Context-dependent genetic benefits of extra-pair mate choice in a socially monogamous passerine
Extra-pair paternity is common in socially monogamous passerines; however, despite considerable research attention, consistent differences in fitness between within-pair offspring (WPO) and extra-pair offspring (EPO) have not been demonstrated. Recent evidence indicates that differences between maternal half-siblings may depend on environmental conditions, but it is unclear whether the influence of paternal genetic contribution should be most apparent under comparatively poor or favourable conditions. We compared phenotypic characteristics of WPO and EPO in 30 mixed-paternity broods of the tree swallow (Tachycineta bicolor) in relation to experimentally increased nest temperature (n = 13 heated nests; 17 control nests) and natural abundance of haematophagous parasites (Protocalliphora spp.). This allowed us to test the hypothesis that genetic benefits of extra-pair mating are environment dependent. EPO grew their ninth primary feathers faster than WPO regardless of nest temperature or parasite load and had significantly longer ninth primary feathers at fledging when parasite abundance was low, and when they were positioned early in the hatching sequence relative to WPO. In contrast, WPO under similar conditions did not differ from EPO in any phenotypic trait measured. These results indicate that the fitness benefits of extra-pair mating are likely to be context dependent, and that genetic effects on some phenotypic traits may be more apparent when conditions are relatively favourable.
KeywordsExtra-pair paternity Good genes Context dependence Fitness consequences Tachycineta bicolor
Extra-pair paternity is common in socially monogamous passerines (Petrie and Kempenaers 1998), and because extra-pair males generally do not provide material benefits, such as territory access or parental care, it is likely that females seek extra-pair fertilizations to obtain indirect benefits, such as high-quality genes for their offspring (Griffith et al. 2002). Despite considerable research attention, however, consistent differences in fitness between within-pair offspring (WPO) and extra-pair offspring (EPO) have not been demonstrated (Griffith et al. 2002). In some species, EPO are more likely to survive (Kempenaers et al. 1997), fledge in a better condition (Sheldon et al. 1997) or exhibit enhanced immune responses (Johnson et al. 2000) relative to their within-pair nest mates. In other species, however, no such differences have been found (e.g. Lubjuhn et al. 1999; Whittingham and Dunn 2001; Kleven and Lifjeld 2004). Consequently, the extent to which female birds select extra-pair mates to obtain indirect genetic benefits for their offspring is not known.
Genetic benefits of extra-pair mate choice should be evident in comparisons of phenotypic characteristics of WPO and EPO, as many traits are both heritable (Ryan 2001) and strongly correlated with future reproductive success or survival in birds (Gebhardt-Henrich and Richner 1998; McCarty 2001). The expression of genetic variation in phenotypic traits, however, commonly varies with environmental conditions (e.g. Merilä 1997; Qvarnström 1999). Thus, the magnitude of phenotypic differences between WPO and EPO that are due to paternal genetic contribution may depend on the conditions of the rearing environment. Context dependence of genetic benefits of mate choice has been demonstrated in a range of taxa, including invertebrates, amphibians and birds (Jia et al. 2000; Welch 2003; Schmoll et al. 2005). To our knowledge, however, only one study has considered context dependence in comparisons of WPO and EPO of socially monogamous birds (Schmoll et al. 2005), and it is unclear whether context-dependent genetic benefits of mate choice should be more evident in low or high stress environments (e.g. Welch 2003; Schmoll et al. 2005). Hence, consideration of genotype-environment interactions may be necessary to identify genetic benefits of extra-pair mating in socially monogamous birds and to determine the range of environmental conditions, under which selection for extra-pair mate choice by females may be maintained in natural populations.
We compared phenotypic characteristics of maternal half-siblings in 30 mixed-paternity broods of tree swallows (Tachycineta bicolor) exposed to varying environmental conditions. We tested the effects of experimentally increased nest temperature, as well as the naturally occurring abundance of Protocalliphora spp., on relative performance of WPO and EPO, as measured by growth and size at fledging. Experimentally heating nests has been shown to enhance growth and size at fledging for mass and ninth primary feathers in tree swallows (Dawson et al. 2005a). This enhanced growth of chicks in heated nests appears to be due to a reduction in thermoregulation costs, allowing greater allocation of resources to growth compared to chicks in unheated nests (Dawson et al. 2005a). In addition, nest-dwelling ectoparasitic haematophagous blow fly (Protocalliphora spp.) larvae have significant negative impacts on nestling survival and condition at fledging in a range of bird species, including cavity nesters (e.g. Merino and Potti 1996; O’Brien et al. 2001). If female tree swallows engage in extra-pair mating to obtain genetic benefits for their offspring and these benefits are context-dependent, we predict that the performance of EPO relative to their within-pair half-siblings will depend on the level of thermal stress or parasite infestation to which nestlings are exposed.
Materials and methods
Study area and study species
We studied tree swallows breeding in nest boxes near Prince George, BC, Canada (53°N, 123°W) from May to August 2004. The site contained 154 nest boxes mounted on fence posts and placed approximately 25 m apart. The study area consists of open agricultural fields mixed with patches of coniferous and deciduous forest, as well as many small wetlands, and is located in a region characterized by a temperate climate, with cool summer temperatures.
Tree swallows are small, migratory, socially monogamous passerines (Robertson et al. 1992) that exhibit among the highest rates of extra-pair paternity identified in birds: extra-pair young have been detected in up to 90% of broods, with as many as 69% of all offspring resulting from extra-pair fertilizations (Barber et al. 1996; Conrad et al. 2001; Whittingham and Dunn 2001). Females do not gain direct benefits from extra-pair mates, yet no consistent indirect benefits of extra-pair mating, in terms of offspring quality or survival, have been identified in this species (Whittingham and Dunn 2001; Whittingham et al. 2003).
Field and experimental methods
Beginning in May, we visited nests every 2 days to determine clutch initiation date and clutch size. Near expected hatch dates, nests were visited every day to determine date of hatch. Adults were captured in nest boxes with a swing-door trap. We banded adults and 14-day-old nestlings with standard aluminium leg bands and collected blood samples (approximately 50 μl) by puncturing the brachial vein.
Nests were paired by hatching date and brood size, and one nest from each pair was randomly assigned to either the heating treatment or control group. Using methods described in Dawson et al. (2005a,b), we experimentally increased temperature of treatment nests (n = 15) by installing small heating pads in nests when young were 4 days old until they were 16 days of age. These pads, which were changed every 2 days, produced heat when the iron powder they contained was exposed to air and oxidized. Control nests (n = 18) contained pads in which the iron powder had been fully oxidized, so heat was not produced. Although we did not measure nest temperatures in this study, this heating treatment has been shown to increase nest temperature by up to 10°C relative to control nests (Dawson et al. 2005a). Treatment and control nests, therefore, represented low and high thermal stress conditions, respectively. Once chicks had fledged, nesting material was collected from each box and stored in sealed bags at room temperature for 14 days. We then sifted through the nesting material to remove all pupae of Protocalliphora. This procedure has been used to quantify parasite abundance in previous studies of blow flies (e.g. Hurtrez-Boussès et al. 1997; Dawson et al. 2005b), with the number of pupae representing the minimum level of infestation. Because the abundance of Protocalliphora increased with brood size (see “Results”), we used the number of pupae per nestling as a relative measure of parasite load in each nest.
Parents may have responded to the nest temperature manipulation or the relative abundance of nest-dwelling ectoparasites by adjusting provisioning rates (Tripet and Richner 1997), which could directly influence nestling performance. We measured parental provisioning rates by placing a digital recorder and small microphone in nest boxes to record begging of chicks. Recordings of at least 1 h (1.46 ± 0.08 h SE, n = 55 recordings from 33 nests) were obtained when chicks were 12–14 days old; most nests were recorded on more than 1 day within this interval, which allowed us to calculate an average provisioning rate for each nest. The number of bouts of begging accurately represents the number of parental visits to nest boxes (E. L. O’Brien, unpublished data), and because parent tree swallows rarely visit nest boxes without feeding chicks (McCarty 2002; Whittingham et al. 2003), the frequency of parental visits is an accurate measure of the rate of food delivery to nestlings (McCarty 2002).
We measured nestling mass, tarsus and ninth primary feather length every 2 days from ages 4 to 16 days. When chicks were 9 days old, we tested the T cell-mediated immune response (CMI) by injecting 30 μl of 1 mg/ml phytohaemagglutinin-P (Sigma) in phosphate buffered saline intradermally in the right wing web. For each nestling, thickness of the wing web was measured four times using a thickness gauge (nearest 0.01 mm) immediately before and 24.07 h (±0.01 SE, n = 172) after injection; the means of these four measurements were then used in all subsequent analyses. We determined CMI as the difference in thickness of the wing web before and after injection (Smits et al. 1999). Growth rate constants for morphological traits were calculated using the logistic model for growth of mass, Gompertz model for growth of tarsus and a linear model for growth of ninth primary. Model selection procedures followed Dawson et al. (2005a).
Blood samples were stored in 1 ml of Queen’s lysis buffer (Seutin et al. 1991) at 4°C. We isolated genomic DNA by proteinase K-phenol/chloroform extraction, and DNA was subsequently resuspended in Tris–EDTA (TE) buffer and stored at −20°C. We used the following three microsatellite loci for paternity analysis: HrU6 (Primmer et al. 1995), HrU10 (Primmer et al. 1996) and IBI MP5-29 (Crossman 1996). Polymerase chain reaction (PCR) amplification was carried out in 10-μl volumes containing 50–100 ng genomic DNA, 100 μM deoxyribonucleotide triphosphates, 2.5–3.0 mM MgCl2, 1× PCR buffer (Invitrogen), 0.2 μM forward (labelled) and reverse primers and 0.35 units of Taq DNA polymerase (Invitrogen). PCR reactions were performed using an MJ Research Peltier thermal cycler under the following conditions: for HrU6, 1 cycle at 94°C for 2 min, followed by 40 cycles at 94°C for 30 s, 58°C for 30 s and 72°C for 1 min and one final extension step at 72°C for 4 min. Conditions for HrU10 and IBI MP5-29 were identical except we used annealing temperatures of 59°C and 57°C, respectively. PCR products were analysed using a Beckman-Coulter CEQ 8000 automated sequencer.
Variability of microsatellite loci for 111 adult tree swallows
We used paired t-tests to compare mean values of phenotypic traits for EPO and WPO in each nest containing offspring of mixed paternity (30 nests; see “Results”). This allowed us to initially determine whether EPO differed from their maternal half-siblings across all nests. To assess whether environmental conditions influenced relative performance of half-siblings, we used analysis of covariance to test for effects of the heating treatment as a fixed factor, and Protocalliphora abundance (number of pupae/nestling), provisioning rate (feeds/hour/nestling), brood size and hatching date as covariates, on the relative performance of EPO and WPO. Position in the hatching sequence affects relative nestling performance (e.g. growth or size at fledging) in many altricial birds, including tree swallows (Clotfelter et al. 2000), and both hatch order and asynchrony can be estimated by relative body mass soon after hatch (Clotfelter et al. 2000). To account for within-brood phenotypic variation due to position in the hatching sequence, we therefore included difference in mass at day 4 (average mass of EPO − average mass of WPO) as an additional covariate representing relative hatch order. Thus, values >0 indicate that, on average, EPO hatched earlier than WPO, whereas values <0 indicate that WPO hatched earlier than EPO. Similarly, for each phenotypic trait, we used the difference between EPO and WPO (average EPO value–average WPO value) as the dependent variable; thus, values >0 indicate EPO were larger or grew faster than WPO, whereas values <0 indicate the opposite pattern. Initial models tested all main effects and interactions. Terms that were not significant (P > 0.10) were then removed by backwards stepwise elimination until only significant terms remained, or a significant interaction was detected. We used SPSS (Norušis 2000) software for all analyses. Means are presented ±1 standard error (SE). All tests were two tailed, and results were considered significant at alpha < 0.05 level.
Results of two-tailed paired t-tests comparing phenotypic traits of extra-pair and within-pair offspring in mixed-paternity broods of tree swallows
Mean difference (±SE)
Ninth primary growth (mm/day)
0.12 ± 0.05
Mass growth (g/day)
0.02 ± 0.01
Tarsus growth (mm/day)
0.01 ± 0.02
Ninth primary at day 16 (mm)
0.88 ± 0.49
Mass at day 16 (g)
0.27 ± 0.24
Tarsus at day 16 (mm)
0.02 ± 0.06
T cell-mediated immune response (mm)
0.02 ± 0.05
The heating treatment did not influence provisioning rates (heated nests: 5.7 ± 0.7 feeds/chick, control nests: 4.7 ± 0.3; t = 1.44, df = 26, P = 0.16), nor did heated nests differ from controls in parasite abundance (heated: 6.7 ± 1.1 Protocalliphora/nestling, controls: 9.1 ± 1.4; t = 1.25, df = 26, P = 0.22). All experimental nests (heated and controls) that contained offspring of mixed paternity were infested with Protocalliphora (excluding one nest, for which parasite abundance was not determined). The average number of Protocalliphora detected was 45.6 ± 32.2 (range, 3–122; n = 29). This level of infestation is similar to that reported in nests of tree swallows in other study populations (e.g. Rogers et al. 1991; Dawson et al. 2005b). Nests with larger broods contained more Protocalliphora pupae (r = 0.48, P = 0.009, n = 29), so all subsequent analyses used Protocalliphora/nestling as a covariate representing level of parasite infestation.
Expression of genetic variation in phenotypic traits often varies with environmental conditions (Qvarnström 2001), and genetic benefits of extra-pair mate choice in socially monogamous birds may therefore not be evident in all environments. Our results support this hypothesis, as EPO positioned early in the hatching sequence produced longer flight feathers than their within-pair half-siblings in nests that contained few parasitic blow fly larvae, whereas WPO that hatched early did not show a similar advantage under low parasite loads (Fig. 1). Wing length is a strong predictor of timing of fledging in tree swallows (Michaud and Leonard 2000) and other aerial insectivores (Martins 1997); thus, in some rearing environments, EPO may reach fledging condition earlier than WPO. Fitness benefits of longer flight feathers may also extend beyond the fledging period in swallows. For example, wing length at fledging is positively associated with probability of recruitment (McCarty 2001), and males with longer flight feathers attract social mates of higher quality and are more likely to sire extra-pair young (E. L. O’Brien, unpublished data). Finally, females who exhibit superior flight performance, a behaviour influenced by traits such as wing length, lay more eggs and initiate clutches earlier in the breeding season (Bowlin and Winkler 2004). For female tree swallows, the fitness benefits of producing EPO in some contexts (e.g. when parasite abundance is low) may therefore be substantial, and the context dependence of these benefits may explain why previous studies have failed to find significant effects (e.g. Whittingham and Dunn 2001).
For traits such as body size or growth, phenotypic expression of genetic variation may be enhanced in favourable environments, where nutrient limitation is minimal (Hoffman and Merilä 1999). Studies of heritability (Qvarnström 1999) and mate choice (Welch 2003) have supported this prediction, and the results of our study suggest that some genetic benefits of extra-pair mating in socially monogamous birds may similarly be realized under low stress conditions (i.e. low parasite load). However, the alternative hypothesis that stressful conditions should promote the expression of heritable phenotypic variation (Hoffman and Merilä 1999) has also received empirical support (e.g. Sheldon et al. 2003; Schmoll et al. 2005), and this discrepancy suggests that selection on phenotypic variation may not be consistently strongest under either high- or low-stress conditions. Whether unfavourable environments constrain or enhance expression of heritable variation in phenotypic traits may depend on the stressor and traits considered (see Malausa et al. 2005). For example, high levels of parasitism may enhance expression of genes associated with immune resistance while reducing expression of growth-promoting loci; under these conditions, variation in traits such as immune response should be more apparent and differences in body size or growth reduced (e.g. Bize et al. 2003). For tree swallows, neither parental provisioning rates nor manipulation of nest microclimate affected relative performance of maternal half-siblings (in terms of growth or size at fledging), whereas phenotypic differences between EPO and WPO were evident when parasite abundance was low. Because we did not record nest temperatures in this study, we cannot exclude the possibility that the heating treatment did not have the desired effect on microclimate of nests. However, this method has previously been shown to increase temperatures in nests of tree swallows in the same geographic area (Dawson et al. 2005a), and we believe it is more likely that unusually mild weather during the year of this study (unpublished data) resulted in all nestlings experiencing relatively low thermal stress. It is therefore possible that abiotic factors, such as nest microclimate, can influence the relative performance of half-siblings in some environments; however, our results suggest that stresses due to environmental factors, such as feeding conditions or nest temperature, may have less potential to influence within-brood variation in phenotypic traits of nestlings in this species than factors such as parasites. Future studies investigating the fitness benefits of extra-pair paternity in socially monogamous birds should therefore examine phenotypic responses of offspring exposed to a wide range of conditions, representing both poor and favourable environments.
Preferential allocation of limited resources to growth of morphological traits that have high functional priority occurs in a range of taxa, including mammals (e.g. Festa-Bianchet et al. 2004) and birds (Ashton and Armstrong 2002; Bize et al. 2003). In aerial insectivores, allocation to growth of wing feathers may be particularly important for fledging (Martins 1997) and adult survival (Brown and Brown 1998). Differences in the tendency to preferentially allocate resources to wing growth may therefore represent an important source of variation in nestling quality. Across all environmental conditions and regardless of relative hatching order, EPO of tree swallows grew wing feathers faster than their WPO half-sibs (Table 2) and, under favourable conditions, had considerably longer ninth primaries at fledging (Fig. 1). This suggests that EPO had a greater tendency to preferentially allocate resources to wing growth than WPO in our study population. More generally, these results highlight the potential importance of examining heritability of resource allocation strategies, and how these strategies influence phenotypic traits under varying levels of resource availability.
Female birds may adjust yolk androgen investment (Gil et al. 1999), egg or clutch size (Cunningham and Russell 2000; Parker 2003) according to the attractiveness of their mates. Thus, phenotypic differences between WPO and EPO may reflect differential allocation of resources among offspring of different sires in addition to effects of paternal genetic contribution (Sheldon 2000). We are unaware of any evidence, however, that female birds are capable of adjusting relative investment to specific eggs within a clutch (see Birkhead et al. 2000) except in relation to laying order (e.g. Saino et al. 2001). Because EPO of tree swallows are not consistently positioned early or late in the hatching sequence (this study; see also Whittingham et al. 2003), differential allocation of maternal resources based on laying order would not allow females to preferentially invest in EPO in this species. Similarly, although male nestlings of tree swallows are larger than their female siblings by day 12 (Whittingham et al. 2003), EPO are not more likely to be male (Whittingham and Dunn 2001), and phenotypic differences between WPO and EPO can therefore not be attributed to differences in sex allocation according to paternity. Phenotypic differences between maternal half-siblings are also unlikely to have resulted from biased distribution of parental feedings: allocation of provisioning effort among nestlings by both parents is generally equitable in tree swallows (Leonard and Horn 1996) regardless of the presence or abundance of EPO in the brood (Whittingham et al. 2003). It is therefore unlikely that phenotypic differences between WPO and EPO in this study can be attributed to differential allocation of resources to EPO.
Expression of heritable variation in phenotypic traits varies with environmental conditions (e.g. Merilä 1997; Qvarnström 1999). Hence, the influence of paternal genetic contribution on performance of WPO and EPO should not necessarily be evident in all contexts. Results of this study support the hypothesis that genetic benefits of extra-pair mating by female tree swallows are environment dependent and, contrary to the findings of Schmoll et al. (2005), these benefits were realized under comparatively favourable conditions (low parasite abundance). Clearly, future tests of ‘good genes’ models of female extra-pair mate choice should consider environmental variation in comparisons of maternal half-sibling performance not only to identify genetic benefits of extra-pair paternity but also to determine the range of environmental conditions under which selection for extra-pair mate choice by females is maintained.
We thank Stephen Pruett-Jones and two anonymous reviewers for comments on a previous version of the manuscript. Anita Castle and Greg Sanders granted property access, and Dan Baxter, Pierre-Paul Bitton and Ben Schonewille provided assistance in the field. We are grateful to Brent Murray for providing access to lab space and equipment, as well as the UNBC molecular genetics lab and Mary Stapleton for advice and discussions concerning genetic analyses. Funding was provided by a Discovery Grant to Russell D. Dawson from the Natural Sciences and Engineering Research Council of Canada. Additional funding was provided by the Canada Foundation for Innovation and British Columbia Knowledge Development Fund. Our research protocols were approved by the Animal Care and Use Committee of UNBC.
- Crossman CC (1996) Single locus DNA profiling in the tree swallow Tachycineta bicolor: a comparison of methods. MSc thesis, Queen’s University, CanadaGoogle Scholar
- Gebhardt-Henrich SG, Richner H (1998) Causes of growth variation and its consequences for fitness. In: Starck JM, Ricklefs RE (eds) Avian growth and development. Oxford University Press, Oxford, pp 324–339Google Scholar
- Jamieson A (1994) The effectiveness of using codominant polymorphic allelic series for (1) checking pedigrees and (2) distinguishing full-sib pair members. Anim Genet 25:19–23Google Scholar
- McCarty JP (2002) The number of visits to the nest by parents is an accurate measure of food delivered to nestlings in tree swallows. J Field Ornithol 73:9–14Google Scholar
- Norušis MJ (2000) SPSS 11.0 Guide to data analysis. Prentice Hall, Upper Saddle RiverGoogle Scholar
- Robertson RJ, Stutchbury BJ, Cohen RR (1992) Tree swallow (Tachycineta bicolor). In: Poole A, Stettenheim P, Gill F (eds) The birds of North America, no. 11. The Academy of Natural Sciences, Philadelphia, PAGoogle Scholar
- Rogers CA, Robertson RJ, Stutchbury BJ (1991) Patterns and effects of parasitism by Protocalliphora sialia on tree swallow nestlings. In: Loye JE, Zuk M (eds) Bird-parasite interactions: ecology, evolution and behaviour. Oxford University Press, Oxford, pp 123–139Google Scholar