Extra-Pair Paternity in the Common Murre Free

Abstract

The Common Murre (Uria aalge; Common Guillemot in Europe) is a colonial, socially monogamous seabird. Common Murres typically breed in physical contact with one or more neighbors at densities of up to 70 pairs m⁻² (Gaston and Jones 1998). The Common Murre was among the first bird species in which routine extra-pair copulation behavior was observed and described (Norrevang 1958). Subsequent studies of this species showed that extra-pair copulations can occur in two ways: they can be either forced on females by males, or more rarely they can be initiated by females (Birkhead et al. 1985, 1987, Hatchwell 1988). Forced extra-pair copulation attempts were much less likely to result in cloacal contact than extra-pair copulations solicited by females, most of which were successful (Hatchwell 1988). In two previous studies, the proportion of all successful copulations that were extra-pair was 2–5% (Birkhead et al. 1985, Hatchwell 1988). Hatchwell (1988) also showed that the likelihood of either a female or a male Common Murre being involved in an extra-pair copulation increased with local breeding density. In addition to breeding at high densities, Common Murres are typical of many marine birds in that they have enduring pair bonds (Gaston and Jones 1998) and biparental care that is essential for successful breeding; each partner takes on an almost equal share of incubation, brooding, and chick-feeding (Gaston and Jones 1998). The levels of extra-pair paternity in bird species where the combined efforts of each parent are essential for chick-rearing are typically rather low (Birkhead and Møller 1996).

The aim of this study was to determine the level of extra-pair paternity of Common Murres at the same colony at which behavioral observations of extra-pair copulation had previously been conducted (Birkhead 1978, Hatchwell 1988).

Methods

The study was conducted at Skomer Island, Wales (51°45′N, 5°17′W) during three breeding seasons (1996–1998). At a single subcolony known as the Amos, we caught a sample of adult birds and their single chick and individually marked them with Darvic bands engraved with large numbers and a British Trust for Ornithology metal band. From each bird captured, we took a sample of blood for paternity analysis. Blood was stored in 96% ethanol for later analysis. Common Murres do not build a nest and chicks are relatively mobile, so for several successive days after marking chicks we observed them from a distance of 50 m using a telescope in order to match adults with their chicks. Chicks seen with the same adults, or fed by the same adults on two or more days were assumed to be the offspring of those adults. This criterion was necessary because Common Murre chicks are occasionally brooded (but rarely fed) by other adults, referred to as alloparents (Birkhead and Nettleship 1984). We obtained data for 42 chicks with both putative parents, and 45 chicks with one parent (35 with the putative father, 10 with the putative mother). In 21 instances, chicks from the same putative male parent were analyzed in successive years (16 in two years, 5 in all three years of the study). Common Murres are sexually monomorphic, and color marked adults were sexed either by observing copulation early in the season, or by molecular methods, as follows.

Sex Identification

Sex identification is based on the amplification of a portion of two sex-linked genes (Griffiths et al. 1998). The first gene is CHD-Z and is located on the Z chromosome (Griffiths and Korn 1997), is common to both sexes (male ZZ; female WZ), and forms a positive control. The second gene, CHD-W, is from the female-specific W chromosome and is used to identify sex. The polymerase chain reaction (PCR) amplification was carried out in a volume of 10 μl where the final reaction conditions were 50 mM KCl, 10 mM Tris.Cl pH9 (25°C), 1.5 mM MgCl₂, 0.1% Triton X-100, 200 μM of each dNTP, 1 μg of each primer, and 0.15 units Taq polymerase (Promega, Madison, Wisconsin). Between 50 and 250 ng of genomic DNA was used as template and the PCR was performed in a Biometra thermal cycler. An initial denaturation step at 94°C for 1 min 30 sec was followed by 30 cycles at 48°C for 45 sec, 72°C for 45 sec, and 94°C for 30 sec. A final cycle of 48°C for 1 min and 72°C for 5 min finished the program. The PCR products were separated by electrophoresis in a 3% agarose gel stained with ethidium bromide. Although this method has been comprehensively tested (Griffiths et al. 1998), we confirmed its specific accuracy by correctly sexing all 20 females and 22 males which we had previously sexed by observation of copulations.

Paternity Analyses

Five to ten μg of purified Hae III-digested DNA was loaded on to 20 × 40 cm 0.8% agarose gels in 1 × TBE buffer. The gels were electrophoresed at 1.2 V cm⁻¹ for about 40 hr and blotted onto Nfp (Amersham Pharmacia, Uppsala, Sweden) nylon membranes (Burke and Bruford 1987, Bjørnstad and Lifjeld 1997). The minisatellite probe per (Shin et al. 1985) was radioactively labelled with Redivue αdCTP using the Prime-a-Gene labeling kit (Promega). The hybridization procedure followed the Amersham protocol for multilocus probes that is supplied with the membrane. Filters were autoradiographed with one intensifying screen at −80°C for 1–4 days using Kodak BioMax MS film.

Results

The mean band-sharing coefficient for social partners, which were assumed to be unrelated individuals, was 0.213 ± 0.014 (maximum = 0.359). The mean band-sharing coefficient for siblings from consecutive years (because clutch and brood-size were one) was 0.643 ± 0.047 (minimum = 0.433).

For the 42 cases where a chick could be compared with both putative parents, the number of novel fragments per offspring showed a discontinuous and bimodal distribution with offspring having either three or fewer, or five or more novel fragments. Twenty-three chicks had at least one novel fragment: 19 had between one and three and 4 had five or more novel bands.

The mutation rate was estimated to be 0.019 (calculated using Westneat's 1990 method), and the probability of scoring a mutated fragment was estimated to be 0.597. Assuming that mutations occur independently of each other, the probability of scoring n novel fragments = pⁿ; the probability of scoring five or more novel fragments due to mutation was therefore P ≤ 0.076.

All offspring (n = 38) with up to three novel fragments had a band-sharing coefficient of 0.35 or higher with both putative parents (Fig. 1: mother-offspring mean band-sharing coefficient = 0.59 ± 0.01, range 0.41–0.76; father-offspring mean band-sharing coefficient = 0.53 ± 0.01, range 0.358–0.737). We assumed that these chicks were all related to their putative parents. Four offspring with more than three novel fragments shared a similar proportion of bands with their putative mother (mean band-sharing coefficient = 0.627 ± 0.06, range 0.46–0.70) as those with fewer novel fragments, suggesting that they were related to their social mother. However, band-sharing with the putative father was much lower in this group of four offspring (mean band-sharing coefficient = 0.22 ± 0.04, range 0.14–0.31; Fig. 1) and similar to the band-sharing coefficient between pair members (above). We therefore assumed that these were extra-pair offspring, genetically unrelated to their social father.

Figure 1.

Relationship between the number of novel fragments and band-sharing coefficients with (a) the social mother and (b) the social father. The dashed lines indicate the criteria for excluding parentage, and solid circles indicate cases where parentage was excluded

Open in new tabDownload slide

For those cases (n = 45) where only one of the putative parents was known, band-sharing coefficients were used to assign parentage. Where parentage could be assigned with certainty (complete families with ≤3 novel fragments), the lowest recorded band-sharing coefficient was 0.358 (n = 38). We therefore defined chick-parent band-sharing values of below 0.35 as mismatched parentage. In all 10 cases where we established the degree of band-sharing between offspring and the putative mother, the values were above this limit. Of the 35 cases where offspring were compared with putative fathers, there were 2 (band-sharing coefficients of 0.25) where the band-sharing coefficient was lower than 0.35, which we also considered to be cases of extra-pair paternity.

In all 52 cases where genetic relatedness of putative mothers and offspring was assessed (42 cases with both putative parents and 10 with putative mothers only), the genetic and social parentage matched. Of the total of 77 cases where paternity was analyzed, we found 6 (7.8%) instances where the social father did not match the genetic father. The 95% confidence interval on this estimate was 1.1–19.3% (Rohlf and Sokal 1981).

Discussion

The level of extra-pair paternity in the Common Murre on Skomer Island, 7.8%, is similar to the values reported for several other socially monogamous marine birds (Table 1). These results are consistent with an emerging pattern suggesting that extra-pair paternity is generally much lower in seabirds and other non-passerines than in passerine birds (Westneat and Sherman 1997).

Table 1.

Open in new tab

Levels of extra-pair paternity (% EPP) in seabirds. Confidence intervals are calculated from sample sizes ([cf2]n[cf1] = number of families) following Rohlf and Sokal (1981)

Table 1.

Open in new tab

Levels of extra-pair paternity (% EPP) in seabirds. Confidence intervals are calculated from sample sizes ([cf2]n[cf1] = number of families) following Rohlf and Sokal (1981)

The extra-pair offspring in the Common Murre could arise from either forced or unforced extra-pair copulations, that is, extra-pair copulations solicited by females. Hatchwell (1988) showed that during the presumed fertile period, 89.5% of 1,316 copulation attempts were pair copulations, whereas 9.9% were forced extra-pair copulations, and 0.6% were unforced extra-pair copulations. The proportions of these that appeared to be successful (resulted in cloacal contact, and we presume, insemination) were 80%, 6%, and 100%, respectively. Therefore, of every 100 successful copulations, 98.4 are likely to be successful pair copulations and just 1.6 successful extra-pair copulations. We found that 7.8% of all offspring were extra-pair, suggesting that extra-pair copulations were disproportionately successful. However, the confidence intervals on our estimate of extra-pair paternity are such that the apparent disparity between incidence of extra-pair copulations and extra-pair paternity must be treated with caution. If the disparity is genuine, there are two explanations. First, although there was no significant difference in the overall timing of pair copulations and extra-pair copulations relative to the day of egg-laying (Hatchwell 1988), forced extra-pair copulations did appear to be better timed (Colegrave et al. 1995), occurring significantly closer to egg-laying and hence ovulation (day −9.13 ± 1.5, n = 8) than unforced extra-pair copulations (day −16.00 ± 2.3, n = 8; Mann-Whitney U-test, P = 0.04). Second, because males that forced extra-pair copulations tended to be those whose partners were absent from the colony (Hatchwell 1988), they may have had fewer recent pair copulations and hence inseminated more sperm than paired males (Birkhead et al. 1995). However, a critical test of the relative success of forced and unforced extra-pair copulations would require considerably larger sample sizes than we obtained in the present study.

We are extremely grateful to the West Wales Wildlife Trust and the Countryside Council for Wales (CCW) for permission to work on Skomer Island. Successive Skomer wardens and their partners, in particular Steve and Anna Sutcliffe and Simon Smith and Christine Barton, provided excellent support. We are also grateful to many individuals, and especially Sherry Parrott, for their efforts during our long-term study of Common Murres on Skomer Island. We are grateful to H. Hoi and H. Winkler for facilities at the Konrad Lorenz Institute, Vienna. We thank Fiona Hunter, Terry Burke, and two referees for helpful comments on the manuscript.

Literature Cited

Author notes