Behav Ecol Sociobiol (2003) 54:30–35 DOI 10.1007/s00265-003-0596-1
ORIGINAL ARTICLE
Hannu Pys
Parasitic common goldeneye (Bucephala clangula) females lay preferentially in safe neighbourhoods Received: 6 July 2002 / Revised: 25 January 2003 / Accepted: 15 February 2003 / Published online: 27 March 2003 Springer-Verlag 2003
Abstract Nest predation has been suggested as an explanation of the adaptive significance and evolution of conspecific brood parasitism, an alternative reproductive tactic pursued by females in several animal taxa. I used new nest boxes that contained only decoy eggs and were erected on lakes differing in real nest predation risk to test this hypothesis in the common goldeneye (Bucephala clangula), a hole-nesting duck. I used broken eggs to simulate predation risk of the boxes to determine if parasites having no previous experience with the boxes discriminate between seemingly safe and risky nest sites. Parasites laid eggs in the experimental boxes independently of the simulated predation risk, suggesting that they do not use broken eggs or nest disarray as indicators of predation intensity. Parasites preferred experimental boxes on lakes where real nest predation risk was low, supporting the nest predation risk hypothesis. Assuming that females in high risk areas have had experience of nest predation, they may take this into account in selecting host nests. Keywords Bucephala clangula · Conspecific brood parasitism · Experience · Nest predation
Introduction Conspecific brood parasitism is an alternative reproductive tactic pursued by females in many animal taxa, especially in birds (Rohwer and Freeman 1989; Sayler 1992; Yom-Tov 1980, 2001). Several hypotheses have been suggested to explain its adaptive significance and evolution (Eadie et al. 1988; Sayler 1992). Here I focus Communicated by J. Dickinson H. Pys ()) Finnish Game and Fisheries Research Institute, Joensuu Game and Fisheries Research, Kauppakatu 18-20, 80100 Joensuu, Finland e-mail:
[email protected] Tel.: +358-20-5751400 Fax: +358-20-5751409
on nest predation risk, an hypothesis that has received little empirical attention until recently. Originally, based on Gillespie’s (1977) finding that natural selection favors genotypes with the smallest variance in reproductive success, Rubenstein (1982) showed with computer simulations that distributing eggs in several nests increases the likelihood that at least some offspring will survive to independence, lending credence to the risk spreading hypothesis (e.g. Brown and Brown 1989; Petrie and Møller 1991; Sayler 1992). The risk spreading hypothesis was subsequently refuted by Bulmer’s (1984) simulations, and, in general, is not considered important in the evolution of conspecific brood parasitism (Petrie and Møller 1991). Recently, a more behavioral approach has been taken to investigate the role of nest predation risk in conspecific brood parasitism. Pys (1999a) found in the common goldeneye (Bucephala clangula), a hole-nesting duck, that parasites prefer to lay eggs in nests of low predation risk, probably using previous experience of nest-sitespecific predation as a cue. This variant of the nest predation risk hypothesis does not invoke the idea of risk spreading per se, but assumes that parasites are able to recognize safe nest sites and lay parasitic eggs accordingly. In his comment on Robertson’s (1998) study of conspecific brood parasitism in common eiders (Somateria mollissima), Ruxton (1999) similarly and independently hypothesized that parasites can detect variation in predation risk between nest sites and preferentially lay in nests where the risk is intrinsically lower. Brown and Brown (1991) found evidence that parasitic cliff swallows (Hirundo pyrrhonota) are able to locate nests that are particularly successful at producing fledglings, though ectoparasitism, rather than nest predation, was considered in their study. Here I report on experiments, addressing at two scales the importance of nest predation risk in conspecific brood parasitism in goldeneyes. By using new nest boxes that contained decoy eggs (i.e. no host at all) I was able to exclude the role of host quality, including the degree of relatedness (see Andersson and hlund 2000), in deter-
31
mining parasitic laying frequencies. First, by using broken eggs, I simulated predation risk of the boxes to determine if parasites having no previous experience of the boxes can discriminate between seemingly safe and risky nest sites. Second, Pys (1999a) showed that predation risk of goldeneye nests was spatially correlated beyond individual nest sites. Similarly, Dow and Fredga (1983) found in goldeneye that the rate of predation showed a progressive decline with increasing distance from nest boxes in which clutches had been depredated. Parasites may therefore use their experience of lake-specific nest predation and adjust parasitic egg laying at the between-lake level accordingly. Hence, in the second experimental setting I used the between-lake variation in real nest predation risk to study if parasites prefer to lay eggs in boxes situating in safe environments.
Methods The study was carried out on 15 small lakes (mean=6.1 ha, SD=6.6, range 1.0–23.2 ha) in southeast Finland (6135' N, 2940' E) in an area dominated by pine (Pinus sylvestris) or mixed pine, birch (Betula spp.), and spruce (Picea abies) forests. Old nest boxes were available for goldeneyes at all lakes and each lake was suitable for nesting (at least one nesting attempt recorded between 1992 and 1998; see Pys and Pys 2002). Experimental nests In early April 1999, two new nest boxes, the entrances of which were closed with a door, were erected in a visible place at the shoreline of each lake (15 pairs of boxes in all). At each lake, one of the boxes was randomly assigned as ‘risky’, the other one being ‘safe’. The bottom of each box was covered with a 3-cm layer of sawdust. Mean distance between the boxes at a lake was 170 m (SD=54.0). Goldeneye females frequently parasitize the nests of other hole-nesting duck species (see Eadie et al. 1995), indicating that parasitic females are not particularly discriminating about the appearance of eggs. Moreover, parasitic Barrow’s goldeneye (Bucephala islandica) (Eadie 1989) and wood duck (Aix sponsa) (Wilson 1993) females lay eggs in nests containing only decoy eggs. I took advantage of these findings and used chicken eggs dyed with silk-painting colours (manufactured Furia cold-reactive color pigments turquoise F-1307 and brown F-1160) to mimic the colour of goldeneye eggs. In 1999, the risky nest box was provided with one broken and two unbroken eggs lying in disorder on the bottom and the safe nest box was provided with three unbroken eggs positioned tidily in the middle of the bottom. In 2000, the experimental set-up was identical (the same 15 pairs of boxes as in 1999), except that the risky nest box (same risky boxes as in 1999) was provided with two broken and one unbroken eggs; the aim of the change was to increase the appearance of vulnerability of the risky nest and to test the response of parasites to this. The eggs were pressed softly in the sawdust but not covered with it. The broken egg was emptied and about one third of the shell was removed to simulate a goldeneye egg destroyed by pine marten (Martes martes) or mink (Mustela vison), the most important predators of goldeneye eggs in the study area (Pys et al. 1997). In general, birds appear to recognize damaged eggs in their nests and respond to them (Kemal and Rothstein 1988; Mallory et al. 2000). The experiment lasted 35 days in both years, starting immediately after egg laying in real nests began (on 26 April in 1999 and on 25 April in 2000). On the first day of the experiment in both years, the door was removed from the entrance hole of each box and the decoy eggs were put in the boxes as described above; the experiment started on the same day at all lakes. Experimental boxes
were checked at an interval of 1.5 days (SD=0.10; range: 1–3 days); boxes on the same lake were always checked on the same visit. On each visit, I recorded whether or not the box had been visited (sawdust scratched, position of the eggs changed, broken egg(s) smashed to pieces, or eggs robbed) and the clutch parasitized (1 real goldeneye eggs laid). I removed parasitic eggs from the nest and restored the experimental setting to its original condition on each visit; broken eggs were replaced if they were smashed to pieces and shell fragments were removed. In all, 31 parasitic eggs were laid in the experimental nests in 1999 and 63 in 2000. On the last day of the experiment in both years, the door was put back on the entrance hole of each box in order to minimize the information goldeneye females could have about the safety of the boxes. Robbing of decoy eggs was occasional (recorded in 5 out of 702 nest-box visits in 1999 and in 14 out of 734 nest-box visits in 2000) and did not differ between risky and safe boxes at lakes (1999, Wilcoxon paired-sample test, T=6, n=number of lakes=4, P>0.50; 2000, T=14, n=6, P>0.50). Based on Eadie (1989), Pys (1999a) and Pys et al. (2001a) and following the criterion verified by Pys et al. (2001a) to identify parasitized nests in the common goldeneye, I used the between-egg variation in length, width and weight (weight calculated as in Pys 1999a and Pys et al. 2001a) of eggs within each experimental nest to estimate the number of females that had laid parasitic eggs in the nest. For each experimental nest in both years, I calculated the relative timing of the start of parasitic egg laying (hereafter, start of parasitic laying; number of days from the start of the experiment until the first goldeneye egg appeared in the nest, i.e. range 1–35; if a nest was not parasitized a value of 36 was given), total number of parasitic eggs laid and the number of females that laid parasitic eggs. Note that because the experiment was started on the same day in all lakes in both years, and because here I am interested only in the relative differences in the start of parasitic laying between risky and safe boxes within lakes, and between lakes in general, the inclusion of non-parasitized nests with a score of 36 does not cause a bias in the results. Lake-specific nest predation risk I used both observational and experimental data to determine overall nest predation risk for each lake (for details of the experimental data, see Pys et al. 1997, 2001b; Pys 1999a). Previous comparisons between observational and experimental data indicate that they give similar nest predation rates (Pys et al. 1997, 2001b; Pys 1999a). Hence, for each lake, I pooled observational and experimental data from 1992–2000 from all nest boxes, except the experimental decoy nests used in this study, and calculated the overall real lake-specific predation risk as the proportion (%) of nests depredated (mean=61.4%, SD=25.0, range 20.7–100%). Sample size varied between lakes (mean=19.3 nests per lake, SD=13.2), but this variation did not bias the predation risk estimate (lake-specific nest predation risk vs lake-specific sample size, r=0.043, n=15, P>0.50); note that this finding does not mean that nesting common goldeneye females were not selective with respect to nest predation risk (see below). Furthermore, based on data from eight of the 15 lakes during three periods of equal duration (i.e. 1992–1994, 1995–1997, 1998–2000, using only lakes that had 5 observations for each of the three periods), predation risk ranking was consistent between years across different lakes (Kendall’s coefficient of concordance, W=0.764, c2=16.04, df=7, P0.20 >0.20
10.5 25.0 26.5
6 10 10
>0.50 >0.50 >0.50
spective powers estimated with bootstrapping by setting the probability level of type I error at a=0.05, using the observed differences as effect sizes and 100,000 bootstrap samples in each comparison; see Thomas and Juanes 1996 for an analogous approach to estimate statistical power for Wilcoxon paired-sample test). However, if I set the probability level of type I and type II errors at a=b=0.05 (Toft and Shea 1983), meaning a statistical power of 0.95, and consider the observed differences biologically important I would need enormous sample sizes to detect statistically significant differences at P=0.05: start of parasitic laying, 280–830 nest-box pairs; number of parasitic eggs laid per box, 210–9,210 nest-box pairs; number of parasitic females per box, 240–7,500 nest-box pairs (sample sizes estimated with bootstrapping technique by using 100,000 bootstrap samples in each case). Considering the start of parasitic laying, survival regression did not reveal a difference between risky and safe nests (Tarone-Ware method; 1999: c2=1.76, df=1, P=0.184; 2000: c2=1.11, df=1, P=0.291). Neither did the magnitude of between-year change of the response variables (value of a variable in 2000 minus value of the same variable in 1999) differ between risky and safe nests (start of parasitic laying, T=31.0, n=11, P>0.50; number of parasitic eggs laid, T=16.0, n=11, P>0.10; number of parasitic females, T=20.5, n=9, P>0.50; tied pairs ignored in all tests). Parasitic laying and real nest predation risk Because parasitic laying did not vary in response to the experimental treatment, the following analyses are based on lake-specific means calculated over treatments and years. The three variables describing different aspects of parasitic laying all showed a response to real nest
33
Fig. 1 Relationship of: a the start of parasitic laying, b the number of parasitic eggs laid and c the number of females that laid parasitic eggs in experimental nests with nest predation risk. Each point represents a lake-specific mean (dependent variables; mean of 2 years and two treatment types) or composite value (independent variable; proportion (%) of nests robbed between1992 and 2000). Dependent variables were first controlled for the number of potential parasites by regressing each of them against the mean number of potential parasites per lake, and residuals of these regressions were used in the final analyses; residuals of the dependent variable in b were log transformed before the final analysis
predation risk (Fig. 1): the higher the lake-specific nest predation risk the later the start of parasitic laying (r2=0.27, F1,13=4.75, P=0.048), the fewer parasitic eggs were laid (r2=0.39, F1,13=8.24, P=0.013) and the smaller the number of females that laid parasitically (r2=0.42, F1,13=9.39, P=0.009) in the experimental nests.
Discussion Parasitic common goldeneye females did not respond to simulated risk when laying in boxes where they did not
have previous experience of nest success. In contrast with this finding, parasites responded to real nest predation risk by laying more frequently in experimental boxes on lakes where the overall nest-predation rate was low. The power of the statistical tests dealing with the simulated predation risk was low, though retrospective power estimated with the data used to test the null hypothesis and the observed effect size is of limited use (Steidl et al. 1997). However, because the distributions of observations overlapped so much between simulated risky and safe boxes (Table 1), and because huge sample sizes would have been needed to find statistically significant differences, I conclude that no biologically important response to the simulated risk occurred in the present experiment setting. Furthermore, none of the 14 tests dealing with the effect of simulated predation risk was statistically significant, whereas all of the three tests addressing the importance of real nest predation risk were significant. This finding also supports the conclusion that previous experience of nest predation risk is an important factor in parasitic laying in the common goldeneye. If we assume that females in high risk areas have had experience of nest predation, they may take this into account in selecting host nests. Similarly, Pys (1999a) found that nest parasitism was more frequent in those nest sites that were not depredated in the previous years than in nest sites that were depredated or control nest sites that had not been used previously. This suggests that parasites used previous nest-site-specific experience of nest success as a cue. An improvement of the present study, as compared with Pys (1999a), is that the role of host quality was experimentally ruled out. It has been suggested in other species that parasites may preferentially select high-quality hosts providing nests of low predation risk for parasitism (Ruxton 1999). Furthermore, the present study extends to a wider spatial scale the idea that nest predation and individual experience with predation are important factors in conspecific nest parasitism in goldeneyes. This finding fits with the fact that nest predation risk in goldeneyes is correlated beyond individual nest sites (Dow and Fredga 1983; Pys 1999a). Nest-site prospecting (Eadie and Gauthier 1985; Zicus and Hennes 1989; Pys et al. 1999) might provide a behavioural mechanism to get information on site-specific predation risk in common goldeneyes; a multi-year experiment addressing the role of this behaviour in conspecific nest parasitism is currently under way (H. Pys, unpublished data). Analogously with the findings in the common goldeneye, Brown and Brown (1991) found that colonially breeding, parasitic cliff swallows selected host nests that ultimately had the lowest infestations of deleterious ectoparasites. They suggested that parasitic cliff swallows apparently assessed the nests around them and used nest age early in the season as a cue to select high-quality host nests. The importance of experience in the decision making of nesting individuals was highlighted in another experiment showing that common goldeneye females likely have not evolved an ability to assess predation risk when
34
selecting among new, previously unoccupied nest sites (Pys et al. 2001b). Site-related experience of own (Switzer 1997) or conspecific reproductive success (Danchin et al. 1998) appears to be an important general determinant in habitat selection and reproductive decisions of individuals. Concerning the role of experience of nest predation in conspecific brood parasitism, McRae (1997, 1998) found in the moorhen (Gallinula chloropus) that some females responded directly to nest loss by laying their next egg parasitically while others adopted a mixed strategy of laying parasitically before renesting, both being within-season responses to nest predation (see Pys 1999a). A further experiment in which nest predation was manipulated during parasitic egg laying is currently in progress and will address more closely the role of own experience in the common goldeneye (H. Pys, unpublished data). Parasitic laying in the experimental nests was surprisingly frequent, implying that decoy nests provide a useful tool to study conspecific brood parasitism, at least in some species (see also Eadie 1989; Wilson 1993), especially if one wants to control for host quality (see below). Are there methodological artefacts that could explain the frequent parasitic laying in the experimental nests in this study? One potential artefact, nest site availability cannot explain the patterns I observed, because dozens of empty, but previously used, boxes were available in the study area (see also Pys 1999a; Pys and Pys 2002) and these boxes remained unoccupied in both years. A second possible artefact is that the higher incidence of parasitic laying in the low predation sites could be a simple demographic consequence of the greater success of low predation sites. For example, if females return to and lay on the lakes where they were born, we might expect a higher frequency of laying on low risk lakes simply because of female philopatry. Parasitic common goldeneye females are reported to lay in nests 1–2 km apart (Andersson and hlund 2000; hlund and Andersson 2001). Also, new data based on protein fingerprinting analyses of parasitic eggs from the present common goldeneye population reveal that, within the same season, a given parasitic female may lay eggs in nests on different lakes several kilometers apart (five parasitic females that used 2–3 lakes each in 2001; female-specific distance between lakes, mean=3.3 km, SD=1.8 km, range 1.4– 5.9 km; H. Pys, J. Rutila, K. Lindblom and J. Sorjonen unpublished data). This suggests that natal philopatry of parasitic females is unlikely to bias the results. Neither does the lack of nest defense of the experimental nests cause an artefact, because common goldeneye host females do not defend the nest during the laying period (see also Eadie et al. 1995). In line with this Andersson and hlund (2000) mention that their video recordings showed that during the laying period hosts usually visit the nest only every other day. Similarly, in 1999–2000 I encountered a female in the nest box on only 10 out of 167 visits in which egg laying was underway and progressed to incubation (i.e. the nest had a host).
These 10 observations most probably included both hosts and parasites. Of course, hosts stay in the nest for a while during each egg-laying event but the period is too short to function as appreciable nest defense. A possible methodological artefact in the comparison between simulated risky and safe nests might be that, because I did not add any other cues than the eggs to indicate an active nest (e.g. down or new eggs), potential parasites may have considered both nest types deserted and, therefore, equally poor. However, because parasites readily laid eggs in the experimental nests, this possibility seems negligible. In addition, common goldeneye females often lay parasitic eggs in the deserted nests of conspecifics and seem not to be able to recognize whether a nest is active and has a host or not (Grenquist 1963; Eriksson and Andersson 1982; Pys 1999b). Hence, a nest with only eggs seems to provide a sufficient stimulus for parasitically laying females. In the present experiment, I was able to exclude the role of relatedness between hosts and parasites. This is an important aspect because, based on protein fingerprinting analyses of common goldeneye eggs, Andersson and hlund (2000) found that host and parasite are often more closely related than expected by chance. They concluded that genetic relatedness and kin discrimination play a role in conspecific brood parasitism in goldeneyes. As far as the decisions of parasites are concerned, the finding that parasites frequently laid in experimental nests that did not have a host suggests that genetic relatedness and kin recognition may not play an important role in conspecific brood parasitism in common goldeneyes. However, an experiment in which parasites are given a choice between kin and non-kin hosts would provide a more conclusive test of the idea. In general, findings of the importance of relatedness in conspecific brood parasitism are divergent. Host-parasite relatedness has been found to be high in the moorhen, but in this species parasites do not preferentially choose relatives as hosts and high relatedness is explained solely by philopatry (McRae and Burke 1996). Lyon and Eadie (2000) mention that a study based on DNA fingerprinting did not reveal evidence of relatedness between hosts and parasites in the Barrow’s goldeneye. Furthermore, parasitic wood duck females even avoid parasitizing close relatives (Semel and Sherman 2001). In conclusion, this study provides experimental support for nest predation as an explanation for the laying behaviour of parasitic females in the common goldeneye. As parasites did not respond to simulated nest predation of which they did not have previous experience, an important element in parasitic laying probably is the experience, own or derived from conspecifics’ nesting attempts, of site-specific predation risk. Furthermore, somewhat diverging results of this study and of that by Anderson and hlund (2000) highlight how important it is to study reproductive tactics and behavioural decisions of individuals of a given species under different ecological conditions. Because the study of conspecific brood parasitism may reveal important insights into the evolution of alternative reproductive tactics and avian breeding
35
systems in general (Andersson and hlund 2000; Lyon and Eadie 2000), further studies addressing the role of different ecological factors and host-parasite relatedness are needed to fully understand this system. Acknowledgements I thank Malte Andersson for discussions about brood parasitism in goldeneyes, Mauri Pesonen for writing the bootstrapping program and Markku Milonoff, Vesa Ruusila, John Eadie and an anonymous referee for comments on the manuscript.
References hlund M, Andersson M (2001) Female ducks can double their reproduction. Nature 414:600–601 Andersson M, hlund M (2000) Host-parasite relatedness shown by protein fingerprinting in a brood parasitic bird. Proc Natl Acad Sci USA 97:13188–13193 Brown CR, Brown MB (1989) Behavioural dynamics of intraspecific brood parasitism in colonial cliff swallows. Anim Behav 37:777–796 Brown CR, Brown MB (1991) Selection of high-quality host nests by parasitic cliff swallows. Anim Behav 41:457–465 Bulmer MG (1984) Risk avoidance and nesting strategies. J Theor Biol 106:529–535 Cohen J (1988) Statistical power analysis for the behavioral sciences. Erlbaum, Hillsdale, N.J. Danchin E, Boulinier T, Massot M (1998) Conspecific reproductive success and breeding habitat selection: implications for the study of coloniality. Ecology 79:2415–2428 Dow H, Fredga S (1983) Breeding and natal dispersal of the goldeneye, Bucephala clangula. J Anim Ecol 52:681–695 Eadie JM (1989) Alternative reproductive tactics in a precocial bird: the ecology and evolution of brood parasitism in goldeneyes. PhD thesis, University of British Columbia, Vancouver Eadie JM, Gauthier G (1985) Prospecting for nest sites by cavitynesting 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 Eadie JM, Mallory ML, Lumsden HG (1995) Common goldeneye Bucephala clangula. In: Poole A, Gill E (eds) The birds of North America, No 170. The Academy of Natural Sciences, Philadelphia, pp 1–32 Eriksson MOG, Andersson M (1982) Nest parasitism and hatching success in a population of goldeneyes Bucephala clangula. Bird Study 29:49–54 Gillespie JH (1977) Natural selection for variances in offspring numbers: a new evolutionary principle. Am Nat 111:1010–1014 Grenquist P (1963) Hatching losses of common goldeneyes in the Finnish Archipelago. In: Sibley CG (ed) Proceedings of the Thirteenth International Ornithological Congress. The American Ornithologist’s Union, Baton Rouge, pp 685–689 Kemal RE, Rothstein SI (1988) Mechanisms of avian egg recognition: adaptive responses to eggs with broken shells. Anim Behav 36:175–183 Koskimies P, Visnen RA (1991) Monitoring bird populations. A manual of methods applied in Finland. Zoological Museum, Finnish Museum of Natural History, Helsinki Lyon BE, Eadie JM (2000) Family matters: kin selection and the evolution of conspecific brood parasitism. Proc Natl Acad Sci USA 97:12942–12944 Mallory ML, Rendell WB, Robertson RJ (2000) Responses of birds to broken eggs in their nests. Condor 102:673–675 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
McRae SB, Burke T (1996) Intraspecific brood parasitism in the moorhen: parentage and parasite-host relationships determined by DNA fingerprinting. Behav Ecol Sociobiol 38:115–129 Petrie M, Møller AP (1991) Laying eggs in others’ nests: intraspecific brood parasitism in birds. Trends Ecol Evol 6:315–320 Pys H (1999a) Conspecific nest parasitism is associated with inequality in nest predation risk in the common goldeneye (Bucephala clangula). Behav Ecol 10:533–540 Pys H (1999b) Association between conspecific nest parasitism and the timing of breeding in the common goldeneye Bucephala clangula: an alternative interpretation. Ornis Fenn 76:89–92 Pys H, Pys S (2002) Nest-site limitation and density dependence of reproductive output in the common goldeneye Bucephala clangula: implications for the management of cavity-nesting birds. J Appl Ecol 39:502–510 Pys H, Rask M, Nummi P (1994) Acidification and ecological interactions at higher trophic levels in small forest lakes: the perch and the common goldeneye. Ann Zool Fenn 31:397–404 Pys H, Milonoff M, Virtanen J (1997) Nest predation in holenesting birds in relation to habitat edge: an experiment. Ecography 20:329–335 Pys 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 Pys H, Runko P, Ruusila V, Milonoff M (2001a) Identification of parasitized nests by using egg morphology in the common goldeneye: an alternative to blood sampling. J Avian Biol 32:79–82 Pys H, Ruusila V, Milonoff M, Virtanen J (2001b) Ability to assess nest predation risk in secondary hole-nesting birds: an experimental study. Oecologia 126:201–207 Robertson GJ (1998) Egg adoption can explain joint egg-laying in common eiders. Behav Ecol Sociobiol 43:289–296 Rohwer FC, Freeman S (1989) The distribution of conspecific nest parasitism in birds. Can J Zool 67:239–253 Rubenstein DI (1982) Risk, uncertainty and evolutionary strategies. In: King’s College Sociobiology Group (ed) Current problems in sociobiology. Cambridge University Press, Cambridge, pp 91–111 Ruxton GD (1999) Are attentive mothers preferentially parasitised? Behav Ecol Sociobiol 46:71–72 Sayler RD (1992) Ecology and evolution of brood parasitism in waterfowl. In: Batt BDJ, Afton AD, Anderson MG, Ankney CD, Johnson DH, Kadlec JA, Krapu GL (eds) Ecology and management of breeding waterfowl. University of Minnesota Press, Minneapolis, pp 290–322 Semel B, Sherman PW (2001) Intraspecific parasitism and nest-site competition in wood ducks. Anim Behav 61:787–803 SPSS (1997) SYSTAT for Windows, version 7.0. SPSS, Evanston, Ill. Steidl RJ, Hayes JP, Schauber E (1997) Statistical power analysis in wildlife research. J Wildl Manage 61:270–279 Switzer PV (1997) Past reproductive success affects future habitat selection. Behav Ecol Sociobiol 40:307–312 Thomas L, Juanes F (1996) The importance of statistical power analysis: an example from Animal Behaviour. Anim Behav 52:856–859 Toft CA, Shea PJ (1983) Detecting community-wide patterns: estimating power strengthens statistical inference. Am Nat 122:618–625 Wilson SF (1993) Use of wood duck decoys in a study of brood parasitism. J Field Ornithol 64:337–340 Yom-Tov Y (1980) Intraspecific nest parasitism in birds. Biol Rev 55:93–108 Yom-Tov Y (2001) An updated list and some comments on the occurrence of intraspecific nest parasitism in birds. Ibis 143:133–143 Zar JH (1996) Biostatistical analysis, 3rd edn. Prentice-Hall, Upper Saddle River, N.J. Zicus MC, Hennes SK (1989) Nest prospecting by common goldeneyes. Condor 91:807–812
