Abstract

Inbreeding depression has been hypothesized to drive the evolution of mating systems and dispersal. Some studies have shown that inbreeding strongly affects survival and/or fecundity, but other studies suggest that fitness consequences of inbreeding are less detrimental or more complex. We studied consequences of mating with a relative in a population of great tits ( Parus major ) with a high local recruitment rate. Genotypic information from microsatellite markers was used to calculate coefficients of kinship, and fitness was measured as seasonal and lifetime reproductive success. We show that mating with a relative affects seasonal reproductive success, as was found in other studies of the same species. However, these effects do not result in a lifetime fitness reduction, suggesting that individuals may have scope for avoidance of inbreeding after inbreeding depression. Several explanations are proposed as compensatory mechanisms. Although individuals are more likely to divorce after experiencing inbreeding depression, we show that divorce alone cannot explain the compensation for inbreeding depression in subsequent breeding attempts in our study. We conclude that the costs of mating with a relative in the short term do not necessarily imply lifetime fitness consequences.

Inbreeding depression has been hypothesized to drive the evolution of mating systems ( Brooker et al., 1990 ) and dispersal ( Greenwood et al., 1978 ; Wolff, 1992 ). These models assume that inbreeding is costly and is avoided in natural populations. This has been challenged: inbreeding may not be as costly as supposed, and many species inbreed intensely in nature ( Shields, 1982 ). Although some studies have shown that inbreeding affects survival and/or fecundity ( Keller, 1998 ; Thornhill, 1993 ) and may ultimately lead to population extinction ( Bijlsma et al., 2000 ; Hartt and Haefner, 1995 ; Nieminen et al., 2001 ; Saccheri et al., 1998 ), other studies suggest that lifetime fitness consequences of inbreeding are less detrimental and/or more complex (e.g., Rowley et al., 1986 ).

Several mechanisms could reduce lifetime fitness consequences of inbreeding. Because fitness is often strongly affected by environmental factors (e.g., Naef-Daenzer and Keller, 1999 ; van Noordwijk et al., 1995 ), including parental quality in species with parental care, inbreeding depression can be compensated if inbreeding individuals are associated with better environments or have better parental abilities. Alternatively, spurious inbreeding depression could be observed if inbreeding individuals are associated with worse environments or have worse parental abilities. Inbreeding levels and environmental factors may interactively affect fitness ( Bijlsma et al., 1999 ; Chen, 1993 ; Dahlgaard and Loeschke, 1997 ; Heschel and Paige, 1995 ; Jimenez et al., 1994 ; Keller et al., 1994 ; Miller, 1994 ; Pray et al., 1994 ). Inbreeding depression in one life stage does not necessarily result in effects in later life stages: hatching failure can reduce sib competition and positively affect subsequent survival ( Kempenaers et al., 1996 ; van Noordwijk and Scharloo, 1981 ). Finally, the cost of mating with a relative may not affect lifetime reproductive success because different mates are chosen for different matings ( Brooker et al., 1990 ; Kempenaers et al., 1998 ; Madsen et al., 1992 ). If the above-mentioned mechanisms reduce or neutralize the cost of inbreeding, the argument that inbreeding depression drives the evolution of mating systems and natal dispersal should be reconsidered.

Although the genetic mechanisms underlying inbreeding depression may be studied most easily under controlled conditions ( Lacy et al., 1996 ; Schultz and Willis, 1995 ), in the light of its relevance for the evolution of mating systems and dispersal, it is interesting to study the consequences of mating with a relative in natural populations where mating decisions and environmental conditions are not under control. The shortage of studies in natural populations addressing these issues certainly follows from logistical difficulties with obtaining information on the coefficient of kinship of mates (i.e., the probability that two alleles at an autosomal locus, one drawn randomly from each individual, are identical by descent; Jacquard, 1974 ). The coefficient of kinship can be calculated from pedigrees ( Lynch and Walsh, 1998 ), but if for a substantial part of individuals in the population social parents differ from biological parents or are unknown (e.g., for immigrants), one has to make assumptions about it ( Keller et al., 1994 ; van Noordwijk and Scharloo, 1981 ; van Tienderen and van Noordwijk, 1988 ), and the consequences of inbreeding cannot be tested reliably. Molecular markers provide a tool to circumvent this problem. Lynch (1988 , 1990 ) advises against using the band-sharing coefficient (the fraction of bands shared by two individuals; Wetton et al., 1987 ) because it overestimates the coefficient of kinship to an extent that is unique and unknown for each individual. This upward bias is larger for distant than for close relationships and larger when marker alleles are common than when they are rare ( Lynch, 1988 , 1990 ). Consequently, studies testing for a relation between fitness and the band-sharing coefficient ( Bensch et al., 1994 ; Kempenaers et al., 1996 ) necessarily overestimate inbreeding depression.

With the discovery of neutral markers that provide locus-specific genotypic information ( Queller et al., 1993 ; Tautz, 1989 ) and with the development of appropriate statistical methods ( Li et al., 1993 ; Lynch and Ritland, 1999 ; Queller and Goodnight, 1989 ; Ritland, 1996 ), unbiased estimators of the coefficient of kinship can be obtained for all individuals in a population. Marker-based estimators show considerable measurement error due to sampling variance ( Lynch and Ritland, 1999 ; Van de Casteele et al., 2001 ), but this can be reduced by using information from as many and as polymorphic loci as possible. With these marker-based estimators the negative consequences of inbreeding can be tested with more confidence in populations with high levels of immigration and extrapair paternity.

We studied consequences of mating with a relative in a population of great tits with a high local recruitment rate ( Matthysen et al., 2001 ). Genotypic information from microsatellite markers was used to calculate coefficients of kinship. First, we showed that mating with a relative affects seasonal reproductive success. Second, we showed that effects of inbreeding on seasonal reproductive success do not result in effects on lifetime reproductive success. Finally, we tested which mechanisms could cause compensation between breeding attempts.

METHODS

Study area and data collection

We obtained breeding data for the great tit from a study area in northern Belgium consisting of 13 deciduous (mainly oak, Quercus robur ) woodlots ranging in size between 1 and 12 ha in a matrix of farmland and residential areas. In all woodlots nest-boxes are available at a constant density of about nine per hectare (see Nour et al., 1998 , for more details). Between woodlots few nest-boxes are provided by the public, and breeding density is low (Van de Casteele, unpublished data). We determined laying date (= date of first egg), clutch size, and hatching date (= date first egg hatched) from at least weekly visits to occupied nest-boxes. Parents were captured while feeding their 8-day-old young each year from 1994 on, and their body weight (to the nearest 0.1 g) and tarsus (to the nearest 0.01 mm) and wing length (to the nearest 0.5 mm) were measured. At some nests the identity and/or morphological measurements of at least one parent were not known due to clutch desertion or brood loss before the nestlings were 8 days old or because the capture attempt was unsuccessful. Capture rate of parents was 98% for females and 90% for males (Matthysen, unpublished data). We ringed nestlings at least from 1984 on. When 15 days old their body weight (= fledging weight) and tarsus (= fledging tarsus) length were measured from 1996 on (in 1999 only the mean weight per brood was measured). We collected blood samples from the parents for microsatellite analysis during the breeding seasons of 1996 and 1997 by puncturing the brachial vein and stored the samples at −80°C.

Microsatellite data

We successfully scored 315 different individuals for 9 microsatellite loci ranging in polymorphism between 2 and 53 alleles. Genotypes of both pair members were known for 154 pairs. For more details about the microsatellite loci and the laboratory protocol, refer to Van de Casteele et al. (2001) .

Kinship estimates

Several unbiased estimators of the coefficient of relatedness (i.e., the expected fraction of alleles in the genome that two individuals have identical by descent; Roughgarden, 1996 ) have been developed ( Li et al., 1993 ; Lynch and Ritland, 1999 ; Queller and Goodnight, 1989 ; Ritland, 1996 ). The coefficient of relatedness divided by two equals the coefficient of kinship. Many factors affect the relative performance of estimators of the coefficient of relatedness ( Van de Casteele et al., 2001 ). We used a weighted similarity index because it performed best for the set of loci we studied in the great tit (for details, see Van de Casteele et al., 2001 ) and divided it by two to get estimates of the coefficient of kinship.

Appropriately estimating allele frequencies is crucial to calculating estimates of coefficients of kinship. First, the presence of genetic substructure in a population would jeopardize the use of the estimators ( Queller and Goodnight, 1989 ; Ritland, 1996 ), and second, allele frequencies should equal the gene frequencies in the gene pool from which alleles were randomly drawn during the formation of the pedigree ( Ritland, 1996 ). We used a model-based clustering method for inferring population structure from multilocus genotype data implemented in the software package structure ( Pritchard et al., 2000 ). With this method an appropriate value for K , the number of subpopulations, is determined that best describes the genotypic data. A series of runs of the Gibbs sampler ( K = 1–12, the number of different plots where data were collected) with a burn-in period of 10,000 and 10,000 Monte Carlo Markov chain repetitions ( Pritchard, 2000 ) revealed no genetic structure in our data for either year (1996 and 1997). This result was consistent over different runs for the same K. This implies that allele frequencies should be estimated from the pooled study sites. Because allele frequencies did not differ significantly between years and to increase sample size, they were estimated from the set of all genotyped individuals, pooled over the 2 years, counting each individual only once if it was sampled in both years ( Van de Casteele et al., 2001 ). Marker-based estimates of coefficients of kinship were calculated using programs written in Mathematica 3.0 (Wolfram Research; //www.wolfram.com/ ) by TVC. In the remainder of the text we use the term “kinship” to refer to the coefficient of kinship.

Statistical analyses

Effects of mating with a relative on seasonal reproductive success

Inbreeding depression can be tested as a regression of reproductive success on kinship ( Falconer and Mackay, 1996 ; Lynch and Walsh, 1998 ). The following components of seasonal reproductive success (SRS) were studied: standardized laying date (= day the first egg was laid minus the yearly mean laying date of first clutches), clutch size, hatching rate (= ratio of number of hatchlings on clutch size), nestling survival (= number of fledglings per hatchling), fledging success (= ratio of number of fledglings on clutch size), fledging size (= mean tarsus length of nestlings on day 15), fledging condition (= mean residual per brood from a regression of fledging weight on fledging tarsus), number of fledglings (= number of young fledged), number of recruits (= number of fledglings that was captured as a breeding individual the next breeding season or at any later time), and recruitment rate (= ratio of number of recruits on number of fledglings). We studied only first broods. For all analyses of SRS the following data were disregarded: nests where at least one parent was unknown, uncompleted or unbrooded clutches, destroyed clutches (by squirrels, weasels, or humans), clutches deserted before the end of incubation, partially or completely destroyed broods (by woodpeckers, weasels, or humans), nests containing young from other bird species (blue tits), and a small number of broods used in a cross-fostering experiment. For analyses of nestling survival, fledging success and the number of fledglings analyses were restricted to clutches that produced at least one fledgling ( n = 2 disregarded nests).

In a first analysis we tested what variables other than kinship affected SRS. Several environmental variables were considered that were ecologically relevant and/or were shown to affect SRS in other studies: clutch size (e.g., Horak et al., 1997 ), standardized laying date ( Verhulst and Tinbergen, 1991 ; Verhulst et al., 1995 ), midparent tarsus length ( Garnett, 1981 ; Gebhardt-Henrich and van Noordwijk, 1991 ), midparent condition ( Gebhardt-Henrich and van Noordwijk, 1991 ), male age ( Dhondt, 1971 ; Krebs, 1971 ), female age ( Dhondt, 1989 ; Perrins and Moss, 1974 ), year (e.g., Verhulst, 1995 ), and the interaction between year and standardized laying date (e.g., Horak et al., 1997 ; Verhulst, 1995 ). To make the data set as large as possible, breeding data were used from 1996 to 1999, but in 1999 tarsi of pulli were not measured, resulting in no available data for mean fledging tarsus and condition for that year. To improve normality, we transformed dependent variables: clutch size was log( x )-transformed, the number of fledglings and the number of recruits were transformed as log( x + 0.5), and hatching rate, nestling survival, fledging success, and recruitment rate of fledglings were transformed as arcsine√ x. A linear model was fitted starting from a model with all independent variables and removing all nonsignificant terms in inverse order of significance. Pseudoreplication from multiple records for the same individual in different years was taken into account by adding a random effect both for the male and the female. We used data from 505 breeding records for 358 different males and 363 different females.

In a second analysis we tested inbreeding depression as a regression of SRS on kinship. For these analyses data on kinship were available for 1996 and 1997, and of the 2 years only the earliest breeding record of an individual was used. Inbreeding depression was tested in a one-way analysis by regressing SRS on kinship and in a multiway analysis using a linear model with kinship and environmental variables as independent variables. For the latter analysis a model was constructed containing environmental variables that were significant in the first analysis plus interactions of kinship with year, laying date, male age, and female age. Then we removed nonsignificant terms belonging to the latter set of variables from the model in inverse order of significance and removed interactions before main effects, but environmental variables that were significant in the first analysis were kept in the model regardless of their significance. We made the following predictions: (1) if inbreeding depression is masked by environmental effects, then kinship would be nonsignificant in the one-way analysis but significant in the multiway analysis; (2) if inbreeding effects are spurious (i.e., are not causal but an effect of a correlated environmental variable), then kinship would be significant in the one-way analysis but nonsignificant in the multiway analysis; (3) if inbreeding depression depends on environmental factors, then inbreeding depression would be nonsignificant in the one-way analysis but significant in interaction with an environmental variable in the multiway analysis.

We used a logistic model for the analysis of survival data because, when genetic effects of different loci are independent but interact multiplicatively among loci, a linear decline of a trait with the coefficient of kinship is expected on a logarithmic scale ( Charlesworth and Charlesworth, 1987 ). For other variables a linear model was used with transformed data.

Effects of mating with a relative in the first breeding attempt on lifetime reproductive success

We measured lifetime reproductive success (LRS) as the lifetime number of fledglings (= number of fledglings produced over an individual's lifetime) and as the lifetime number of recruits (= number of recruits produced over an individual's lifetime). The proportion of local recruits in the first year breeding cohort (= all recruits born in our study area) was 48% and 40% for males and females, respectively ( Matthysen et al., 2001 ). Log( x + 0.5)-transformed LRS data were regressed on kinship in the first breeding attempt. The analysis was performed separately for first-year females and first-year males.

Compensation for inbreeding depression

For several reasons a negative effect of mating with a relative does not necessarily result in negative effects on LRS: individuals that experienced inbreeding depression at a first breeding attempt may (1) survive better than individuals mated with nonrelatives, (2) increase SRS in subsequent breeding attempts by breeding with a different, less related mate, (3) increase SRS by dispersing to better quality territories, or (4) increase SRS through an increase in the number of extrapair offspring (from less related or better quality mates), or a combination of these.

Parental survival to the next breeding season was determined from recapture of an individual the next breeding season; therefore it is a combination of survival and capture probability which is close to 1. The remating status of individuals was defined as follows: “faithful,” both pair members survived to and bred together at the next breeding attempt; “divorced,” both pair members survived to the next breeding season but were not faithful; “widowed,” an individual survived to the next breeding season but there was no evidence that the partner from the previous year was still alive; “died,” there was no evidence that an individual survived to the next breeding season. Breeding dispersal was defined as the distance in meters between the breeding locations of two subsequent breeding seasons and was log( x + 0.5)-transformed in all analyses. Because we only had microsatellite data of parents, we could not determine if individuals improve SRS after inbreeding depression by having more extrapair offspring.

For all analyses we used the SAS software version 6.12. Linear models were tested with Proc Mixed, which uses generalized least squares to estimate slopes and is particularly superior to ordinary least squares for unbalanced designs ( Littell et al., 1996 ). When random factors were present in a model, denominator degrees of freedom for tests of fixed effects were calculated using Satterthwaite's formula, and the NoBound option was used to allow estimation of negative within-subject variances ( Littell et al., 1996 ). We tested logistic models with Proc Genmod with a binomial error structure and a logit link function.

RESULTS

Population level of kinship

The mean kinship did not differ significantly between years ( F1,121 = 0.51, p =.47) and was not significantly different from zero (1996 and 1997 pooled, mean ± SE: 0.0075 ± 0.0083, t = 0.90, df = 122, p =.37).

Effect of mating with a relative on SRS

We tested which variables other than kinship affected SRS. The results are presented in Table 1 . For the analysis of the number of fledglings, no convergence of the iterative procedure to fit the mixed model was achieved unless nests where no young fledged were disregarded. Larger clutches were associated with a lower mean fledging condition and a higher number of fledglings, suggesting a trade-off between offspring quality and number (see Nur, 1984 ; Smith et al., 1989 ; Stearns, 1992 ). Laying date (as a main effect or in interaction with year) affected clutch size and mean fledging condition (see, e.g., Verhulst and Tinbergen, 1991 ). Parents in lower condition had a larger clutch size, hatching rate, nestling survival, fledging success, and number of fledglings, suggesting that body weight is adjusted to or affected by reproductive effort ( Norberg, 1981 ; Sanz and Moreno, 1995 ; Woodburn and Perrins, 1997 ). Mean parent tarsus length and condition affected mean fledging tarsus length and condition, respectively, reflecting a heritable component to both traits (see Garnett, 1981 ; Gebhardt-Henrich and van Noordwijk, 1991 ; van Noordwijk et al., 1988 ). Male age affected laying date, recruitment rate, and the number of recruits (see Perrins and McCleery, 1985 ). Female age significantly affected laying date and clutch size (see Dhondt, 1989 ; Perrins and McCleery, 1985 ; Perrins and Moss, 1974 ). Year effects were found for hatching rate, nestling survival, fledging success, and the mean fledging tarsus.

Results of the tests for inbreeding depression in a one-way analysis are presented in Table 2 . Negative effects of mating with a relative were found for hatching rate, mean fledging tarsus, nestling survival, fledging success, and number of fledglings. After applying the sequential Bonferroni procedure for multiple tests ( k = 10, pcrit =.0051 for the most significant test; Sokal and Rohlf, 1995 ), effects remained significant except for mean fledging tarsus. The strongest effect was found for hatching rate. Although the coefficient of kinship negatively affects the number of fledglings and has no positive effect on the recruitment rate, we did not find a significant relation between the number of recruits and kinship.

Results of the tests for inbreeding depression in a multiway analysis are presented in Table 2 . A significant interaction of an environmental variable with kinship was found for clutch size, nestling survival, fledging success, and the number of fledglings. After sequential Bonferroni correction ( k = 10), only the interaction of kinship and laying date remained significant for nestling survival and fledging success: the negative interaction indicates that the negative effects of mating with a relative are more pronounced for late clutches. For laying date, hatching rate, mean fledging condition, mean fledging tarsus length, recruitment rate of fledglings, and the number of recruits, the effect of kinship was similar to the results for the one-way analysis. When kinship was calculated from all but one locus, the results were similar, suggesting that the observed effects were not caused by individual marker loci.

To summarize, we found negative effects of mating with a relative on SRS, but we found no evidence that negative effects are masked or caused by environmental effects or that inbreeding effects depend on environmental conditions.

Effects of mating with a relative in the first breeding attempt on LRS

We found no significant relation between LRS and kinship at the first breeding attempt for females (lifetime number of fledglings: slope = 0.10 ± 0.85, F1,81 = 0.01, p =.90; lifetime number of recruits: slope = 0.44 ± 0.93, F1,81 = 0.22, p =.64) or males (lifetime number of fledglings: slope = −0.30 ± 0.76, F1,87 = 0.16, p =.69; lifetime number of recruits: slope = −0.25 ± 0.91, F1,87 = 0.08, p =.78). There is a discrepancy between the effects of mating with a relative on SRS and LRS, implying that there is compensation for inbreeding depression in subsequent breeding attempts. The fact that the sign of the slopes is positive for females and negative for males suggests that compensation is stronger for females than for males.

What causes this compensation? One explanation is that yearly survival is higher for individuals with high kinship than for individuals with low kinship. There was no significant relation between survival (0/1) and kinship (only first-year individuals; females: 34 out of 83 survived, χ 2 = 2.29, df = 1, p =.13, slope = 3.67 ± 2.48; males: 41 out of 89 survived: χ 2 = 0.03, df = 1, p = 0.87, slope = −0.36 ± 2.25; year and year × kinship not significant in any analysis: p >.10).

Because survival differences could not explain the discrepancy, it has to be due to a stronger increase in recruitment from subsequent breeding attempts for individuals with a high kinship than for individuals with a low kinship. Because kinship strongly affects hatching rate, we compared the probability of an increase in SRS (number of fledglings and number of recruits) in a subsequent breeding attempt for individuals that did and did not experience hatching failure in a previous breeding attempt (only the earliest breeding record of an individual was used). To make the data set as large as possible, data for 4 years (1996-1999) were used. The probability of an increase in SRS depended on hatching failure in the previous breeding attempt for females (Fisher's Exact test; n = 89, fledglings: 10/15 vs. 23/74, p <.02; recruits: 8/15 vs. 15/74, p < 0.02) but not for males ( n = 81, fledglings: 7/14 vs. 29/67, p =.77; recruits: 3/14 vs. 21/67, p =.54). Thus, females but not males succeeded in increasing SRS after a breeding attempt with hatching failure. This suggests that compensation was stronger for females than for males.

Divorce and inbreeding

We tested whether individuals were more likely to divorce after experiencing inbreeding depression. Individuals with a different remating status differed in kinship in the year before remating, but this was only a trend for females (females: F3,107 = 2.65, p <.06; males: F3,107 = 3.13, p <.03; all comparisons with divorce, p <.05, except divorced vs. faithful, p <.13). However, there was a strong difference between years: kinship differed significantly between individuals with a different remating status in 1996 ( Figure 1 ; females: F3,62 = 5.19, p <.01; males: F3,62 = 6.12, p <.001): individuals that would divorce in the next breeding season had a higher kinship than individuals of any other remating status (all comparisons with divorce, p <.05; all other comparisons, p >.05). In 1997 this difference did not exist (females: F3,41 = 0.94, p =.43; males: F3,41 = 0.15, p =.93). The decision to divorce may depend more strongly on inbreeding depression than on kinship because kinship estimates show considerable measurement error or because individuals are not able to assess kinship directly. Therefore, we tested whether the decision to divorce depended on hatching failure. Data for 4 years (1996–1999) were combined, and only the earliest breeding record of an individual was used.

For individuals that were known to be alive in 2 consecutive years (divorced or faithful), 5 of 7 (71%) individuals that experienced hatching failure in their first recorded breeding attempt divorced, compared to 9 of 35 (26%) that experienced no hatching failure (Fisher's Exact test; p <.04). There was a difference in hatching rate between individuals with a different remating status (females: n = 220, χ 2 = 13.92, df = 3, p <.01; males: n = 220, χ 2 = 19.88, df = 3, p <.001): both for females and males, individuals that would divorce in the next breeding season had a lower hatching rate than any other remating status ( Figure 2 ; females: all comparisons with divorced, p <.05; males: all comparisons with divorced, p <.05, except widowed vs. divorced, p =.20). Thus, divorce was related to kinship with the mate or, in a larger data set, hatching rate. This suggests that mate fidelity between breeding seasons is affected by hatching failure at the previous breeding attempt.

Next we tested whether individuals benefited from divorce in terms of a reduced kinship and improved SRS. We tested whether divorced individuals significantly reduced kinship from 1996 to 1997 and compared the change with widowed individuals. The change in kinship was tested as an interaction between year and remating status (divorced or widowed) in a mixed model with kinship as a dependent variable, year and remating status as fixed factors, and individual as a random factor. The change in kinship was significant for divorced females and differed significantly between divorced and widowed females ( n = 13, year × remating status: F1,11 = 5.41, p =.04; divorced: change = −0.16 ± 0.06, p <.02; widowed: change = −0.01 ± 0.03, p =.69, see Figure 1 ). The change in kinship was almost significant for divorced males and differed significantly between divorced and widowed males ( n = 10, year × remating status: F1,8 = 7.62, p <.03; divorced: change = −0.18 ± 0.08, p <.07; widowed: change = 0.10 ± 0.05, p =.11, see Figure 1 ).

To test whether divorcees improved SRS more than faithful or widowed birds after a breeding attempt with hatching failure, we analyzed the change in the number of fledglings and the number of recruits. The probability of an increase in the number of fledglings did not depend on remating status (Fisher's Exact test; females: p <.06; males: p =.27, see Table 3 ). The probability of an increase in the number of recruits did not depend on remating status (Fisher's Exact test; females: p =.50; males: p = 0.33; see Table 3 ). Thus, although individuals were more likely to divorce after inbreeding depression and succeeded in obtaining mates with lower kinship, divorce alone cannot explain compensation for inbreeding depression over an individual's lifetime.

Breeding dispersal and inbreeding

A final possible compensatory mechanism could be breeding dispersal (to better territories). Because breeding dispersal did not depend on remating status (females: F2,86 = 0.08, p =.92, males: F2,78 = 0.19, p =.83), we could test whether breeding dispersal was affected by kinship or hatching rate regardless of remating status. We found no significant relation between breeding dispersal and kinship (Pearson correlation; females: r =.23, n = 47, p =.12; males: r =.27, n = 48, p =.07) or hatching rate (Pearson correlation: females: r = −.13, n = 89, p =.23; males: r = −.00, n = 81, p =.97).

DISCUSSION

In this study we tested the fitness consequences of having a relative as a social mate. We found strong negative effects on SRS but no effects on LRS. This suggests that there is compensation for inbreeding depression in subsequent breeding attempts.

The effects we found on SRS of being mated with a relative are comparable to results of some other studies of natural bird populations: a strong effect of the kinship of the parents on hatching rate (see Bensch et al., 1994 ; Kempenaers et al., 1996 ; van Noordwijk and Scharloo, 1981 ), fledging success (see Greenwood et al., 1978 ), and the number of fledglings ( Brown and Brown, 1998 ). However, we found no significant inbreeding depression on fledgling survival, contrary to some other studies ( Brown and Brown, 1998 ; Daniels and Walters, 2000 ). As all bird studies of inbreeding depression ( Crnokrak and Roff, 1999 ), our study was correlational: we can never be sure that the observed effects are the consequence of a correlation between kinship and an unmeasured variable reflecting environmental conditions (e.g., territory quality) or parental phenotypic quality (e.g., parental care). However, there was no evidence that the negative consequences of mating with a relative are masked by environmental conditions or by an interaction between environmental conditions and kinship, or that the observed inbreeding effect would be spurious.

Our results should be more reliable than most other studies of inbreeding depression in natural bird populations because the estimators we used provide unbiased estimates of kinship in the presence of extrapair parentage and for individuals with an unknown origin. Most other studies use pedigree-based kinship estimates and make assumptions about it for a considerable number of individuals or use the band-sharing coefficient, which is a biased estimator of kinship ( Lynch, 1988 , 1990 ). Two issues could cast doubt on the reliability of our results, but we believe we can easily reject them. First, as we mentioned in the introduction, kinship estimators show considerable measurement error that depends on the number and polymorphism of scored microsatellite loci ( Lynch and Ritland, 1999 ; Van de Casteele et al., 2001 ). One should be careful in applying statistical methods that do not take measurement error on independent variables into account ( Bollen, 1989 ). In univariate regression models, measurement error on independent variables leads to attenuated slopes ( Bollen, 1989 ), implying that the tests in our one-way analyses are conservative. Still, we found strong effects on SRS. Second, if the social parents of a brood are not the biological parents, this induces additional error on the fitness estimates. In other great tit populations the frequency of extrapair paternity has been shown to range between 3.5% and 15% of nestlings ( Blakey, 1994 ; Gullberg et al., 1992 ; Krokene et al., 1998 ; Lubjuhn et al., 1999 ; Strohbach et al., 1998 ; Verboven and Mateman, 1997 ), but intraspecific brood parasitism is absent ( Blakey, 1994 ; Gullberg et al., 1992 ; Kempenaers et al., 1995 ; Strohbach et al., 1998 ) or rare (0.2%; Verboven and Mateman, 1997 ). We have no data on extrapair parentage in our population, but these studies suggest that fitness is accurately estimated from attended nests for females and less so for males depending on the occurrence of extrapair paternity. Therefore, we conclude that there is inbreeding depression on SRS.

The consequences of inbreeding on SRS were strong up to fledging but weaker for recruitment. However, we found no relation between recruitment rate and kinship. Therefore, compensatory differences in recruitment between inbred and outbred nests ( van Noordwijk and Scharloo, 1981 ) probably cannot explain the lack of significance of the relation for the number of recruits. Rather, additional noise on the relation between the number of fledglings and the number of recruits should explain this. The number of recruits is without doubt estimated with much lower accuracy than the number of fledglings, which have a capture probability of 1, because recruits are missed in the study area or dispersed out of the study area. This obviously reduces the power of finding a significant relation.

Few studies have looked at lifetime consequences of mating with a relative. Keller (1998 ; Keller et al., 1994 ) showed that an individual's LRS can be strongly reduced as a consequence of being inbred in an island population. In our study lifetime fitness consequences of mating with a relative were weak, both when LRS was measured with high accuracy (lifetime number of fledglings) as when it was measured with lower accuracy (lifetime number of recruits). This is in contrast with the strong effects on SRS.

The compensation of inbreeding depression over an individual's lifetime cannot be explained by a higher number of subsequent breeding attempts for related than for unrelated pairs: we found no evidence for a relation between adult local survival and kinship. Rather, compensation should be explained by a stronger increase in recruitment for related pairs than for unrelated pairs in subsequent breeding attempts. This was the case for females but not for males, which suggests that females but not males succeed in compensating for inbreeding depression. Alternatively, this finding is a consequence of the fact that the male's (change in) SRS is measured with lower accuracy than the female's due to increased extrapair paternity.

Of all possible explanations for compensation between successive breeding attempts, we found evidence for one: that individuals avoid inbreeding and inbreeding depression through divorce. Our results are in part consistent with the incompatibility hypothesis which states that individuals would benefit from divorce when their combined qualities result in reduced fitness ( Choudhury, 1995 ). Both females and males had a higher kinship and a lower hatching rate before divorce than any other remating status. The relation between the decision to divorce and SRS at the previous breeding attempt is not straightforward when other studies are taken into account: Dhondt et al. (1996) and Kempenaers et al. (1998) found no relation, Perrins and McCleery (1985) found that clutch size but not the number of fledglings was smaller for divorcees than for faithful individuals, implying that fledging success was higher for divorcees, and, in an experimental study, Lindén (1991) showed that individuals were more likely to remain faithful when SRS was experimentally increased and less likely to remain faithful when SRS was decreased. However, only in the study of Kempenaers et al. (1998) was SRS measured as hatching rate. Perhaps more studies should focus on hatching rate as a measure of SRS, as hatching failure is a consistent cost to inbreeding in bird species ( Daniels and Walters, 1999 ). Moreover, experimental manipulation of unhatched eggs would reveal whether hatching failure is used as cue to decide to divorce.

Predictions about the consequences of divorce were not all fulfilled. Our results showed that individuals effectively reduced kinship by divorcing compared to any other remating status. However, we failed to show that individuals benefited from divorce compared to faithful or widowed individuals. Divorced females tended to improve SRS after hatching failure more than faithful and widowed females when measured as the number of fledglings. No effect was found for males or for the number of recruits of females. Kempenaers et al. (1998) showed that hatching rate increased more after hatching failure for divorcees than for faithful blue tits ( Parus caeruleus ). For great tits, hatching rate has not been studied in relation to divorce, but neither Perrins and McCleery (1985) , nor Dhondt et al. (1996) found evidence for improved SRS after divorce.

We conclude that compensation for inbreeding depression between successive breeding attempts cannot be explained by inbreeding avoidance through divorce alone. Individuals could use divorce as one of several ways to improve SRS. It has been suggested that promiscuity is a strategy to avoid the negative consequences of inbreeding ( Brooker et al., 1990 ; Madsen et al., 1992 ; Stockley et al., 1993 ). Neither Kempenaers et al. (1996) , nor Keller and Arcese (1998) found evidence for such a mechanism in other species. However, in the absence of other mechanisms of kin recognition, individuals could use hatching failure as a cue of kinship and decide to engage more in extrapair parentage in a subsequent attempt. In that case, individuals could use divorce and promiscuity as alternative or complementary strategies (see Ramsay et al., 2000 ) to avoid the negative consequences of inbreeding. A detailed study of extrapair parentage in relation to kinship, hatching failure, and divorce would reveal such a pattern. A final possible mechanism causing compensation for inbreeding depression in a subsequent breeding attempt is a trade-off between SRS of successive breeding attempts ( Daan and Tinbergen, 1997 ; Gustafsson and Sutherland, 1988 ). The fact that there is a relation between hatching rate and parental body condition, for example ( Table 1 ), possibly indicates such a cost of reproduction.

Whatever causes compensation, our results indicate that, although there may be fitness costs to inbreeding on a short term, this does not lead to lifetime costs: individuals have scope for avoidance of inbreeding depression after one breeding attempt. Hence, if there are any differences in lifetime reproductive success between individuals with different natal dispersal strategies, it is unlikely that they are caused by differences in kinship.

Figure 1

Kinship in two subsequent breeding seasons for individuals with a different remating status. Filled circles: 1996, first recorded breeding attempt; open circles: 1997, after remating. Means and standard errors were calculated from generalized linear models (see text for details)

Figure 1

Kinship in two subsequent breeding seasons for individuals with a different remating status. Filled circles: 1996, first recorded breeding attempt; open circles: 1997, after remating. Means and standard errors were calculated from generalized linear models (see text for details)

Figure 2

Hatching rate of first recorded breeding attempt before remating for individuals with a different remating status. Means and standard errors were calculated from generalized linear model with hatching rate (untransformed) as a dependent and remating status as an independent variable

Figure 2

Hatching rate of first recorded breeding attempt before remating for individuals with a different remating status. Means and standard errors were calculated from generalized linear model with hatching rate (untransformed) as a dependent and remating status as an independent variable

Table 1

Generalized linear model of seasonal reproductive success with variables other than kinship as independent variables.

 Seasonal reproductive success 
Independent variables Standardized laying date  Clutch size a  Hatching rate b  Nestling survival b  Fledging success b 
Clutch size — — F1,393 = 2.77  F1,306 = 2.59  F1,313 = 0.43  
Standardized lay date — F1,419 = 37.61 *** F1,383 = 0.02  F1,291 = 0.15  F1,314 = 0.38  
Midparent tarsus length F1,341 = 0.02  F1,339 = 0.39  F1,361 = 2.24  F1,292 = 0.13  F1,330 = 0.44  
Midparent condition F1,408 = 1.65  F1,405 = 5.85 * F1,365 = 8.04 ** F1,291 = 4.65 * F1,308 = 9.14 ** 
  (−0.031 ± 0.013) (−0.040 ± 0.014) (−0.051 ± 0.024) (−0.077 ± 0.025) 
Age male F1,449 = 7.67 ** F1,344 = 0.00  F1,418 = 0.26  F1,419 = 0.49  F1,417 = 1.53  
Yearling (0.93 ± 0.26)     
Adult (−0.61 ± 0.29)     
Age female F1,439 = 19.82 *** F1,376 = 7.39 ** F1,421 = 1.03  F1,401 = 0.07  F1,398 = 0.00  
Yearling (1.08 ± 0.26) (2.24 ± 0.01)    
Adult (−0.86 ± 0.29) (2.27 ± 0.01)    
Year — F3,296 = 1.30  F3,380 = 8.03 *** F3,362 = 7.70 *** F3,358 = 3.57 * 
1996   (1.513 ± 0.016) (1.396 ± 0.028) (1.352 ± 0.030) 
1997   (1.533 ± 0.015) (1.284 ± 0.027) (1.252 ± 0.028) 
1998   (1.431 ± 0.017) (1.460 ± 0.029) (1.331 ± 0.031) 
1999   (1.511 ± 0.015) (1.419 ± 0.025) (1.370 ± 0.026) 
Year × standardized lay date — F3,302 = 2.91 * F3,406 = 0.69  F3,412 = 1.76  F3,408 = 1.19  
1996  (−0.019 ± 0.005)    
1997  (−0.006 ± 0.002)    
1998  (−0.008 ± 0.003)    
1999  (−0.014 ± 0.003)    
 Seasonal reproductive success 
Independent variables Standardized laying date  Clutch size a  Hatching rate b  Nestling survival b  Fledging success b 
Clutch size — — F1,393 = 2.77  F1,306 = 2.59  F1,313 = 0.43  
Standardized lay date — F1,419 = 37.61 *** F1,383 = 0.02  F1,291 = 0.15  F1,314 = 0.38  
Midparent tarsus length F1,341 = 0.02  F1,339 = 0.39  F1,361 = 2.24  F1,292 = 0.13  F1,330 = 0.44  
Midparent condition F1,408 = 1.65  F1,405 = 5.85 * F1,365 = 8.04 ** F1,291 = 4.65 * F1,308 = 9.14 ** 
  (−0.031 ± 0.013) (−0.040 ± 0.014) (−0.051 ± 0.024) (−0.077 ± 0.025) 
Age male F1,449 = 7.67 ** F1,344 = 0.00  F1,418 = 0.26  F1,419 = 0.49  F1,417 = 1.53  
Yearling (0.93 ± 0.26)     
Adult (−0.61 ± 0.29)     
Age female F1,439 = 19.82 *** F1,376 = 7.39 ** F1,421 = 1.03  F1,401 = 0.07  F1,398 = 0.00  
Yearling (1.08 ± 0.26) (2.24 ± 0.01)    
Adult (−0.86 ± 0.29) (2.27 ± 0.01)    
Year — F3,296 = 1.30  F3,380 = 8.03 *** F3,362 = 7.70 *** F3,358 = 3.57 * 
1996   (1.513 ± 0.016) (1.396 ± 0.028) (1.352 ± 0.030) 
1997   (1.533 ± 0.015) (1.284 ± 0.027) (1.252 ± 0.028) 
1998   (1.431 ± 0.017) (1.460 ± 0.029) (1.331 ± 0.031) 
1999   (1.511 ± 0.015) (1.419 ± 0.025) (1.370 ± 0.026) 
Year × standardized lay date — F3,302 = 2.91 * F3,406 = 0.69  F3,412 = 1.76  F3,408 = 1.19  
1996  (−0.019 ± 0.005)    
1997  (−0.006 ± 0.002)    
1998  (−0.008 ± 0.003)    
1999  (−0.014 ± 0.003)    

In parentheses are parameter estimates and their standard errors; parameter estimates for nonsignificant interactions are not shown.

a log( x )-transformed.

b Arcsine √x-transformed.

c log( x + 0.5)-transformed.

d Too many likelihood evaluations (no convergence) when all data are used; therefore, nests where no young fledged were disregarded.

* p <.05;

** p <.01;

*** p <.001.

Table 1

, extended.

Seasonal reproductive success 
Mean fledgling condition Mean fledgling tarsus  No. of fledglings c , d  Recruitment rate of fledglings b  No. of recruits c 
F1,214 = 9.82 ** F1,253 = 1.09  F1,333 = 198.83 *** F1,408 = 0.56  F1,419 = 0.26  
(−0.089 ± 0.028)  (0.105 ± 0.007)   
F1,245 = 1.42  F1,316 = 1.90  F1,305 = 0.54  F1,351 = 0.10  F1,368 = 0.09  
F1,235 = 3.38  F1,300 = 151.70 *** F1,323 = 1.73  F1,368 = 1.54  F1,369 = 1.47  
 (0.592 ± 0.048)    
F1,219 = 32.11 *** F1,243 = 2.40  F1,296 = 5.33 * F1,249 = 0.19  F1,263 = 0.17  
(0.396 ± 0.070)  (−0.042 ± 0.018)   
F1,270 = 1.96  F1,272 = 0.00  F1,409 = 0.04  F1,487 = 9.67 ** F1,500 = 7.96 ** 
   (0.139 ± 0.013) (−0.262 ± 0.040) 
   (0.198 ± 0.014) (−0.098 ± 0.043) 
F1,272 = 1.68  F1,275 = 1.12  F1,413 = 0.55  F1,399 = 0.12  F1,407 = 0.01  
F2,200 = 1.16  F2,262 = 4.15 * F3,363 = 1.82  F3,416 = 0.53  F3,422 = 0.68  
 (−10.47 ± 0.85)    
 (−10.49 ± 0.85)    
 (−10.38 ± 0.85)    
— —    
F2,238 = 10.19 *** F2,306 = 1.08  F3,400 = 0.75  F3,399 = 0.48  F3,407 = 0.38  
(0.074 ± 0.024)     
(0.012 ± 0.013)     
(−0.047 ± 0.016)     
— —    
Seasonal reproductive success 
Mean fledgling condition Mean fledgling tarsus  No. of fledglings c , d  Recruitment rate of fledglings b  No. of recruits c 
F1,214 = 9.82 ** F1,253 = 1.09  F1,333 = 198.83 *** F1,408 = 0.56  F1,419 = 0.26  
(−0.089 ± 0.028)  (0.105 ± 0.007)   
F1,245 = 1.42  F1,316 = 1.90  F1,305 = 0.54  F1,351 = 0.10  F1,368 = 0.09  
F1,235 = 3.38  F1,300 = 151.70 *** F1,323 = 1.73  F1,368 = 1.54  F1,369 = 1.47  
 (0.592 ± 0.048)    
F1,219 = 32.11 *** F1,243 = 2.40  F1,296 = 5.33 * F1,249 = 0.19  F1,263 = 0.17  
(0.396 ± 0.070)  (−0.042 ± 0.018)   
F1,270 = 1.96  F1,272 = 0.00  F1,409 = 0.04  F1,487 = 9.67 ** F1,500 = 7.96 ** 
   (0.139 ± 0.013) (−0.262 ± 0.040) 
   (0.198 ± 0.014) (−0.098 ± 0.043) 
F1,272 = 1.68  F1,275 = 1.12  F1,413 = 0.55  F1,399 = 0.12  F1,407 = 0.01  
F2,200 = 1.16  F2,262 = 4.15 * F3,363 = 1.82  F3,416 = 0.53  F3,422 = 0.68  
 (−10.47 ± 0.85)    
 (−10.49 ± 0.85)    
 (−10.38 ± 0.85)    
— —    
F2,238 = 10.19 *** F2,306 = 1.08  F3,400 = 0.75  F3,399 = 0.48  F3,407 = 0.38  
(0.074 ± 0.024)     
(0.012 ± 0.013)     
(−0.047 ± 0.016)     
— —    
Table 2

Generalized linear model of seasonal reproductive success with kinship as an independent variable.

 Seasonal reproductive success 
Independent variables  Standardized laying date ( n = 111)   Clutch size a ( n = 111)   Hatching rate ( n = 111)   Nestling survival b ( n = 109)   Fledging success b ( n = 109)  
One-way analysis      
Kinship F1,109 = 1.57  F1,109 = 0.22  χ2 = 18.52 *** χ2 = 12.88 *** χ2 = 24.22 *** 
 (−4.40 ± 3.52) (−0.08 ± 0.16) (−9.91 ± 2.20) (−4.19 ± 1.15) (−5.16 ± 1.04) 
Multiway analysis      
Kinship F1,107 = 1.53  F1,101 = 0.66  χ2 = 16.90 ***  χ 2 = 1.41   χ 2 = 3.82  
 (−4.26 ± 3.44) (−0.13 ± 0.16) (−9.54 ± 2.23) (−5.97 ± 1.39) (−7.18 ± 1.25) 
Kinship × female age F1,103 = 0.01  F1,97 = 0.05   χ 2 = 0.05   χ 2 = 0.40   χ 2 = 0.58  
First year      
Adult      
Kinship × male age F1,104 = 0.06  F1,99 = 1.10   χ 2 = 0.49   χ 2 = 4.60 *  χ 2 = 4.97 * 
First year    (−5.60 ± 1.34) (−6.80 ± 1.20) 
Adult    (−0.35 ± 3.74) (−1.06 ± 3.52) 
Kinship × year F1,105 = 1.68  F1,98 = 0.13   χ 2 = 0.08   χ 2 = 0.83   χ 2 = 3.52  
1996      
1997      
Kinship × standardized laying date — F1,101 = 4.10 * (0.10 ± 0.05)   χ 2 = 0.91  χ2 = 34.53 *** (−2.45 ± 0.44)  χ2 = 19.45 *** (−1.59 ± 0.38)  
 Seasonal reproductive success 
Independent variables  Standardized laying date ( n = 111)   Clutch size a ( n = 111)   Hatching rate ( n = 111)   Nestling survival b ( n = 109)   Fledging success b ( n = 109)  
One-way analysis      
Kinship F1,109 = 1.57  F1,109 = 0.22  χ2 = 18.52 *** χ2 = 12.88 *** χ2 = 24.22 *** 
 (−4.40 ± 3.52) (−0.08 ± 0.16) (−9.91 ± 2.20) (−4.19 ± 1.15) (−5.16 ± 1.04) 
Multiway analysis      
Kinship F1,107 = 1.53  F1,101 = 0.66  χ2 = 16.90 ***  χ 2 = 1.41   χ 2 = 3.82  
 (−4.26 ± 3.44) (−0.13 ± 0.16) (−9.54 ± 2.23) (−5.97 ± 1.39) (−7.18 ± 1.25) 
Kinship × female age F1,103 = 0.01  F1,97 = 0.05   χ 2 = 0.05   χ 2 = 0.40   χ 2 = 0.58  
First year      
Adult      
Kinship × male age F1,104 = 0.06  F1,99 = 1.10   χ 2 = 0.49   χ 2 = 4.60 *  χ 2 = 4.97 * 
First year    (−5.60 ± 1.34) (−6.80 ± 1.20) 
Adult    (−0.35 ± 3.74) (−1.06 ± 3.52) 
Kinship × year F1,105 = 1.68  F1,98 = 0.13   χ 2 = 0.08   χ 2 = 0.83   χ 2 = 3.52  
1996      
1997      
Kinship × standardized laying date — F1,101 = 4.10 * (0.10 ± 0.05)   χ 2 = 0.91  χ2 = 34.53 *** (−2.45 ± 0.44)  χ2 = 19.45 *** (−1.59 ± 0.38)  

One-way analysis includes only kinship and multiway analysis includes kinship, environmental variables significant in Table 1 (not shown), plus main effects of standardized laying date, female age, male age and year (not shown) and their interactions with kinship. In parentheses are parameter estimates and their standard errors; parameter estimates for nonsignificant interactions are not shown. Terms that remained significant after sequential Bonferroni correction are in bold.

a log( x )-transformed.

b Nests where no young fledged were disregarded.

c log( x + 0.5)-transformed.

* p < 0.5;

** p <.01;

*** p <.001.

Table 2

, extended.

Seasonal reproductive success 
Mean fledgling condition ( n = 106)   Mean fledgling tarsus ( n = 106)   No. of fledglings b , c ( n = 109)   Recruitment rate of fledglings ( n = 109)   No. of recruits c ( n = 111)  
     
F1,104 = 0.37  F1,104 = 4.08 * F1,107 = 8.49 **  χ 2 = 0.04  F1,109 = 0.92  
(−0.42 ± 0.69) (−0.77 ± 0.38) (−0.87 ± 0.30) (−0.33 ± 1.61) (−0.62 ± 0.65) 
F1,97 = 0.10  F1,101 = 4.18 * F1,100 = 10.40 **  χ 2 = 0.03  F1,107 = 1.36  
(−0.22 ± 0.69) (−0.71 ± 0.35) (−0.84 ± 0.26) (−0.31 ± 1.66) (−0.73 ± 0.62) 
F1,95 = 3.27  F1,95 = 0.12  F1,99 = 1.19   χ 2 = 1.38  F1,103 = 0.77  
     
     
F1,92 = 0.70  F1,94 = 0.08  F1,95 = 0.36   χ 2 = 0.08  F1,102 = 0.23  
     
     
F1,93 = 0.87  F1,98 = 0.86  F1,97 = 1.22   χ 2 = 0.58  F1,106 = 3.14  
     
     
F1,91 = 0.20  F1,99 = 1.88  F1,100 = 7.73 **  χ 2 = 0.05  F1,101 = 0.16  
  (−0.24 ± 0.09)   
Seasonal reproductive success 
Mean fledgling condition ( n = 106)   Mean fledgling tarsus ( n = 106)   No. of fledglings b , c ( n = 109)   Recruitment rate of fledglings ( n = 109)   No. of recruits c ( n = 111)  
     
F1,104 = 0.37  F1,104 = 4.08 * F1,107 = 8.49 **  χ 2 = 0.04  F1,109 = 0.92  
(−0.42 ± 0.69) (−0.77 ± 0.38) (−0.87 ± 0.30) (−0.33 ± 1.61) (−0.62 ± 0.65) 
F1,97 = 0.10  F1,101 = 4.18 * F1,100 = 10.40 **  χ 2 = 0.03  F1,107 = 1.36  
(−0.22 ± 0.69) (−0.71 ± 0.35) (−0.84 ± 0.26) (−0.31 ± 1.66) (−0.73 ± 0.62) 
F1,95 = 3.27  F1,95 = 0.12  F1,99 = 1.19   χ 2 = 1.38  F1,103 = 0.77  
     
     
F1,92 = 0.70  F1,94 = 0.08  F1,95 = 0.36   χ 2 = 0.08  F1,102 = 0.23  
     
     
F1,93 = 0.87  F1,98 = 0.86  F1,97 = 1.22   χ 2 = 0.58  F1,106 = 3.14  
     
     
F1,91 = 0.20  F1,99 = 1.88  F1,100 = 7.73 **  χ 2 = 0.05  F1,101 = 0.16  
  (−0.24 ± 0.09)   
Table 3

Change in seasonal reproductive success from first recorded to subsequent breeding attempt in relation to remating status for individuals that experienced hatching failure in the first attempt.

Change in number of fledglings 
 

 

 
Change in number of recruits 
 

 

 
Sign Divorced Faithful Widowed Sign Divorced Faithful Widowed 
Females 
> 0 > 0 
≤ 0 ≤ 0 
Males 
> 0 > 0 
≤ 0 ≤ 0 
Change in number of fledglings 
 

 

 
Change in number of recruits 
 

 

 
Sign Divorced Faithful Widowed Sign Divorced Faithful Widowed 
Females 
> 0 > 0 
≤ 0 ≤ 0 
Males 
> 0 > 0 
≤ 0 ≤ 0 

Figures indicate number of individuals per category. sign: > 0: positive change; ≤ 0: no or negative change.

We thank Frans Fierens and Frank Adriaensen for help with fieldwork and database management, Harrie Bickle, Lorenz Heer, Ken Otter, Jonathan Griffith, and Terry Burke for sharing primers or unpublished sequences, and Stefan Van Dongen for discussion. Two anonymous referees provided valuable comments on an earlier version of the manuscript. This work was financed by projects G-0111–97 and G-0166–00 of the Fund for Scientific Research – Flanders (Belgium), and a specialization grant of the IWT – Flanders (Belgium) to T.V.C.

REFERENCES

Bensch S, Hasselquist D, von Schantz T,
1994
. Genetic similarity between parents predicts hatching failure: nonincestuous inbreeding in the great reed warbler?
Evolution
 
48
:
317
-326.
Bijlsma R, Bundgaard J, Boerema AC,
2000
. Does inbreeding affect the extinction risk of small populations?: predictions from Drosophila .
J Evol Biol
 
13
:
502
-514.
Bijlsma R, Bundgaard J, Van Putten WF,
1999
. Environmental dependence of inbreeding depression and purging in Drosophila melanogaster .
J Evol Biol
 
12
:
1125
-1137.
Blakey JK,
1994
. Genetic evidence for extra-pair fertilizations in a monogamous passerine, the great tit Parus major .
Ibis
 
136
:
457
-462.
Bollen KA,
1989
. Structural equations with latent variables, 1st ed. New York: Wiley.
Brooker MG, Rowley I, Adams M, Baverstock PR,
1990
. Promiscuity: an inbreeding avoidance mechanism in a socially monogamous species?
Behav Ecol Sociobiol
 
26
:
191
-199.
Brown JL, Brown ER,
1998
. Are inbred offspring less fit? Survival in an natural population of Mexican jays.
Behav Ecol
 
9
:
60
-63.
Charlesworth D, Charlesworth B,
1987
. Inbreeding depression and its evolutionary consequences.
Annu Rev Ecol Syst
 
18
:
237
-268.
Chen X,
1993
. Comparison of inbreeding and outbreeding in hermaphroditic Arianta arbustorum (L.) (land snail).
Heredity
 
71
:
456
-461.
Choudhury S,
1995
. Divorce in birds: a review of the hypotheses.
Anim Behav
 
50
:
413
-429.
Crnokrak P, Roff DA,
1999
. Inbreeding depression in the wild.
Heredity
 
83
:
260
-270.
Daan S, Tinbergen JM,
1997
. Adaptation of life histories. In: Behavioural ecology: an evolutionary approach, 4th ed (Krebs JR, Davies NB, eds). Oxford: Blackwell; 311–333.
Dahlgaard J, Loeschke V,
1997
. Effects of inbreeding in three life stages of Drosophila buzzatii after embryos were exposed to a high temperature stress.
Heredity
 
78
:
410
-416.
Daniels SJ, Walters JR,
1999
. Inbreeding depression and its effects on natal dispersal in wild birds.
Intl Ornithol Congr
 
22
:
2492
-2498.
Daniels SJ, Walters JR,
2000
. Inbreeding depression and its effect on natal dispersal in red-cockaded woodpeckers.
Condor
 
102
:
482
-491.
Dhondt AA,
1971
. Some factors influencing territory in the great tit, Parus major L.
Giervalk
 
61
:
125
-135.
Dhondt AA,
1989
. The effect of old age on the reproduction of great tits Parus major and blue tits Parus caeruleus .
Ibis
 
131
:
268
-280.
Dhondt AA, Adriaensen F, Plompen W,
1996
. Between- and within-population variation in mate fidelity in the great tit. In: Partnerships in birds: the study of monogamy (Black JM, ed). Oxford: Oxford University Press; 235–248.
Falconer DS, Mackay TFC,
1996
. Introduction to quantitative genetics, 4th ed. Essex: Longman.
Garnett MC,
1981
. Body size, its heritability and influence on juvenile survival among great tits, Parus major .
Ibis
 
123
:
31
-41.
Gebhardt-Henrich SG, van Noordwijk A,
1991
. Nestling growth in the great tit I. Heritability estimates under different environmental conditions.
J Evol Biol
 
3
:
341
-362.
Greenwood PJ, Harvey PH, Perrins CM,
1978
. Inbreeding and dispersal in the great tit.
Nature
 
271
:
52
-54.
Gullberg A, Tegelström H, Gelter HP,
1992
. DNA fingerprinting reveals multiple paternity in families of great and blue tits ( Parus major and P. caeruleus ).
Hereditas
 
117
:
103
-108.
Gustafsson L, Sutherland WJ,
1988
. The costs of reproduction in the collared flycatcher Ficedula albicollis .
Nature
 
335
:
813
-815.
Hartt L, Haefner JW,
1995
. Inbreeding depression effects on extinction time in a predator-prey system.
Evol Ecol
 
9
:
1
-9.
Heschel MS, Paige KN,
1995
. Inbreeding depression, environmental stress, and population size variation in scarlet gilia (Ipomopsis aggregata).
Conserv Biol
 
9
:
126
-133.
Horak P, Mänd R, Ots I,
1997
. Identifying targets of selection: a multivariate analysis of reproductive traits in the great tit.
Oikos
 
78
:
592
-600.
Jacquard A,
1974
. The genetic structure of populations, 2nd ed. Berlin: Springer-Verlag.
Jimenez JA, Hughes KA, Alaks G, Graham L, Lacy RC,
1994
. An experimental study of inbreeding depression in a natural habitat.
Science
 
266
:
271
-273.
Keller LF,
1998
. Inbreeding and its fitness effects in an insular population of song sparrows ( Melospiza melodia ).
Evolution
 
52
:
240
-250.
Keller LF, Arcese P,
1998
. No evidence for inbreeding avoidance in a natural population of song sparrows ( Melospiza melodia ).
Am Nat
 
152
:
380
-392.
Keller LF, Arcese P, Smith JNM, Hochachka WM, Stearns SC,
1994
. Selection against inbred song sparrows during a natural population bottleneck.
Nature
 
372
:
356
-357.
Kempenaers B, Adriaensen F, Dhondt AA,
1998
. Inbreeding and divorce in blue and great tits.
Anim Behav
 
56
:
737
-740.
Kempenaers B, Adriaensen F, Van Noordwijk AJ, Dhondt AA,
1996
. Genetic similarity, inbreeding and hatching failure in blue tits: are unhatched eggs infertile?
Proc R Soc Lond B
 
263
:
179
-185.
Kempenaers B, Pinxten R, Eens M,
1995
. Intraspecific brood parasitism in two tit Parusspecies: occurrence and responses to experimental parasitism.
J Avian Biol
 
26
:
114
-120.
Krebs JR,
1971
. Territory and breeding density in the great tit, Parus major .
L. Ecology
 
52
:
2
-22.
Krokene C, Rigstad K, Dale M, Lifjeld JT,
1998
. The function of extrapair paternity in blue tits and great tits: good genes or fertility insurance?
Behav Ecol
 
9
:
649
-656.
Lacy RC, Alaks G, Walsh A,
1996
. Hierarchical analysis of inbreeding depression in Peromyscus polionotus .
Evolution
 
50
:
2187
-2200.
Li CC, Weeks DE, Chakravarti A,
1993
. Similarity of DNA fingerprints due to chance and relatedness.
Hum Hered
 
43
:
45
-52.
Lindén M,
1991
. Divorce in great tits — chance or choice? An experimental approach.
Am Nat
 
138
:
1039
-1048.
Littell RC, Milliken GA, Stroup WW, Wolfinger RD,
1996
. SAS system for mixed models. Cary, North Carolina: SAS Institute.
Lubjuhn T, Strohbach S, Brün J, Gerken T, Epplen JT,
1999
. Extra-pair paternity in great tits ( Parus major ) — a long term study.
Behaviour
 
136
:
1157
-1172.
Lynch M,
1988
. Estimation of relatedness by DNA fingerprinting.
Mol Biol Evol
 
5
:
584
-599.
Lynch M,
1990
. The similarity index and DNA fingerprinting.
Mol Biol Evol
 
7
:
478
-484.
Lynch M, Ritland K,
1999
. Estimation of pairwise relatedness with molecular markers.
Genetics
 
152
:
1753
-1766.
Lynch M, Walsh B,
1998
. Genetics and analysis of quantitative traits, 1st ed. Sunderland, Massachusetts: Sinauer Associates.
Madsen T, Shine R, Loman J, Hakansson T,
1992
. Why do female adders copulate so frequently?
Nature
 
355
:
440
-441.
Matthysen E, Adriaensen F, Dhondt AA,
2001
. Local recruitment of great and blue tits ( Parus major , P. caeruleus ) in relation to study plot size and degree of isolation.
Ecography
 
24
:
33
-42.
Miller PS,
1994
. Is inbreeding more severe in a stressful environment?
Zoo Biol
 
13
:
195
-208.
Naef-Daenzer B, Keller LF,
1999
. The foraging performance of great and blue tits ( Parus major and P. caeruleus ) in relation to caterpillar development, and its consequences for nestling growth and fledging weight.
J Anim Ecol
 
68
:
708
-718.
Nieminen M, Singer MC, Fortelius W, Schops K, Hanski I,
2001
. Experimental confirmation that inbreeding depression increases extinction risk in butterfly populations.
Am Nat
 
157
:
237
-244.
Norberg RA,
1981
. Temporary weight decrease in breeding birds may result in more fledged young.
Am Nat
 
118
:
838
-850.
Nour N, Currie D, Matthysen E, Van Damme R, Dhondt AA,
1998
. Effects of habitat fragmentation on provisioning rates, diet and breeding success in two species of tit (great tit and blue tit).
Oecologia
 
114
:
522
-530.
Nur N,
1984
. The consequences of brood size for breeding blue tits II. Nestling weight, offspring survival and optimal brood size.
J Anim Ecol
 
53
:
497
-517.
Perrins CM, McCleery RH,
1985
. The effect of age and pair bond on the breeding success of great tits Parus major .
Ibis
 
127
:
306
-315.
Perrins CM, Moss D,
1974
. Survival of young great tits in relation to age of female parent.
Ibis
 
116
:
220
-224.
Pray LA, Schwartz JM, Goodnight CJ, Stevens L,
1994
. Environmental dependency of inbreeding depression: implications for conservation biology.
Conserv Biol
 
8
:
562
-568.
Pritchard JK,
2000
. Documentation for structure software. http://pritch.bsd.uchicago.edu/ .
Pritchard JK, Stephens M, Donnelly P,
2000
. Inference of population structure using multilocus genotype data.
Genetics
 
155
:
945
-959.
Queller DC, Goodnight KF,
1989
. Estimating relatedness using genetic markers.
Evolution
 
43
:
258
-275.
Queller DC, Strassman JE, Hughes CR,
1993
. Microsatellites and kinship.
Trends Ecol Evol
 
8
:
285
-288.
Ramsay SM, Otter KA, Mennill DJ, Ratcliffe LM, Boag PT,
2000
. Divorce and extrapair mating in female black-capped chickadees ( Parus atricapillus ): separate strategies with a common target.
Behav Ecol Sociobiol
 
49
:
18
-23.
Ritland K,
1996
. Estimators for pairwise relatedness and individual inbreeding coefficients.
Genet Res
 
67
:
175
-185.
Roughgarden J,
1996
. Theory of population genetics and evolutionary theory: an introduction, 1st ed. London: Prentice-Hall.
Rowley I, Russell E, Brooker M,
1986
. Inbreeding: benefits may outweigh costs.
Anim Behav
 
34
:
939
-941.
Saccheri I, Kuussaari M, Kankare M, Vikman P, Fortelius W, Hanski I,
1998
. Inbreeding and extinction in a butterfly metapopulation.
Nature
 
392
:
491
-494.
Sanz JJ, Moreno J,
1995
. Mass loss in brooding female pied flycatchers Ficedula hypoleuca : no evidence for reproductive stress.
J Avian Biol
 
26
:
313
-320.
Schultz ST, Willis JH,
1995
. Individual variation in inbreeding depression: the roles of inbreeding history and mutation. Genetics: 1209–1223.
Shields WM,
1982
. Philopatry, inbreeding, and the evolution of sex. Albany: State University of New York Press.
Smith HG, Kallander H, Nilsson JA,
1989
. The trade-off between offspring number and quality in the great tit Parus major .
J Anim Ecol
 
58
:
383
-401.
Sokal RR, Rohlf FJ,
1995
. Biometry, 3rd ed. New York: Freeman.
Stearns SC,
1992
. The evolution of life histories, 1st ed. Oxford: Oxford University Press.
Stockley P, Searle JB, MacDonald DW, Jones CS,
1993
. Female multiple mating behaviour in the common shrew as a strategy to reduce inbreeding.
Proc R Soc Lond B
 
254
:
173
-179.
Strohbach S, Curio E, Bathen A, Epplen JT, Lubjuhn T,
1998
. Extrapair paternity in the great tit ( Parus major ): a test of the good genes hypothesis.
Behav Ecol
 
9
:
388
-396.
Tautz D,
1989
. Hypervariability of simple sequences as a general source for polymorphic DNA markers.
Nucleic Acids Res
 
17
:
6463
-6471.
Thornhill NW,
1993
. The natural history of inbreeding and outbreeding: theoretical and empirical perspectives, 1st ed. Chicago: University of Chicago Press.
Van de Casteele T, Galbusera P, Matthysen E,
2001
. A comparison of microsatellite-based pairwise relatedness estimators.
Mol Ecol
 
10
:
1539
-1549.
van Noordwijk AJ, McCleery RH, Perrins CM,
1995
. Selection for the timing of great tit breeding in relation to caterpillar growth and temperature.
J Anim Ecol
 
64
:
451
-458.
van Noordwijk AJ, Scharloo W,
1981
. Inbreeding in an island population of the great tit.
Evolution
 
35
:
674
-688.
van Noordwijk AJ, Van Balen JH, Scharloo W,
1988
. Heritability of body size in a natural population of the great tit ( Parus major ) and its relation to age and environmental conditions during growth.
Genet Res
 
51
:
149
-162.
van Tienderen PH, van Noordwijk AJ,
1988
. Dispersal, kinship and inbreeding in an island population of the great tit.
J Evol Biol
 
1
:
117
-137.
Verboven N, Mateman AC,
1997
. Low frequency of extra-pair fertilizations in the great tit ( Parus major ) revealed by DNA fingerprinting.
J Avian Biol.
 
28
:
231
-239.
Verhulst S,
1995
. Reproductive decisions in great tits, an optimality approach (PhD thesis). Groningen, The Netherlands: Rijksuniversiteit Groningen.
Verhulst S, Tinbergen JM,
1991
. Experimental evidence for a causal relationship between timing and success of reproduction in the great tit Parus m. major .
J Anim Ecol
 
60
:
269
-282.
Verhulst S, Van Balen JH, Tinbergen JM,
1995
. Seasonal decline in reproductive success of the great tit: variation in time or quality?
Ecology
 
76
:
2392
-2403.
Wetton JH, Carter RE, Parkin DT, Walters D,
1987
. Demographic study of a wild house sparrow population by DNA fingerprinting.
Nature
 
327
:
147
-149.
Wolff JO,
1992
. Parents suppress reproduction and stimulate dispersal in opposite-sex juvenile white-footed mice.
Nature
 
359
:
409
-410.
Woodburn RJW, Perrins CM,
1997
. Weight change and te body reserves of female blue tits, Parus caeruleus , during the breeding season.
J Zool
 
243
:
789
-802.