Nest parasitism in the barnacle goose: evidence from protein fingerprinting and microsatellites

Animal Behaviour 78 (2009) 167–174

Contents lists available at ScienceDirect

Animal Behaviour journal homepage: www.elsevier.com/locate/yanbe

Nest parasitism in the barnacle goose: evidence from protein fingerprinting and microsatellites Sofia Anderholm a, *, Rupert C. Marshall a, b,1, Henk P. van der Jeugd c, d, 2, Peter Waldeck a, Kjell Larsson e, 3, Malte Andersson a a

Department of Zoology, University of Gothenburg Institute of Biological, Environmental and Rural Sciences, Aberystwyth University SOVON Dutch Centre for Field Ornithology d Vogeltrekstation Dutch Centre for Avian Migration and Demography e Department of Biology, Gotland University b c

a r t i c l e i n f o Article history: Received 10 July 2008 Initial acceptance 3 October 2008 Final acceptance 20 April 2009 Published online 30 May 2009 MS. number: 08-00442R Keywords: adoption albumen fingerprinting barnacle goose best of bad job Branta leucopsis brood parasitism egg female alternative reproductive tactic microsatellite profiling

Geese are often seen as one of nature’s best examples of monogamous relationships, and many social pairs stay together for life. However, when parents and young are screened genetically, some chicks do not match their social parents. Although this has often been explained as adoption of foreign young after hatching, conspecific nest parasitism is another possibility. We used nondestructive egg albumen sampling and protein fingerprinting to estimate the frequency and success of nest parasitism in a Baltic Sea population of barnacle geese, Branta leucopsis. Among the 86 nests for which we had the most complete information, 36% were parasitized, and 12% of the eggs were parasitic. Almost 80% of the parasitic eggs were laid after the host began incubation. Hatching of these eggs was limited to the few cases where the host female incubated longer than normally because her own eggs failed to hatch. Conspecific nest parasitism in this population therefore seems mainly to be an alternative reproductive tactic of lower fitness than normal nesting. Comparison with DNA profiling of chicks (with 10–14 microsatellites) and other evidence confirmed the suitability of protein fingerprinting for analysis of nest parasitism. It can often provide more data than microsatellites, if eggs are albumen-sampled soon after being laid, before most losses occur. Ó 2009 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

In spite of their apparently monogamous family life, goose families are not always composed of an adult female, her mate and their common genetic offspring. Often one or more chicks has genetic parents other than the social ones, male as well as female. Whether these extrapair goslings represent adoption of young or conspecific nest parasitism (CNP) is harder to determine. In CNP, parasites lay eggs in the nests of other females (hosts) of the same species, without sharing incubation or rearing of young. This is

* Correspondence: S. Anderholm, Department of Zoology, University of Gothenburg, Box 463, SE-405 30 Gothenburg, Sweden. E-mail address: sofi[email protected] (S. Anderholm). 1 R. C. Marshall is now at the Institute of Biological Sciences, Aberystwyth University, Aberystwyth, Ceredigion SY23 3DA, U.K. 2 H. P. van der Jeugd is at the SOVON Dutch Centre for Field Ornithology, Rijksstraatweg 178, NL-6573 DG Beek-Ubbergen, The Netherlands and Vogeltrekstation Dutch Centre for Avian Migration and Demography, NIOO-KNAW, Heteren, The Netherlands. 3 K. Larsson is at the Department of Biology, Gotland University, SE-621 67 Visby, Sweden.

a widespread alternative reproductive tactic (Oliveira et al. 2008) among animals, being found in insects, fish and more than 200 species of birds (e.g. Brockmann 1993; Wisenden 1999; Yom-Tov 2001; Tallamy 2005). It is especially common among waterfowl, in extreme cases with parasitism in 95% of all nests (Semel & Sherman 1986). The advantages for the parasitic female seem obvious: by laying costs of offspring care on others, she can use the saved energy for investment in more eggs or higher survival (Brown & Brown 1998; Åhlund & Andersson 2001; Hartke et al. 2006). For the host, extra eggs can lower hatching success and future fecundity (Visser & Lessells 2001; Hanssen et al. 2003, 2005; Milonoff et al. 2004). However, if costs are low in precocial species, they may be overcome if extra young also add benefits such as dilution of predation risk (Eadie et al. 1988), or social dominance for a larger family (Loonen et al. 1999). There may also be inclusive fitness gains if host and parasite, or adopter and adoptee, are closely related (Andersson 1984, 2001; Lo´pez-Sepulcre & Kokko 2002). Host–parasite relatedness has been found in some insects and waterfowl (Andersson &

0003-3472/$38.00 Ó 2009 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.anbehav.2009.04.011

168

S. Anderholm et al. / Animal Behaviour 78 (2009) 167–174

Åhlund 2000; Loeb 2003; Roy Nielsen et al. 2006; Andersson & Waldeck 2007). To what extent adoptions are accidental or tactical choices is debated, as are fitness consequences of adoptions (Kalmbach 2006). Scanning of goose families with DNA techniques cannot distinguish between CNP and adoption of young. To identify CNP unambiguously, we used protein fingerprinting of egg albumen (Andersson & Åhlund 2001) together with microsatellite genotyping of hatchlings and adults in a Baltic Sea population of barnacle goose, Branta leucopsis, determining the maternity of eggs and chicks. In this precocial species with strong female natal philopatry, families often contain extrapair young (Choudhury et al. 1993; Larsson et al. 1995). However, it is not clear to what extent they represent CNP or adoption of young. We aimed to clarify the occurrence of CNP. In addition, combining protein fingerprinting of eggs with microsatellite typing of chicks enabled us to compare for the first time these different molecular methods for analysis of brood parasitism. METHODS Study Area and Population The barnacle goose colony on three small islands, Laus holmar off the east coast of Gotland, Sweden (57170 N, 18 450 E), has been studied for more than 20 years. Since the first breeding attempt in 1971 the colony has increased rapidly (Larsson et al. 1988; Larsson & Forslund 1994; Black et al. 2007). Adult and juvenile geese were captured in mid-July each year on moulting localities by a rounding-up technique. Captured birds were individually marked to allow individual identification at a distance. In some years, blood samples (0.15 ml) from captured adults were taken from the wing vein and stored in ethanol. In cases where blood could not be sampled, a growing wing feather was taken from adult birds at capture. There are no indications that marking or sampling affects the birds negatively. The present study of brood parasitism was done in 2003 on Storholmen (39.7 ha), one of the three islands hosting the goose colony. Of 1860 pairs on Laus holmar in 2003, 1400 pairs nested on Storholmen. Research was approved by the Swedish Board of Agriculture. Sampling and Nest Success During the breeding season (22 April to 14 June) we searched the area for new nests, checking them daily until 3 days after the last egg in the nest was added and incubation had started (except for 26–27 April and 16–23 May). We numbered new eggs with a nontoxic felt pen and sampled 0.3 ml of albumen through a 1 mm hole drilled 5–10 mm from the narrow end of the egg. The hole was sealed with cyanoacrylate glue (Loctite Superattack, Loctite Sweden AB, Go¨teborg, Sweden), with a droplet of activator (Loctite TAK PAK 7452) added to accelerate hardening. Samples were kept at 20  C until electrophoresis. This procedure does not affect egg hatchability (Andersson & Åhlund 2001; Waldeck et al. 2004). While we sampled albumen, the pair usually remained in the neighbourhood and returned to the nest soon after we left. Albumen proteins are genetically variable among females, and their electrophoretic band patterns are useful both for identification of parasitic eggs and for estimation of host–parasite relatedness (Andersson & Åhlund 2000, 2001). An egg can be sampled up to a week after being laid, until the albumen becomes too viscous and easily contaminated with yolk. We noted differences in colour and shape between eggs in the nest, measured egg length and breadth with callipers, and noted any disappearances of eggs, usually caused by gull or fox predation. Foxes do not usually occur on Laus holmar, but during the 2003

goose breeding season a fox resided on the study island. We scored a nest as successful if one or more eggs hatched. In cases where visited nests contained hatching eggs, we marked the emerging egg tooth with coloured nail varnish (IsaDora, www.isadora.com nontoxic) to be able to assign hatched chicks to individual eggs. We drew 25 ml of blood from the tarsal vein of hatched chicks. Hatching is synchronous and chicks usually stay in the nest for about 24 h. If they had already left we searched the nest for vascularized egg membranes, which can also be useful for microsatellite analyses. Albumen Analyses Albumen samples were analysed with isoelectric focusing (IEF) electrophoresis, using precast polyacrylamide gels with a fixed pH range (Immobiline DryPlate, GE Healthcare, www.gehealthcare. com see Andersson & Åhlund 2001 for details). Proteins applied to such gels migrate towards and come to rest at their isoelectric points, leading to high resolution of protein bands, which can also be compared between gels. These were stored dry at 20  C and rehydrated in a solution designed to maximize band numbers and sharpness (Andersson & Åhlund 2001). The pH ranges used were 4–7, 4.2–4.9 and 4.5–5.4. Gels were loaded with 7 ml of crude albumen and run on an Amersham Biosciences Multiphor II System (GE Healthcare) with Electrophoresis power supply EPS 3501 for 6–14 h at 3000 V, 1 mA, 3 W and a cooling temperature of 10  C. DNA Analyses DNA was isolated from blood, egg membranes, feathers and tissue using spin columns (Sigma, www.sigmaaldrich.com). We developed a panel of microsatellite primers to facilitate identification of true parents of hatchlings, using published primers from a range of geese, other waterfowl and passerines. PCR was done on a PTC-200 DNA Engine (MJ Research Inc., now Bio-rad Laboratoruis, www.biorad.com), comprising an initial denaturing step (94  C– 60 s) followed by 40 cycles of 94  C (35 s), the annealing temperature (49–63  C: see below; 40 s) and 72  C (40 s). A final 2 min at 72  C completed the run. Each 10 ml reaction contained 10–100 ng of sample DNA, 1 ml of 10 NH4 buffer (Bioline, www.bioline.com), 2 mM of MgCl, 0.2 mM of each dNTP, 0.75 mM of each primer and 1 unit Biotaq polymerase (Bioline). Products were visualized by running on a 6% denaturing acrylamide gel and stained with silver (Bassam et al. 1991). We tested in total 32 microsatellites, of which 14 were polymorphic and possible to score, making them suitable for parentage determination. Details of these 14 microsatellite loci are listed in Table 1. Three loci (Sfim5, Fields & Scribner 1997; Oxy6 and Oxy13 ˜ oz-Fuentes et al. 2005) produced three or four extra bands in Mun the same size range as the main product, making scoring of alleles unreliable. We overcame this by adding a further three or more bases (taken from the original sequence in Genbank www.ncbi. nlm.nih.gov) to both Forward and Reverse primers to increase specificity. This editing completely removed the additional products/bands, but it did not improve scoring of other markers with excessive stuttering. In two cases (Oxy6 and Oxy13) the original primer sequence had a ‘PIGtail’ (Brownstein et al. 1996) which was removed by adding the extra bases. Final primer sequences are listed in Table 2, ‘a’ denoting amendment from the original. Tested on a panel of 70 presumed unrelated adult barnacle geese (54 females and 16 males), these 14 loci had a probability of 0.9983 of excluding an unrelated individual from paternity/maternity if the other parent was known (Chakravarti & Li 1983; Marshall et al. 1998). Five loci (CAUD012, Sfim11, Smo11, Smo8 and Ase46) were linked with the other nine but are still useful for parentage

S. Anderholm et al. / Animal Behaviour 78 (2009) 167–174

169

Table 1 Characteristics of 14 polymorphic microsatellites for the barnacle goose, Branta leucopsis, tested on a panel of 70 presumed unrelated adults ID

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Locus

K

Sfim5a* Apl11 Oxy13a* TTUCG-5y APH02y Oxy6a* Smo1y APH16 TTUCG-1 CAUD012 Sfim11 Smo11 Smo8 Ase46

22 14 14 14 14 8 7 6 4 6 6 4 2 2

N

68 70 70 69 60 69 66 49 52 68 70 70 70 54

Readability

Heterozygosity

1¼easy, 5¼difficult

O

E

2 1 2 1 3 2 2 4 4 1 3 2 1 2

0.971 0.843 0.714 0.551 0.667 0.754 0.364 0.633 0.635 0.368 0.514 0.486 0.429 0.463

0.918 0.845 0.779 0.906 0.881 0.776 0.719 0.674 0.643 0.350 0.514 0.458 0.502 0.475

P, HWE test



0.847 0.649 0.510 <0.001 <0.001 0.197 <0.001 0.416 0.457 0.425 0.104 0.153 0.239 1.000

55 62 53 62 53 62 48 57 59 50 63 60 49 61

C

Source

Fields & Scribner 1997 Denk et al. 2004 ˜ oz-Fuentes et al. 2005 Mun Cathey et al. 1998 Maak et al. 2003 ˜ oz-Fuentes et al. 2005 Mun Paulus & Tiedemann 2003 Maak et al. 2003 Cathey et al. 1998 Huang et al. 2005 ¨ st et al. 2005 O Paulus & Tiedemann 2003 Paulus & Tiedemann 2003 Richardson et al. 2000

ID ¼ identity; numbers in bold indicate unlinked loci; K ¼ no. of alleles; N ¼ no. of individuals; O ¼ observed heterozygosity; E ¼ expected heterozygosity. HwE ¼ Harly– Weinberg equilibrium; C ¼ annealing temperature. * Amended primer; see Table 2 details. y Null alleles suspected.

determination. Three loci (Smo1, TTUCG-5, APH02 deviated significantly from Hardy–Weinberg equilibrium (Genepop on-theweb [3.1b] http://wbiomed.curtin.edu.au/genepop; Raymond & Rousset 1995), suggesting the presence of null alleles. Identification of Parasites by Protein Fingerprinting All bands consistently identifiable as present or absent on the gels were scored for all individuals. The female with most eggs in the clutch is here called the ‘host’. In goldeneye ducks, Bucephala clangula, the female laying most eggs incubated the clutch in all 12 cases tested (Andersson & Åhlund 2001). Eggs were scored as parasitic if differing from host eggs in three or more bands. Relaxing this criterion for parasitism to a difference in only one band did not change the number of parasitic eggs scored. In four cases, two different females laid the same number of eggs in a nest, but the assignment of either female as host or parasite does not affect our conclusions. Identification of Parasites by Microsatellites Social parents and chicks were compared at a minimum of 10 microsatellite loci. We identified parasites by allele mismatch: offspring were scored as parasitic when two or more loci (not affected by null alleles) mismatched with alleles of the social mother. The six offspring mismatching their social mother at one locus (and not because of null alleles) were only one or two alleles apart, indicating probable mutations. These mismatches were also most common at loci with a high number of alleles.

Table 2 Amended primer sequences Locus

Primer sequences used (50 –>30 )

Sfim5a

F: AATTTCATTCAAAATAAGACAAGA R: TGAATTTGCTCTGTTTGGTTTA F: ACCTGCAGTCGGGCGTCATCAAGATTCTGGGATTCAAAC R: *ACTCGTTAAAAATGGGCTCTTGGAAGG F: AGCTTGGCCAGTCGGGCGTCATCAGGAATCAATGAGATTAG R: TTGTATGGGGTGCTGCTTCTGAGTAG

Oxy6a Oxy13a

The original primer sequence is given with extra bases underlined and deleted bases with lines struck through them. F ¼ Forward; R ¼ Reverse. * The 4th additional base reversed a mismatch between the sequence in the original paper (G) and the sequence on Genbank (C).

Where DNA is also available from the male, mismatches may be observed between offspring and gander, caused by extrapair matings by the female or by a male involved in extrapair matings allowing his extrapair mate to lay eggs in his nest. However, we identified no such case of extrapair parentage (45 chicks in 14 nests). DNA samples were only available for some of the adults (54 females and 16 males), while the number of potential parents is far greater (1400 nesting pairs on the island). This, and strong female natal philopatry making some females closely related (Larsson & Forslund 1994; van der Jeugd et al. 2002) will render the likelihoodbased approach of CERVUS (Kalinowski et al. 2007) unable to assign parents with any confidence. However, the calculated CERVUS LOD score for the offspring and its social mother (the incubating female) can be used as an indication of whether she is a likely parent or not. A positive LOD score (log-likelihood value) indicates that she is more likely to be the true mother than a random individual from the population. A negative LOD score on the other hand indicates that the social mother is less likely to be the true mother than a random individual from the population (Marshall et al. 1998; Hill & Post 2005). Here, individuals scored as parasitic by mismatching analysis had negative LOD scores (11 loci unaffected by null alleles; error rate; 0.03). The cases with a mismatch one or two alleles apart from the incubating female’s, indicating mutations, were associated with positive LOD scores, supporting our scoring of parasites. The single individual deviating from this pattern mismatched on a locus affected by null alleles (both incubating female and offspring were heterozygous for different alleles) as well as on another locus, but had a positive LOD score when the null-allele locus was excluded from the CERVUS analysis. To be conservative, we did not exclude the incubating female as the true mother of this individual. Sample Sizes Altogether, 165 nests spread evenly over the island were sampled for a total of 672 eggs. To enable parasitism detection (which requires two or more eggs) as well as determination of which female laid most eggs in parasitized nests, we omitted 25 nests containing one or two albumen-sampled eggs, and only analysed nests with three or more sampled eggs. However, we included three cases where we had both albumen from two eggs and DNA for female and chicks in the nest, allowing comparison of

170

S. Anderholm et al. / Animal Behaviour 78 (2009) 167–174

the two methods (below). From colour rings we knew the identity of 107 of the incubating females and in 39 cases also of their males; 47 females and 16 males were also sampled for blood, and we had feathers from seven more females. Because some eggs were depredated, unfertilized or rotten, and some newly hatched chicks left the nest before being sampled, DNA was available for only 224 chicks (from blood, egg membranes, tissue samples from dead chicks and embryos), about one-third as many as the albumensampled eggs. Of the 140 nests analysed with protein fingerprinting, two became sites of egg dumping by several females, with total hatching failure for the females starting the nest. In one nest, four of 18 eggs hatched, none of which were laid by the starting female. The data set also includes nine nests depredated during the female’s laying period, four nests where daily checks were ended before clutch completion, nine nests where not all eggs laid during the female’s laying period were sampled, and another 30 nests that disappeared between clutch completion and hatching. The 25 nests not analysed with protein fingerprinting (see above) all disappeared before hatching. All these nests were excluded from further analyses, except for albumen band scoring and comparison with microsatellite analyses. In the 86 remaining nests, 26 contained a total of 39 eggs that were added after incubation began and we had ended daily checks of the nest; 21 of these eggs could not be sampled. However, eggs added after clutch completion (at least 3 days after the previous egg was added to the nest) are apparently always parasitic. All seven late eggs (in six nests) where the embryo could be microsatelliteprofiled and compared with the host female were parasitic. This was also the case for the 17 albumen-sampled eggs in the same category (14 nests). We therefore scored all late added eggs that could not be sampled as parasitic. We used SPSS 13.0 (SPSS, Chicago, IL, U.S.A.) and StatXact 6.0 (Cytel Inc., Cambridge, MA, U.S.A.) software for statistical analyses. All tests are two tailed, and means are given  SD. RESULTS Albumen Band Scoring In the analysis of albumen band patterns, 72 bands were consistently identifiable as present or absent on the four gel types. Of these, 40 redundant bands were excluded, showing the same occurrence among individuals as one or more other bands on the same or other gel types. The 32 remaining bands were distributed with on average 9.53  2.65 bands per individual and an average band frequency of 0.30  0.25 (N ¼ 187 individuals). Of all 140 nests analysed with protein fingerprinting, 37 contained eggs from multiple females, some from more than one additional female, increasing the number of individuals to 187. A few of these may be the same female appearing at more than one nest, but this is unlikely to bias markedly the band frequency estimation, the only purpose for which this larger data set is used. Differences in staining quality between gels prevented additional bands (beyond the 72 Mentioned above) from being scorable on some individual gels. These additional bands, however, can still be useful in comparing females with identical patterns of the 72 scored bands. Indications that eggs with identical protein patterns may still have come from different females are: (1) two eggs laid in the nest on the same day; (2) the total number of eggs exceeding the maximum a female usually lays (ca. 7); (3) eggs added late, several days after the host starts incubation; and (4) if there are large differences in egg morphology (see e.g. Yom-Tov 1980; MacWhirter 1989; Petrie & Møller 1991; Alisauskas & Ankney 1992). Our data

set contains 82 individuals with band patterns identical to one or more other individuals. These 82 individuals can be partitioned into 29 groups with identical band patterns within the group (but not between groups). Altogether, 36 of these females (in 16 groups) are probably different individuals based on the criteria above, but within the groups they cannot be told apart based on their patterns of standard bands. However, comparisons of additional bands, egg sizes, laying dates and locations suggest that one female laid a parasitic egg before she started laying in her own nest, and two females each laid parasitically in two different nests. However, since only 12% of the population was sampled, we cannot be certain that these are the same females laying in different nests, nor can we estimate the full occurrence of these behaviours. Parasitism Frequency A total of 86 nests survived until hatching (but did not necessarily produce any hatchlings) and had all eggs sampled except for some of the late added eggs, thus giving us nearly complete information on the nest. Of these 86, 31 nests were parasitized (36%), and 12% of the eggs (50 of 428 eggs) were parasitic if we count the late added, unsampled eggs as parasitic (21 eggs in 17 nests). These are probably the best estimates of parasitism frequency we can obtain from the protein-fingerprinting results. However, they are likely to underestimate parasitism, since some host–parasite pairs may have identical protein band patterns and thus go undetected. Comparing protein fingerprinting results with microsatellite profiling (see below) indicates that host and parasite sometimes do have indentical albumen band patterns, but not very often. Of the 31 parasitized nests, 17 had one parasitic egg, 10 had two, three had three and one had four parasitic eggs added to it (in addition to five host eggs, some of which hatched). In three of the nests with two or more parasitic eggs sampled, two parasites laid one egg each in the nest, whereas the other five such nests appeared to involve only one. There may, however, have been more parasites in a nest as late, unsampled (parasitic) eggs were not included. Using microsatellites we profiled 60 adults and 113 offspring, which comprised the 41 nests where we had DNA from at least one parent and one hatchling (96) or embryo (17). The incubating female was excluded as the mother, that is, an offspring was scored as parasitic, in nine cases (8%) among eight broods (20%). However, because of depredated, unfertilized or rotten eggs, not all offspring were DNA-profiled in all nests, leading to biased estimates. In particular, late added (and hence parasitic) eggs were often not included in the sample. Compensating for unsampled eggs, we estimated the total frequency of parasitic eggs to be 12%, a similar result as with protein fingerprinting (see Appendix). Whereas we can estimate the proportion of parasitic eggs in the 41 nests used in the microsatellite analysis, these data are not suitable for estimating the proportion of parasitized nests. This is because the 41 nests contained 74 timely eggs that were not microsatellite-profiled. Because some of these 74 eggs were probably parasitic, the proportion of parasitized nests may be underestimated. We therefore refrained from estimating the proportion of parasitized nests based on microsatellite data. Clutch Size For the 86 nests with complete information, mean clutch size in single-female nests was 4.45  0.90 (N ¼ 55), significantly lower than the mean for clutches including parasitic eggs: 5.90  1.54 (N ¼ 31; t test: t84 ¼ 4.8, P < 0.01). Excluding parasitic eggs, mean host clutch size was 4.49  1.24 (N ¼ 31), not significantly different from single-female nests (t test: t84 ¼ 0.71, P ¼ 0.48). Among

S. Anderholm et al. / Animal Behaviour 78 (2009) 167–174

171

nests that were parasitized during the host’s laying period, the host’s clutch size was 4.11  1.62 (N ¼ 9), not significantly different from the 4.43  0.95 (N ¼ 77) eggs in nests not parasitized during the host’s laying period (non-normal data; two-sample permutation test: P ¼ 0.39). The largest clutch laid by a single female was six eggs (nine nests).

hatched chicks were parasitic. The corresponding estimates for protein fingerprinting were five parasitic hatchlings in 83 broods (6%) and six of 260 hatchlings (2%). Among the 28 parasitized nests in the sample of 140 nests where we had full information on laying order, two nests were taken over by another female after the first female laid one or more eggs.

Timing of Parasitism

Protein Fingerprinting versus Microsatellite Profiling

Parasitic eggs were often late: 39 of 50 eggs in the 31 parasitized clutches with complete information were laid after the host laid her last egg. Hatching success for the late eggs was low; only three of all 39 late added parasitic eggs hatched (two nests), but only because the host female’s own eggs failed to hatch, making her incubate for longer than normally. None of the 29 late added eggs hatched from nests where the host’s own eggs hatched (25 nests). Comparing all eggs that survived up to hatching in 24 parasitized nests where we had information on hatching status of all eggs (based on nests as independent sample units), we found that hatching success was lower for parasitic eggs (Fig. 1). The mean percentage of hatched eggs per nest was 16% for parasitic eggs and 70% for host eggs (Wilcoxon test: T ¼ 219, N ¼ 21, three ties, P < 0.01). For parasitic eggs laid during the host’s laying sequence, the mean percentage of eggs hatched per nest was 50%, compared to 78% for host eggs. This difference is not significant, but the sample is too small to allow a definite conclusion (Wilcoxon test: T ¼ 16, N ¼ 6, one tie, P ¼ 0.3). Based on total egg numbers instead of nest means, hatching success among the 24 nests was 56% for parasitic eggs laid during the host’s sequence, 4% for late parasitic eggs, and 71% for host eggs (Fig. 1). As a consequence of the late added eggs, the frequency of hatchlings from parasitic eggs was low. Of all 113 microsatelliteprofiled chicks in the 41 broods mentioned (see Parasitism frequency), 17 hatchlings (in 12 nests) disappeared before we could sample them. In the 29 nests where we had DNA from all hatched chicks, two (7%) produced parasitic offspring and two of 80 (3%)

Because of high nest predation (46% of 140 nests were partly or completely depredated), 30% of the albumen-sampled eggs were lost before hatching, preventing DNA sampling. Among 104 offspring in 40 clutches where we had both egg albumen and DNA from blood, tissue or egg membranes, both methods independently identified five offspring (one in each of five different clutches) as parasitic and 98 offspring as belonging to the host. There was disagreement in one case: microsatellite profiling identified one parasitic event (the chick mismatching the female at six loci) that was not detected with protein fingerprinting. Identification of the female with the most eggs in the nest as the host was verified in each of 10 nests with multiple females where we had the DNA profile of the incubating female. In all cases the incubating female’s DNA was compatible with the DNA profile of the chicks that hatched from eggs laid by the female identified as the host by protein fingerprinting. These results add to the evidence that the female with most eggs in the nest is usually the one who incubates the clutch (Andersson & Åhlund 2001).

100

Percentage of eggs

80

60

DISCUSSION Nest Parasitism or Chick Adoption? Whether extrapair goslings in families of geese are usually the result of added foreign eggs or of chick adoptions is often unclear. Both alternatives are known to occur (Choudhury et al. 1993; Forslund & Larsson 1995; Larsson et al. 1995), but the difficulty of distinguishing between them has prevented estimation of the frequency of each category. Using protein fingerprinting combined with microsatellite profiling we found that nest parasitism is common in this population of barnacle geese, accounting for 12% of eggs and occurring in about 36% of nests. However, among hatchlings, this percentage was reduced to 2–3%, and only 6–7% of hatched broods contained parasitic goslings (protein fingerprinting and microsatellites, respectively). In a previous study of this population, Larsson et al. (1995) found that 17% of young at fledging were extrapair offspring. The low parasitism frequency found here at hatching indicates that a considerable proportion of the 17% extrapair fledglings may have been adoptions. However, CNP can vary greatly between years in goose populations (Lank et al. 1990; Larsson et al. 1995).

40 Frequency and Timing of Parasitism

20

0

Host eggs

Timely parasitic eggs

Late parasitic eggs

Figure 1. Mean hatching success (%) of eggs in 24 parasitized nests with known hatching status of all eggs (depredated eggs excluded). The figure shows the outcome (grey: unhatched; white: hatched) of host eggs (N ¼ 99), timely parasitic eggs (laid within the host’s laying sequence, N ¼ 9) and late parasitic eggs (laid after the host’s last egg, N ¼ 26).

Most parasitic eggs were added late, after the host had begun incubation. In the lesser snow goose, Chen caerulescens, hatchability fell to zero for eggs added 4 days or more after incubation started (Davies & Cooke 1983). In our sample only three of 50 late eggs hatched. All these came from nests where the host female’s own eggs for unknown reasons failed to hatch, with the consequence that the host female incubated the clutch for longer. This suggests that nest parasitism in barnacle geese is often a ‘best of a bad job’ tactic (Forslund & Larsson 1995), rather than an alternative tactic of as high or higher fitness than normal nesting, as in some other waterfowl (Åhlund & Andersson 2001; Roy Nielsen et al. 2006). In snow geese, Lank et al. (1990) estimated that a minimum of 6.6% of

172

S. Anderholm et al. / Animal Behaviour 78 (2009) 167–174

eggs were parasitic. As in our barnacle goose population, many of the parasitic eggs were added after the host started incubation, and the frequency of parasitic hatchlings was 5.6% (Lank et al. 1989). Parasitic eggs laid after the host female’s laying sequence, with low hatching success as a consequence, were common also in a population of moorhens, Gallinula chloropus. These eggs were attributed to a salvage strategy where females that had their nests destroyed turned to parasitism (McRae 1998). Tactics and Costs of Parasitism Who are the parasites and the hosts, and what are the benefits and costs associated with these roles? The identity of a parasite can sometimes be revealed by protein fingerprinting combined with individual marking. In goldeneye ducks, this approach showed that some females increased their reproductive success by laying parasitically before starting a nest of their own (Åhlund & Andersson 2001). In this population of barnacle geese, we only found one case where a female probably laid parasitically in addition to having a nest of her own, and two cases where females laid parasitically in more than one nest. However, our chances of detecting such tactics were low, since we only sampled 12% of all nests in the population and only analysed nests containing more than two eggs. Because some females had identical band patterns, there is also a possibility of mistaken identity, even if place and timing of laying suggest otherwise. The timing of parasitism in barnacle geese suggests that it is not a superior tactic in most cases, but an alternative that may provide at least some reproductive success if a female has no nest of her own, for example because of nest predation or competition over nest sites. Nest take-over, which might reflect competition over nest sites (Robertson 1998; Waldeck & Andersson 2006) or males, was found in two of 28 mixed clutches. A parasitic egg added late during the host’s incubation period probably adds little extra cost of incubation for the host, and hatching failure of late parasitic eggs precludes costs of rearing extra young. However, parasitism can be costly for the host, as was evident in the two nests that became targets of egg dumping by several females (see Methods). Unhatched parasitic eggs might also delay the nest exodus for the family, reducing energy reserves of goslings and parents. Another possible cost is reduction of the host’s own clutch when it is parasitized during her laying sequence (Andersson & Eriksson 1982; Andersson 1984; Lyon 1998). This reduction is here estimated at 0.3 eggs (7%) and is not statistically significant. In other species, additional eggs are important. In eider ducks, Somateria mollissima, with similar clutch size and breeding habitat as barnacle geese, females with nests to which eggs were added in a randomized experiment laid fewer eggs (Erikstad & Bustnes 1994). In Barrow’s goldeneyes, Bucephala islandica, parasitic eggs increase rates of nest desertion (Jaatinen et al., in press). Other evidence that parasitism may be costly for the host in geese is that parasitism attempts occasionally lead to severe fights and attacks by the host female on the parasite (Forslund & Larsson 1995). Protein Fingerprinting versus Microsatellite Profiling Estimates of CNP in waterfowl have often been based on determination of daily addition of more than one egg per day, and on differences in egg morphology (e.g. Lank et al. 1989; Robertson et al. 1992; Po¨ysa¨ et al. 2001). Alone, such methods may severely underestimate parasitism (Ådahl et al. 2004; Gronstol et al. 2006), but suitable combinations can detect much of it (Eadie et al., in press). In spite of daily nest visits during the laying period, with few days missed, we only once found two eggs added to a nest on the same day. Albumen sampling on the other hand identified 32 parasitic eggs laid during the host’s laying period.

Careful sampling of albumen for protein fingerprinting does not reduce egg hatchability (Andersson & Åhlund 2000; Waldeck et al. 2004), and it allows large sample sizes because eggs can be sampled soon after being laid, before predation or other egg losses occurs. Tested in 21 colour-ringed and video-filmed goldeneye females, all of which had unique albumen band patterns, protein fingerprinting correctly identified each of 100 eggs from the 21 females, many of them being laid parasitically (Andersson & Åhlund 2001). We have no means here of matching albumen protein patterns to an individual female’s tissue or DNA to identify her; observations of egg laying by individually marked females (Åhlund & Andersson 2001) were not possible. Limited variation among albumen proteins can render the resolution of individuals too low to reveal all parasitic events. This may be the case in our population, which was recently founded by relatively few individuals, followed by gene flow from other larger populations (Larsson et al. 1988). As therefore expected, many females have similar or identical albumen band patterns, and DNA minisatellite variability is also reduced compared to populations less affected by founding events (Tegelstro¨m & Sjo¨berg 1995). In addition there is strong female philopatry, 80% of females returning to their birth colony, and related females often nesting near each other (van der Jeugd et al. 2002). Microsatellite and protein analysis performed similarly as regards detection of parasitism. In 98 of 104 offspring analysed by both methods, neither method suggested parasitism, indicating that the proportion of false positives was low. On the other hand, false negatives may occur if host and parasite are often genotypically similar, as seems to be the case here. There was agreement between microsatellites and protein fingerprinting in five of the parasitic events. In addition, microsatellite profiling identified one chick as parasitic where protein fingerprinting found none. Although this may seem a high error rate for protein fingerprinting, in reality it is lower. In 17 cases we had albumen samples from late laid eggs, which being late were almost certainly parasitic. In all 17 cases we blind-scored the egg as parasitic with protein fingerprinting, and only thereafter became aware of the egg’s status as late. Including these late parasitic eggs, protein fingerprinting therefore correctly identified 18 of 19 parasitic eggs. A major advantage of protein fingerprinting is that it usually gives much larger samples than DNA methods based on offspring blood or tissue samples. Here, a predation rate of almost 30% of all albumen-sampled eggs, and difficulties of identifying and sampling blood from the incubating female, reduced the 140 nests available for protein fingerprinting to only 41 broods that could be used for microsatellite analysis. This includes membranes from hatched eggs that were taken when chicks could not be sampled for blood. Furthermore, sampling DNA at hatching gives no information on, for example, the laying order of eggs. Even if eggs are labelled with individual numbers directly after laying and survive until hatching, it may be difficult to assign hatched chicks or egg membranes to individual eggs when hatching takes place synchronously, within a single day. Albumen sampling soon after the egg is laid also avoids most of the bias that can arise with DNA techniques because of nest predation or other hatching failure that may affect host eggs and parasitic eggs differently, before blood is sampled from chicks (also see Andersson & Åhlund 2001). Protein fingerprinting is therefore suitable for analysis of conspecific nest parasitism. Conclusions Adoption of young is not the only source of extrapair offspring in established families of barnacle geese. Conspecific nest parasitism is a common alternative reproductive tactic in this population, occurring in an estimated 36% of all nests. About 80% of the

S. Anderholm et al. / Animal Behaviour 78 (2009) 167–174

parasitic eggs, however, were laid too late during the host female’s incubation period to hatch. Our results therefore show that conspecific nest parasitism in the barnacle goose is usually not successful, but a ‘best of a bad job’ tactic (Forslund & Larsson 1995). Comparison with the results from microsatellites confirms that protein fingerprinting is a suitable method for analysis of nest parasitism, often providing much larger samples. Acknowledgments We thank the Sheffield Molecular Genetics Facility (NERC, U.K.) and S. Maak for primer samples, the Swedish Research Council (K.L. M.A.) the Wallenberg foundation (M.A.) and Wilhelm and Marina Lundgren’s foundation (PW) for financial support of this work. We also thank the referees for helpful suggestions on the manuscript. References ¨ m, J., Ruxton, G. D., Arnold, K. E., Begg, A. & Begg, T. 2004. Can Ådahl, E., Lindstro intraspecific brood parasitism be detected using egg morphology only? Journal of Avian Biology, 35, 360–364. Åhlund, M. & Andersson, M. 2001. Female ducks can double their reproduction. Nature, 414, 600–601. Alisauskas, R. T. & Ankney, C. D. 1992. The cost of egg laying and its relationship to nutrient reserves in waterfowl. In: Ecology and Management of Breeding Waterfowl (Ed. by B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec & G. L. Krapu), pp. 30–61. Minneapolis: University of Minnesota Press. Andersson, M. 1984. Brood parasitism within species. In: Producers and Scroungers: Strategies of Exploitation and Parasitism (Ed. by C. J. Barnard), pp. 195–228. London: Croom Helm. Andersson, M. 2001. Relatedness and the evolution of conspecific brood parasitism. American Naturalist, 158, 599–614. Andersson, M. & Åhlund, M. 2000. Host–parasite relatedness shown by protein fingerprinting in a brood parasitic bird. Proceedings of the National Academy of Sciences, U.S.A., 97, 13188–13193. Andersson, M. & Åhlund, M. 2001. Protein fingerprinting: a new technique reveals extensive conspecific brood parasitism. Ecology, 82, 1433–1442. Andersson, M. & Eriksson, M. O. G. 1982. Nest parasitism in goldeneyes Bucephala clangula: some evolutionary aspects. American Naturalist, 120, 1–16. Andersson, M. & Waldeck, P. 2007. Host–parasite kinship in a female-philopatric bird population: evidence from relatedness trend analysis. Molecular Ecology, 16, 2797–2806. Bassam, B. J., Caetanoanolles, G. & Gresshoff, P. M. 1991. Fast and sensitive silver staining of DNA in polyacrylamide gels. Analytical Biochemistry, 196, 80–83. Black, J. M., Prop, J. & Larsson, K. 2007. Wild Goose Dilemmas: Population Consequences of Individual Decisions in Barnacle Geese. Groningen: Branta Press. Brockmann, H. J. 1993. Parasitizing conspecifics: comparisons between hymenoptera and birds. Trends in Ecology & Evolution, 8, 2–4. Brown, C. R. & Brown, M. B. 1998. Fitness components associated with alternative reproductive tactics in cliff swallows. Behavioral Ecology, 9, 158–171. Brownstein, M. J., Carpten, J. D. & Smith, J. R. 1996. Modulation of non-templated nucleotide addition by taq DNA polymerase: primer modifications that facilitate genotyping. Biotechniques, 20, 1004–1010. Cathey, J. C., DeWoody, J. A. & Smith, L. M. 1998. Microsatellite markers in Canada geese (Branta canadensis). Journal of Heredity, 89, 173–175. Chakravarti, A. & Li, C. 1983. The effect of linkage on paternity calculations. In: Inclusion Probabilities in Parentage Testing (Ed. by R. H. Walker), pp. 411–422. Arlington, Virginia: American Association of Blood Banks. Choudhury, S., Jones, C. S., Black, J. M. & Prop, J. 1993. Adoption of young and intraspecific nest parasitism in barnacle geese. Condor, 95, 860–868. Davies, J. C. & Cooke, F. 1983. Intraclutch hatch synchronization in the lesser snow goose. Canadian Journal of Zoology, 61, 1398–1401. Denk, A. G., Gautschi, B., Carter, K. & Kempenaers, B. 2004. Seven polymorphic microsatellite loci for paternity assessment in the mallard (Anas platyrhynchos). Molecular Ecology Notes, 4, 506–508. Eadie, J. M., Kehoe, F. P. & Nudds, T. D. 1988. Pre-hatch and post-hatch brood amalgamation in North American Anatidae: a review of hypotheses. Canadian Journal of Zoology, 66, 1709–1721. Eadie, J. M., Smith, J. N. M., Zadworny, D., Kuhnlein, U. & Cheng, K. 2009. Probing parentage in parasitic goldeneyes: an evaluation of methods to detect brood parasitism. Journal of Avian Biology, 40. Erikstad, K. E. & Bustnes, J. O. 1994. Clutch size determination in common eiders: an egg removal and egg addition experiment. Journal of Avian Biology, 25, 215–218. Fields, R. L. & Scribner, K. T. 1997. Isolation and characterization of novel waterfowl microsatellite loci: cross-species comparisons and research applications. Molecular Ecology, 6, 199–202.

173

Forslund, P. & Larsson, K. 1995. Intraspecific nest parasitism in the barnacle goose: behavioural tactics of parasites and hosts. Animal Behaviour, 50, 509–517. Gronstol, G., Blomqvist, D. & Wagner, R. H. 2006. The importance of genetic evidence for identifying intra-specific brood parasitism. Journal of Avian Biology, 37, 197–199. Hanssen, S. A., Erikstad, K. E., Johnsen, V. & Bustnes, J. O. 2003. Differential investment and costs during avian incubation determined by individual quality: an experimental study of the common eider (Somateria mollissima). Proceedings of the Royal Society of London, Series B, 270, 531–537. Hanssen, S. A., Hasselquist, D., Folstad, I. & Erikstad, K. E. 2005. Cost of reproduction in a long-lived bird: incubation effort reduces immune function and future reproduction. Proceedings of the Royal Society of London, Series B, 272, 1039–1046. Hartke, K. M., Grand, J. B., Hepp, G. R. & Folk, T. H. 2006. Sources of variation in survival of breeding female wood ducks. Condor, 108, 201–210. Hill, C. E. & Post, W. 2005. Extra-pair paternity in seaside sparrows. Journal of Field Ornithology, 76, 119–126. Huang, Y. H., Tu, J. F., Cheng, X. B., Tang, B., Hu, X. X., Liu, Z. L., Feng, J. D., Lou, Y. K., Lin, L., Xu, K., Zhao, Y. L. & Li, N. 2005. Characterization of 35 novel microsatellite DNA markers from the duck (Anas platyrhynchos) genome and cross-amplification in other birds. Genetics Selection Evolution, 37, 455–472. ¨ st, M., Waldeck, P. & Andersson, M. In press. Clutch desertion in Jaatinen, K., O Barrow’s goldeneyes (Bucephala islandica): effects of non-natal eggs, the environment and host female characteristics. Annales Zoologici Fennici. van der Jeugd, H. P., van der Veen, I. T. & Larsson, K. 2002. Kin clustering in barnacle geese: familiarity or phenotype matching? Behavioral Ecology, 13, 786–790. Kalinowski, S. T., Taper, M. L. & Marshall, T. C. 2007. Revising how the computer program CERVUS accommodates genotyping error increases success in paternity assignment. Molecular Ecology, 16, 1099–1106. Kalmbach, E. 2006. Why do goose parents adopt unrelated goslings? A review of hypotheses and empirical evidence, and new research questions. Ibis, 148, 66–78. Lank, D. B., Mineau, P., Rockwell, R. F. & Cooke, F. 1989. Intraspecific nest parasitism and extra-pair copulation in lesser snow geese. Animal Behaviour, 37, 74–89. Lank, D. B., Rockwell, R. F. & Cooke, F. 1990. Frequency-dependent fitness consequences of intraspecific nest parasitism in snow geese. Evolution, 44, 1436–1453. Larsson, K. & Forslund, P. 1994. Population dynamics of the barnacle goose Branta leucopsis in the Baltic area: density-dependent effects on reproduction. Journal of Animal Ecology, 63, 954–962. Larsson, K., Forslund, P., Gustafsson, L. & Ebbinge, B. S. 1988. From the high arctic to the baltic: the successful establishment of a barnacle goose Branta leucopsis population on Gotland, Sweden. Ornis Scandinavica, 19, 182–189. ¨ m, H. & Forslund, P. 1995. Intraspecific nest parasitism and Larsson, K., Tegelstro adoption of young in the barnacle goose: effects on survival and reproductive performance. Animal Behaviour, 50, 1349–1360. Loeb, M. L. G. 2003. Evolution of egg dumping in a subsocial insect. American Naturalist, 161, 129–142. Loonen, M. J. J. E., Bruinzeel, L. W., Black, J. M. & Drent, R. H. 1999. The benefit of large broods in barnacle geese: a study using natural and experimental manipulations. Journal of Animal Ecology, 68, 753–768. Lo´pez-Sepulcre, A. & Kokko, H. 2002. The role of kin recognition in the evolution of conspecific brood parasitism. Animal Behaviour, 64, 215–222. Lyon, B. E. 1998. Optimal clutch size and conspecific brood parasitism. Nature, 392, 380–383. Maak, S., Wimmers, K., Weigend, S. & Neumann, K. 2003. Isolation and characterization of 18 microsatellites in the Peking duck (Anas platyrhynchos) and their application in other waterfowl species. Molecular Ecology Notes, 3, 224–227. McRae, S. B. 1998. Relative reproductive success of female moorhens using conditional strategies of brood parasitism and parental care. Behavioral Ecology, 9, 93–100. MacWhirter, R. B. 1989. On the rarity of intraspecific brood parasitism. Condor, 91, 485–492. Marshall, T. C., Slate, J., Kruuk, L. E. B. & Pemberton, J. M. 1998. Statistical confidence for likelihood-based paternity inference in natural populations. Molecular Ecology, 7, 639–655. ¨ ysa¨, H., Runko, P. & Ruusila, V. 2004. Brood rearing costs affect Milonoff, M., Po future reproduction in the precocial common goldeneye Bucephala clangula. Journal of Avian Biology, 35, 344–351. ˜ oz-Fuentes, V., Gyllenstrand, N., Negro, J. J., Green, A. J. & Vila`, C. 2005. MicroMun satellite markers for two stifftail ducks: the white-headed duck, Oxyura leucocephala, and the ruddy duck, O-jamaicensis. Molecular Ecology Notes, 5, 263–265. Oliveira, R. F., Taborsky, M. & Brockmann, H. J. 2008. Alternative Reproductive Tactics: an Integrative Approach. New York: Cambridge University Press. ¨ st, M., Vitikainen, E., Waldeck, P., Sundstro ¨ m, L., Lindstro ¨ m, K. A. I., O Hollmen, T., Franson, J. C. & Kilpi, M. 2005. Eider females form non-kin broodrearing coalitions. Molecular Ecology, 14, 3903–3908. Paulus, K. B. & Tiedemann, R. 2003. Ten polymorphic autosomal microsatellite loci for the eider duck Somateria mollissima and their cross-species applicability among waterfowl species (Anatidae). Molecular Ecology Notes, 3, 250–252. Petrie, M. & Møller, A. P. 1991. Laying eggs in others nests: intraspecific brood parasitism in birds. Trends in Ecology & Evolution, 6, 315–320. ¨ ysa¨, H., Runko, P., Ruusila, V. & Milonoff, M. 2001. Identification of parasitized Po nests by using egg morphology in the common goldeneye: an alternative to blood sampling. Journal of Avian Biology, 32, 79–82.

174

S. Anderholm et al. / Animal Behaviour 78 (2009) 167–174

Raymond, M. & Rousset, F. 1995. Genepop (Version-1.2): population-genetics software for exact tests and ecumenicism. Journal of Heredity, 86, 248–249. Richardson, D. S., Jury, F. L., Dawson, D. A., Salgueiro, P., Komdeur, J. & Burke, T. 2000. Fifty Seychelles warbler (Acrocephalus sechellensis) microsatellite loci polymorphic in Sylviidae species and their cross-species amplification in other passerine birds. Molecular Ecology, 9, 2226–2231. Robertson, G. J. 1998. Egg adoption can explain joint egg-laying in common eiders. Behavioral Ecology and Sociobiology, 43, 289–296. Robertson, G. J., Watson, M. D. & Cooke, F. 1992. Frequency, timing and costs of intraspecific nest parasitism in the common eider. Condor, 94, 871–879. Roy Nielsen, C. L., Semel, B., Sherman, P. W., Westneat, D. F. & Parker, P. G. 2006. Host–parasite relatedness in wood ducks: patterns of kinship and parasite success. Behavioral Ecology, 17, 491–496. Semel, B. & Sherman, P. W. 1986. Dynamics of nest parasitism in wood ducks. Auk, 103, 813–816. Tallamy, D. W. 2005. Egg dumping in insects. Annual Review of Entomology, 50, 347–370. ¨ m, H. & Sjo ¨ berg, G. 1995. Introduced Swedish Canada geese (Branta Tegelstro canadensis) have low levels of genetic variation as revealed by DNA fingerprinting. Journal of Evolutionary Biology, 8, 195–207. Visser, M. E. & Lessells, C. M. 2001. The costs of egg production and incubation in great tits (Parus major). Proceedings of the Royal Society of London, Series B, 268, 1271–1277. Waldeck, P. & Andersson, M. 2006. Brood parasitism and nest takeover in common eiders. Ethology, 112, 616–624. ¨ st, M. & Andersson, M. 2004. Brood parasitism in a popWaldeck, P., Kilpi, M., O ulation of common eider (Somateria mollissima). Behaviour, 141, 725–739. Wisenden, B. D. 1999. Alloparental care in fishes. Reviews in Fish Biology and Fisheries, 9, 45–70. Yom-Tov, Y. 1980. Intraspecific nest parasitism in birds. Biological Reviews, 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.

APPENDIX : COMPENSATION FOR SAMPLING BIAS OF LATE EGGS We could not sample all eggs for microsatellite profiling, and correct here for the bias that will otherwise arise in the estimates of parasitism frequency. Among the 180 timely eggs laid in the 41 nests included in the microsatellite analyses, 20 eggs became depredated, 49 did not hatch (10 of which were sampled) and 111 hatched (96 of which were sampled). Among the 106 sampled chicks, two were parasitic. We here assume the proportion of parasitic eggs to be the same among the remaining 74 not sampled eggs, that is, 74  2/106 ¼ 1.4 parasitic egg. Thus, the corrected number of parasitic eggs among

the timely eggs is estimated at 2 þ 1.4 ¼ 3.4, that is, 1.9% of the 180 timely eggs in those nests. All 20 late eggs (laid > 3 days after the previous host egg) analysed by microsatellites (N ¼ 3), protein fingerprinting (13) or both (4) were parasitic, as expected (see main text). Below, we therefore assume that all late eggs are parasitic. Of the 21 late eggs, 19 were still present in the nests at hatching (although all failed to hatch). Another two late eggs were depredated before hatching. It is possible that some late eggs were added after we ended daily nest checks, but were depredated before we resumed checks at hatching (‘x’ unknown late eggs). Assuming that predation rates are identical for timely and late eggs, we can compensate for the unknown late eggs that disappeared before hatching. Then, the number of late eggs disappearing before hatching is the number of known depredated, late eggs (2) plus the unknown number of late eggs (‘x’), i.e. ED ¼ 2 þ x. Similarly, the total number of late eggs is the sum of known late eggs (21) and the unknown number of late eggs (‘x’), i.e. EL ¼ 21 þ x. Since 1/9 of all timely eggs were depredated (20/180), we have ED/EL ¼ 1/9, and hence (2 þ x)/(21 þ x) ¼ 1/9, which gives x ¼ 3/8 ¼ 0.4. Thus, the total number of late parasitic eggs in the nests subjected to microsatellite analysis is estimated at 21 þ x ¼ 21.4 out of a total of 180 þ 21.4, i.e. 10.6% late parasitic eggs in these nests. The total percentage of parasitic eggs, timely or late, is estimated at 21.4 þ 3.4 ¼ 24.8 out of 201.4, or 12.3%. Late eggs were undersampled also with protein fingerprinting. Among the 389 timely eggs, 11 were parasitic (2.8%). There were also 39 late eggs, making up 9.1% among the 389 þ 39 eggs in these nests. Counting the late eggs as parasitic and adding them to the 11 timely parasitic eggs gives a total parasitism frequency of (11 þ 39)/ 428 ¼ 50/428 ¼ 11.7%. These estimates of late parasitism assume that all late eggs were parasitic. Perhaps this is sometimes not so, in which case our estimates of late parasitism will be slightly exaggerated, but probably not by much, as all 20 tested late eggs were parasitic. In conclusion, microsatellites as well as protein fingerprinting show that parasitism during the host’s laying sequence was rare, comprising only 2–3% of the timely eggs, whereas late parasitism was relatively common, estimated at 9–11% of all eggs.