Behavioral Ecology Advance Access originally published online on August 4, 2008
Behavioral Ecology 2008 19(6):1200-1207; doi:10.1093/beheco/arn086
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Correlates of the occurrence of inbreeding in a wild bird population
Department of Zoology, Edward Grey Institute, University of Oxford, OX1 3PS, UK
Address correspondence to M. Szulkin. E-mail: marta.szulkin{at}zoo.ox.ac.uk.
Received 12 November 2007; revised 16 June 2008; accepted 19 June 2008.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Inbreeding occurs when close relatives mate. In most populations, inbreeding is uncommon, with the result that we know little about the characteristics of those individuals that inbreed. However, such information may be invaluable for understanding the causes of inbreeding and inbreeding avoidance and the consequences of inbreeding. We investigated the characteristics of closely inbreeding individuals using 43 years of data from a wild great tit (Parus major) population in order to 1) characterize close inbreeding in terms of the family relationships involved; 2) compare inbreeding with noninbreeding individuals; and 3) determine whether close inbreeding was subsequently avoided by those that experience it. We found an uneven distribution of different types of close inbreeding, with sib mating most frequent and father–daughter matings least frequent. We did not find any evidence for clear predictors of inbreeding. By comparing the reproductive success of individuals that both inbred and outbred in their lifetimes, we demonstrate that inbreeding depression is solely attributable to the mixing of relatives’ genes and is thus independent of any tendency for low-quality parental genotypes or phenotypes to inbreed. Surprisingly, mating with close kin reduced the probability of divorce in the following reproductive season, while brood or fledging failure did not predict for remating. Overall, our results suggest that inbreeding is not a response to lowered prospects and that despite the substantial fitness cost of inbreeding, its avoidance in this population may be based on general mechanisms such as dispersal rather than specific kin avoidance.
Key words: divorce, great tit, inbreeding, inbreeding depression, parental quality, Parus major, pedigree, remating, Wytham.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Mating with close relatives leads to inbreeding in offspring, which increases genome-wide homozygosity and the expression of deleterious recessive alleles (Lynch and Walsh 1998
). The resulting decline in offspring fitness is called inbreeding depression. Inbreeding depression has been documented in farm and laboratory animals (Charlesworth D and Charlesworth B 1987; Dahlgaard and Loeschcke 1997; Smith et al. 1998), and also in wild animal populations (Keller and Waller 2002; Kruuk et al. 2002; Szulkin et al. 2007), where it is often found to be more severe than in captivity (Crnokrak and Roff 1999). Studying inbreeding in captive and/or laboratory conditions allows us to ask a particular kind of question as to the importance and consequence of inbreeding. For example, laboratory studies provide a great setting to investigate subtle mechanisms of allelic purging (Barrett and Charlesworth 1991), patterns of deleterious allele expression (McCune et al. 2004), and inbreeding by environment interactions (Bijlsma et al. 1999; Szulkin and Sheldon 2007). In contrast, studies of inbreeding in wild populations are particularly useful if we want to investigate the relevance of inbreeding in an ecological and evolutionary context (Endler 1986; Keller and Waller 2002). By focusing on the effect of inbreeding on lifetime fitness in a natural setting, we gain more insight into the extent to which inbreeding might act as a selective force for the evolution of mating systems, kin recognition, or dispersal, and by analyzing the types of mating events that involve inbreeding, we can generate hypotheses as to the mechanisms that lead to inbreeding and its avoidance.
However, estimating the extent of inbreeding depression in natural populations, composed of natural mating assemblages, is problematic: in many studies, no independent information is available about the fitness of the parents involved in inbreeding events. It has been suggested that those individuals that inbreed may be constrained to choose relatives as mates because their own fitness prospects are poor (Kruuk et al. 2002; Reid et al. 2006; Szulkin and Sheldon 2006). As a consequence, progeny with different levels of inbreeding may be the products of different groups of parents. Thus, if the individuals that inbreed are themselves low on the fitness scale, the magnitude of inbreeding depression in their offspring may be overestimated (Lynch and Walsh 1998). In addition, because one individual's close relative is another's nonrelative, systematic differences between in- and outbreeding individuals would suggest the operation of kin discrimination mechanisms, which might be adjusted with respect to individual state. Although there have been reports of the type of familial relations involved in close inbreeding (Bulmer 1973; Greenwood et al. 1978; van Noordwijk and Scharloo 1981; Wheelwright et al. 2006; Szulkin et al. 2007), to our knowledge only one study (Reid et al. 2008) has reported detailed analyses of the characteristics of inbreeding individuals relative to the outbreeding population. In an analysis of insular song sparrows Melospiza melodia, Reid et al. (2008) showed that closely inbreeding individuals (siring offspring with f 
0.125) were phentoypically distinct from outbreeding individuals for several phenotypic traits, such as tarsus length and hatch date, suggesting that closely inbreeding individuals may involve specific subsets of individuals within the population. Such nonindependence between phenotype and relatedness of breeding pairs may therefore interfere with the assumptions of randomness of inbreeding with regards to individual quality and consequently bias estimates of inbreeding loads (Reid et al. 2008).
Using a wild population of great tits (Parus major) as a model species, our study aims to explore the characteristics of individuals involved in mating with close kin, siring inbred offspring. To do so, we investigate a series of morphological and life-history traits that may indicate individual quality. We describe the type of close inbreeding observed in the population and ask whether individuals involved in closely inbred matings (f = 0.25: brother–sister, mother–son, father–daughter matings) differ phenotypically from the remaining outbreeding population, and whether the different types of close inbreeding vary in their reproductive success. Second, we ask whether close inbreeding is predictable on the basis of population structure. Further, we aim to tease apart the relative importance of parental quality and that of the expression of dominance effects, observed in offspring in the form of inbreeding depression. In the absence of experimentally controlled matings, a simple way to determine whether inbreeding depression is a property of individuals, or matings, is to compare the decline in reproductive success of individuals that both inbreed and outbreed over their lifetime. Finally, we ask whether individuals that did inbreed try to compensate for inbreeding depression experienced in one breeding season by "divorcing" and thus increasing the frequency of remating with unrelated kin. Our overall aim was thus to gain an understanding of the extent to which inbreeding with close relatives was predictable by characteristics of individuals or populations that might be associated with lowered fitness prospects.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
The present study is based on an analysis of mating and life-history data collected over 43 continuous years (1964–2006), from the population of great tits (Parus major) in Wytham, near Oxford (51°46'N1°20'W). Wytham Woods is a continuous semideciduous forest of approximately 388 hectares, divided into 9 sectors, where just more than 1000 nest-boxes suitable for great tits are distributed at variable densities. Each year all boxes were checked at weekly intervals and standard breeding parameters (egg-laying date, clutch size, hatch date, and brood size) measured. Parents were captured at the nest, their ring details recorded, whereas nestlings were ringed and measured while in the nest. Details of general field methods can be found in McCleery et al. (2004)
and Szulkin et al. (2007).
Pedigree building
Despite high rates of immigration (47% and 40% of females and males, respectively, breeding in any year within Wytham are born outside the wood [McCleery et al. 2004]), pooling information on great tit breeding events from the entire study period allows us to build a substantial pedigree of the population. For the purpose of pedigree building, we used information on all great tit breeding events occurring in Wytham and its vicinity between 1958 and 2006 where at least one parent could be linked to the previous breeding generation. In cases where one of the breeding parents was not identified (22% of males and 9% of females were unknown in the dataset of 9360 breeding events constituting the pedigree), the breeding event was included in the pedigree, but the unknown parent was assigned a unique identification number specific to that breeding event alone. Breeding attempts where predation events or experimental manipulations such as cross-fostering, brood reduction, or egg swaps had occurred (termed "interference" thereafter) were excluded from subsequent analysis of fitness parameters. Additionally, all breeding events for which cross-fostering occurred, but where the biological parents could not be identified due to historically missing records, were removed from the pedigree; their descendents, however, were included and assumed to have unknown and unrelated parents.
The coefficient of inbreeding f was estimated from pedigree data using Pedigree Viewer version 5.1a (available at: http://www-personal.une.edu.au/
bkinghor/pedigree.htm). Inbreeding coefficients are given relative to the founder population: it is assumed that the founding population and all immigrant birds are unrelated (Keller and Waller 2002). In addition to males and females of unknown identity linked to particular breeding events (see above), 76 391 unique individuals were included in the pedigree, which extended for up to 35 generations, and had a median pedigree depth of 4 generations.
The pedigree inferred in this study relies on the assumption that the social father is the genetic father of all offspring. However, rates of extrapair paternity in this population range from 16–19% (Blakey 1994, Patrick S, unpublished data) and may thus introduce some error in some of the pedigree links, most likely decreasing estimates of inbreeding depression in broods qualified as inbred. One aim of this study was thus to verify the extent to which inbreeding depression indeed varies among different types of close inbreeding (i.e., brother–sister, mother–son, father–daughter matings).
Dataset restrictiveness
As emphasized by Marshall et al. (2002), when studying inbreeding in wild populations it is necessary to identify not only inbreeding events themselves but also those events for which particular types of inbreeding can be detected. Consequently, the lower the inbreeding level one is interested in, the greater the number of ancestors required to enable the detection of any potential inbreeding. Because our aim here was to compare the fitness of highly inbred offspring relative to noninbred birds, we used a minimal degree of dataset restrictiveness: for a breeding event to be included in the dataset, both parents needed to be known based on their ring number; grandparental identity was not taken into account. We thus focused on 2 categories of inbreeding: "inbreeding" (close inbreeding) refers to a mating between 2 individuals that would produce offspring with an inbreeding coefficient of f = 0.25 and is equivalent to a brother–sister, mother–son, or father–daughter mating. "Outbreeding" is here defined as a mating between 2 individuals that would yield an offspring inbreeding coefficient f < 0.25. Because we were only interested in the contrast between outbred and closely inbred mating events, mating events corresponding to those between double first cousins (f = 0.125) or any other breeding events with a non-null inbreeding coefficient (as calculated by Pedigree Viewer) were still considered as outbred. The distinction between those matings involving first-order relatives and all other matings is biologically justified, particularly when the question of inbreeding avoidance and any putative response to inbreeding is concerned, because all first-order relatives have the opportunity to learn each others’ characteristics through association at the nest, with the exception of full-sibs born in different years; however, there were no cases of inbreeding involving such birds in this study. For any less related individuals, any kin discrimination would need to involve alternative mechanisms of discrimination.
Such low dataset restrictiveness enables us to maximize the dataset of individuals that both inbreed (f = 0.25) and outbreed (f < 0.25) in their lifetime. It is theoretically possible that 2 closely related (r = 0.5) immigrant birds may breed together, yielding offspring with f = 0.25; such breeding events would be classified by us as outbred given that no information is available about their ancestry. However, if such cases do occur occasionally, they would be classified as outbred and would have very little influence in a sample size of 6737 outbred (f < 0.25) matings.
Individual characteristics
We used models with normal, Poisson, and binomial error distribution with identity, logarithm, and logit links, respectively. Because data fitted to models with few parameters, such as the Poisson and binomial distributions, often display greater or lesser variability than predicted by the statistical model, the dispersion parameter 
is estimated to control for over- or underdispersion of the data. In the case of Poisson and binomial distributions, we used an estimated value of the dispersion parameter whenever > 1 to adjust the residual degrees of freedom (df) to the value of the residual variance (Genstat, VSN International, Hertfordshire, UK). Whenever underdispersion occurred, the dispersion parameter was set at 1.
For all analyses where fitness estimates were not involved, we used a dataset where both parents were known for each breeding event, independently of whether the brood was involved in any experimental manipulation or experienced predation. Thus, for characterization of the types of inbreeding event (brother–sister, father–daughter, mother–son) and their spatial correlates, we used a dataset of 6791 breeding events, of which 54 (0.8%) were cases resulting in producing offspring with an inbreeding coefficient of f = 0.25. This translated in 0–4 cases of close inbreeding yearly and represented 0–3.7% of all breeding events where both parents were known in any given year. To compare parental characteristics, and reproductive success at the fledging and recruitment stage for broods sired by related and unrelated mates, we used a dataset where any interference to the brood (such as cross-fostering or other experimentation) was excluded and consisted of 6226 breeding events, of which 48 matings were qualified as close inbreeding (f = 0.25).
To investigate the effect of individual birth characteristics on the likelihood of ever inbreeding (at f = 0.25), we ran a generalized linear model where inbreeding in its lifetime was fitted as a binomial response variable (0/1); hence, each individual was only represented once in the dataset. Fitted explanatory variables were birth year, hatch date, and mass at fledging, as well as interactions between sex and mass at fledging and hatch date and mass and fledging. To test for differences in terms of adult morphometrics between outbreeding versus inbreeding individuals, we used mixed models with normal errors, where the type of mating (with close kin or unrelated partner), sex, wood sector, and year of the breeding event were fitted as fixed effects. Parental (individual) identity was fitted as a random effect to control for multiple breeding.
Whether there is a difference with respect to age while in- or outbreeding was tested using different datasets depending on the type of inbreeding event. Because all parents from parent–offspring matings by definition mated at least twice in their lifetime, their log-transformed age at inbreeding was compared with adult individuals (e.g., at least 2 years old) who mated at least twice in their lifetime and whose exact age was known. Offspring from parent–offspring matings and siblings from brother–offspring matings were tested against the entire dataset of outbreeding events. In all cases, we modeled effects on age using a general linear model (GLM) with normal errors, where year of the breeding event and woodland sector were fitted as additional explanatory variables.
Fitness differences between different types of close inbreeding
Differences in reproductive success in terms of number of fledged and recruited individuals between inbreeding and outbreeding individuals were tested using mixed models with normal and Poisson errors, respectively. Breeding events involved in manipulations and experiments were excluded. Pair identity was fitted as random effect to control for multiple breeding within the pair, and woodland sector, year of the breeding event, and the type of mating ("1"—inbred and "0"—outbred) were fitted as explanatory variables. To test for the effect of different types of close inbreeding on reproductive success, a much smaller dataset was involved and thus did not allow the use of mixed models. Instead, reproductive success for pairs that bred more than once together was averaged per pair; we used a linear regression with either normal or Poisson errors (for number of fledged and recruited individuals, respectively). Breeding events involved in manipulations and experiments were excluded. The type of relatedness (father–daughter, brother–sister, mother–son) was fitted as a categorical variable.
Inbreeding and population characteristics
To test whether inbreeding is related to the density of breeding pairs in the population, we initially carried out our analyses at a sector level: breeding density (per hectare) in each sector and for each year was fitted as explanatory variable in a GLM with normal errors, and the density of inbreeding events (per sector per year) acted as response variable. However, breeding densities in Wytham are strongly influenced by the varying nest-box density across the woodland: nest-box density explains 66% of the variance in the breeding density at a sector-year level, when corrected for year (adjusted R2, P < 0.001, n = 374). These differences in breeding densities between woodland sectors allowed us to additionally test whether breeding density is a predictor of the likelihood of inbreeding. We therefore further used a GLM with binomial proportions where the number of breeding pairs (at a sector per year level) was fitted as the denominator and the number of inbreeding pairs as the numerator of the response variable. Explanatory variables used were woodland sector, breeding pair density, and their interaction; their effect on the likelihood of inbreeding was also tested independently from each other. Because overall population density may affect the likelihood of inbreeding as a greater number of relatives are potentially alive, we further tested relationships between the number of inbreeding pairs and the overall population size in year n, as well as in relation to the population size in year n – 1 using a GLM with normal errors.
Remating
Sequentially in- and outbreeding individuals
In the data available for investigation (1964–2006), we identified 25 individuals that both in- and outbred in their lifetime. Their reproductive success was calculated as the number of recruited offspring per breeding attempt. If an individual bred more than once in the same category, its reproductive success was calculated as the average number of offspring recruited from each breeding attempt within each breeding category. None of the breeding events in this analysis were subject to any interference. Within-individual differences in reproductive success were tested using a Wilcoxon matched-pairs test; those results were further compared with a cross-sectional dataset where all individuals in the population who bred at least twice in their lifetime were included. The effect of inbreeding in the cross-sectional dataset was analyzed with a minimal set of variables so as not to limit the dataset; we used a generalized linear mixed model (GLMM) with Poisson errors, where inbreeding (0/1), year of the breeding event, and sector were fitted as fixed effects and parental identity was fitted as a random effect.
Divorce and inbreeding
We restricted our analyses to individuals breeding at least twice and whose breeding partner from the previous year was known to have survived to breed the following year n + 1 (N = 1018 pairs). Broods subjected in year n to experimental interference (such as cross-fostering, brood size manipulation, etc.) were excluded from the dataset.
If a pair bred with each other in year n and n + 1, it was classified as faithful; pairs that bred with each other in year n and then changed partners in year n + 1 were defined as divorcing. Analyses were carried out at the breeding event level (e.g., whether the pair divorced in the subsequent year or not). We fitted divorce (0/1) as a binomial response variable and asked whether year, sector, and different reproductive characteristics such as clutch size, brood size, number of fledged individuals, clutch failure, brood failure, or inbreeding (0/1) influenced parental decision on divorcing. Because inbreeding causes reduced hatching and fledging success (Szulkin et al. 2007), we asked whether clutch and brood failure can be used as a general cue for genetic similarity and/or inbreeding (the difference between the 2 would be that in the case of genetic similarity, the offspring would inherit the same allele(s) at both loci, independently of whether the parents were in fact related or not), as described by Kempenaers et al. in 1996; clutch failure was defined as [1 – (brood size/clutch size)] x 100; we further extended the notion of clutch failure to brood failure, calculated as the percentage decrease in reproductive success from clutch stage till fledging [1 – (number of fledged/clutch size)] x 100. We expect that clutch/brood failure would be strong indicators for genetic incompatibility, cuing birds to find a better mate in the following breeding season by divorcing its current mate.
All statistical analyses were carried out using GenStat Version 10.2 supplied by VSN International.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Characteristics of inbreeding individuals
Out of 6791 breeding events, made up by 6068 unique pairs of birds where both parents were known and thus close inbreeding (f = 0.25) could be identified, there were 42 pairs that mated incestuously, 9 of them in repeated years, and in one case for up to 4 consecutive years. Of the close inbreeding pairs found, 27 (64%) involved brother–sister, 4 (10%) father–daughter and 11 (26%) mother–son matings. In cases where offspring fitness was investigated, the dataset was restricted to breeding events where no manipulations or experiments were carried out on the brood and consisted of 6226 breeding events, out of which 48 were qualified as close inbreeding and were the result of 37 different inbreeding pairs.
Inbreeding and outbreeding individuals did not differ in terms of their original hatch date (deviance ratio = 0.18, P = 0.669, n = 4995) or mass at fledging (deviance ratio = 1.40, P = 0.236, n = 4995), nor were they influenced by an interaction between sex and mass at fledging (deviance ratio = 0.15, P = 0.698, n = 4995) or hatch date and mass at fledging (deviance ratio = 1.11, P = 0.291, n = 4995). Similarly, we did not find any differences in adult morphometrics, as inbreeding individuals did not differ from outbreeding individuals in mass (Wald = 0.86, P = 0.355, n = 9845), nor wing length (Wald = 0.01, P = 0.910, n = 10649) when caught while feeding young during the breeding season. A summary of parental morphometrics is presented in Table 1.
|
Age patterns among inbreeding and outbreeding individuals varied: on average, siblings from brother–sister matings were 1.56 years old (0.1 standard error [SE], n = 68) and were younger than the rest of the outbreeding population, whose average age was 1.74 years old (0.01 SE, n = 11 281; dropping terms from model: siblings vs outbreeding population: F1,11 259 = 4.18, P = 0.041, regression coefficient: –0.13 [0.06 SE]; effect of wood sector on parental age: F8,11 266 = 5.89, P < 0.001; effect of breeding year on parental age: F41,11 299 = 12.24, P < 0.001). However, this result may be expected from the fact that an individual is more likely to have siblings alive early in life. Offspring from parent–offspring matings did not differ in terms of age when inbreeding relative to the rest of the outbreeding population (mean age when inbreeding 1.85 [0.2 SE, n = 20]; offspring vs outbreeding population: F1,11 211 = 0.12, P = 0.731). Finally, when inbreeding parents from parent–offspring matings were compared with the outbreeding population where only birds older than 1 year of age and who bred at least twice in their lifetime were taken into account, we found that the 2 groups did not differ significantly from each other in terms of mean age (F1,3996 = 2.30, P = 0.13, regression coefficient: 0.12 [0.08 SE]; average age of inbreeding parents: 3.31 [0.37 SE, n = 16], average age of the restricted outbreeding population: 2.80 [0.016 SE, n = 4033]). We did not find any particular pattern in the sequence of in- and outbreeding throughout an individual's life, which furthermore was limited by small sample size: overall, 16 out of 25 individuals who both in- and outbred in their lifetime were found to outbreed first, and 1 out of 10 siblings and 3 out of 4 offspring involved in parent–offspring matings were found to outbreed first.
As expected and shown elsewhere with a smaller dataset (Szulkin et al. 2007), individuals that mated with unrelated partners had a higher reproductive success relative to those that mated incestuously, both at the fledging as at the recruitment stage (Table 2). In contrast, there were no differences in fledging or recruitment success between the different types of close inbreeding (brother–sister, mother–son, father–daughter; Table 2).
|
Inbreeding and population characteristics
The number of inbreeding events (per hectare) increased in areas with high pair density (F1,362 = 28.64, P < 0.001), which is likely simply to be a function of the larger sample of pairs. However, the relationship between breeding density and the likelihood of inbreeding between sectors and across years is not as clear-cut: a GLM with binomial proportions suggests that woodland sector by itself does not influence the likelihood of inbreeding (deviance ratio = 1.38, P = 0.199, n = 374), nor does breeding pair density (deviance ratio = 2.32, P = 0.128, regression coefficient: 0.323 (0.209 SE), n = 374). The interaction between the 2 variables was significant however (deviance ratio = 2.04, df = 8, P = 0.038, n = 374) and suggests that the effect of density on the likelihood of inbreeding may vary spatially. When the entire forest was considered as a unit, we found no relationship between the number of inbreeding events and the overall breeding population size when measured across all years of the dataset (F1,41 = 2.55, P = 0.118), nor when the number of inbreeding events in year n is compared with the number of breeding events—and thus population size—in year n – 1 (F1,41 = 0.80, P = 0.375).
Reproductive success of individuals that in- and outbred in their lifetime
If individuals that inbreed are themselves low on the fitness scale, the level of inbreeding depression in their offspring may be overestimated (Lynch and Walsh 1998), and we would predict that the reproductive success of such individuals would be low, both when inbreeding and when outbreeding. To test for such possibility, we compared the reproductive success of individuals that both in- and outbred in their lifetimes and compared it with a cross-sectional analysis of the entire dataset of males and females who bred at lest twice in their lifetime.
We identified 25 individuals that mated at least once in their lifetime with both an unrelated partner and a close relative (12, 4, and 9 individuals involved in mother–son, father–daughter, and brother–sister matings, respectively). Parental reproductive success was measured by the number of offspring that recruited to the breeding population in subsequent years. Individuals that in- and outbred during their lifetime suffered a substantial reduction, by 49%, in recruitment success while inbreeding (mean 0.45 ± 0.11 SE, n = 25 while inbreeding and mean 0.88 ± 0.15 SE, n = 25 when outbreeding; Wilcoxon matched-pairs test = 20, P = 0.016; Figure 1a).
|
Of the 25 individuals that both inbred and outbred, there were 10 cases (5 pairs) where both members of the pair in- and outbred (e.g., a brother and a sister that bred together, subsequently "divorced" and then repaired). To test whether such repeated measures bias our estimates of reproductive success in the 2 breeding categories, we restricted our dataset using 3 approaches: 1) by excluding all males for which their female breeding partner was also present in the sample (thus reducing the sample size to n = 20), 2) by excluding all females whose male breeding partner was present in the dataset (reducing the sample size to n = 20), and 3) by excluding at random either the male or female from all breeding pairs where both members both in- and outbred (5 pairs recorded, reducing the sample size to n = 17). In all 3 cases, reducing the dataset resulted in very similar estimates of the fitness decline caused by inbreeding—the reduction in fitness was found to be 46%, 52%, and 49% in approaches 1), 2) and 3), respectively.
Values of reproductive success of birds that both in- and outbreed in their lifetime overlap entirely with estimates of reproductive success for males and females who bred at least twice in their lifetime in a cross-sectional dataset, where close inbreeding reduced recruitment success by 49% for males and 54% for females (GLMM with Poisson errors, males: Wald1 = 5.70, P = 0.017, n = 3226; females: Wald1 = 6.98, P = 0.008, n = 3356; Figure 1b,c). Our results thus suggest that the reduced fitness of inbreeding birds is true inbreeding depression (i.e., resulting from the expression of deleterious recessive alleles) and is not confounded by individual differences in fitness.
Divorce as a response to inbreeding
We identified 1018 pairs for which both pair members survived at least a year past their first breeding. In 32% of cases, the birds bred with a different partner while their previous partner was still alive. We found divorce to be weakly, but significantly, influenced by clutch size, with lower clutch size increasing the likelihood of divorce. However, neither brood size, the number of fledged individuals, or clutch or brood failure, measuring the pair's reproductive success relative to their original clutch size, influenced the likelihood of divorcing (Table 2); hence, there was little evidence that potential cues of inbreeding led to elevated rates of divorce. Interestingly, inbreeding did impact on the pair's decision to divorce, but in the opposite direction to that which might be expected: of 11 inbred pairs, only 1 divorced, and inbreeding was a significant predictor of divorce likelihood in the full model (Table 3).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
In the analysis of a long-term dataset of mating events in wild great tits, we show that individuals involved in inbreeding differed very little from outbreeding birds—we did not find support for any difference in terms of birth characteristics or morphology at breeding nor were there any clear-cut differences in terms of age. There was also no clear trend for the likelihood of inbreeding to be related to population density. However, we show elsewhere that a strong predictor of close inbreeding, for both parent–offspring and sib–sib matings, is limited natal dispersal (Szulkin and Sheldon 2008
). Out of the types of close inbreeding inferred (brother–sister, mother–son, father–daughter), father–daughter matings are those that are most vulnerable to the effects of extra pair paternity, as the social male may be entirely unrelated to its daughter offspring. However, the different types of close inbreeding did not differ in terms of reproductive success: all were detrimentally affected by inbreeding depression when compared to outbreeding individuals, suggesting that rates of pedigree error may be equivalent between the different types of close inbreeding. While investigating remating events, we found that clutch size predicts the likelihood of divorce (negative trend), but not clutch failure or brood failure, which is usually associated with inbreeding depression. Surprisingly, close inbreeding was found to influence divorce rates in the opposite direction to what was intuitively expected, as inbreeding pairs were less likely to divorce.
Our results complement previously found patterns showing that individuals mating with close kin do not significantly differ in their timing of breeding, clutch size, or mean egg weight (Szulkin et al. 2007). By comparing reproductive success of individuals that both inbred and outbred in their lifetimes, we demonstrate that estimates of inbreeding depression are not confounded by the fact that individuals that inbreed have low fitness in all mating events. The similarity between the within-individual decrease in reproductive success caused by inbreeding, and that across individuals at the population level, allows us to reject the possibility that estimates of inbreeding depression are confounded by low parental quality. Our findings clearly support the interaction between the genomes of relatives as the cause of inbreeding depression in wild populations. Hence, differences between inbred and outbred matings can only be seen at the postzygotic stage, where deleterious alleles are more likely to be expressed in homozygous individuals, and result in reduced rates of fledged and recruited individuals when those were sired by 2 closely related individuals.
Van Noordwijk and Scharloo (1980) suggested that some individuals may be more likely to be involved in inbreeding than others. For example, if a large number of siblings from a successful brood survive, an individual's likelihood of mating with kin, assuming random mating, may increase substantially. There is no support from our analyses for the idea that inbreeding individuals may be of better quality than those outbreeding. In our study system, the probability that 2 individuals or more of opposite sex recruited from one brood was found to be approximately 10%. The likelihood of having siblings from different broods alive does not greatly inflate this value, given that the probability for a parent to mate twice with the same partner is 30%. Thus, it is theoretically possible that fitter individuals have a greater likelihood of encountering kin and breed together. However, based on the overlapping estimates of reproductive success for individuals that both in- and outbreed relative to a cross-sectional analysis of the population, and given that no difference in fledging mass nor hatch date between in- and outbreeding individuals was observed, our results do not suggest that inbreeding individuals may in fact be of better quality, or worse, than those outbreeding.
Our results are in contrast with a recent study published by Reid et al. (2008) showing that closely inbreeding insular song sparrows often differ from outbreeding individuals in terms of phenotypic traits such as hatch date, tarsus length, or age. Importantly, the song sparrow population also exhibits low levels of immigration and an inherent correlation whereby more inbred individuals are more likely to sire inbred offspring by mating with more closely related mates than expected at random (Reid et al. 2006). Differences in the observed effects, as reported by our study and those of Reid et al. (2006, 2008), show that while inbreeding individuals may indeed represent a specific subset of the entire breeding population, such structuring does not have to be widespread across populations; our study provides no support for different phenotypic quality of inbreeding and outbreeding great tits, as validated by a phenotypic analysis of breeding birds, but also through a longitudinal and cross-sectional comparison of reproductive success while in- and outbreeding.
We did not find any record of mating between full-sibs born in different years. We compared this value with null distributions of all possible brother–sister matings, where each bird was mated with a bird of opposite sex in a given year, based on different scenarios of mate availability. Depending on whether each bird is mated to all members of opposite sex alive in a given year in the population, or when only birds breeding for the first time, those who divorced or were widowed were available as mates in a given year, the generated frequencies of full-sib matings born in different years varied from 2.3 to 1.8 pairs for the entire duration of the study, respectively. Given that the observed number of full-sib mating born in different years were indeed lower (0 out of 27) than values generated by 2 models of random mating, the possibility of close inbreeding based on some form of experience-based attraction cannot be discarded; at the same time, however, such small expected values do not offer enough power to provide a definite answer.
Individuals mating with close kin were unequally represented by parent–offspring and sibling matings. As in Wheelwright et al. (2006), investigating inbreeding in Savannah Sparrows, Passerculus sandwichensis, and Bulmer (1973) and Greenwood et al. (1978), looking at much more restricted datasets of the Wytham great tit (up to 1975), we found more than twice as many cases of brother–sister matings compared with parent–offspring cases of inbreeding. Wheelwright et al. (2006) suggested that such unequal distribution of the different types of close inbreeding may be due to sex- and age-specific tendencies for active kin recognition and inbreeding avoidance. Although we would not discard this hypothesis, passive mechanisms of inbreeding avoidance, which do not require active kin recognition, would probably generate similar patterns of close inbreeding, at least in passerine birds. First, given the female-biased natal dispersal observed in many passerine bird species (Bulmer 1973; Pusey 1987), and in this population in particular (Greenwood et al. 1978; Szulkin and Sheldon 2008), we expect females to disperse further than males. Given limited breeding dispersal (i.e., dispersal between breeding events) in the great tit (Greenwood and Harvey 1982), and the larger natal dispersal radius in females relative to males, the likelihood of a mother–son encounter should be greater than that of a father–daughter mating. The high numbers of brother–sister matings found, relative to parent–offspring close inbreeding, may be influenced by the fact that great tit siblings disperse in similar directions, which results in siblings breeding at closer distances than randomly chosen unrelated birds (Matthysen et al. 2005); this may therefore increase the likelihood of breeding with each other, relative to parents, although the relative frequencies of these events will be influenced by many other processes. Further work on patterns of mating relative to individual relatedness is needed in order to definitively support, or reject, the possibility of active inbreeding avoidance in great tits (Szulkin 2007).
We observed a clear pattern for more inbreeding occurring in high-density areas within our study area. But there was no clear-cut effect of breeding density on the likelihood of inbreeding: instead, an interaction between breeding pair density and woodland sector was found to influence the likelihood of inbreeding. Given that individuals ending up inbreeding often do not disperse far (Greenwood et al. 1978; Daniels and Walters 2000; Szulkin and Sheldon 2008), and juvenile tits often disperse in a similar direction (Matthysen et al. 2005), a higher nest-box density may favor the pairing of closer relatives. Clearly however, breeding density alone cannot be responsible—instead, other features of the environments in different woodland sectors may influence how breeding density affects the likelihood of inbreeding. It is worth noting that a tendency for a positive effect of density on dispersal rates is often found among bird and mammal species (Matthysen 2005), in consequence of which individuals born in high-density areas would be expected to disperse further, instead of breeding in the same sectors. Given that dispersal, as a complex trait, is also likely to influence the occurrence of inbreeding, more work is needed to understand the relationship between inbreeding and breeding density.
The divorce rate found in our study with 42 years of data of great tit breeding events (32%) is very similar to that found by Perrins and McCleery (1985), who found a 33% divorce rate from an analysis of 23 years of data from the same population. The fact that clutch size influences the likelihood of divorce, as found in our study, was also reported by Dhondt and Adriaensen (1994) and Linden (1991). Similarly as in the study of Kempenaers et al. (1998), we found that brood failure does not influence divorce rates. Interestingly, inbreeding individuals were less likely to divorce their partner than outbreeding birds. Given that inbreeding individuals disperse less than the population average (Szulkin and Sheldon 2008), we suggest that in some nests, "abnormal" imprinting may occur, leading some offspring to disperse little and perhaps form particularly stable pair bonds relative to other outbreeding pairs. In the light of available evidence showing that clutch size (a parentally determined trait immune to inbreeding depression) influences divorce, but that the latter does not change with clutch or brood failure, or brood size, or the number of fledged individuals, we conclude that there is little evidence that divorce is used as a mean of inbreeding avoidance in a context of lifelong reproductive efforts. Although great tits could compensate for the fact of staying with a close relative by seeking extrapair fertilizations, preliminary data on paternity levels among inbreeding pairs suggest reduced levels of extrapair paternity among inbreeding pairs (Szulkin 2007). Thus, at the moment we do not have evidence suggesting that extrapair fertilizations may be used to compensate for reduced rates of divorce.
By exploiting a series of comparisons at both cross-sectional and longitudinal scales, we found no indication that relatives mate because of particular phenotypic constraints, nor evidence that birds try to compensate for mating with a relative by divorcing. We show that inbreeding depression can most parsimoniously be attributed entirely to the effects of related individuals’ genes being combined, resulting in the expression of deleterious recessive alleles and inbreeding depression in offspring. Although the question of whether great tits in our population actually avoid mating with kin at all still needs to be further investigated, we suggest that rather than actively avoiding with kin, passive inbreeding avoidance through dispersal is the most efficient behavior for avoiding mating with a relative (Szulkin and Sheldon 2008).
|
FUNDING |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
PhD funding from the Christopher Wlech Trust and the Queen's College, Oxford.
|
ACKNOWLEDGEMENTS |
|---|
We sincerely thank Robin McCleery for help with divorce data analysis and Teddy Wilkin for providing data on Wytham Woods breeding densities. We thank Jane Reid as well as 2 anonymous referees for excellent comments. We also sincerely thank the many generations of fieldworkers collecting breeding data who made this study possible.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Barrett SCH, Charlesworth D. Effects of a change in the level of inbreeding on the genetic load. Nature (1991) 352:522–524.[CrossRef][Web of Science][Medline]
Bijlsma R, Bundgaard J, Van Putten WF. Environmental dependence of inbreeding depression and purging in Drosophila melanogaster. J Evol Biol (1999) 12:1125–1137.[CrossRef][Web of Science]
Blakey JK. Genetic evidence for extra pair fertilizations in a monogamous passerine, the great tit Parus major. Ibis (1994) 136:457–462.[CrossRef]
Bulmer MG. Inbreeding in the great tit. Heredity (1973) 30:313–925.[CrossRef][Web of Science][Medline]
Charlesworth B, Charlesworth D. The genetic basis of inbreeding depression. Genet Res (1999) 74:329–340.[CrossRef][Web of Science][Medline]
Charlesworth D, Charlesworth B. Inbreeding depression and its evolutionary consequences. Ann Rev Ecol Syst (1987) 18:237–268.[CrossRef][Web of Science]
Crnokrak P, Roff DA. Inbreeding depression in the wild. Heredity (1999) 83:260–270.[CrossRef][Web of Science][Medline]
Dahlgaard J, Loeschcke V. Effects of inbreeding in three life stages of Drosophila buzzatii after embryos were exposed to a high temperature stress. Heredity (1997) 78:410–416.[CrossRef][Web of Science][Medline]
Daniels SJ, Walters JR. Inbreeding depression and its effects on natal dispersal in red-cockaded woodpeckers. Condor (2000) 102:482–491.
Dhondt AA, Adriaensen F. Causes and effects of divorce in the blue tit Parus caeruleus. J Anim Ecol (1994) 63:979–987.[CrossRef]
Endler JA. Natural Selection in the Wild (1986) Princeton: Princeton University Press.
Greenwood PJ, Harvey PH. The natal and breeding dispersal of birds. Annu Rev Ecol Syst (1982) 13:1–21.[CrossRef][Web of Science]
Greenwood PJ, Harvey PH, Perrins CM. Inbreeding and dispersal in the great tit. Nature (1978) 271:52–54.[CrossRef][Web of Science]
Keller LF, Waller DM. Inbreeding effects in wild populations. Trends Ecol Evol (2002) 17:230–241.[CrossRef]
Kempenaers B, Adriaensen F, Dhondt AA. Inbreeding and divorce in blue and great tits. Anim Behav (1998) 56:737–740.[CrossRef][Web of Science][Medline]
Kempenaers B, Adriaensen F, van Noordwijk AJ, Dhondt AA. Genetic similarity, inbreeding and hatching failure in blue tits: are unhatched eggs infertile? Proc R Soc Lond B Biol Sci (1996) 263:179–185.[CrossRef]
Kruuk LEB, Sheldon BC, Merila J. Severe inbreeding depression in collared flycatchers (Ficedula albicollis). Proc R Soc Lond B Biol Sci (2002) 269:1581–1589.[CrossRef][Medline]
Linden M. Divorce in great tits—chance or choice—an experimental approach. Am Nat (1991) 138:1039–1048.[CrossRef][Web of Science]
Lynch M, Walsh B. Genetics and Analysis of Quantitative Traits (1998) Sunderland, MA: Sinauer Associates.
Marshall TC, Coltman DW, Pemberton J, Slate J, Spalton JA, Guinness FE, Smith JA, Pilkington JG, CluttonBrock TH. Estimating the prevalence of inbreeding from incomplete pedigrees. Proc R Soc Lond B Biol Sci (2002) 269:1533–1539.[CrossRef][Medline]
Matthysen E. Density-dependent dispersal in birds and mammals. Ecography (2005) 28:403–416.[Medline]
Matthysen E, Van de Casteele T, Adriaensen F. Do sibling tits (Parus major, P. Caeruleus) disperse over similar distances and in similar directions? Oecologia (2005) 143:301–307.[CrossRef][Web of Science][Medline]
McCleery RH, Pettifor RA, Armbruster P, Meyer K, Sheldon BC, Perrins CM. Components of variance underlying fitness in a natural population of the great tit Parus major. Am Nat (2004) 164:E62–E72.[CrossRef][Web of Science][Medline]
McCune AR, Houle D, McMillan K, Annable R, Kondrashov AS. Two classes of deleterious recessive alleles in a natural population of zebrafish, Danio rerio. Proc R Soc Lond B Biol Sci (2004) 271:2025–2033.[CrossRef][Medline]
Perrins CM, McCleery RH. The effect of age and pair bond on the breeding success of great tits Parus major. Ibis (1985) 127:306–315.[CrossRef]
Pusey AE. Sex-biased dispersal and inbreeding avoidance in birds and mammals. Trends Ecol Evol (1987) 2:295–299.[CrossRef]
Reid JM, Arcese P, Keller LF. Intrinsic parent-offspring correlation in inbreeding level in a song sparrow (Melospiza melodia) population open to immigration. Am Nat (2006) 168:1–13.[CrossRef][Web of Science]
Reid JM, Arcese P, Keller LF. Individual phenotype, kinship, and the occurrence of inbreeding in song sparrows. Evolution (2008) 62:887–899.[CrossRef][Web of Science][Medline]
Smith LA, Cassell BG, Pearson RE. The effects of inbreeding on the lifetime performance of dairy cattle. J Dairy Sci (1998) 81:2729–2737.[Abstract]
Szulkin M. Inbreeding and its avoidance in a wild bird population [DPhil thesis] (2007) Oxford: University of Oxford.
Szulkin M, Garant D, McCleery RH, Sheldon BC. Inbreeding depression along a life-history continuum in the great tit. J Evol Biol (2007) 20:1531–1543.[CrossRef][Web of Science][Medline]
Szulkin M, Sheldon BC. Inbreeding: when parents transmit more than genes. Curr Biol (2006) 16:R810–R812.[CrossRef][Web of Science][Medline]
Szulkin M, Sheldon BC. The environmental dependence of inbreeding depression in a wild bird population. PLoS ONE (2007) 2:e1027.[CrossRef]
Szulkin M, Sheldon BC. Dispersal as a means of inbreeding avoidance in a wild bird population. Proc R Soc Lond B Biol Sci (2008) 275:703–711.[CrossRef]
van Noordwijk AJ, Scharloo W. Inbreeding in an island population of the great tit. Evolution (1981) 35:674–688.[CrossRef][Web of Science]
Wheelwright NT, Freeman-Gallant CR, Mauck RA. Asymmetrical incest avoidance in the choice of social and genetic mates. Anim Behav (2006) 71:631–639.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||