| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Behavioral Ecology Vol. 10 No. 5: 533-540
© 1999 International Society for Behavioral Ecology
Conspecific nest parasitism is associated with inequality in nest predation risk in the common goldeneye (Bucephala clangula)
Finnish Game and Fisheries Research Institute, Evo Game Research Station, Kaitalammintie 75, FIN-16970 Evo, Finland
Address correspondence to H. Pöysä. E-mail: hannu.poysa{at}rktl.fi .
Received 23 November 1998; revised 25 January 1999; accepted 8 February 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Previous studies of the role of nest predation in conspecific nest parasitism have not taken into account the possibility that predation risk may not be randomly distributed among nest sites and that breeding individuals may use different cues to assess the risk and adjust their reproductive tactic between years accordingly. Especially in cavity-nesting species, the role of nest predation in conspecific nest parasitism has been downplayed, while the role of nest site limitation has been highlighted. Using both observational and experimental data, I show that in the common goldeneye (Bucephala clangula), a cavity-nesting species in which conspecific nest parasitism is common, predation risk varies considerably between nest sites and does not follow a random expectation. The inequality in predation risk between nest sites also showed up in the occurrence of parasitized nests in an experimental setup. Nests parasitized in year t were more frequent in those nest sites that were not depredated during the previous nesting attempt in year t - n than in nest sites that were depredated and in control nest sites that had not been used for nesting before. A nest site addition experiment revealed that conspecific nest parasitism was not associated with nest site limitation. My findings give support for the hypothesis that nest predation is an important ecological factor explaining conspecific nest parasitism in goldeneyes.
Key words: between-year behavioral response, Bucephala clangula, conspecific nest, parasitism, nest predation, nest site limitation, parasitic egg laying.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Intraspecific nest parasitism is common in birds (Rohwer and Freeman, 1989
;
Yom-Tov, 1980) and has been
reported as well in other taxa (examples in
McRae, 1998). Several
hypotheses have been proposed to explain its adaptive significance and
evolution (Andersson, 1984;
Eadie et al., 1988;
Petrie and Møller,
1991; Sayler,
1992). Adaptive explanations of nest parasitism can be grouped
into two broad categories (e.g., Eadie et
al., 1988): (1) conditional (salvage) strategy hypotheses and (2)
alternative reproductive strategy hypotheses. Eadie and Fryxell
(1992) studied the conditions
under which intraspecific brood parasitism would be an evolutionarily stable
strategy (ESS). In a conditional ESS, parasitic females always have lower
success than nesting females, whereas in a mixed ESS, fitness payoffs of
parasitic and nesting females are frequency dependent. Eadie and Fryxell's
(1992) simulations showed that
it may be difficult to distinguish between these alternatives in the field
because results consistent with either a conditional ESS or a mixed ESS can be
obtained by varying population density.
Even though modeling approaches like that based on the ESS theory provide
insight into the evolution of parasitism, more empirical work is needed to
identify the ecological conditions under which nest parasitism may have
evolved. Four proximate explanations have been suggested to promote nest
parasitism under the conditional strategy hypothesis
(Eadie et al., 1988) : nest
loss, availability of nest sites, female energetics and body condition, and
female age or experience. The roles of nest site availability and female
characteristics have been studied extensively, but nest loss has received less
attention (e.g., Sayler,
1992). In particular, considering the fact that nest predation is
the major source of nesting mortality in birds
(Martin, 1988;
Nilsson, 1984;
Ricklefs, 1969) and the
important role it may play in life-history evolution of birds
(Julliard et al., 1997;
Martin, 1995;
Martin and Clobert, 1996;
Sæther, 1996), it is
surprising that nest predation has received so little attention in the study
of nest parasitism until recently.
The original nest loss hypothesis deals with a within-season response:
females that have lost their nests before laying a full clutch lay the
remaining eggs parasitically (Hamilton and
Orians, 1965; Yom-Tov,
1980). Indeed, recent manipulative studies using nest destruction
have succeeded in inducing parasitic laying in females of several species
(Emlen and Wrege, 1986;
Feare, 1991;
Haramis et al., 1984;
McRae, 1998;
Stouffer and Power, 1991),
though not in all cases (Rothstein,
1993). The importance of nest loss in the evolution of nest
parasitism is still suspect (e.g., Lyon,
1993). However, the idea that nest predation risk could favor
parasitism has a more general time dimension than just the immediate response
during laying. Payne (1977)
and Rubenstein (1982)
suggested that when nest predation is common, nest parasitism increases the
likelihood that at least some offspring will survive to independence, also
known as the "risk spreading" hypothesis (e.g.,
Petrie and Møller,
1991). A thorough exploration of this idea provides not only the
study of immediate individual responses to nest predation, but also knowledge
of predation risk of individual nest sites and between-year response of laying
females to that risk. In line with this idea, Petrie
(1986) suggested that, where
there are differences between moorhen (Gallinula chloropus)
territories, brood parasitism may have been favored by the selection pressure
of predation. However, there have been no attempts to relate the occurrence of
nest parasitism to predation risk of individual nest sites in the moorhen or
in other species (but see Robertson,
1998).
I can see two main reasons for the scarcity of empirical studies directly
addressing the role of nest predation in parasitism, in particular the role of
inequalities in predation risk between nest sites. First, Bulmer
(1984) theoretically showed
that, contrary to Rubenstein's
(1982) finding, the selection
pressure for risk spreading is weak, except when predation risk is very high.
However, an important point ignored both in Rubenstein's
(1982) and Bulmer's
(1984) models is that nests may
not be depredated at random. In other words, predation risk may vary
considerably between nest sites, and individuals may have evolved an ability
to predict the risk or simply use their previous experience to assess the risk
and adjust their reproductive tactics accordingly. This can be an important
factor, especially in long-lived species in which not only within-season but
also between-year responses to predation risk matter. The second reason is
practical and connected with the first one: it is usually difficult to
directly study both the predation risk of and the occurrence of parasitism in
individual nest sites in the field.
In this study I took a new approach and focus on inequalities in predation
risk between nest sites and its effect on the occurrence of nest parasitism.
My study species is the common goldeneye (Bucephala clangula), a
cavity-nesting duck in which intraspecific nest parasitism is usual
(Dow and Fredga, 1984;
Eadie, 1989;
Eriksson and Andersson, 1982;
Grenquist, 1963). Goldeneyes
are ideal for this kind of study because they readily accept nest-boxes, the
same nest-boxes are used for several years by same or different females, and
nest sites are selected only by females
(Dow and Fredga, 1983,
1985;
Pöysä
et al., 1997,
1999). Also, nest predation
rate is relatively high in goldeneyes
(Pöysä
et al., 1997, and references therein), and individuals readily
change their nest site after unsuccessful breeding
(Dow and Fredga, 1983). Most
important, nest site prospecting, presumably in preparation for the next
breeding season, has been demonstrated in goldeneye females
(Eadie and Gauthier, 1985;
Pöysä
et al., 1999; Zicus and
Hennes, 1989), and there are indications that females may use the
outcome of other females' nesting attempts in the preceding year as a cue in
nest site selection (Dow and Fredga,
1985).
My study had the following aims. First, I used both observational and
experimental data to show that predation risk varies considerably between nest
sites. Second, with a nest-box addition experiment, I studied the occurrence
of parasitism in individual nest sites (year t) in relation to the
occurrence and outcome of previous (year t-n) breeding attempts in
the same nest sites. Third, with the same experimental setup, I also studied
the effect of nest site availability on nest parasitism. It has been suggested
that nest site limitation is an important ecological factor responsible for
parasitism in Anatidae, especially in goldeneyes
(Eadie, 1991).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
The study was conducted in southeast Finland (61°35' N, 29°40' E) in the breeding seasons 1992-1998. The 8 x 12-km study area is dominated by pine (Pinus sylvestris) or mixed (pine, birch Betula spp., and spruce Picea abies) forests fragmented with lakes of varying size (mean size of lakes included was 5.4 ha, SD = 8.6, n = 30; a map of the study area is given in Pöysä, 1996a
).
Nest-box addition experiment
Fourteen nest-boxes on six lakes had been available for goldeneyes for
several years before 1992. New, previously unoccupied nest-boxes were erected
in three phases: 15 boxes in spring 1992 before the break-up of ice cover and
arrival of migrating goldeneyes (together making 29 boxes on 18 lakes in the
breeding season 1992), another 15 boxes in winter 1992-1993 (together 44 boxes
on 22 lakes in 1993), and a further 20 boxes in winter 1993-1994 (together 64
boxes on 30 lakes in 1994 onward). Some of the old boxes available before 1992
were removed or replaced with a new box, causing variation in the number of
boxes actually available each year (Table
1). To avoid crowding of nest-boxes, the study area was extended
so that the distance between lakes farthest apart increased from 6.7 km in
1992 to 11.8 km in 1994. The sequential addition of nest-boxes enabled me to
create a setup in which there was in a particular year t available
occupied nest-boxes that varied in terms of the status in the previous year,
in particular to get enough control nests (i.e., new nest-boxes that did not
have previous nesting attempts; see "Statistical analyses"
below).
|
I determined the occupation (occupied, not occupied) of the nest-boxes by
goldeneye females and the fate of each nesting attempt (successful, deserted,
depredated) by visiting each box three to four times between early May and
early June each year; eggs of deserted nests were removed after the nesting
period each year. Identification of parasitic laying involved more frequent
visiting in some cases and was continued until laying was finished (see
"Identification of parasitized nests" below). The date on which
the first egg of a clutch was laid (egg-laying date) was determined by
back-dating, using the criterion that it takes 1.32 days to lay one egg
(Fredga and Dow, 1983). I
recorded timing of parasitic laying whenever possible, and if more than 1 day
elapsed between successive visits, I assumed that laying occurred in the
middle of the period. I measured length and width of each egg close to the end
of incubation or, in the case of deserted and/or parasitized nests, after
laying was finished.
Nest predation experiment and observations
The nest predation experiment was carried out in each year between 1993 and
1998, and the procedure follows in detail that used in
Pöysä et al.
(1997,
1999), except that data from
year 1998 were included. I used 34 nest-boxes erected at 15 lakes (the boxes
and lakes were also included in the nest-box addition experiment) in all
years, except in 1993 when only 16 boxes were used (the other 18 boxes used in
this experiment were erected in winter 1993-1994; see above). The experiments
were started between 23 and 29 June. One domestic hen's egg was placed in each
nest-box (nest-boxes in which a breeding attempt was in progress were not used
in the experiment in that year). I checked nest-boxes after 21 days and
classified them as non-depredated if the egg was unbroken in the nest-box or
depredated if the egg was broken or missing. I calculated the predation risk
of each nest-box as the proportion of years in which the artificial nest was
depredated. The number of years when the experiment was carried out varied
between boxes (range 4-6 years), but predation risk values did not vary with
that number (Spearman rank correlation, rs = -.075,
p >.50, n = 34). Based on comparisons with other studies
done with artificial nests (in boxes or natural tree holes) and natural
goldeneye nests (in boxes or natural tree holes), the artificial nests used in
the experiment give comparable nest predation rates (see comparisons in
Pöysä
et al., 1997). The pine marten (Martes martes) has been
identified as the major nest predator of common goldeneyes in my study area
and in other goldeneye populations
(Pöysä
et al., 1997, and references therein).
Observational data of nest-box—specific predation risk was gathered
from all nesting attempts (at least one egg laid) between 1992 and 1998 (only
one nesting attempt per box per year). I classified a nesting attempt as
depredated if all eggs were taken (this usually was the case), or if at least
one egg disappeared and the nest was deserted, and non-depredated if ducklings
successfully left the nest-box or no eggs disappeared even if the nest was
deserted. Each nesting attempt was followed until the ducklings left the
nest-box (within 48 h of hatching) or, in deserted nests, at least for the
time it takes to lay and incubate a normal clutch. I calculated nest predation
risk as in artificial nests and included boxes that had at least three nesting
attempts (range, three to six attempts). Contrary to experimental data,
predation risk values were higher in nest-boxes with smaller number of
breeding attempts (rs = -.599, p <.005,
n = 22), simply reflecting the fact that females that fail to nest
successfully move to a new box the next year, leaving the previous box
unoccupied (Dow and Fredga,
1983).
Because 5 of the 34 nest-boxes in which predation risk was studied experimentally were included in the observational study of nest predation risk, I excluded the boxes from the experimental data. In all, I used experimental data of 29 boxes and observational data of other 22 boxes to study the distribution of predation risk. In addition to nest-site—specific predation risk, I studied whether predation risk is spatially concentrated. For each nest-box, separately for experimental and observational data, I determined if it was depredated or not in a given year and if the nearest active nest-box (artificial nest in the experimental data and nesting attempt in the observational data) was depredated or not in the same year. Each nest-box pair was included only once.
Identification of parasitized nests
Eadie (1989) developed an
objective criterion to identify parasitized nests on the basis of egg
morphology in goldeneyes. The criterion is based on standardized (z
score) measures of length, width, and weight of eggs. Eadie
(1989) performed cluster
analysis on these measures for all eggs in each nest, using complete linkage
clustering with Euclidean distance as the distance metric. He found that the
Euclidean distance between the two most dissimilar eggs (maximum Euclidean
distance, MED) in nonparasitized nests rarely exceeded 2.5 and was never
greater than 3.0 (mean = 1.74, n = 29), whereas MEDs in parasitized
nests were almost always greater than 2.5 (mean = 3.85, n = 57)
(nests were identified as parasitized or nonparasitized using other criteria;
see Eadie, 1989). Eadie
concluded that an MED of 2.5 can be used as an objective criterion to identify
parasitized nests; in other words, any nest for which the MED is greater than
2.5 can be classified as parasitized, whereas nests with an MED less than 2.5
can be classified as nonparasitized.
I tested the applicability of Eadie's
(1989) criterion in my
goldeneye population using 22 nests (completed clutches) for which I was able
to determine if they were parasitized (17 nests) or not (5 nests). If two or
more eggs appeared in a nest within 24 h during laying, or if new eggs
appeared in the nest during incubation, the nest was
parasitized—reliable criteria commonly used in birds
(MacWhirter, 1989), including
goldeneyes (Eadie, 1989). The
sample size for the test was constrained by high nest depredation rate (see
Table 1) and my intention to
minimize disturbing nesting attempts during laying. Nonetheless, the 22
clutches used in the test were from all the years between 1993 and 1998 and
represent a random sample of all possible nesting attempts with respect to
whether the laying sequence data were obtained in the same nest-boxes in
previous years [laying sequence obtained also in previous years in 32%
(n = 22) of cases in the test sample and in 14% (n = 22) of
cases in a random sample; G = 2.010, df = 1, p >.20; the
random sample included the same number of nests for each year between 1993 and
1998 than did the test sample]. Based on the measures of length and width, I
calculated the weight of each egg using the equation developed by Rohwer
(1988) for waterfowl. I
followed in detail Eadie's
(1989) procedure to calculate
the MED for each nest. Mean MED was 2.01 (SD = 0.55) in nonparasitized nests
and 4.86 (SD = 1.63) in parasitized nests. As Eadie
(1989) reported, MED increased
with clutch size (r =.500, df = 20, p =.02), and parasitized
clutches were larger than nonparasitized clutches (t = 3.683, df =
20, p =.001). However, the difference in MED between nonparasitized
and parasitized nests was significant even when clutch size was controlled for
(ANCOVA, using clutch size as covariate; F = 6.13, df = 1, 19,
p =.02). As in Eadie's
(1989) sample, all MEDs of
nonparasitized nests were <3.00; all but one of the MEDs in parasitized
nests were >3.00 (see Figure
1). However, both in nonparasitized and parasitized nests, the
mean MED was higher in my study than in Eadie's
(1989) study (see above). In
conclusion, I decided to use an MED of 3.00 as a conservative criterion to
identify parasitized nests. A post-hoc test showed that 95.5% (21/22) of the
nests used in the MED test were classified correctly with this criterion.
|
I included only nests with completed clutches to study the occurrence of nest parasitism. To minimize disturbance and the risk of nest desertion, I tried to measure the eggs as close to hatching as possible. Because nest predation was high, this meant that I was not able to measure all nests in a given year (i.e., nests were depredated before measuring; Table 1). In all, I was able to classify 84 nests as parasitized or nonparasitized using the MED criterion. [Note that eight of the nests from year 1992 could not be used in all analyses because the outcome of the nesting attempt in 1991 was not followed (i.e., boxes that were available before 1992)].
I do not have data of the frequency of conspecific nest parasitism in
natural nest sites for comparison. However, Eadie
(1989) compared frequencies of
nest parasitism between nest-boxes and natural cavities in the same area in
Canada and concluded that they were similar.
Statistical analyses
To study if nest predation is distributed among nest sites at random, I
compared the observed variance in nest predation in experimental nests with
the expected variance under a binomial distribution (see
Sokal and Rohlf, 1981). I
calculated the observed variance in predation risk for each number of years
studied (4-6 years) based on the observed depredation events; for instance, a
nest-box used for 5 years in the experiment may have been depredated in any
number of years between 0 and 5. I calculated the expected binomial
probability of predation risk using pooled data from all experimental nests
(n = 151 box-year observations); it was p =.576 (probability
of being depredated) and thus q = 0.424 (probability of not being
depredated). I used these values to obtain the expected variance in predation
risk for each number of years. I compared the difference between the observed
and expected variance in predation risk with F tests, calculated an
exact p value for each comparison, and finally combined the
p values of different comparisons using the chi-square method
(Sokal and Rohlf, 1981). The
dependence of predation risk values on the number of breeding attempts (years)
in the observational data (see "Nest predation experiments and
observations" above) did not allow a corresponding test; sample sizes to
estimate the probabilities of being depredated for each number of breeding
attempts (years) separately were too small.
The focus of this study was to examine the occurrence of parasitism in a given nest site in year t in relation to the status and outcome of the nesting attempt in the same site in year t - 1, in particular whether the nest was depredated during the previous nesting attempt. Therefore, I divided the 76 nests for which the occurrence of parasitism in year t was determined (see "Identification of parasitized nests" above) into the three groups (data pooled from different years): (1) previous nesting attempt was depredated (n = 23 nests); (2) previous nesting attempt was not depredated (n = 30; this group included both successful and deserted previous nesting attempts; a deserted nest with eggs indicates a safe nest-box with respect to predation); (3) control (n = 23; no previous nesting attempts in the nest-box in any year). Previous nesting attempt and control refer to year t - 1, except in the following cases: group 1, in five cases to year t - 2 (i.e., a nesting attempt was depredated 2 years earlier), in two cases to year t - 3, and in two cases to year t - 4 (in the years between the nest-box was not occupied); group 2, in four cases to year t - 2 [i.e., a nesting attempt was successful (not depredated) 2 years earlier; in the years between the nest-box was not occupied]; group 3 in nine cases the new nest-box had been unoccupied for 2 years and in one case for 6 years (i.e., a nest-box was available, but no previous nesting attempts occurred). For the sake of consistency, hereafter previous nesting attempt in a nest-box and control nest refers to year t - n in all cases if not otherwise stated.
Because the occurrence of parasitism per se (parasitized or nonparasitized) in a given nest-box in year t was not dependent on the occurrence of parasitism in the same box in year t - 1 (binomial test, p =.50, n = 16 boxes in which the occurrence of parasitism was determined in 2 successive years with a nesting attempt), data from a given box for different years were considered independent. However, each new nest-box could provide only one control observation because any laying female may have derived experience or information of nest predation risk in the box after the first nesting attempt.
The analyses were made with SYSTAT procedures
(Wilkinson 1992). In
G tests I used William's correction for continuity as recommended by
Sokal and Rohlf (1981). All
significance levels are two-tailed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Variation of predation risk between nest sites
Both observational and experimental data showed that predation risk varied considerably between nest sites (Figure 2). In the observational data, 9.1% of boxes were never depredated, whereas 9.1% of boxes were depredated in all years, the corresponding proportions in the experimental data being 14% and 17%, respectively. Distribution of predation risk values did not differ between observational and experimental data (Kolmogorov-Smirnov two-sample test, maximum difference = 0.254, p =.36). In the experimental data the observed variance in nest predation risk was significantly greater than the expected (random) variance under a binomial distribution (
2 =
26.334, df = 6, p <.001). Note that a corresponding test was not
possible with observational data (see "Statistical analyses"
above).
|
Predation risk was spatially concentrated. Considering the observational data, if a given nest-box was depredated, its nearest neighbor had a 65.7% (n = 35) probability of being depredated in the same year, but if a nest-box was not depredated, its nearest neighbor had a 70.6% (n = 34) chance of avoiding predation. Thus, in 68.1% of cases, the fate of nearest neighbors was the same (G = 9.127, df = 1, p <.01). The corresponding probabilities in the experimental data were 87.9% (n = 33) for depredated nests and 73.3% (n = 30) for non-depredated nests, meaning that in 81.0% of cases the fate of nearest neighbors was the same (G = 25.473, df = 1, p <.001).
Occurrence of parasitism in relation to the outcome of the previous
nesting attempt
My prediction was that, if the outcome of the previous nesting attempt
(specifically, predation risk of a nest site) affects the probability of the
nest site being parasitized in the next nesting attempt, nest sites that were
not depredated during the previous nesting attempt (year t - n) will
be parasitized more frequently than nest sites that were depredated. The
proportion of parasitized nests was higher in nest sites that were not
depredated in year t - n (83.3%, n = 30) than in control
nest sites (34.8%, n = 23; G = 13.087, df = 1, p
<.002) and in nest sites that were depredated in year t - n
(52.1%, n = 23; G = 5.834, df = 1, p <.05). The
proportion of parasitized nests in year t did not differ between nest
sites that were depredated in year t - n and control nest sites
(G = 1.377, df = 1, p >.20).
Parasitism as a within-season response to nest predation
I do not have direct data to address the response of laying females to nest
loss within the same season, but data from 1998 are complete enough for an
indirect exploration. Overall, nest predation rate was the lowest in that
year, but the proportion of parasitized nests was still rather high
(Table 1). In most nesting
attempts in 1998 (83%, n = 18) egg-laying date was before 16 May.
Also, most (54%, n = 24) of the identified parasitic egg layings took
place before 16 May, whereas all of the nest predation events in 1998
(n = 6) took place after 16 May, the difference in timing being
significant (Fisher's Exact test, p =.02). The difference in timing
suggests that most of the parasitic egg laying in 1998 did not occur as a
within-season response to nest loss.
Occurrence of parasitism and nest site availability
Due to high predation rate, I was not able to determine the occurrence of
parasitism in all nesting attempts in a given year (see
Table 1). The frequency of
parasitism in those nests that I was able to measure for parasitism may not be
based on representative data in all years, and therefore all between-year
comparisons are not relevant. However, I consider data from years 1992 and
1998 unbiased, because (1) a high proportion of all nesting attempts was
determined for parasitism in both years (73% in 1992 and 94% in 1998), and (2)
there was no difference in egg-laying dates between nesting attempts for which
parasitism was determined and nesting attempts for which it was not determined
in either year (Mann-Whitney U tests; 1992, U = 15.00,
p =.82, n1 = 11, n2 = 3;
1998, U = 9.00, p =.92, n1 = 17,
n2 = 1).
There were more nest-boxes available for goldeneyes in 1998 than in 1992 (Figure 3). It should be noted that this was the case both with all boxes and with boxes that were used at least once (i.e., acceptable) between 1992 and 1998; about 58% of acceptable boxes were occupied in 1992, the corresponding proportion being only 33% in 1998. If nest site availability per se explains the occurrence of parasitism in goldeneyes, I expected that the proportion of parasitized nests would be higher in 1992 than in 1998. However, the proportion of parasitized nests was lower, not higher, in 1992 (36.4%, n = 11) than in 1998 (64.7%, n = 17), though the difference was not significant (Fisher's Exact test, p =.25). In general, the occurrence of nest parasitism in my study population could not be associated with nest site availability per se because parasitism was observed in all years, despite the fact that a lot of empty but acceptable nest-boxes were available (Table 1).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Earlier studies of conspecific nest parasitism have ignored two basic evolutionary and ecological aspects associated with nest predation and individuals' response to it. First, predation risk may not be randomly distributed among nest sites, and individuals may use different cues to assess the risk and adjust their reproductive tactic accordingly. Second, within-season response to nest loss may not be the only route to parasitism, but individuals may respond to nest predation by changing their tactic between years. I showed that, indeed, predation risk varied considerably between nest sites and did not follow a random expectation. Predation risk also seemed to be spatially concentrated, an additional implication of nonrandom nest predation in common goldeneyes. This was not an unexpected result considering the fact that the pine marten, a long-living species that also uses tree cavities for breeding, is the major nest predator of common goldeneyes (see Pöysä et al., 1997
). Furthermore, parasitized nests were more frequent
in those nest sites that were not depredated during the previous nesting
attempt than in nest sites that were depredated and in control nest sites. It
should be emphasized that, in my approach focusing on nest sites, female
characteristics possibly associated with nest parasitism
(Eadie et al., 1988;
Sayler, 1992) do not confound
the results because there are no a priori reasons to expect that a female's
body condition, for example, determines whether she will lay parasitically in
a nest site that was not depredated during the previous nesting attempt or in
a nest site that was depredated. The findings of this study give strong
support for the hypothesis that nest predation is an important ecological
factor influencing conspecific nest parasitism in goldeneyes.
In agreement with my results, Robertson
(1998) found in common eiders
(Somateria mollissima), a species nesting on the ground without
fidelity to specific nest sites, that there are differences between nests in
vulnerability to predation. Nests in which parasitic eggs were laid on the
same day or on the previous day as the host had a lower predation risk than
nests where no parasitic eggs were laid on the day the host initiated her
clutch. Robertson (1998)
suggested that female eiders prospecting for a nest site may preferentially
choose a site that already contains an egg because the egg indicates that the
site offers protection against predation. Differing from my hypothesis,
however, Robertson's idea concerns a kind of within-season response to
perceived predation risk, though the fate of a parasite's own nesting was
evidently not a factor inducing parasitic laying (see below). Indeed,
Robertson (1998) hypothesized
that egg adoption is a behavioral tactic exhibited by the putative host rather
than by the putative parasite.
I did not find a difference in the frequency of parasitism between nest
sites that were depredated during the previous nesting attempt and control
nest sites that had not been used for nesting before. This suggests that
parasitically laying females did not differentiate between depredated and
previously unused boxes. Similarly, when studying nest-site selection in
goldeneye females, Dow and Fredga
(1985) found that females
changing a nest site between years preferred previously occupied boxes but
seemed not to differentiate between previously depredated and empty boxes.
Parasitically laying females did not have earlier experience or any other
information of predation risk in the control nest sites, except that
previously empty nest sites had been accepted by other females in that year.
Therefore, it was a reasonable tactic at least not to prefer those nest sites
in parasitic laying. In line with this reasoning, we found that goldeneye
females selecting a nest site are not able to assess predation risk of new,
previously unoccupied nest sites
(Pöysä et al., in
preparation). All females certainly did not have perfect knowledge of the
outcome of the previous nesting attempt in depredated nest sites and therefore
laid parasitically in those nests.
The study of individual females was beyond the topic of this study.
However, Eadie (1989) found
that parasitism was not a fixed strategy in goldeneyes but that females
switched tactics between years (i.e., nested on their own in one year and laid
parasitically in another year). Eadie
(1989) did not report the
outcome of the nesting attempt of females that switched from normal to
parasitic laying. In any case, the flexibility of individual laying behavior
documented by Eadie (1989) and
the nest site prospecting behavior of goldeneye females
(Eadie and Gauthier, 1985;
Pöysä
et al., 1999; Zicus and
Hennes, 1989) suggest that unsuccessful females parasitized
previously successful nest sites. Females that have failed in nesting (nest
lost or deserted) usually stay in the area for several weeks
(Pöysä, personal
observation). During that period they may gain information, not only for nest
site selection (Dow and Fredga,
1985), but about the outcome of other females' nesting attempts in
preparation for the next season's parasitic laying. Nest site fidelity is high
in successful goldeneye females (Dow and
Fredga, 1983;
Pöysä et al.,
unpublished data) so it should be advantageous for unsuccessful females to use
such kind a tactic.
The traditional nest loss hypothesis deals with the response of laying
females to nest loss within the same season
(Hamilton and Orians, 1965;
Yom-Tov, 1980). I was able to
address this possibility indirectly using data from year 1998. The fact that
most parasitic egg laying took place before nest predation occurred suggests
that nest loss during laying was not an important factor explaining parasitism
in my study population, though within-season response to nest loss may also
exist. Neither did Eadie (1989)
find support for the within-season nest loss hypothesis in five goldeneye
females subjected to experimental nest destruction. I suggest that the rather
high frequency of parasitized nests in 1998 was due to high nest predation
rates in previous years (see Table
1), not due to nest predation in that year. It is noteworthy that
during the years of heavy nest predation (1993-1996), the mean number of
nesting attempts per year was 24 (range 18-29), whereas only 18 nesting
attempts were recorded in 1998. However, the number of breeders on the lakes
of this study did not show a corresponding decrease; on the contrary, there
was an increasing trend from 1993-1996 to 1998
(Pöysä,
unpublished data from waterfowl pair counts done in the mid May each year; see
Pöysä,
1996b). These observations suggest that many of the putative
breeders did not establish nests of their own but switched to parasitic laying
in 1998.
Contrary to Eadie (1991), I
did not find support for the nest site availability hypothesis. Clearly, nest
site limitation per se could not explain the occurrence of conspecific nest
parasitism in my study population. Also, Eadie
(1991) acknowledged that nest
limitation cannot be the sole explanation for parasitism in the goldeneye
population he studied because parasitic laying occurred even when suitable
empty nest sites were available. Similarly, in his review of conspecific nest
parasitism in Anatidae, Sayler
(1992) concluded that nest
site limitation cannot be the only factor explaining high levels of parasitic
laying, even among cavity-nesting species (see also
Rohwer and Freeman, 1989).
Eadie (1991) based his
conclusion partly on an experiment in which he increased and decreased the
number of available nest-boxes, though his sample size did not allow
statistical tests. Eadie (1991)
found that, when he removed some nest-boxes from one lake in the previous
year, the proportion of parasitized nests in the remaining boxes on the lake
seemed to increase in the next breeding season, while no corresponding change
occurred in the control lake. Eadie interpreted this finding to support the
nest-site availability hypothesis. The finding is in accordance with that
hypothesis, but there is an alternative explanation. It is also possible that
females that bred earlier in the removed boxes assessed those nest-boxes that
were not removed as safe, and thus laid parasitically in the remaining
nest-boxes (i.e., analogous to laying parasitically in nests with low
predation risk). In other words, increased parasitism may not have been a
response to nest site limitation per se, but a response to the loss of a
female's own nest site and the possibility to lay parasitically in a safe nest
site on the same lake.
In conclusion, I have shown, for the first time, that inequality of
predation risk between nest sites is an important ecological factor affecting
the between-year variation in the occurrence of conspecific nest parasitism.
It has been shown earlier that nest predation during laying may induce
parasitic laying, supporting the nest loss hypothesis as a within-season
response. Recent studies by McRae
(1997,
1998) in the moorhen provide
an excellent example of within-season response. McRae showed that an increase
in nest predation directly influenced the rate of brood parasitism and
demonstrated experimentally that at least some females lay parasitically after
their own clutches were removed during laying. The findings of my study and
earlier studies, together covering both between-season and within-season
responses, strongly suggest that nest predation risk is an important
evolutionary factor behind conspecific nest parasitism. Clearly, more
attention should be paid in future studies to inequalities in predation risk
between nest sites and to between-year variation in the behavioral response of
individuals to nest predation, especially in long-lived, parasitically laying
species.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
I thank John Eadie for introducing me to the MED method to identify parasitized nests and for access to his unpublished work. For comments on the manuscript, I thank Markku Milonoff, Vesa Ruusila, and two anonymous reviewers.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Andersson M, 1984. Brood parasitism within species. In: Producers and scroungers: strategies of exploitation and parasitism (Barnard CJ, ed). London: Croom Helm;195 -228.
Bulmer MG, 1984. Risk avoidance and nesting strategies. J Theor Biol 106:529-535.
Dow H, Fredga S, 1983. Breeding and natal dispersal of the goldeneye, Bucephala clangula. J Anim Ecol 52:681-695.
Dow H, Fredga S, 1984. Factors affecting reproductive output of the goldeneye duck Bucephala clangula. J Anim Ecol 53:679-692.
Dow H, Fredga S, 1985. Selection of nest sites by a hole-nesting duck, the goldeneye Bucephala clangula.Ibis 127:16-30.
Eadie JM, 1989. Alternative reproductive tactics in a precocial bird: the ecology and evolution of brood parasitism in goldeneyes (PhD dissertation). Vancouver: University of British Columbia.
Eadie JM, 1991. Constraint and opportunity in the evolution of brood parasitism in waterfowl. In: Twentieth International Ornithological Conference, Christchurch, New Zealand, 2-9 December 1990 (Bell BD, Cossee RO, Flux JEC, Heather BD, Hitchmough RA, Robertson CJR, Williams MJ, eds). Wellington, New Zealand: Hutcheson, Bowman & Stewart Ltd; 1031-1040.
Eadie JM, Fryxell JM, 1992. Density dependence, frequency dependence, and alternative nesting strategies in goldeneyes.Am Nat 140:621 -641.
Eadie JM, Gauthier G, 1985. Prospecting for nest sites by cavity-nesting ducks of the genus Bucephala. Condor 87:528-534.
Eadie JM, Kehoe FP, Nudds TD, 1988. Pre-hatch and post-hatch brood amalgamation in North American Anatidae: a review. Can J Zool 66:1709-1721.
Emlen ST, Wrege PH, 1986. Forced copulations and intra-specific parasitism: two costs of social living in the white-faced bee-eater. Ethology 71:2-29.
Eriksson MOG, Andersson M, 1982. Nest parasitism and hatching success in a population of goldeneyes Bucephala clangula.Bird Study 29:49-54.
Feare CJ, 1991. Intraspecific nest parasitism in starlings Sturnus vulgaris: effects of disturbance on laying females.Ibis 133:75-79.
Fredga S, Dow H, 1983. Annual variation in the reproductive performance of goldeneyes. Wildfowl 34:120-126.
Grenquist P, 1963. Hatching losses of common goldeneyes in the Finnish Archipelago. In: Thirteenth International Ornithological Congress, Ithaca, New York, 17-24 June 1962 (Sibley CG, ed). Baton Rouge, Louisiana: The American Ornithologist's Union;685 -689.
Hamilton WJ, III, Orians GH, 1965. Evolution of brood parasitism in altricial birds. Condor 67:361-382.
Haramis GM, Alliston WG, Richmond ME, 1984. Dump nesting in the wood duck traced by tetracycline. Auk 100:729-730.
Julliard R, McCleery RH, Clobert J, Perrins CM, 1997. Phenotypic adjustment of clutch size due to nest predation in the great tit.Ecology 78:394-404.
Lyon BE, 1993. Conspecific brood parasitism as a flexible female reproductive tactic in American coots. Anim Behav 46:911-928.
MacWhirter RB, 1989. On the rarity of intraspecific brood parasitism. Condor 91:485-492.
Martin TE, 1988. Processes organizing open-nesting bird assemblages: competition or nest predation? Evol Ecol 2:37-50.
Martin TE, 1995. Avian life history evolution in relation to nest sites, nest predation, and food. Ecol Monogr 65:101-127.
Martin TE, Clobert J, 1996. Nest predation and avian life history evolution in Europe versus North America: A possible role of humans? Am Nat 147:1028-1046.
McRae SB, 1997. A rise in nest predation enhances the frequency of intraspecific brood parasitism in a moorhen population. J Anim Ecol 66:143-153.
McRae SB, 1998. Relative reproductive success of
female moorhens using conditional strategies of brood parasitism and parental
care. Behav Ecol
9:93-100.
Nilsson SG, 1984. The evolution of nest-site selection among hole-nesting birds: the importance of nest predation and competition.Ornis Scand 15:167-175.
Payne RB, 1977. The ecology of brood parasitism in birds. Annu Rev Ecol Syst 8:1-28.[Web of Science]
Petrie M, 1986. Reproductive strategies of male and female moorhens (Gallinula chloropus). In: Ecological aspects of social evolution (Rubenstein DI, Wrangham RW, eds). Princeton, New Jersey: Princeton University Press;43 -63.
Petrie M, Møller AP, 1991. Laying eggs in others' nests: intraspecific brood parasitism in birds. Trends Ecol Evol 6:315-320.
Pöysä H, 1996a. Factors affecting the probability of extinction of local Anas species populations. In: Anatidae 2000, An International Conference on the Conservation, Habitat Management and Wise Use of Ducks, Geese and Swans, Strasbourg, France, 5-9 December 1994 (Birkan M, van Vessem J, Havet P, Madsen J, Trolliet B, Hainard R, eds). Paris: Office National de la Chasse;547 -555.
Pöysä H, 1996b. Population estimates and the timing of waterfowl censuses. Ornis Fenn 73:60-68.
Pöysä H, Milonoff M, Ruusila V, Virtanen J, 1999. Nest-site selection in relation to habitat edge: experiments in the common goldeneye. J Avian Biol 30:79-84.
Pöysä H, Milonoff M, Virtanen J, 1997. Nest predation in hole-nesting birds in relation to habitat edge: an experiment. Ecography 20:329-335.
Ricklefs RE, 1969. An analysis of nesting mortality in birds. Smithson Contrib Zool 9:1-48.
Robertson GJ, 1998. Egg adoption can explain joint egg-laying in common eiders. Behav Ecol Sociobiol 43:289-296.
Rohwer FC, 1988. Inter- and intraspecific relationships between egg size and clutch size in waterfowl.Auk 105:161-176.
Rohwer FC, Freeman S, 1989. The distribution of conspecific nest parasitism in birds. Can J Zool 67:239-253.
Rothstein SI, 1993. An experimental test of the Hamilton-Orians hypothesis for the origin of avian brood parasitism.Condor 95:1000-1005.
Rubenstein DI, 1982. Risk, uncertainty and evolutionary strategies. In: Current problems in sociobiology (King's College Sociobiology Group, ed). Cambridge: Cambridge University Press; 91-111.
Sæther B-E, 1996. Evolution of avian life histories—does nest predation explain it all? Trends Ecol Evol 11:311-312.
Sayler RD, 1992. Ecology and evolution of brood parasitism in waterfowl. In: Ecology and management of breeding waterfowl (Batt BDJ, Afton AD, Anderson MG, Ankney CD, Johnson DH, Kadlec JA, Krapu GL, eds). Minneapolis: University of Minnesota Press;290 -322.
Sokal RR, Rohlf FJ, 1981, Biometry, 2nd ed. San Francisco: Freeman.
Stouffer PC, Power HW, 1991. Brood parasitism by starlings experimentally forced to desert their nests. Anim Behav 41:537-538.
Wilkinson L, 1992. SYSTAT: the system for statistics. Evanston, Illinois: SYSTAT, Inc.
Yom-Tov Y, 1980. Intraspecific nest parasitism in birds. Biol Rev 55:93 -108.
Zicus MC, Hennes SK, 1989. Nest prospecting by common goldeneyes. Condor 91:807-812.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Poysa Public information and conspecific nest parasitism in goldeneyes: targeting safe nests by parasites Behav. Ecol., May 1, 2006; 17(3): 459 - 465. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Ruusila, H. Poysa, and P. Runko Costs and benefits of female-biased natal philopatry in the common goldeneye Behav. Ecol., November 1, 2001; 12(6): 686 - 690. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||