## Abstract

Few studies of avian mating systems have identified the sires of extrapair young, and hence it has been difficult to determine the scale at which reproductive interactions occur. For instance, females may be free to copulate with any male in the population (a “global” scale of interactions), or females may be restricted to copulating only with males on neighboring territories (a “local” scale). The scale of such interactions has important consequences for an understanding of the evolutionary causes and consequences of extrapair fertilizations. We used five hypervariable microsatellite loci and multilocus DNA fingerprinting to examine parentage of more than 400 nestling black-throated blue warblers (Dendroica caerulescens). Extrapair fertilizations were common, and the microsatellite markers allowed us to identify the sires for 89% of the young analyzed. Most identified extrapair sires were males on neighboring or nearby territories, and most nestlings for whom we could not identify a sire came from territories at the edge of the study plot. Thus, reproductive interactions appear to be more local than global in this population. Extrapair fertilizations contributed significantly to total variation in male reproductive success. However, the standardized variance in male reproductive success (0.68-0.74) was not substantially greater than that for females (0.53-0.60), and the contribution of extrapair fertilizations (9-14%) was much lower than the contribution of within-pair fertilizations (75-77%). This suggests that the local scale of reproductive interactions may limit variation in male reproductive success and hence the opportunity for selection.

Due to the large number of DNA fingerprinting studies conducted during the past 15 years (Westneat and Webster, 1994), it is now clear that social monogamy does not imply genetic monogamy: in many birds, females often copulate with males other than their social partners. As a consequence, extrapair fertilizations (EPF) are a common and widespread feature of avian mating systems. Now that this fact is firmly established, attention has turned to identifying the social, genetic, and ecological factors that may promote or hinder extrapair mating (e.g., Hoi-Leitner et al., 1999; Johnsen et al., 1998; Petrie et al., 1998; Westneat and Gray, 1998), and to the selective consequences of extrapair mating (reviewed in Møller and Ninni, 1998).

An understanding of the causes and consequences of extrapair mating depends critically on the scale at which extrapair interactions occur. For example, if females of a territorial species interact and copulate only with males on neighboring territories (a “local” scale of interaction), the variance in male mating success and intensity of sexual selection generated by EPF (Webster et al., 1995) are likely to be relatively low. In contrast, if copulations can occur between individuals from territories that are widely separated (a “global” scale of interaction), then a very small number of males could potentially monopolize mating success (e.g., Dunn and Cockburn, 1998) and traits affecting the ability to obtain or prevent EPF could be subject to very strong selection. Similarly, breeding synchrony has been suggested as a factor having a strong influence on EPF frequency (Stutchbury, 1998; Stutchbury and Morton, 1995), but it is not clear at what scale such synchrony should be measured: the synchrony among all females in the population, or the synchrony only among females on adjacent territories (Chuang et al., 1999)?

The scale at which extrapair interactions occur can be determined by identifying the sires of extrapair young: if local interactions are important, then sires will come primarily from territories adjacent to the female's own. Although some multilocus DNA fingerprinting studies have successfully identified a number of extrapair sires (Hasselquist et al., 1995; Stutchbury et al., 1997; Westneat, 1993; Yezerinac et al., 1995), this approach is laborious and the sires of many young are often not identified. This is because it is difficult to compare each nestling to a large number of potential sires using multilocus blotting techniques (Webster and Westneat, 1998; Westneat and Webster, 1994). Single-locus markers, such as microsatellites, are much more powerful in identifying extrapair sires, because each individual's genotype can be catalogued and compared to all others (Ellegren, 1992; Primmer et al., 1995).

We have studied the social and genetic mating system of a population of a migratory songbird, the black-throated blue warbler (Dendroica caerulescens), breeding in New Hampshire, USA. Our multilocus DNA fingerprinting studies (Chuang et al., 1999) have shown that EPF are common in this population and that the frequency of extrapair young is affected by local breeding synchrony (i.e., the synchrony among females on adjacent territories) but not by population breeding synchrony. This suggests that females interact and copulate with males on nearby territories, but not with males from distant territories. In this article we employ microsatellite markers to identify the sires of extrapair young and directly test the importance of local interactions. We used microsatellites to examine the parentage of 342 nestlings sampled from 97 broods (five to 28 females per season, 60 females total) during the 1995-1998 breeding seasons, and verify the ability of microsatellites to identify cases of EPF by comparison to results of our earlier DNA fingerprinting analyses. Using the combined microsatellite and multilocus results (413 nestlings from 117 broods), we describe general patterns of EPF across the 4 years of this study, determine the locations of extrapair sires relative to the young that they sire, and estimate the effects of EPF on variance in male reproductive success. Finally, to determine whether particular male traits affect a male's ability to obtain EPF, we present preliminary comparisons of morphological traits of extrapair sires to traits of the males that they cuckold.

## MATERIALS AND METHODS

### Field methods

We monitored breeding on a 100 ha study plot at the Hubbard Brook Experimental Forest, West Thornton, New Hampshire, USA (see Holmes et al., 1992) during the years 1995 to 1998. All adults breeding on this plot (e.g., n = 75 in 1998) were color-banded for individual identification. At the time of banding we measured the adults (tarsus and wing length, wing spot size, measured from coverts to distal tip, and mass) and collected a small (ca. 20 μ l) blood sample for genetic analyses. We also collected blood samples from other adults captured incidentally near the main study plot each year. We determined territory boundaries by mapping aggressive interactions among males, as well as male singing behavior, and we found all warbler nests built on these territories. We also recorded instances when we encountered an individual outside of the area considered to be its territory. Each nest was visited approximately once every 3 days.

When nestlings reached 6 days of age, we banded them and collected a blood sample. We considered sampling young at an earlier age to be too risky. Therefore, our methods allow an estimate of the number of extrapair young that survive to fledging (about age 8-10 days; Holmes, 1994), rather than an estimate of the number of eggs fertilized by extrapair males. The mean ± SD brood size of sampled nests was 3.59 ± 0.73 (range 1 to 5), and we sampled 3.41 ± 0.84 nestlings per brood (range 1 to 5); the difference between these two figures was due to partial brood loss between hatching and day six (n = 16 nests). An additional 109 nests received eggs but were not sampled (97 failed, mostly due to predation, and 12 fledged before the young could be sampled).

### Microsatellite methods

We used variation at five microsatellite loci to determine paternity of nestlings (Table 1). Two of these loci, Dpμ 01 and Dpμ 16, were isolated from the genome of the yellow warbler (D. petechia) by Dawson et al. (1997). The remaining three loci, Dca 24, Dca 28, and Dca 32, were isolated from a D. caerulescens genomic library enriched for simple sequence repeats following the method of Hammond et al. (1998).

Table 1

Microsatellite loci used for parentage analyses

Locus   Ta (°C)   Primers
aSequences for Dpμ primers were reported previously by Dawson et al. (1997).
Dca 24   55   F-TGGGAGCCAGGAGAAGTTGTTTG
R-CGGGGATCNTCTGTAGGTCGAAT
Dca 28   64   F-CTTCACAACCACAGTAAACC
R-CAAATTCTTGCAGTCATAGC
Dca 32   60   F-GGACACAAGCACATCACAATC
R-CCCATGCNTTCCACANACTCT
Dpμ 01a  60   F-TGGATTCACACCCCAAAATT
R-AGAAGTATATAGTGCCGCTTGC
Dpμ 16a  60   F-ACAGCAAGGTCAGAATTAAA
R-AACTGTTGTGTCTGAGCCT
Locus   Ta (°C)   Primers
aSequences for Dpμ primers were reported previously by Dawson et al. (1997).
Dca 24   55   F-TGGGAGCCAGGAGAAGTTGTTTG
R-CGGGGATCNTCTGTAGGTCGAAT
Dca 28   64   F-CTTCACAACCACAGTAAACC
R-CAAATTCTTGCAGTCATAGC
Dca 32   60   F-GGACACAAGCACATCACAATC
R-CCCATGCNTTCCACANACTCT
Dpμ 01a  60   F-TGGATTCACACCCCAAAATT
R-AGAAGTATATAGTGCCGCTTGC
Dpμ 16a  60   F-ACAGCAAGGTCAGAATTAAA
R-AACTGTTGTGTCTGAGCCT

We determined the microsatellite genotypes of all adults and nestlings captured on the study plot during the 1997 and 1998 breeding seasons. Parentage had already been analyzed by multilocus DNA fingerprinting for the 1995 and 1996 family groups (Chuang et al., 1999). In addition, we determined the microsatellite genotypes of all males sampled in 1995 and 1996 as well as females and nestlings from most (15 of 17) 1995 and 1996 family groups that showed mixed parentage in the earlier study.

To score individual genotypes, we amplified genomic DNA from each individual in a 10 μl PCR reaction that contained 100 μM dNTP (each), 0.25 μM primers (each), 1.5 μCi 33-P dATP (NEN™ Life Science Products), and our standard PCR reaction mix (1-3 units Taq DNA polymerase, 3.0 mM MgCl2, 50 mM KCl, 10 mM Tris-HCl). This reaction was subjected to 30 cycles of 94°C for 60 s, X°C for 60 s, and 72°C for 45 s, where the annealing temperature (X) depended on the primers used (Table 1). PCR products were size-sorted via electrophoresis in a 6% denaturing polyacrylamide gel containing 7 M urea at a constant temperature of 50°C (see Strassmann et al., 1996). We also ran a standard sequencing reaction of M13 DNA (also labeled with 33-P dATP) as a size reference for each set of PCR reactions run on a gel. After electrophoresis, the gel was dried and exposed to autoradiography film for 2-4 days.

We scored the size of PCR fragments for each individual by comparing its band(s) to the reference M13 sequence. We calculated the frequency of each allele (xi) from the total population of adults genotyped (Σ xi = 1.00), and calculated the expected frequency of heterozygotes (he) as:

1
$\ h_{e}=1-{\Sigma}(x_{i})^{2}\$
This expected frequency was then compared to the observed frequency of heterozygotes (ho). A significant difference between he and ho suggests the presence of a null (i.e., nonamplifying) allele, and in these cases we estimated the frequency of the null allele (r) under the assumption that we had no null homozygotes (Brookfield, 1996):
2
$\ r=(h_{e}-h_{o})/(1+h_{e})\$

### Determination of parentage

In our previous DNA fingerprinting analyses (Chuang et al., 1999), all 125 nestlings sampled (from 38 broods) showed high band-sharing with their social mothers, indicating that intraspecific brood parasitism is rare in this population. Therefore, for each nestling we assumed the social mother was the biological mother, and used the microsatellite genotypes to identify the sire: potential sires were the male or males who possessed the nestling's nonmaternal allele at each locus. Thus, a female “matched” a nestling at a particular locus if she possessed either of the nestling's alleles, and a male “matched” the offspring if he possessed the nestling's remaining nonmaternal allele.

We calculated the average probability of paternal exclusion (Pej) for each polymorphic locus. This is the probability, averaged over all alleles at the jth locus, that a randomly chosen nonsire male will not possess the paternal allele found in an offspring (i.e., will not match), given that the mother of the offspring is known with certainty. This probability can be estimated as (Jamieson, 1994):

3
$\ \begin{array}{rl}P_{ej}=&1-2{\Sigma}(x_{i})^{2}+{\Sigma}(x_{i})^{3}+2{\Sigma}(x_{i})^{4}-3{\Sigma}(x_{i})^{5}\\&+2({\Sigma}(x_{i})^{2})^{2}+3{\Sigma}(x_{i})^{2}{\Sigma}(x_{i})^{3}\end{array}\$
We also calculated the total probability of exclusion (Pet), which is the probability that a randomly-chosen male will not possess the paternal allele of an offspring at one or more of the loci surveyed (i.e., it is equal to one minus the probability that the male will match at all loci surveyed):
4
$\ P_{et}=1-{\Pi}(1-P_{ej})\$
Two loci showed evidence of null alleles (see Results). We took this into account when assigning paternity: for these two loci, a male and nestling were considered to match if the male appeared to be homozygous and the nestling appeared to be homozygous for the maternal allele.

We calculated the standardized variance (total variance divided by mean squared, see Arnold and Wade, 1984) in female and male reproductive success. For males, we calculated the standardized variance of both apparent reproductive success (A), defined as the number of young produced on a male's territory, and genetic reproductive success (T), defined as the number of young sired. Following Webster et al. (1995), we also broke down male genetic reproductive success into three component parts:

5
$\ \mathrm{Svar}(T)=\mathrm{Svar}(W)+\mathrm{Svar}(E)+2\mathrm{Cov}(W,E)\$
where Svar is the standardized variance, Cov is the covariance, W is the number of young a male sires with his social mate(s), and E is the number of young he sires with his extrapair mate(s).

## RESULTS

### Analyses of parentage

All five microsatellites used in this study proved highly variable, with 11 to 30 alleles, high levels of observed heterozygosity, and high probabilities of exclusion (Table 2). The combined probability that a nonsire would not match the paternal alleles of a nestling at all five loci was Pet ≥ 0.9998, and we were able to identify extrapair young and their sires with high confidence. Two loci, Dpμ 01 and Dca 24, showed a significant deficiency of heterozygotes, suggesting a relatively high frequency of null alleles at these loci (Table 2). This is unlikely to have biased our results, as we accounted for possible null alleles when scoring these loci (see Methods), and the total probability of exclusion was high even when either locus was excluded (Pet ≥ 0.9985).

Table 2

Variability of five microsatellite loci among 106 adult black-throated blue warblers sampled during the 1998 breeding season

Heterozygosity
Locus   No. alleles   Expected (he)   Observed (ho)   Frequency of null allele (r)  Pej
Results for the 1997 adults (n = 93) were almost identical. Pej, probability of exclusion.
aSignificantly fewer heterozygotes observed than expected under Hardy-Weinberg equilibrium (χ2 tests, df = 1, p <.05).
Dca 24  26   0.928   0.485a  0.230   0.855
Dca 28  30   0.944   0.943   0.001   0.888
Dca 32  19   0.886   0.925   0.000   0.773
Dpμ 01  25   0.944   0.887a  0.030   0.887
Dpμ 16  11   0.785   0.759   0.014   0.583
Heterozygosity
Locus   No. alleles   Expected (he)   Observed (ho)   Frequency of null allele (r)  Pej
Results for the 1997 adults (n = 93) were almost identical. Pej, probability of exclusion.
aSignificantly fewer heterozygotes observed than expected under Hardy-Weinberg equilibrium (χ2 tests, df = 1, p <.05).
Dca 24  26   0.928   0.485a  0.230   0.855
Dca 28  30   0.944   0.943   0.001   0.888
Dca 32  19   0.886   0.925   0.000   0.773
Dpμ 01  25   0.944   0.887a  0.030   0.887
Dpμ 16  11   0.785   0.759   0.014   0.583

Nestlings sampled during the 1995 and 1996 breeding seasons had been previously analyzed by multilocus DNA fingerprinting (Chuang et al., 1999). We determined the microsatellite genotypes for 56 of these nestlings in the present study. As expected, 55 of these nestlings matched their social mothers at all five loci (the single exception was a nestling that matched at four of five loci), regardless of whether DNA fingerprinting analyses indicated the nestling was produced by EPF or not (Figure 1a,b). Similarly, of the 23 nestlings that matched their social fathers in the DNA fingerprinting analyses, 20 matched their social fathers at all five microsatellite loci and one matched at four of five loci (Figure 1a). The remaining two nestlings did not match their sires at three and four of the loci, respectively. Both of these nestlings were “borderline” in the fingerprinting analyses, in that each showed relatively high sharing with the social male (≥ 50%) but had two novel fragments (nestlings showing three or more novel fragments were considered to have resulted from EPF in the earlier study). Of the 33 nestlings that did not match their sires in the DNA fingerprinting analyses, 31 were mismatched at two or more microsatellite loci (Figure 1b). The remaining two nestlings matched their social fathers at four and five loci, respectively. Thus, multilocus and microsatellite analyses agree with each other in large part, but with a small number of discrepancies (see Discussion). These results support the assumption that intraspecific brood parasitism is rare in this population (but see next paragraph), and that most cases of microsatellite mismatches between nestlings and their social parents will be due to extrapair fertilizations.

Figure 1

Microsatellite genotype comparisons between nestlings and their social mothers (white bars) and between nestlings and their social fathers (black bars): (a) 1995 and 1996 nestlings that DNA fingerprinting analyses indicated were sired by their social fathers (n = 33 nestlings); (b) 1995 and 1996 nestlings that DNA fingerprinting analyses indicated were sired by extrapair males (n = 23 nestlings); and (c) nestlings from the 1997 and 1998 breeding seasons (n = 283 comparisons to social mothers and 280 comparisons to social fathers).

Figure 1

Microsatellite genotype comparisons between nestlings and their social mothers (white bars) and between nestlings and their social fathers (black bars): (a) 1995 and 1996 nestlings that DNA fingerprinting analyses indicated were sired by their social fathers (n = 33 nestlings); (b) 1995 and 1996 nestlings that DNA fingerprinting analyses indicated were sired by extrapair males (n = 23 nestlings); and (c) nestlings from the 1997 and 1998 breeding seasons (n = 283 comparisons to social mothers and 280 comparisons to social fathers).

We determined the microsatellite genotypes of all individuals sampled during the 1997 and 1998 breeding seasons, and compared the genotypes of nestlings to those of their social parents (Figure 1c). In the vast majority (97.2%) of comparisons between nestlings and their presumed mothers (n = 283), the nestling matched the female (i.e., had an allele that matched one of the alleles present in the female) at all loci (Figure 1c). In an additional six cases the nestling matched its presumed mother at all but one locus. Five of these six cases (four from a single nest) were at locus Dca 28, and were situations in which both the female and nestling appeared homozygous for different alleles. This suggests that a null allele may have existed at this locus, although its frequency would have been very low (Table 2) and so would be unlikely to affect our analyses. The remaining case of a single mismatch likely represents a case of mutation or laboratory artifact. However, one nestling did not match its social mother at three loci, and another did not match at five loci. Both of these nestlings were from the same nest, and the other two nestlings from this nest did match the female at all loci. This suggests that a very low level of intraspecific brood parasitism may exist in this population, although the possibility of mislabeled samples cannot be ruled out. Even if present, a low level of parasitism is unlikely to affect the results given below.

Comparisons between the 1997 and 1998 nestlings and their social fathers (n = 280) yielded a somewhat different pattern (Figure 1c). In many comparisons the nestling matched its social father at all five loci (n = 216) or at four of the five loci (n = 3). In 61 comparisons, however, the nestling and its social sire were mismatched at two or more loci; two of these also matched poorly with the social mother (see above), where the remaining 59 represent cases of extrapair fertilization. At two additional nests (six nestlings total) we had a DNA sample for the nestlings and social father, but no sample for the social mother; in all six cases the male possessed an allele found in the nestling at all loci, indicating that the social father was the sire of each.

### Identification of extrapair sires

We identified sires for 88.6% of the 413 nestlings analyzed, including the sires for approximately half of the extrapair young (Table 3). Our success at identifying extrapair sires was highest in 1997 and 1998 (sires identified for 62.3% of 53 EPF offspring), when sampling efforts were most extensive. In all but one case the extrapair sire possessed the nestling's paternal allele at all loci scored. In the final case a male and nestling matched at four of five loci and had alleles of very similar size at the fifth (Dca 24, 208 versus 210 bp); this male also matched two other extrapair young from the same nest. Only one nestling matched two extrapair males at all loci; one of these males also matched the second EPF offspring from the same nest, and this male was therefore designated to be the sire. Three additional comparisons were somewhat ambiguous, with an extrapair male matching a nestling at four of five loci in each case. To be as conservative as possible, we did not assign these nestlings to the extrapair males.

Table 3

Frequency of extrapair fertilization (EPF) young across years

Year   % EPF young (n)   % broods with EPF young (n)a  % EPF young matched with sires (n)b
1995 and 1996 results based on combined DNA fingerprinting and microsatellite analyses, 1997 and 1998 results based on microsatellite analyses only.
aComplete broods only included in n; partially sampled/analyzed broods excluded.
bComplete broods only included in n (number of EPF young).
1995   19.6 (51)   31.3 (16)   60.0 (10)
1996   33.8 (77)   47.6 (21)   34.6 (26)
1997   17.4 (138)   26.3 (38)   94.4 (18)
1998   23.8 (147)   35.7 (42)   45.7 (35)
Total   23.0 (413)   34.2 (117)   53.9 (89)
Year   % EPF young (n)   % broods with EPF young (n)a  % EPF young matched with sires (n)b
1995 and 1996 results based on combined DNA fingerprinting and microsatellite analyses, 1997 and 1998 results based on microsatellite analyses only.
aComplete broods only included in n; partially sampled/analyzed broods excluded.
bComplete broods only included in n (number of EPF young).
1995   19.6 (51)   31.3 (16)   60.0 (10)
1996   33.8 (77)   47.6 (21)   34.6 (26)
1997   17.4 (138)   26.3 (38)   94.4 (18)
1998   23.8 (147)   35.7 (42)   45.7 (35)
Total   23.0 (413)   34.2 (117)   53.9 (89)

In 68.0% of the cases in which we identified an extrapair sire (n = 25 males), the nestling was located on a territory adjacent to that of its extrapair sire, and in most of the remaining cases the sire was less than two territories away (Figures 2 and 3). Moreover, 70.2% of the young whose sires were not identified (n = 47) were produced on territories at the edge of the study plot, where some neighboring males had not been sampled. There were, however, four cases in which a male was separated from his extrapair young by two or more territories (e.g., male 352 in Figure 2), and another four cases in which young with an unidentified sire were found in a nest two or more territories from the edge of the study plot (e.g., territory of male 006 in Figure 2). Excluding territories at the edge of the study plot and limiting the analyses to 1997 and 1998 (when all males breeding on the plot were sampled), we identified the sires for 76.5% of all extrapair nestlings (n = 34). Thus, Figure 3 gives a general picture of the distribution of distances between extrapair offspring and their sires, but likely excludes a few cases in which the nestling and sire are separated by several territories. These results indicate that most, but not all, extrapair sires are males on nearby territories (usually adjacent neighbors), thus strongly supporting the role of local interactions in producing EPF.

Figure 2

Location of extrapair sires during the 1997 breeding season. Curved lines show estimated territory boundaries, and numbers are identification number of select males. Arrows connect the territory of the extrapair sire (base of arrow) with the territory where he sired young (arrow tip). For example, male 317 sired young in the nests of two males—003 and 331, where male 003 sired young in the nest of male 007. One male (006) had a nestling in his nest that did not match with any sampled male; this nestling was apparently sired by an unsampled male (indicated by “?”) from off the study plot.

Figure 2

Location of extrapair sires during the 1997 breeding season. Curved lines show estimated territory boundaries, and numbers are identification number of select males. Arrows connect the territory of the extrapair sire (base of arrow) with the territory where he sired young (arrow tip). For example, male 317 sired young in the nests of two males—003 and 331, where male 003 sired young in the nest of male 007. One male (006) had a nestling in his nest that did not match with any sampled male; this nestling was apparently sired by an unsampled male (indicated by “?”) from off the study plot.

Figure 3

The number of territories separating the territory of an extrapair sire from the territory of his extrapair nestling(s). Cases in which a male sired more than one nestling in the same nest are counted only once.

Figure 3

The number of territories separating the territory of an extrapair sire from the territory of his extrapair nestling(s). Cases in which a male sired more than one nestling in the same nest are counted only once.

### Patterns of extrapair paternity and variation in reproductive success

The frequency of extrapair fertilizations was high in all years of this study (Table 3), with nearly 1/4 of all sampled young having extrapair sires and over 1/3 of all sampled nests containing such young. The frequency of EPF offspring was unaffected (23.5%) by exclusion of the 16 nests showing partial brood loss. The frequency of EPF offspring varied significantly among years (χ2 = 7.88, df = 3, p =.049), but post-hoc comparisons showed that the only significant pairwise difference was between 1996 and 1997 (the years with the highest and lowest frequency of EPF offspring, respectively; χ2 = 7.43, df = 1, p =.006).

The proportion of extrapair young per nest varied from 0.0 to 1.0, with significantly more broods having many EPF than would be expected by chance (Figure 4). Of all broods analyzed, 16 contained a single extrapair nestling and 26 contained multiple extrapair nestlings. We identified the extrapair sires for 16 of the broods containing multiple EPF offspring—in all but one case a single male had sired all of the extrapair young in the brood. The one exception was a brood of four consisting of three nestlings sired by one extrapair male and a single nestling sired by a different extrapair male.

Figure 4

The number of extrapair young per brood in broods of four young (the most common size, n = 72 broods). White bars show the observed number of broods observed to contain the specified number of EPF. Black bars show the number of broods expected under a poisson distribution. The observed distribution differed significantly from expectation (χ2 = 50.27, df = 3, p <.01). Other brood sizes showed a similar excess of nests with many EPF, but sample sizes were too small for statistical analysis.

Figure 4

The number of extrapair young per brood in broods of four young (the most common size, n = 72 broods). White bars show the observed number of broods observed to contain the specified number of EPF. Black bars show the number of broods expected under a poisson distribution. The observed distribution differed significantly from expectation (χ2 = 50.27, df = 3, p <.01). Other brood sizes showed a similar excess of nests with many EPF, but sample sizes were too small for statistical analysis.

Our sampling efforts were most complete in 1997 and 1998, when we obtained blood samples from nearly all nestlings that survived to an age of 6 days. This extensive sample allowed us to calculate variance in male reproductive success for both males and females (Table 4). The standardized variance in male genetic reproductive success (T) was approximately 45% greater than variance in apparent reproductive success (number of young per male territory, A). Moreover, success in siring extrapair offspring (E) contributed 9-14% of the total variance in T, and a positive covariance between within-pair success (W) and extrapair success (E) contributed a similar amount (correlation between W and E, 1997 and 1998 combined, r =.23, p =.06). Thus, EPF increased variance in male reproductive success. Nevertheless, variance in male reproductive success was only 23-29% greater than variance in female reproductive success in this socially monogamous population.

Table 4

Variation in male and female reproductive success (RS) in 1997 and 1998

1997 (n = 34 males)   1998 (n = 33 males)
Mean   Variance   St. var. (%)   Mean   Variance   St. var. (%)
Standardized variance (St. Var.) is variance divided by mean-squared. For males, measures are given for both apparent reproductive success (number of young produced on a male's territory) and genetic reproductive success (number of young sired). Measures of apparent male reproductive success and female reproductive success differ due to low levels of polygyny and bachelor males. Male genetic reproductive success is also broken down into three component parts, and numbers in parentheses give the percentage of standardized variance (T) explained by each factor. For the covariance term. the standardized variance has been weighted by a factor of two (see equation 5).
Female RS   3.26   5.61   0.529   3.84   8.89   0.602
Male RS, apparent (A)   3.56   6.06   0.478   4.42   9.88   0.505
Male RS, genetic (T)   3.32   7.56   0.684   3.71   10.21   0.740
Within-pair (W)   2.82   5.79   0.524 (76.6)   3.35   7.69   0.557 (75.3)
Extrapair (E)   0.50   1.05   0.095 (13.8)   0.46   0.90   0.065 (8.8)
Covariance (W, E)    0.35   0.064 (9.3)    0.72   0.104 (14.0)
1997 (n = 34 males)   1998 (n = 33 males)
Mean   Variance   St. var. (%)   Mean   Variance   St. var. (%)
Standardized variance (St. Var.) is variance divided by mean-squared. For males, measures are given for both apparent reproductive success (number of young produced on a male's territory) and genetic reproductive success (number of young sired). Measures of apparent male reproductive success and female reproductive success differ due to low levels of polygyny and bachelor males. Male genetic reproductive success is also broken down into three component parts, and numbers in parentheses give the percentage of standardized variance (T) explained by each factor. For the covariance term. the standardized variance has been weighted by a factor of two (see equation 5).
Female RS   3.26   5.61   0.529   3.84   8.89   0.602
Male RS, apparent (A)   3.56   6.06   0.478   4.42   9.88   0.505
Male RS, genetic (T)   3.32   7.56   0.684   3.71   10.21   0.740
Within-pair (W)   2.82   5.79   0.524 (76.6)   3.35   7.69   0.557 (75.3)
Extrapair (E)   0.50   1.05   0.095 (13.8)   0.46   0.90   0.065 (8.8)
Covariance (W, E)    0.35   0.064 (9.3)    0.72   0.104 (14.0)

### Male characteristics and extrapair success

We were unable to identify any male characteristics that were associated with EPF. First, male age class was not related to the probability of being cuckolded (n = 47, log likelihood = -30.14, χ2 = 2.56, p =.917). Similarly, males who obtained EPF did not differ significantly in age from the males that they cuckolded (paired sign test, n = 25 comparisons, p >.999). Finally, none of the morphological traits that we measured differed significantly between extrapair sires and the males that they cuckolded (Table 5).

Table 5

Comparison of morpological traits of extrapair sires to those of cuckolded males

Trait   Extrapair sires (mean ± s.e.)   Cuckolded males (mean ± s.e.)  p (n)a
aPaired t tests comparing the extrapair sire to the male he cuckolded, n = number of paired comparisons.
Tarsus (mm)   18.94 ± 0.16   19.20 ± 0.12   0.185 (25)
Wing cord (mm)   63.40 ± 0.30   63.58 ± 0.32   0.554 (23)
Mass (g)   9.61 ± 0.10   9.68 ± 0.11   0.187 (18)
Wing spot size (mm)   12.06 ± 2.65   11.08 ± 0.58   0.326 (18)
Trait   Extrapair sires (mean ± s.e.)   Cuckolded males (mean ± s.e.)  p (n)a
aPaired t tests comparing the extrapair sire to the male he cuckolded, n = number of paired comparisons.
Tarsus (mm)   18.94 ± 0.16   19.20 ± 0.12   0.185 (25)
Wing cord (mm)   63.40 ± 0.30   63.58 ± 0.32   0.554 (23)
Mass (g)   9.61 ± 0.10   9.68 ± 0.11   0.187 (18)
Wing spot size (mm)   12.06 ± 2.65   11.08 ± 0.58   0.326 (18)

## DISCUSSION

### Effectiveness of identifying extrapair sires

Although many studies have used multilocus DNA fingerprinting to analyze parentage in birds, relatively few studies have used microsatellites (Webster and Westneat, 1998), and fewer still have assessed the degree of congruence between these two types of genetic markers. We analyzed the parentage of 56 nestlings using both techniques and found that they reached similar conclusions in all but four cases. In three of these cases the nestling was near the cutoff for classifying the nestling as resulting from EPF, and the fourth case may have arisen because we used a conservative rule for determining whether bands matched in the earlier fingerprinting study. Thus, although these results corroborate the effectiveness of both techniques in identifying EPF offspring, caution should be used in borderline cases.

With that caveat in mind, microsatellites nevertheless proved highly effective in identifying both extrapair young and their sires. We were able to identify the extrapair sires in over half of the cases of EPF, and in only one case did more than one male match a nestling at all five loci. Moreover, when nestlings produced on territories at the edge of the study plot were excluded (because some of their potential sires had not been sampled), we were able to assign sires to the majority of extrapair young. Thus, as suggested by earlier studies (e.g., Ellegren, 1992; Primmer et al., 1995), microsatellite markers appear to be a highly effective approach to the analysis of parentage in birds.

### Locations of extrapair sires

In this study population, the majority of extrapair sires were males on adjacent territories (i.e., neighboring males). Very few nestlings were sired by males from distant territories, and the majority of nestlings with unidentified sires were from territories at the edge of the study plot. Other studies of socially monogamous birds have identified extrapair sires, and in most (e.g., Hasselquist et al., 1995; Kempenaers et al., 1992; Sheldon and Ellegren, 1999; Stutchbury et al., 1997; Yezerinac et al., 1995) but not all (Dunn and Cockburn, 1998; Dunn et al., 1994) of these studies the identified sires were usually neighboring males.

Thus, the emerging picture of mating systems in territorial passerines is that most (but not all) extrapair sires come from neighboring territories. This has two important implications for the study of avian mating systems. First, because reproductive interactions appear to act on a local scale, variance in total male reproductive success is likely to be lower than would be the case if such interactions acted on a more global scale (see below). For example, if females actively choose extrapair mates, any given female will have a relatively limited pool of potential sires to choose among (i.e., her neighbors). Second, studies attempting to identify the factors affecting frequency of EPF should examine those factors at a local scale. For example, in assessing the effects of breeding synchrony on EPF (e.g., Birkhead and Biggins, 1987; Kempenaers, 1997; Saino et al., 1999; Stutchbury, 1998; Weatherhead, 1997; Westneat and Gray, 1998), the synchrony among adjacent neighbors may be more important than the synchrony among all females in the population (Chuang et al., 1999).

### Effects of EPF on variance in male reproductive success

The standardized variance in reproductive success measures the “ opportunity for selection,” which is the maximum strength of selection (Arnold and Wade, 1984). In populations with low variation in reproductive success, the difference between successful and unsuccessful breeders will be slight, and therefore selection on traits associated with success will be weak. In contrast, in populations with high variation in reproductive success, successful individuals will produce sub-stantially more offspring than unsuccessful individuals, and traits that contribute to successful breeding will be strongly favored by selection.

It is for this reason that sexual dimorphism in socially monogamous birds has proved somewhat enigmatic. In a large number of species, males differ substantially from females in morphology (e.g., in plumage coloration), implying that sexual selection is acting strongly on male traits. However, in monogamous species with a balanced adult sex ratio, variance in male mating success should be very low, and variance in male reproductive success should be almost identical to that of females. Under such conditions, how could sexual selection be strong enough to affect male morphology so markedly? Darwin (1871) proposed two solutions to this problem. First, sexual selection may be generated by variation in mate quality (i.e., some males may obtain mates that are much more fecund than are the mates of other males). This proposal has received theoretical (Kirkpatrick et al., 1990) and some empirical (e.g., Møller, 1991, 1994) support, but requires further study. Second, adult sex ratios may be male biased, such that some males may not be able to obtain a mate. Consistent with this suggestion, adult sex ratios of most monogamous birds appear to be male biased (Breitwisch, 1989), and this has been shown to generate selective pressures on male morphology (e.g., Price, 1984).

The recent discovery of extrapair fertilizations has suggested an alternative explanation; if EPF are common, variation in male reproductive success could be substantial even in socially monogamous species. Yet, simulation studies (Webster et al., 1995) have shown that EPF may have little effect on variation in reproductive success, even when they are common. This could happen, for example, if males who sire extrapair young are also cuckolded by other males. In this case EPF would “cancel out” and there would be little reproductive variation among males.

In this study, variance in male reproductive success was greater than variation in number of young produced per territory (i.e., apparent reproductive success), and variation in number of extrapair young sired contributed significantly to total variation in male reproductive success. This result is robust to sampling effort, because we obtained samples (and hence genotypes) for nearly all territorial males on our study site, and previous results indicate that there are few if any floaters in this population (Marra and Holmes, 1997). Interestingly, a positive and fairly substantial covariance existed between within-pair and extrapair success (see also Ketterson et al., 1998). This positive covariance indicates that males who sired many young on their own territories were also successful in siring extrapair young on other territories. Our preliminary analyses, though, have failed to identify any male trait that might be associated with success in obtaining EPF (Table 5). This result should be accepted tentatively, however, because our sample sizes for these comparisons are not large at this point, and we may have failed to measure morphological traits important to female mate choice or intrasexual competition.

Although the frequency and effects of EPF can vary substantially among populations (e.g., Gyllensten et al., 1990), our results suggest that EPF increase variation in reproductive success among males, and hence the opportunity for selection, in this dichromatic species. However, variation in male reproductive success was only slightly greater than variation in female reproductive success even after the effects of EPF had been factored in. This limited effect of EPF did not occur because EPF gains and losses cancel each other out (the covariance between W and E was positive). Rather, it appears to be due to the fact that most EPF occurred on a local scale; that is, because males sired young only on territories near their own, none sired a large number of extrapair young. As a consequence, extrapair success was not biased strongly toward a small subset of highly successful males in this population. Indeed, most of the variation in male reproductive success was generated by variation in within-pair success (Table 4; see also Webster et al., 1995). This component of reproductive success is affected by pairing success, mate quality, predation on nestlings, and numerous other factors in addition to EPF (i.e., whether a male is cuckolded).

Although studies indicate that EPF increase the opportunity for sexual selection in socially monogamous species (see Møller, 1998; Møller and Ninni, 1998), variance in male mating success is typically less than that seen in polygynous species (e.g., Pruett-Jones and Pruett-Jones, 1990; Trail, 1985). Other studies of sexually dichromatic passerines have found little effect of EPF on sexual selection (e.g., Hill et al., 1994; Webster et al., 1995; Westneat, 1993). Thus, it is not yet clear whether EPF underlie the evolution of pronounced sexual dimorphism (e.g., dichromatism) in socially monogamous birds (Møller and Birkhead, 1994), or whether the mechanisms originally proposed by Darwin are better explanations.

We thank T.S. Sillett, J.J. Barg, and numerous assistants from the University at Buffalo and Dartmouth College for assistance in the field. H.L. Gibbs generously provided his preliminary results, including primer sequences. J. Feng, H.A. Murphy, and K.S. Phipps provided invaluable laboratory assistance. We thank M.A. Coffroth, H.R. Lasker, D.J. Taylor, and three anonymous referees for suggestions that improved this manuscript. Fieldwork was conducted in the Hubbard Brook Experimental Forest, which is administered by the U.S. Forest Service, Northeast Research Station, Radnor, Pennsylvania, USA. This study was supported by National Science Foundation grants to the State University of New York at Buffalo and to Dartmouth College.

## References

Arnold SJ, Wade MJ,
1984
. On the measurement of natural and sexual selection: theory.
Evolution

38
:
709
-719.
Birkhead TR, Biggins JD,
1987
. Reproductive synchrony and extrapair copulation in birds.
Ethology

74
:
320
-334.
Breitwisch R,
1989
. Mortality patterns, sex ratios, and parental investment in monogamous birds.
Curr Ornithol

6
:
1
-50.
Brookfield JFY,
1996
. A simple new method for estimating null allele frequency from heterozygote deficiency.
Mol Ecol

5
:
453
-455.
Chuang HC, Webster MS, Holmes RT,
1999
. Extra-pair paternity and local synchrony in the black-throated blue warbler.
Auk

116
:
726
-736.
Darwin C,
1871
.
The descent of man, and selection in relation to sex
. Princeton, New Jersey: Princeton University Press.
Dawson RJG, Gibbs HL, Hobson KA, Yezerinac SM,
1997
. Isolation of microsatellite DNA markers from a passerine bird, the yellow warbler, Dendroica petechia, and their use in population studies.
Heredity

78
:
506
-514.
Dunn PO, Cockburn A,
1998
. Costs and benefits of extra-group paternity in Superb fairy-wrens. In:
Avian reproductive tactics: female and male perspectives, ornithological monographs no. 49
(Parker PG, Burley NT, eds). Washington, DC: American Ornithologists' Union;
147
-161.
Dunn PO, Robertson RJ, Michaud-Freeman D, Boag PR,
1994
. Extrapair paternity in tree swallows: why do females mate with more than one male?
Behav Ecol Sociobiol

35
:
273
-281.
Ellegren H,
1992
. Polymerase-chain-reaction (PCR) analysis of micro-satellites—a new approach to studies of genetic relationships in birds.
Auk

109
:
886
-895.
Gyllensten UB, Jakobsson S, Temrin H,
1990
. No evidence for illegitimate young in monogamous and polygynous warblers.
Nature

343
:
168
-170.
Hammond RL, Saccheri IJ, Ciofi C, Coote T, Funk SM, McMillan WO, Bayes MK, Taylor E, Bruford MW,
1998
. Isolation of microsatellite markers in animals. In:
Molecular tools for screening biodiversity: plants and animals
(Karp A, Isaac PG, Ingram DS, eds). London: Chapman and Hall;
279
-285.
Hasselquist D, Bensch S, von Schantz T,
1995
. Low frequency of extrapair paternity in the polygynous great reed warbler, Acrocephalus arundinaceus.
Behav Ecol

6
:
27
-38.
Hill GE, Montgomerie R, Roeder C, Boag P,
1994
. Sexual selection and cuckoldry in a monogamous songbird: implications for sexual selection theory.
Behav Ecol Sociobiol

35
:
193
-199.
Hoi-Leitner M, Hoi H, Romero-Pujante M, Valera F,
1999
. Female extra-pair behaviour and environmental quality in the serin (Serinus serinus): a test of the `constrained female hypothesis'.
Proc R Soc Lond B

266
:
1021
-1026.
Holmes RT, Sherry TW, Marra PP, Petit KE,
1992
. Multiple brooding and productivity of a neotropical migrant, the black-throated blue warbler (Dendroica caerulescens), in an unfragmented temperate forest.
Auk

109
:
321
-333.
Jamieson A,
1994
. The effectiveness of using co-dominant polymorphic allelic series for (1) checking pedigrees and (2) distinguishing full-sib pair members.
Anim Genetics 25
(suppl)
1
:
37
-44.
Johnsen A, Lifjeld JT, Rohde PA, Primmer CR, Ellegren H,
1998
. Sexual conflict over fertilizations: female bluethroats escape male paternity guards.
Behav Ecol Sociobiol

43
:
401
-408.
Kempenaers B,
1997
. Does reproductive synchrony limit male opportunities or enhance female choice for extrapair paternity?
Behaviour

134
:
551
-562.
Kempenaers B, Verheyen GR, Van den Broeck M, Burke T, Van Broeckhoven C, Dhondt AA,
1992
. Extra-pair paternity results from female preference for high-quality males in the blue tit.
Nature

357
:
494
-496.
Ketterson ED, Parker PG, Raouf SA, Nolan V, Ziegenfus C, Chandler CR,
1998
. The relative impact of extra-pair fertilizations on variation in male and female reproductive success in dark-eyed Juncos (Junco hyemalis). In:
Avian reproductive tactics: female and male perspectives, ornithological monographs no. 49
(Parker PG, Burley NT, eds). Washington, DC: American Ornithologists' Union;
81
-101.
Kirkpatrick M, Price T, Arnold SJ,
1990
. The Darwin-Fisher theory of sexual selection in monogamous birds.
Evolution

44
:
180
-193.
Marra PP, Holmes RT,
1997
. Avian removal experiments: do they test for habitat saturation or female availability?
Ecology

78
:
947
-952.
Møller AP,
1991
. Preferred males acquire mates of higher phenotypic quality.
Proc R Soc Lond B

245
:
179
-182.
Møller AP,
1994
.
Sexual selection and the barn swallow
. Oxford: Oxford University Press.
Møller AP,
1998
. Sperm competition and sexual selection. In:
Sperm competition and sexual selection
(Birkhead TR, Møller AP, eds). San Diego: Academic Press;
55
-90.
Møller AP, Birkhead TR,
1994
. The evolution of plumage brightness in birds is related to extrapair paternity.
Evolution

48
:
1089
-1100.
Møller AP, Ninni P,
1998
. Sperm competition and sexual selection: a meta-analysis of paternity studies in birds.
Behav Ecol Sociobiol

43
:
345
-358.
Petrie M, Doums C, Møller AP,
1998
. The degree of extra-pair paternity increases with genetic variability.
Proc Natl Acad Sci USA

95
:
9390
-9395.
Price TD,
1984
. Sexual selection on body size, territory and plumage variables in a population of Darwin's finches.
Evolution

38
:
327
-341.
Primmer CR, Møller AP, Ellegren H,
1995
. Resolving genetic relation-ships with microsatellite markers: a parentage testing system for the swallow Hirundo rustica.
Mol Ecol

4
:
493
-498.
Pruett-Jones SG, Pruett-Jones MA,
1990
. Sexual selection through female choice in Lawes' parotia, a lek-mating bird of paradise.
Evolution

44
:
486
-501.
Saino N, Primmer CR, Ellegren H, Møller AP,
1999
. Breeding synchrony and paternity in the barn swallow (Hirundo rustica).
Behav Ecol Sociobiol

45
:
211
-218.
Sheldon BC, Ellegren H,
1999
. Sexual selection resulting from extrapair paternity in collared flycatchers.
Anim Behav

57
:
285
-298.
Strassmann JE, Solís CR, Peters JM, Queller DC,
1996
. Strategies for finding and using highly polymorphic DNA microsatellite loci for studies of genetic relatedness and pedigrees. In:
Molecular zoology: advances, strategies and protocols
(Ferraris JD, Palumbi SR, eds). New York: Wiley;
163
-180.
Stutchbury BJM,
1998
. Female mate choice of extra-pair males: breeding synchrony is important.
Behav Ecol Sociobiol

43
:
213
-215.
Stutchbury BJM, Morton ES,
1995
. The effect of breeding synchrony on extra-pair mating systems in songbirds.
Behaviour

132
:
675
-690.
Stutchbury BJM, Piper WH, Neudorf DL, Tarof SA, Rhymer JM, Fuller G, Fleischer RC,
1997
. Correlates of extra-pair fertilization success in hooded warblers.
Behav Ecol Sociobiol

40
:
119
-126.
Trail PW,
1985
. The intensity of selection: intersexual and interspecific comparisons require consistent measures.
Am Nat

126
:
434
-439.
1997
. Breeding synchrony and extra-pair mating in red-winged blackbirds.
Behav Ecol Sociobiol

40
:
151
-158.
Webster MS, Pruett-Jones S, Westneat DF, Arnold SJ,
1995
. Measuring the effects of pairing success, extra-pair copulations and mate quality on the opportunity for sexual selection.
Evolution

49
:
1147
-1157.
Webster MS, Westneat DF,
1998
. The use of molecular markers to study kinship in birds: techniques and questions. In:
Molecular approaches to ecology and evolution
(DeSalle R, Schierwater B, eds). Boston: Birkhäuser;
7
-35.
Westneat DF,
1993
. Polygyny and extrapair fertilizations in eastern red-winged blackbirds (Agelaius phoeniceus).
Behav Ecol

4
:
49
-60.
Westneat DF, Gray EM,
1998
. Breeding synchrony and extrapair fertilizations in two populations of red-winged blackbirds.
Behav Ecol

9
:
456
-464.
Westneat DF, Webster MS,
1994
. Molecular analyses of kinship in birds: interesting questions and useful techniques. In:
Molecular ecology and evolution: approaches and applications
(Schierwater B, Streit B, Wagner GP, DeSalle R, eds). Basel: Birkhäuser;
91
-126.
Yezerinac SM, Weatherhead PJ, Boag PT,
1995
. Extra-pair paternity and the opportunity for sexual selection in a socially monogamous bird (Dendroica petechia).
Behav Ecol Sociobiol

37
:
179
-188.