Dispersal is an important mechanism used to avoid inbreeding. However, dispersal may only be effective for part of an individual's lifespan since, post-dispersal individuals that breed over multiple reproductive events may risk mating with kin of the philopatric sex as they age. We tested this hypothesis in black grouse Tetrao tetrix, and show that yearling females never mated with close relatives whereas older females did. However, matings were not with direct kin suggesting that short-distance dispersal to sites containing kin and subsequent overlap of reproductive lifespans between males and females were causing this pattern. Chick mass was lower when kinship was high, suggesting important fitness costs associated with inbred matings. This study shows that increased inbreeding risk might be a widespread yet rarely considered cost of ageing.
Natal dispersal is a one-way movement of individuals from their location of birth to their breeding location . Natal dispersal rates are typically sex-biased , and the predominant evolutionary forces that account for these patterns include inbreeding avoidance [2,3], competition for resources  or mates , or cooperative behaviour among kin . Natal dispersal typically occurs prior to the first breeding attempt and one reason is to avoid reproducing with kin . Dispersal can be costly, with reduced survival or reproduction of dispersers (e.g. ), so dispersal after the first breeding attempt (breeding dispersal) is usually rarer than natal dispersal rates . Increased inbreeding risk can trigger breeding dispersal , but in many species there is little evidence of active inbreeding avoidance , so that most individuals disperse only once.
The benefits of dispersal as a mechanism for reducing inbreeding depends on dispersal distances and propensities , but benefits may not last a lifetime. Irrespective of the dispersal distance, there is an increased likelihood in dispersers of both sexes of breeding with philopatric kin (i.e. sons or daughters) during subsequent reproductive attempts. Alternatively, individuals with short-distance dispersal may be at risk of mating with kin by joining a site where close relatives are already present. Inbred individuals might have lower fitness-related characteristics [10,11], and so increased risks of mating with close relatives may be an important additional fitness cost of ageing.
To our knowledge, no previous studies have rigorously determined the extent to which inbreeding risk increases over time after individuals' natal dispersal. Therefore, we tested this hypothesis in the black grouse (Tetrao tetrix), a lekking bird with male philopatry and female-biased dispersal [12–14]. In this species, females mate with a single male and males' reproduction is skewed towards dominant 3–5 year old males at leks [15,16]. Since breeding dispersal is rare  and females may not discriminate against relatives during mate choice, females may experience increased risk of mating with relatives in subsequent breeding attempts. To assess the importance of age, inbreeding risk and their impact on fitness, we used 5 years of parentage data from a natural population of black grouse, to examine (i) whether older females mate more often with close relatives and if (ii) increased inbreeding induces inbreeding depression by relating parental kinship to offspring body mass.
2. Material and methods
Black grouse were captured in Central Finland from 2002 to 2006 at one to seven feeding sites. The birds were aged according to the shape of the outmost primaries , weighed, marked with aluminium and colour rings for future identification and morphological measurements taken . Blood (1–2 ml) was taken with a heparinized syringe from the brachial vein, and red blood cells were stored in 70 per cent ethanol at 4°C until DNA extraction. Most broods were sampled on the hatching date, with all broods sampled less than 2 days post-hatching; chicks were weighed to nearest 0.1 g and two capillaries (70 µl) were filled with blood after puncturing one of the brachial wing veins. The red blood cells were kept and extracted as described above.
Genomic DNA was extracted from the red blood cells, genotyped at 11 autosomal microsatellite loci and one sex-specific locus (see  for details). Parentages were assigned using the procedures detailed in Lebigre et al. . We calculated the coefficient of kinship (k) between all parents; k equals half of the relatedness coefficient RQG  and as we did not have pedigree information available, k was calculated as half of RQG generated using SPAGeDI . Close relatives were conservatively defined as k > 0.10 (RQG > 0.20: see ).
Females at capture were aged as yearling (1 year) or older (more than or equal to 2 years). Recapture data of yearling-aged females were low, but repeated captures of females aged 2 years or more allowed us to assign females into four minimum age classes: 1 (exact age), 2 (exact age), more than or equal to 3 (minimum 3 years old) and more than or equal to 4 (minimum 4 years old). Maximum female lifespan is approximately 5–6 years, but generally much lower. We examined the relationship between female age and k of individual matings using a Spearman's rank correlation. Individual matings rather than individual females were used as the sampling unit to account for one case of multiple paternity. No female was present in more than one age class.
We used two linear mixed-effects (lme) models to examine the effect of age (1, 2 or more) and k on log-transformed chick mass (gram). The first was an additive model, with both age and parental relatedness as variables, while the second was an interaction model that contained the interaction between female age and k. In both models, female ring number was a random grouping variable. Effect sizes r and non-central 95% CI were calculated for all results . Models were checked for significant differences from the null model using the likelihood ratio test. All models were run in R , using the nlme package . Inbreeding load B (95% CI) for each female age class was calculated as the slope of the regression of log-transformed chick mass on k.
There was a positive relationship between female age and k (n = 42 matings, Spearman's rank correlation: rs = 0.381, p = 0.012). Matings between close relatives occurred only with females aged 3 years or more (figure 1). Both female age (lme model: r = 0.08 (0.02/0.13), t1,68 = 2.51, p = 0.015) and k (lme model: r = 0.07 (0.01/0.13), t1,511 = −2.19, p = 0.029) significantly affected chick mass (n = 582 chicks, 70 females, full model: χ2 = 9.48, p = 0.009). Older females had significantly heavier chicks, but chick body mass was significantly negatively related to k (figure 2); slopes of the regression lines (inbreeding load: B) of chick mass and k were not significant different between yearling (B = −0.26 (−0.49 to −0.02, intercept = 3.13)) and older females (B = −0.35 (−0.44 to −0.26, intercept = 3.19); t578 = −0.79, p = 0.431). There was no effect of the interaction between age and k (lme model, r = −0.01 (−0.05/0.06), t1,510 = 0.18, p = 0.857) on chick mass, with female age (lme model: r = 0.08 (0.02/0.14), t1,68 = 2.49, p = 0.015) but not k significant in the model (lme model: r = −0.03 (−0.09/0.03), t1,510 = −0.86, p = 0.289; full model: χ2 = 9.47, p = 0.024). There was no significant difference between models (likelihood ratio test: χ2 = 0.88, p = 0.348).
This study demonstrates that the risks of mating with a close relative increase with female age, a pattern that was previously suggested in grouse . Since the majority of females aged two years or more had at least one son in the flock closest to their nesting site  and the majority of copulations are acquired by males at ages 3–5 years , females surviving to ages three or older would be at highest risk of mating with philopatric male kin. However, matings were between second degree relatives (k ∼ 0.125) rather than with first degree relatives (k ∼ 0.25). This suggests that the age-related increase in inbred matings is related to female dispersal to sites with existing kin and overlap of female lifespan with the reproductive lifespans of male kin. Instead, sex differences in survival and sex differences in the age at first reproduction may limit the opportunity for mother–son matings in this species.
Alternatively, matings between relatives could be reduced through kin recognition, e.g. phenotype matching . No matings between mothers and sons are expected if females prefer mating with second degree relatives to mating with sons. However, inbred matings occurred when females mated with already very successful males, suggesting that females prefer dominant males irrespective of their relatedness . Overall, female-biased dispersal combined with age-structured male reproduction are effective mechanisms to reduce direct inbreeding, as no yearling females were closely related to their mate, compared with an overall rate of 12.8 per cent of matings between close relatives (see ).
Chick mass was significantly greater in older females but significantly decreased with increased k in both age classes. This latter effect might be due to reduced maternal investment (e.g. [26,27]) or inbreeding depression . As the level of inbreeding and chick mass are both likely to directly impact chick survival to recruitment [28,29], females that mate with close relatives might have reduced fitness as fewer offspring are likely to survive to adulthood.
In conclusion, this is, to our knowledge, the first study to show age-related inbred matings associated with fitness costs through reduced chick mass. Though inbreeding may have fitness costs, these may be traded-off by the costs of dispersing a second time . Age-specific inbreeding may potentially be found in other species, particularly in systems where females select a new mate each year, or in species with mate fidelity, after being divorced or widowed. Though our study demonstrates an age-specific inbreeding, there is a need for a more detailed examination of the mechanism(s) involved in inbreeding risk and their relationship to female age.
We thank Matti Halonen, Gilbert Ludwig, Tuomo Pihlaja, Elina Virtanen and numerous people over the years who have helped both in the laboratory and the field. Funding was provided by the Academy of Finland (grant nos. 7211271 and 7119165) and the Finnish Center of Excellence in Evolutionary Research (Academy of Finland).
- Received April 5, 2011.
- Accepted May 9, 2011.
- This journal is © 2011 The Royal Society