
foraging theory, which predicts that birds should gain weight as a
compensation mechanism in response to a reduced probability of
feeding when predation risk rises [50,51]. In teals, the dynamic of
food consumption was the same in both disturbed groups despite
pronounced body mass loss in G2. Consequently, the decrease in
food intake should result in decreased body mass in order to gain
efficiency in escape flights [19,23,29]. Improvement of flight
capabilities could be achieved by adjusting the mass of different
organ groups such as digestive compartments, pectoral muscle or
fuel reserves [9,10,52,53]. Nevertheless, it is unlikely that the
weight loss in teals and tufted ducks could be explained solely by a
reduction in the digestive compartment, since the birds never
fasted. However, it would be interesting to check whether
variation of pectoral muscle size could be decoupled from body
mass variations [53]. As suggested by Van den Hout and
colleagues [54] for other waterbirds, variation in pectoral muscle
size could be due to the predator-escape tactic and should be
smaller in gregarious species than in solitary-living species.
Solitary-living species are particularly vulnerable to surprise
attacks, and in this case rely on a speed-based escape that is
made easier by increasing pectoral muscle size. In contrast,
gregarious species can detect predators earlier and prepare
themselves for an escape response facilitated by a decrease in
Figure 3. Autocorrelation functions relative to food intake in teals. Data are for each session, in the three groups. Dashed-lines represent the
limit of the significant differences from 0 for the correlations. For example, for positive correlations, each correlation located above the dashed line is
significantly different from 0 whereas each correlation located below the dashed-line is not significantly different from 0.
doi:10.1371/journal.pone.0022352.g003
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org7 July 2011 | Volume 6 | Issue 7 | e22352

Page 8

body mass [54]. The gregarious behaviour of both species in this
study [55] leads us to assume that muscle size is likely to remain
relatively stable. Therefore, it seems that the body mass loss
observed here was rather the result of a decrease in body reserves,
but this can only be asserted after detailed studies on body
composition variation during disturbance.
Flight mechanics theory and experimental studies show that a
high body mass and associated high wing loading can impair
flight capabilities during predator attacks by decreasing the speed
and the angle at take-off and aerial manoeuvrability [10,12–
15,17,56,57]. Indeed, wing loading is a major issue for flight
performance and is mainly negatively related to flight speed [40].
Interestingly, we showed that body mass loss was positively linked
to initial wing loading and not to final wing loading (Figure 2). In
teals, although body mass loss differed between disturbed groups
and sessions, final wing loadings never differed, and were lower
than in control group birds. In tufted ducks, similar results have
been obtained. It therefore appears that ducks under disturbance
attempt to reach a target wing loading by adjusting body mass
through the control of food intake. Furthermore, final wing
loading did not differ among groups and sessions despite the
different number of take-off flights. Thus, these results do not
support the idea that physical training through repeated flights
would have a major impact on flight performance, but rather
suggest that wing loading at the beginning of an increased risk
period is the main factor driving body mass regulation. This
conclusion is also supported by the fact that body mass loss
reached its highest values during the first part (1–4 days) of each
disturbance session. Overall, disturbed ducks increased their
power margin by 7 to 10 % during disturbance sessions, resulting
in a higher availability of energy for flying. Birds hence have
better manoeuvrability and can climb more easily [27], which
probably also explains why they achieved more take-off flights
throughout the disturbance session as body mass decreased.
Overall, these results concord with those obtained for the mallard
[29] and support the argument for an optimal adjustment of body
mass and flight performance among different duck species in
response to an increase in predation risk. It seems that these are
strategic adjustments rather than an environmentally-induced
response, since we observed the same adjustments in three
different duck species with different body size and ecology.
Environmental conditions, and particularly very harsh winter
conditions, may affect response to predation and starvation risks
and thus have an impact on energy reserve levels and body
masses [58]. The response was nevertheless similar in all three
species, particularly for the final wing loading in a given species
during different periods in winter. Moreover, none of the three
species of ducks encountered similar wintering ambient weather
conditions. Our results are therefore in accordance with the
starvation-predation risk trade-off theory [6,7], suggesting that
this is a general mechanism driving body mass and wing loading
changes under different predation risks (see also [41]). It should
be noted that results obtained when stressing a group of birds
may differ from those observed in individually stressed birds.
Metcalfe and Ure [17] pooled data of a group of alarmed zebra
finches and showed that daily body mass change influences flight
performances. Yet it has been shown in the same species that
mass has little or no effect on flight velocity within the natural
body mass range when birds were individually stressed [59].
Nevertheless, despite the fact that we stressed groups of birds in
our study, all individuals were subjected to the same level of
disturbance since ducks responded to the danger represented by
the rapid approach of the car and it seems that ducks did not
react to each other since some ducks can take-off before the rest
of the group while some others can delay their take-off.
Moreover, all disturbed ducks decreased their body mass to
reach approximately equivalent wing loading in response to the
disturbance. Finally, as ducks generally live in flocks in the wild,
disturbing a flock of ducks closely recreates their behaviour in the
wild.
According to the results obtained in this study, the overall
response to the aforementioned trade-off in a representative set of
Anatidae species (tufted ducks and common teals, this study;
mallards, [29]) is broadly similar in all three species in so far that it
relates to body mass and food intake regulation. However, the
study of large birds with high body reserves indicated a greater
response to an increased predation risk compared to small birds
with low body reserves, indicated by a body mass loss
approximately two times higher in the former (6-16%) than in
the latter (2-5%, [19,22,23]). In such large animals, the power
Figure 4. Mean food intake compared to average intake on the first day of disturbance. Panel a: teals, panel b: tufted ducks. Data were
the mean value 6 SE for the three sessions in control group (grey circles), group 2 (black circles) and group 1 (white circles).1 corresponded to the
first day of disturbance.
doi:10.1371/journal.pone.0022352.g004
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org8 July 2011 | Volume 6 | Issue 7 | e22352

Page 9

margin is relatively low because the power available for flight
increases with body mass at a slower rate than the required power
[27]. As a result, high body mass loss leads to significant gains in
power margin, which should be a key element for increased flight
manoeuvrability during predator attacks in these ducks species.
However, in birds laying down significant amounts of body
reserves, predation risk is minimized without dramatically
increasing starvation risk. Actually, the body mass at the end of
each disturbance session was on average 50% higher than the
values recorded in lean or depleted teals and tufted ducks ([60],
Boos unpublished data), thus leading us to the conclusion that
predation risks could be of greater importance in the regulation of
body mass than the starvation risk. In contrast, in small birds like
passerines, the achievement of the trade-off between both risks is
quite different and, as shown in several studies, body mass loss in
response to enhanced predation risk is more limited [19–25]. In
fact as passerines store limited amounts of body reserves that only
allow short-term resistance to fasting [4,30] the risk of starvation
is higher than predation risk. This phenomenon is amplified by
their high specific metabolism, due to their small size and to the
high proportion of lean mass which is energetically more costly to
maintain than fat mass [34]. Furthermore, in passerines the
power margin and consequently the manoeuvrability are very
high as compared with larger ducks, and body mass reductions
are of minor impact on these parameters (see [59,61–63]). In
contrast, it has been shown that in small migrant birds that build
up relatively large fuel reserves before migration, this increase in
body reserves (up to 67% of lean body mass) results in a reduction
in the take-off angle and/or velocity, leading to a higher
predation risk [12,14,64,65]. During this period in small birds
there is an increase in predation risk and a decrease in that of
starvation. This situation could be comparable to that seen in
birds with high body reserves. This study also shows that the
response to the starvation-predation risk trade-off is broadly the
same in a small and a medium size duck species as in a large one
that manages body reserves in the same way. Therefore, we
suggest that the relative importance of starvation and predation
and the response to the trade-off between these risks are not
Figure 5. Autocorrelation functions relative to food intake in tufted ducks. Data are for each session, in both groups. Dashed-lines
represent the limit of the significantly differences from 0 for the correlations. For example, for positive correlations, each correlation located above
the dashed line is significantly different from 0 whereas each correlation located below the dashed line is not significantly different from 0.
doi:10.1371/journal.pone.0022352.g005
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org9 July 2011 | Volume 6 | Issue 7 | e22352

Page 10

directly related to the size of the species as proposed by Zimmer
et al. [29] but are rather dependent on the strategy of body
reserve management (Figure 6).
To conclude, this study revealed that an increase in predation
risk leads to a strategic body mass loss in the teal and the tufted
duck, most probably in order to reach a more favourable wing
loading. This result echoes previous results obtained in the mallard
[29] and with those observed in passerines [19,21,23]. Thus, the
starvation-predation risk trade-off seems to be one of the major life
history traits underlying optimal body mass adjustment in a variety
of duck species, and we expect that these findings can be
generalized to all birds facing disturbance or predation events.
However, the relationship between the power margin gain and the
decrease in body mass differed between duck species and
passerines, probably because of different strategies in body reserve
management. To confirm these results it will be necessary to
specifically measure the impact of wing loading adjustment on
flight performance and body fuel amount. Overall, body mass
decline under disturbance is not implicitly deleterious when birds,
depending on their body fuel reserves and metabolism changes,
escape more efficiently and thus avoid immediate mortality by
being caught. Finally, as shown in shorebirds [41], the benefits of
wing loading adjustments and associated behaviour changes
should be better addressed when dealing with the impact of
disturbance on the fitness of birds, especially among waterfowl
species.
Supporting Information
Figure S1
to access if the data are random, we performed a Ljung-Box test
Ljung-Box test for food intake in teal. In order
that is a portemanteau test whose null hypothesis is ‘‘data are
random’’. (The name portemanteau test refers to a test that is made
for each lag.) The tests were performed for each group of teal and
each session at each lag. The p-values are drawn in the graphic to
have a better sight on the results. In the control group, for each
session and almost all lags, the null hypothesis can not be rejected so
that we can consider that this data are drawn at random. The
graphics for the group 2 are clearly showing really low p-value so
that the data are not drawn at random so there is a pattern in the
data. It is almost the same as regards group 1 with an exception for
session 3 where most of the p-values are larger than 0.05. However,
this has to be seen in parallel of the ACF for this session. As a matter
of fact none of the correlations are significant but the shape of the
ACF is typical of periodic (so not random) process.
(TIF)
Figure S2
In order to access if the data are random, we performed a Ljung-
Box test. The tests were performed for each group of tufted ducks
and each session at each lag. The p-values are drawn in the
graphic to have a better sight on the results. In the control group at
session 1, for all lags, the null hypothesis can not be rejected so that
we can consider that this data are drawn at random. At session 2,
only the p-value related to lag 1 is lower than 0.05. Thus, except
for the food intake of the next day the data are random. At session
3, p-values related to lag 1 to 3 and from 10 to 14 are lower than
0.05. Indeed, even if the data are not totally random, there is still a
lot of randomness in the drawing of the data. The graphics for the
disturbed group are clearly showing really low p-value so that the
data are not drawn at random so there is a pattern in the data.
(TIF)
Ljung-Box test for food intake in tufted duck.
Figure 6. Hypothetical responses to an increase of predation risk. The diagram considers different types of bird species according to their
theoretical amount of body reserves.
doi:10.1371/journal.pone.0022352.g006
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org 10July 2011 | Volume 6 | Issue 7 | e22352

Page 11

Acknowledgments
We sincerely thank G. Herzberg for his comments on the manuscript and
language editing, and Joanna Munro for language editing. We also thank
Anders Brodin and one other anonymous reviewer, for their helpful
comments on the manuscript.
Author Contributions
Conceived and designed the experiments: CZ MB OP JPR. Performed the
experiments: CZ MB OP JPR. Analyzed the data: CZ MB NP AG OP
JPR. Contributed reagents/materials/analysis tools: CZ MB NP AG OP
JPR. Wrote the paper: CZ MB NP AG OP JPR.
References
1. Stearns SC (1992) The evolution of life histories. New York: Oxford University
Press. 249 p.
2. MacLeod R, MacLeod CD, Learmonth JA, Jepson PD, Reid RJ, et al. (2007)
Mass-dependent predation risk and lethal dolphin porpoise interactions.
Proc R Soc B 274: 2587–2593.
3. Van der Veen IT, Sivars LE (2000) Causes and consequences of mass loss upon
predator encounter: feeding interruption, stress or fit-for-flight? Funct Ecol 14:
638–644.
4. Blem CR (1990) Avian energy storage. Curr Ornithol 7: 59–113.
5. Blem CR (1976) Patterns of lipid storage and utilization in birds. Am Zool 16:
671–684.
6. Lima SL (1986) Predation risk and unpredictable feeding conditions:
determinants of body-mass in birds. Ecology 67: 377–385.
7. McNamara JM, Houston AI (1990) The value of fat reserves and the tradeoff
between starvation and predation. Acta Biotheor 38: 37–61.
8. Brodin A (2001) Mass-dependent predation and metabolic expenditure in
wintering birds: is there a trade-off between different forms of predation? Anim
Behav 62: 993–999.
9. Houston AI, McNamara JM, Hutchinson JMC (1993) General results
concerning the trade-off between gaining energy and avoiding predation. Philos
Trans R Soc Lond B Biol Sci 341: 375–397.
10. Witter MS, Cuthill IC (1993) The ecological costs of avian fat storage. Philos
Trans R Soc Lond B Biol Sci 340: 73–92.
11. Bednekoff PA, Houston AI (1994) Optimizing fat reserves over the entire winter:
a dynamic-model. Oikos 71: 408–415.
12. Burns JG, Ydenberg RC (2002) The effects of wing loading and gender on the
escape flights of least sandpipers (Calidris minutilla) and western sandpipers
(Calidris mauri). Behav Ecol Sociobiol 52: 128–136.
13. Hedenstrom A (1992) Flight performance in relation to fuel load in birds. J Theor
Biol 158: 535–537.
14. Kullberg C, Fransson T, Jakobsson S (1996) Impaired predator evasion in fat
blackcaps (Sylvia atricapilla). Proc R Soc B 263: 1671–1675.
15. Kullberg C, Jakobsson S, Kaby U, Lind J (2005) Impaired flight ability prior to
egg-laying: a cost of being a capital breeder. Funct Ecol 19: 98–101.
16. MacLeod R (2006) Why does diurnal mass change not appear to affect the flight
performance of alarmed birds? Anim Behav 71: 523–530.
17. Metcalfe NB, Ure SE (1995) Diurnal variation in flight performance and hence
potential predation risk in small birds. Proc R Soc B 261: 395–400.
18. Witter MS, Cuthill IC, Bonser RHC (1994) Experimental investigations of mass-
dependent predation risk in the European starling, Sturnus vulgaris. Anim Behav
48: 201–222.
19. Carrascal LM, Polo V (1999) Coal tits, Parus ater, lose weight in response to
chases by predators. Anim Behav 58: 281–285.
20. Cresswell W (1998) Diurnal and seasonal mass variation in blackbirds Turdus
merula: consequences for mass-dependent predation risk. J Anim Ecol 67: 78–90.
21. Gentle LK, Gosler AG (2001) Fat reserves and perceived predation risk in the
great tit, Parus major. Proc R Soc B 268: 487–491.
22. Gosler AG, Greenwood JJD, Perrins C (1995) Predation risk and the cost of
being fat. Nature 377: 621–623.
23. Lilliendahl K (1997) The effect of predator presence on body mass in captive
greenfinches. Anim Behav 53: 75–81.
24. MacLeod R, Barnett P, Clark JA, Cresswell W (2005) Body mass change
strategies in blackbirds Turdus merula: the starvation-predation risk trade-off.
J Anim Ecol 74: 292–302.
25. MacLeod R, Gosler AG, Cresswell W (2005) Diurnal mass gain strategies and
perceived predation risk in the great tit Parus major. J Anim Ecol 74: 956–964.
26. Rands SA, Cuthill IC (2001) Separating the effects of predation risk and
interrupted foraging uponmasschangesinthebluetitParuscaeruleus.ProcRSocB
268: 1783–1790.
27. Pennycuick CJ (2008) Modelling the flying bird. London: Academic Press.
28. Cresswell W, Whitfield DP (2008) How starvation risk in Redshanks Tringa
totanus results in predation mortality from Sparrowhawks Accipiter nisus. Ibis 150:
209–218.
29. Zimmer C, Boos M, Petit O, Robin JP (2010) Body mass variations in disturbed
mallards Anas platyrhynchos fit to the mass-dependent starvation-predation risk
trade-off. J Avian Biol 41: 637–644.
30. Haftorn S (1989) Seasonal and diurnal body-weight variations in titmice, based
on analyses of individual birds. Wilson Bull 101: 217–235.
31. Boos M, Zorn T, Delacour G, Robin JP (2007) Weather and body condition in
wintering Mallards Anas platyrhynchos. Bird Study 54: 154–159.
32. Loesch CR, Kaminski RM, Richardson DM (1992) Endogenous loss of body-
mass by mallards in winter. J Wildl Manag 56: 735–739.
33. Pattenden RK, Boag DA (1989) Effects of body-mass on courtship, pairing, and
reproduction in captive mallards. Can J Zool 67: 495–501.
34. Schmidt-Nielsen K (1990) Animal physiology: Adaptation and environment.
New York: Cambridge University Press. 602 p.
35. Fox AD, King R, Watkin J (1992) Seasonal-variation in weight, body
measurements and condition of free-living teal. Bird Study 39: 53–62.
36. Ryan RR (1972) Body weight and weight changes of wintering diving ducks.
J Wildl Manag 36: 759–765.
37. Baldassarre GA, Bolen EG (1986) Body weight and aspects of pairing
chronology of green-winged teal and northern pintails wintering on the southern
high-plains of Texas. Southwest Nat 31: 361–366.
38. Rave DP, Baldassarre GA (1991) Carcass mass and composition of green-winged
teal wintering in Louisiana and Texas. J Wildl Manag 55: 457–461.
39. Grissom N, Bhatnagar S (2009) Habituation to repeated stress: Get used to it.
Neurobiol Learn Mem 92: 215–224.
40. Norberg UM (1996) Energetics of Flight. In: Carey C, ed. Avian Energetics and
Nutritional Ecology. New York: Chapman & Hall. pp 199–249.
41. Nebel S, Ydenberg R (2005) Differential predator escape performance
contributes to a latitudinal sex ratio cline in a migratory shorebird. Behav Ecol
Sociobiol 59: 44–50.
42. Gosler AG (2001) The effects of trapping on the perception, and trade-off, of
risks in the Great Tit Parus major. Ardea 89: 75–84.
43. MacLeod R, Gosler AG (2006) Capture and mass change: perceived predation
risk or interrupted foraging? Anim Behav 71: 1081–1087.
44. Nudds RL, Bryant DM (2000) The energetic cost of short flights in birds. J Exp
Biol 203: 1561–1572.
45. Videler JJ (2005) Avian Flight. Oxford: Oxford University Press.
46. Marsden SJ (2000) Impact of disturbance on waterfowl wintering in a UK
dockland redevelopment area. Environ Manag 26: 207–213.
47. Nilsson L (1970) Food-seeking activity of south Swedish diving ducks in non-
breeding season. Oikos 21: 145–154.
48. Pedroli JC (1982) Activity and time budget of tufted ducks on Swiss lakes during
winter. Wildfowl 33: 105–112.
49. Tamisier A (1972) Rythmes nycthe ´me ´raux des sarcelles d’hiver pendant leur
hivernage en Camargue. Alauda 10: 107–135 and 235–256.
50. Lilliendahl K (1998) Yellowhammers get fatter in the presence of a predator.
Anim Behav 55: 1335–1340.
51. McNamara JM, Houston AI, Lima SL (1994) Foraging routines of small birds in
winter: a theoretical investigation. J Avian Biol 25: 287–302.
52. Piersma T, Lindstrom A (1997) Rapid reversible changes in organ size as a
component of adaptive behaviour. Trends Ecol Evol 12: 134–138.
53. van den Hout PJ, Piersma T, Dekinga A, Lubbe SK, Visser GH (2006) Ruddy
turnstones Arenaria interpres rapidly build pectoral muscle after raptor scares.
J Avian Biol 37: 425–430.
54. van den Hout PJ, Mathot KJ, Maas LRM, Piersma T (2010) Predator escape
tactics in birds: linking ecology and aerodynamics. Behav Ecol 21: 16–25.
55. Del Hoyo J, Elliot A, Sargatal J (1992) Handbooks of the Birds of the World.
Volume 1: Ostrich to Ducks. Barcelona: Lynx Edicions. 696 p.
56. Hedenstrom A, Alerstam T (1992) Climbing performance of migrating birds as a
basis for estimating limits for fuel-carrying capacity and muscle work. J Exp Biol
164: 19–38.
57. Pennycuick CJ (1989) Bird Flight Performance. Oxford: Oxford University
Press.
58. Verhulst S, Hogstad O (1996) Social dominance and energy reserves in flocks of
Willow Tits. J Avian Biol 27: 203–208.
59. Veasey JS, Metcalfe NB, Houston DC (1998) A reassessment of the effect of
body mass upon flight speed and predation risk in birds. Anim Behav 56:
883–889.
60. Suter W, Vaneerden MR (1992) Simultaneous mass starvation of wintering
diving ducks in switzerland and the netherlands - a wrong decision in the right
strategy. Ardea 80: 229–242.
61. Kullberg C (1998) Does diurnal variation in body mass affect take-off ability in
wintering willow tits? Anim Behav 56: 227–233.
62. Kullberg C, Jakobsson S, Fransson T (1998) Predator-induced take-off strategy
in great tits (Parus major). Proc R Soc B 265: 1659–1664.
63. Van der Veen IT, Lindstrom KM (2000) Escape flights of yellowhammers and
greenfinches: more than just physics. Anim Behav 59: 593–601.
64. Kullberg C, Jakobsson S, Fransson T (2000) High migratory fuel loads impair
predator evasion in Sedge Warblers. Auk 117: 1034–1038.
65. Lind J, Fransson T, Jakobsson S, Kullberg C (1999) Reduced take-off ability in
robins (Erithacus rubecula) due to migratory fuel load. Behav Ecol Sociobiol 46:
65–70.
Starvation-Predation Risk Trade-Off in Ducks
PLoS ONE | www.plosone.org 11July 2011 | Volume 6 | Issue 7 | e22352