Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jul 3.
Published in final edited form as: J Evol Biol. 2026 Mar 2;39(3):394–403. doi: 10.1093/jeb/voaf148

An experimental test for ecologically dependent reproductive isolation across an avian migratory divide

Stephanie A Blain 1,*, Hannah C Justen 2, Kira E Delmore 1
PMCID: PMC13325972  NIHMSID: NIHMS2133757  PMID: 41405419

Abstract

Divergent adaptation can promote ecological speciation if hybrids have reduced fitness because they are poorly adapted to either parental niche. We tested for ecologically dependent, postzygotic isolation between two subspecies of Swainson’s thrushes, which form a migratory divide and hybrid zone in western North America. To do this, we translocated backcrossed and admixed birds from the hybrid zone into the range of each subspecies in the beginning of fall migration. We estimated a proxy for their survival on migration and migratory behaviour using automated radio tracking. Apparent survival of birds in the two environments did not depend on their genomic ancestry, suggesting that Swainson’s thrushes’ divergent adaptation to different fall migration routes does not fit the classic model of ecological speciation. We propose an alternate scenario where ecological selection on migration may interact with intrinsic maladaptation in hybrids to cause hybrid survival on migration. By translocating birds from the same genomic backgrounds into different environments, our experiment also allowed us to distinguish between the effects of environmental relative to genetic contributors to their migratory behaviour. We found evidence that both genetic and environmental factors influence migratory behaviour, as an effect of genomic ancestry on initial migratory trajectories depended on the start location for migration but birds ultimately followed expected routes given their genomic ancestries.

Introduction

Divergent ecological adaptation between species can contribute to the evolution of reproductively isolating barriers between them (Funk et al., 2006; Rundle & Nosil, 2005). Under the ecological speciation hypothesis, reduced hybrid fitness can result not only from intrinsic dysfunction but also if hybrid phenotypes are poorly adapted to the different niches occupied by the parental species (Schluter, 2009). This scenario of ecologically dependent isolation is commonly inferred, but there are few examples in which it has been directly tested. To establish that isolation derives from divergent ecological adaptation between the parental species, one possible test is to estimate fitness of backcrossed hybrids translocated to the environments of both parental species (Rundle & Whitlock, 2001). If first generation hybrids perform poorly in an environment relative to the parental species, this could be a result of either intrinsic or ecological maladaptation. However, if a backcrossed hybrid performs better in the environment of its majority parent species than that of the minor parent species, this demonstrates a role for ecological adaptation in isolation.

Translocation experiments are additionally useful for determining the relative contributions of environmental and genetic variation to variation in phenotypic traits, including those hypothesized to be important for speciation, because they may also be differently expressed in backcrossed hybrids transplanted to different parental species’ environments. Specifically, if phenotypic variation is mediated largely by environmental variation, translocated individuals should exhibit behaviour similar to individuals from their new environment (and vice versa). This may be particularly useful for complex traits that cannot be fully measured or expressed in a laboratory environment (Åkesson, 2003). Seasonal migration, the annual movement of populations between breeding and nonbreeding localities, is a complex phenotype that can only be fully expressed in a natural environment (Alerstam et al., 2003). Seasonal migration has been hypothesized to drive ecological speciation via migratory divides, a form of divergent ecological adaptation in which two lineages evolve different migratory routes (Helbig, 1991; Irwin & Irwin, 2005). When speciation is associated with the evolution of a migratory divide, ecological isolation due to poor hybrid performance on migration is often suggested as a possible contributing factor (Scordato et al., 2020; von Rönn et al., 2016).

Ecological isolation driven by hybrid maladaptation on migration may maintain divergence of between two subspecies of Swainson’s thrush (Catharus ustulatus; coastal and inland subspecies). These subspecies hybridize in western North America, along the Coast and Cascade Mountains. Estimates of sequence divergence from mitochondrial data (0.69%; Ruegg & Smith, 2002) suggest they diverged during the late Pleistocene. Population densities at the center of the hybrid zone are low and geographic clines constructed using phenotypic and genetic data are narrow and coincident, suggesting that reproductive isolation is being maintained by selection against hybrids in the system (Ruegg, 2008). The subspecies differ in song and color but the trait that distinguishes them most is their migratory behavior (Delmore & Irwin, 2014); the coastal subspecies follows a migration route west of the hybrid zone, navigating along the Pacific Coast to Central America, while the inland subspecies begins migration with a more eastward orientation, crossing continental North America and the Caribbean to overwinter in Columbia (Delmore et al., 2012). There is good evidence these differences in migration serve as extrinsic postzygotic isolating barriers. Individual tracking data have shown that, relative to the parental subspecies, hybrids take intermediate migratory routes to parental forms (Delmore & Irwin, 2014). Ecological models indicate these routes take hybrids through areas of reduced habitat suitability that permit less movement (Justen et al., 2021) and estimates of survival indicate that coastal backcrosses and hybrids have reduced survival on migration (Blain et al., 2024; Delmore & Irwin, 2014; H. Justen et al., 2021). Hybrids, backcrosses, and the parental subspecies all survive well in a laboratory environment where ecological stress is limited, providing further evidence that there is an extrinsic component to speciation (Louder et al., 2024).

We tested for ecologically dependent isolation in the migratory divide between Swainson’s thrushes by translocating both genetically intermediate and backcrossed juvenile hybrids from the centre of the hybrid zone to a site within the inland subspecies range and to a site within the coastal range in the beginning of fall migration. Using automated radio telemetry, we quantified detection as a proxy for survival on migration and migratory behaviour. First, we tested for an effect of translocation on survival. Under the ecological speciation hypothesis, we predicted that survival would depend on genomic ancestry, with inland backcrosses exhibiting higher relative survival when translocated east, into the inland range, and coastal backcrosses favoured in the coastal range. Second, we used the translocation experiment to examine the relative effects of genetic and environmental factors on migratory behaviour. If genetic effects contribute to migratory orientation, we predicted that relationships between orientation and ancestry would persist regardless of their release site. However, if orientation is primarily determined by environmental effects, we expected that release site would affect migratory orientation.

Methods

Transplant experiment

Juvenile Swainson’s thrushes were captured in the beginning of fall migration (August to September) in 2020 and 2023 at sites inside the hybrid zone in British Columbia (Table S1). In 2020, 20 birds were transplanted to a site near Lillooet, within the inland subspecies range (Fig. 1). In 2023, 30 birds were transplanted to a site near Squamish, within the coastal subspecies range. The centre of the hybrid zone and the borders of the coastal and inland subspecies ranges were defined based on results from previous geographic cline analyses conducted using these localities, which identified all individuals captured in Squamish as coastal and 82% of individuals captured in Lillooet as inland (Ruegg, 2008). Additional sets of birds were captured and released without being transplanted at sites within the hybrid zone in 2020 (n = 90) and 2023 (n = 97).

Figure 1.

Figure 1.

Open in a new tab

Map showing the hybrid zone and translocation experiment. (Left) Breeding and non-breeding ranges of the Swainson’s thrush, with the coastal subspecies shown in blue and the inland in green. (Right) Black points indicate capture sites, while large, coloured points indicate release sites. 20 birds were translocated to Lillooet, in the inland range, while 30 birds were translocated to Squamish, in the coastal range. Grey shading indicates elevation.

Birds were captured with mist nets, using song playback, within a region of the hybrid zone that birds with a range of ancestries are known to pass through on migration. All bird handling followed protocols approved by Institutional Animal Care and Use Committee at Texas A&M (IACUC 2019–0066) and Environment and Climate Change Canada (10921). We took a blood sample from each bird when they were first captured. These samples were used to extract DNA and construct whole genome resequencing libraries (see below).

Before release, all birds were tagged with radio transmitters (Lotek nanotags NTQB2–3-2). These tags emit uniquely identifiable signals at 166.380 MHz frequency with burst intervals every 25 seconds. They are detected by receiver stations that are part of the Motus Wildlife Tracking System (https://www.motus.org). This system includes radio stations maintained across North America, including a transect of stations that span the hybrid zone between Swainson’s thrushes in southern British Columbia (Justen et al., 2024). These radio station detections were used to estimate both survival and fall migratory behaviours of the experimental birds.

Whole genome resequencing and genomic ancestry

Libraries were prepared, sequenced and used to call SNPs following methods described in Justen et al. (2024). First, we extracted DNA following a standard phenol-chloroform protocol, prepared libraries for Illumina whole genome resequencing, and sequenced them to low coverage (average 4.4×). Next, we trimmed the resulting sequences for quality using trim galore (- - clip_R1 15 - - clip_R2 15 - - three_prime_clip_R1 5 - - three_prime_clip_R2 5; https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/), aligned them to reference genomes from both subspecies of thrush using bwa mem (Li & Durbin, 2009), removed reads that mapped to low quality using samtools (Li et al., 2009), extracted reads that mapped uniquely to both references using samtools, found the intersection of these reads using ngsutilj (https://github.com/compgen-io/ngsutilsj) and limited all subsequent analyses the resulting reads and their positions on the inland reference genome. Finally, we used bcftools (Danecek et al., 2021; Li, 2011) to call an initial set of SNPs (--min-BQ 20, --min-483 MQ 20, %QUAL>500, --skip-variants indels) and imputed genotypes using the hidden Markov model in STITCH (Davies et al., 2016). Imputation requires large numbers of individuals so more than just the birds used in the present study were used in this step (n=1310 thrushes from various projects).

Genomic ancestry was estimated as described in Blain et al. (2024). First, we filtered the dataset to remove SNPs with minor allele frequencies less than 5%, missing data greater than 25%, not in Hardy–Weinberg equilibrium (0.0001), and on the Z and W chromosomes using vcftools (Danecek et al., 2011). We also used applied linkage pruning using plink (50 10 0.1). Next, we identified SNPs that distinguish coastal and inland thrushes using vcftools to estimate Weir-Cockerham’s FST between ten coastal and eight inland birds captured in pure populations adjacent to the hybrid zone. We considered sites with FST values greater than 0.90 divergent between the subspecies and estimate allele frequencies within these parental populations for these loci. Finally, we estimated genomic ancestry with the HIest R package, using allele counts estimated from the imputed genotypes (Fitzpatrick, 2012). Ancestry is an estimate of the proportion of inland alleles, ranging from zero to one. A coastal subspecies bird would have an ancestry of 0, an F1 hybrid would be expected to have an ancestry of 0.5, and an inland subspecies bird would be expected to have an ancestry of 1 (Fig. S1).

Survival

We estimated survival using Motus radio station detections of experimental birds. Detection data were downloaded from the Motus website (www.motus.org) on August 19, 2024. Following recommendations in Motus documentation to retain only high quality detections (Birds Canada, 2024), we filtered the data to remove unreliable detections; those with run lengths ≤ 3 (runs are groups of consecutive detections; short runs often reflect false positive detections, especially around receivers with noise. This filtering step reduced the raw dataset of 38,443 unique detections (with detections of the same tag, on the same day, by the same Motus station grouped together) to a set of 2048 unique detections. We additionally removed detections from five stations in northeastern North America with several faulty detections (LacEdouard-Champs, Lambs Gap, Lockoff, Tadoussac – Andr?, and Triton2). We also excluded all detections east of 70°W (east of the Atlantic coast), detections below 20°N between June and August (south of the continental US during the summer), and above 35°N in December to February (northern US or Canada during the winter), resulting in 1771 unique detections. We then mapped each bird’s detection history to manually identify and remove detections that were biologically improbable. Detections were considered biologically improbable if they required significant backtracking (ex. a detection in Ontario in the fall following a series of detections in the Mississippi River basin and Costa Rica) or route deviations (ex. a detection in Florida between a set of detections in British Columbia). This filtering stage was performed blind to any identifying information about the bird, such as genomic ancestry or release site and resulted in a final set of 1709 unique detections.

To model apparent survival using capture-mark-recapture models, we constructed capture histories by coding the birds as detected (1) or not detected (0) in every period of ten days, for the first 300 days after the bird was tagged. Birds detected after the 300 day period were coded as 1 on the final day. We chose sets of ten days as the unit of time because the data were very sparse with each day as the unit of time, but ten days is short enough to capture birds in different migratory states (i.e., migratory, overwintering, breeding). Treating ten days as the detection period additionally reduces the potential for problems associated with the clustered distribution of Motus, where a bird travelling through a region with a high density of towers could be detected repeatedly over a few days (Cooper et al., 2024). For a second, simplified, set of analyses, we produced a binary survival variable where birds were considered ‘survived’ if they were detected above 40°N in the spring or at any time point the year after the primary 300 day sampling period.

Migratory behaviour

We estimated the direction of a bird’s travel in the beginning of fall migration (migratory orientation). This estimate of fall orientation was measured relative to the bird’s release site. We retained the first detection of each bird, within 300 km of the focal location – a radius that includes the full extent of the stations that were established for the Swainson’s thrush specifically, spanning the entire width of the hybrid zone in British Columbia (Delmore & Easton, 2019). We estimated distances between points with distm from the package geosphere (Hijmans et al., 2024). We then calculated the orientation to the detection location relative to the release site with bearingRhumb. The resulting orientations were measured clockwise relative to a northward direction, resulting in lower orientation values for eastward than westward trajectories and intermediate values for a southward direction.

Statistical analysis

We fit Cormack-Jolly-Seber (CJS) models, a form of capture-mark-recapture model that jointly estimates the contributions of ancestry and release site to both detection probability and apparent survival. Modelling detection probability alongside survival is necessary to account for potential variation across years and migration routes in the density of Motus radio stations and resulting probability of detection, which could otherwise appear to drive patterns of differing survival. We implemented CJS models with crm from the marked package, using the ‘CJS’ model and the hessian method (Laake et al., 2013). First, we fit a CJS model including only experimental, translocated birds, to test for the predicted interaction effect between genomic ancestry and release site on survival. We included time, ancestry, and release site as predictors for detection probability, as these variables might affect the number of radio stations encountered by migrating birds, and ancestry, release site, and their interaction as predictors for apparent survival.

We then fit a second CJS model including both translocated birds and individuals that were not translocated and instead captured and released within the hybrid zone, to determine whether relationships between survival and ancestry in translocated birds differed from those in their original environment. This model included time, ancestry, release site, and release year as predictors for detection probability and ancestry, release site, and their interaction as predictors for survival. Release year is included only in this model because birds were caught and released in the hybrid zone in both translocation years. We additionally used the birds captured and released in Pemberton to evaluate whether any effects of ancestry on survival depended on release year. To do this, we fit a CJS model including only the Pemberton birds with time, ancestry, and release year as predictors of detection probability and ancestry, release year, and the interaction between ancestry and release year as predictors of apparent survival.

We used logistic regression models as a second method with reduced model complexity, relative to the CJS models, to evaluate effects of ancestry, release site, and their interaction on apparent survival. Using only birds that were translocated, we first fit a logit-link generalized linear model with a binary survival response, and ancestry, release sites, and their interaction as predictors. We then fit a second model, this time included birds that were both tagged and released in Pemberton, with a binary survival response, ancestry, release sites, and their interaction as fixed effects, and release year as a random effect.

We evaluated the effect of translocation on migratory behaviour by fitting linear models with orientation in the beginning of fall migration as the response, and release site, genomic ancestry, and their interaction as predictors. First, we fit a linear model only including birds that were translocated to the coastal or inland range. Next, we fit a mixed-effect linear model that additionally included birds that were captured and released within the hybrid zone in the experiment years (Bates et al., 2015; Kuznetsova et al., 2017). In these models, orientation was the response, release site, genomic ancestry, and their interaction were fixed effects, and release year was a random effect. To evaluate whether relationships between ancestry and orientation depended on release year, we fit an additional linear model only including birds that were captured and released in Pemberton, with release year, genomic ancestry, and their interaction as predictors.

Results

Survival

We expected that birds released in the inland range (Lillooet) with more inland ancestry and birds released in the coastal range (Squamish) with more coastal ancestry would have higher relative survival. However, we found no evidence for this predicted interactive effect of release site and genomic ancestry. In a test that included only the translocated birds, we found no effect of release site (φ = 1.19, 95% Confidence Limit = [−2.66, 5.04]), genomic ancestry (φ = 3.55 [−1.55, 8.64]; Fig. 2), or the interaction between release site and ancestry (φ = −1.39 [−7.32, 4.54]) on apparent survival. Detection probabilities were higher for birds released in Squamish than Lillooet (p = 2.81 [1.94, 3.67]) but did not depend on ancestry (p = −0.33 [−2.52, 1.86]). These results were consistent with those from our logistic regression model, which was run with survival as a binary response variable as an alternative model with lower complexity, with the exception that survival was marginally higher in Squamish than Lillooet because differences in detection probabilities were not accounted for (genomic ancestry: χ21 = 1.55, p = 0.21; release site: χ21 = 2.92, p = 0.09; ancestry × release site: χ21 = 0.86, p = 0.35).

Figure 2.

Figure 2.

Open in a new tab

Effect of ancestry on survival did not depend on release site. Each point represents one bird, with shading indicating whether the bird was detected above 40°N in the spring, indicating survival on migration. Birds that were undetected from that time onwards were assumed dead. Ancestry values of 0 indicate fully coastal genotypes while ancestries of 1 indicate fully inland genotypes. Left panels show expectations under different hypotheses. If selection is ecologically dependent (top panel), survival (shading or not) should depend on release site (x-axis) and be higher for birds that have ancestry similar to the range they have been transplanted to. For example, when birds are transplanted to the inland subspecies range (green), those with higher ancestry should survive at higher rates. The opposite should be true for those with lower ancestry values, which should survive better when transplanted into the coastal range. If selection is ecologically independent (bottom panel), survival should not depend on release site.

We repeated this analysis including birds tagged within the hybrid zone (Pemberton), to test whether translocation altered relationships between survival and ancestry that were observed in the hybrids’ original environment. We expected no linear relationship between ancestry and survival in Pemberton, a positive relationship in Lillooet, and a negative relationship in Squamish. Instead, we found no evidence that survival depended on an interaction between release site and ancestry. Once again, survival did not depend on genomic ancestry (φ = 0.0 9 [−0.98, 1.18]; Fig. 2). Relative to Pemberton, there was neither an effect of release in Lillooet (φ = −2.05 [−4.76, 0.66]) or Squamish (φ = −0.95 [−3.43, 1.53]), nor an interaction between release in Lillooet and ancestry (φ = 2.71 [−1.32, 6.74]) or an interaction between release in Squamish and ancestry (φ = 1.61 [−2.03, 5.26]; Fig. 2). Detection probability relative to Pemberton was reduced in Lillooet (p = −1.09 [−1.80, −0.38]) and marginally higher in Squamish (p = 0.35 [0.02, 0.67]). Detection probability was also higher in 2023 compared to 2020 (p = 1.14 [0.86, 1.42]) but did not depend on ancestry (p = 0.25 [−0.49, 0.98]). The results of the logistic regression models were, again, consistent with those from the CJS models (genomic ancestry: χ21 = 1.10, p = 0.29; release site: χ21 = 0.75, p = 0.69; ancestry × release site: χ21 = 1.14, p = 0.57).

To explore potential effects of the different release years, we tested for an interaction between release year and ancestry in an additional CJS model using only the Pemberton birds. We did not find any evidence for effects on apparent survival (release year: φ = −0.23 [−1.52, 1.05], ancestry: φ = −0.24 [−1.80, 1.32], release year x ancestry: φ = 1.03 [−0.96, 3.02]). Detection probability differed between years (p = 1.32 [1.05, 1.58]) but did not depend on ancestry (p = 0.37 [−0.34, 1.09]).

Migratory behaviour

We tested whether orientation in the beginning of migration depends on genetics, translocation environment (release site), or an interaction between genetic and environmental effects. If there were an environmental effect on orientation, we expected birds to have more eastward initial orientations from Lillooet relative to from Squamish in order to match the routes of birds that migrate through those locations. However, we did not find this effect; for birds that were translocated into either the coastal range (Squamish) or the inland range (Lillooet), orientation relative to release site did not depend on where they were translocated (F1, 31 = 1.90, p = 0.18; Fig. S2A). If there were genetic effects on initial orientation in the translocated birds, we would expect to see a negative relationship between orientation and genomic ancestry so that birds with more inland genotypes started migration heading in a more eastward direction. We found that orientation did not depend on genomic ancestry (F1, 31 = 2.56, p = 0.12). However, we did observe a marginally non-significant interaction between release site and genomic ancestry, with a stronger negative relationship between orientation and ancestry for birds released in Lillooet than those released in Squamish (F1, 31 = 3.00, p = 0.09).

We repeated this analysis including birds that were released where they were captured in Pemberton, to determine whether migratory behaviour of translocated birds differed from migratory behaviour of birds released within the hybrid zone. Again, orientation did not vary among release sites (χ22 = 2.37, p = 0.31; Fig. 3). When we included birds released within the hybrid zone, orientation did depend on genomic ancestry (χ21 = 43.88, p < 0.01) and the strength of this effect varied among release sites (χ22 = 7.27, p < 0.01). On average, birds with more inland genomic ancestry started migration following more eastward orientations than those with more coastal ancestry; this trend was stronger in Pemberton than in Lillooet and absent in Squamish (model fitted slopes: Squamish = 0.45, Pemberton = −2.89, Lillooet = −3.39; Fig. 3, S2B). In an additional model including only Pemberton birds, we found that orientation differed between release years (F1, 132 = 4.77, p = 0.03) and depended on genomic ancestry (F1, 132 = 45.48, p = 0.01), but there was no interaction between release year and ancestry (F1, 132 = 2.29, p = 0.13).

Figure 3.

Figure 3.

Open in a new tab

Orientation in the beginning of migration varied among genomic ancestries but not release sites. (A) Trajectories from release site to first detection, within 300 km of the release site. More westward trajectories correspond to higher orientation values while more eastward trajectories result in lower estimates for orientation. For ancestry, values closer to 0 indicate a more coastal genotype while values closer to 1 represent a more inland genotype. (B) Expected relationships between release site and orientation under different hypotheses. Note that the pictured relationships represent one example of what we could expect to observe within our system; a wider diversity of parallel relationships could indicate genetic or environmental factors while any non-parallel interaction norms could indicate genetic by environmental effects. (C) Model-fitted results for relationships between release site and orientation, by ancestry. Points represent the model-predicted mean for a given ancestry value and error bars indicating 95% confidence intervals, estimated using the visreg R package (Breheny & Burchett, 2017). Orientation is a normalized estimate of each bird’s direction in the beginning of migration, measured relative to the site where the bird was released. Higher values indicate a more westward orientation, and lower values indicate a more eastward orientation. For ancestry, ‘coastal backcross’ represents a genomic ancestry of 0.25, ‘hybrid’ represents an ancestry of 0.5, and ‘inland backcross’ represents an ancestry of 0.75.

Comparing full detection histories, birds that were translocated appear to have followed routes consistent with expectations given their migratory genotype (Fig. 4). For example, although detection histories for birds released in Lillooet are relatively sparse, given that the Motus network was less developed in 2020, a bird with mostly coastal ancestry was detected in western Mexico, aligning with the route expected given its ancestry (Delmore et al., 2012). Similarly, birds with mostly inland ancestry that were released in Squamish were primarily detected in central and eastern North America, following routes expected for the inland subspecies.

Figure 4.

Figure 4.

Open in a new tab

Mapping individual migration paths. Each point represents one detection, and lines connect detections, in order, for each bird. Colour indicates genomic ancestry, with coastal = 0 and inland = 1.

Discussion

Through a translocation experiment, we tested for divergent ecological adaptation of fall migration routes as a cause of reproductive isolation between incipient songbird species and evaluated the relative contributions of genetics and the environment in determining migratory behaviour. We suggest that selection against hybrid Swainson’s thrushes may not follow traditional models for ecological speciation but may nonetheless be ecological in nature. Furthermore, we argue that hybrid migratory behaviour may be influenced by elements of both the environment and genetics.

No support for traditional models of ecological selection against hybrids

In contrast to expectations under the ecological speciation hypothesis (Rundle & Whitlock, 2001), we found no effect of translocation on ancestry-specific survival. This finding could suggest that ecology does not play a role in maintaining reproductive isolation between thrushes. We propose two alternatives to this suggestion. First, it is possible that a more nuanced view of ecological selection against hybrids is needed in this case. Previous results, including evidence of highly divergent coastal and inland migration routes, combined with reduced hybrid survival on migration, suggested that reproductive isolation is ecologically dependent in Swainson’s thrushes (Blain et al., 2024; Delmore et al., 2012). Ecology can affect hybrid fitness via an alternative mechanism if hybrids experience some form of intrinsic dysfunction that is exacerbated on migration. Intrinsic incompatibilities can be environmentally dependent, so that they are only observable when hybrids are under ecological stress (Miller & Matute, 2017). Under this scenario, selection against hybrids is still ecological in nature but does not derive directly from ecological divergence between parental forms (Kulmuni & Westram, 2017), meaning that we would not expect to find the relationship between genomic ancestry and survival predicted under the ecological divergence hypothesis.

The suggestion that intrinsic incompatibilities exacerbated by the environment contribute to reproductive isolation is supported by previous work on the Swainson’s thrush; a comparative transcriptomic study showed that although hybrids survive well in a laboratory environment, they exhibit considerable misexpression (under or overexpression compared to parental forms) and these patterns are more common on migration (Louder et al., 2024). Misexpression in hybrids can evolve rapidly and reflect underlying epistatic incompatibilities (Mack & Nachman, 2017; Satokangas et al., 2020). Long distance migratory birds generally exhibit higher mortality on migration than during other life history stages (Newton, 2024; Rushing et al., 2017) and birds are under substantial physiological stress during this time (Jenni-Eiermann et al., 2014). The stress and risks associated with migration may cause hybrid dysfunction to manifest as elevated mortality during migration even if they are able to survive well during the overwintering or breeding seasons.

Second, it is possible that ecological divergence between the Swainson’s thrush subspecies does contribute to reproductive isolation, but not at the beginning of fall migration. We observed several birds following migratory routes that are consistent with expectations given their genomic ancestry, indicating that translocation did not cause the birds to complete migration in the other species’ environment. Early in fall migration, Swainson’s thrushes may have sufficient energy to course correct when translocated outside of their range. There are several other locations during migration and on their annual cycle where selection may act against hybrids. For example, migratory paths may matter the most when they first hit specific ecological barriers, such as southwestern deserts or the Gulf of Mexico (Justen et al., 2021). Where they orient when returning to the breeding grounds from spring migration may be under stronger selection, as birds tend to complete spring migration at a faster pace and will have depleted fat stores and stamina at this stage (McWilliams et al., 2004; Nilsson et al., 2013).

Alternatively, the ecological gradient may be more important for breeding than migration. Despite not being geographically far from each other, Squamish, in the coastal range, is a temperate rainforest bordering an ocean inlet, while Lillooet, in the inland range, is semi-arid and east of the Coast Mountains (Ruegg et al., 2006). This ecological transition coincides with a narrow geographical cline in the Swainson’s thrush and could shape selection on nesting and foraging during the breeding season (Delmore & Irwin, 2014; Ruegg et al., 2006; Ruegg et al., 2006). The hybrid zone could instead be settled over the Coast Mountains if this is a region of low population density, where both subspecies experience reduced survival and reproduction (Barton & Hewitt, 1985). In this scenario, ecological selection against both subspecies could maintain the hybrid zone without directly involving divergent ecological adaptation to alternate migration routes.

It is critical to point out that it is also possible that we did not detect the expected pattern because we did not have the power to do so. This was a unique experiment design (reciprocal transplants are rarely attempted with natural populations of large vertebrates) and may not have included the sample sizes necessary to capture effects of translocation on survival. In addition, stochasticity and ecological processes unrelated to ancestry also affect survival of migrating songbirds (McKellar et al., 2015; Rockwell et al., 2017); these may have introduced variation that eliminated a signal of ancestry by release site effects. However, the logistic regression model quantifying survival as binary variable should require less power than the mark recapture model and still recovered same results. Another caveat in interpreting the results is that the translocations to Pemberton and Lillooet occurred in different years. Migratory survival can vary among years, and could conceivably do so in a way that varies among ancestries, which could have had a confounding effect on our results. Additionally, estimating survival using the Motus network can be challenging, given that the distribution of Motus radio stations is uneven across North America and changed between years of our study (i.e. higher density in 2023; Taylor et al., 2017). We took steps in our analysis to account for this, including grouping detections in each ten day period to prevent overestimation of survival for birds repeatedly detected over a short time when passing near a cluster of towers (Cooper et al., 2024). While we can incorporate differences in detection probabilities between years and release sites in capture mark recapture models, if there is an interaction between the variables of interest (here, ancestry by release site relationships) and detection probability, this cannot be fully accounted for. However, we do not believe tower distributions can account for the observed patterns in our study, as the relationship between ancestry and survival did not change with release year or release site.

Genetic and environmental variation underlie variation in migratory behaviour

Relationships among genotypes, translocation environment, and migratory behaviour were consistent with the interpretation that genetic variation interacts with the environment to predict migratory behavior. Migratory orientations varied across environments, with birds translocated inland starting migration along a more eastward route than those translocated to the coast. Birds also oriented as expected based on their genotypes, with birds exhibiting more inland genomic ancestry starting migration along more eastern routes than those with more coastal ancestry, with a stronger relationship observed for birds translocated east to Lillooet than those translocated west to Squamish. It is important to note that the observed relationship to ancestry was limited to when birds released in the hybrid zone were included in the analysis. Accordingly, it is possible that this result relates to limited power when these birds were not included. Future work increasing the sample size of translocated birds is essential to evaluating this suggestion. We might see a greater effect of the environment in Lillooet and Pemberton in our present study because the birds have more options in terms of orientation there. Squamish is close enough to the ocean that birds following more coastal routes are forced to orient south even if their inherent orientations would push them further west, while birds released in Lillooet have space to begin migrating with a wider range of potential orientations.

Relationships we observed here between genomic background and orientation within the hybrid zone, alongside high heritability of migratory orientation (Justen et al., 2024), support a role for genetics in determining how Swainson’s thrushes orient on migration. Migratory behaviours in other songbirds have been found to be genetically determined (Liedvogel et al., 2011; Merlin & Liedvogel, 2019), with lab experiments on European blackcaps demonstrating a genetic basis for migratory orientation (Helbig, 1991), while quantitative genetic approaches applied to a set of songbirds suggest that migratory behaviours are highly heritable and that heritability is consistent across environments (Pulido & Berthold, 2003). Still, other evidence supports a role for environmental effects on migratory behaviours in songbirds, including timing of fall migration and time spent at stopover sites (Calvert et al., 2012; Schmaljohann et al., 2012; Smith & McWilliams, 2014).

Despite variation in orientations in the beginning of migration, birds that survived migration were generally detected along routes that would be expected given their genotype. While we did not have a sufficient density of data to statistically test for variation among individual routes, it does appear that translocated birds are ultimately able to course-correct. This aligns with evidence that birds can navigate on migration from translocation experiments with adult cuckoos, another species of long distance, nocturnal migrants (Willemoes et al., 2015). In southward translocations of European starlings, juveniles did not adjust their position on fall migration but did orient correctly on spring migration (Perdeck, 1958; Piersma et al., 2020). Navigation mechanisms are not entirely clear, but there is evidence that songbirds integrate magnetic, polarized light, and olfactory cues (Chernetsov et al., 2017; Holland et al., 2009; Muheim et al., 2006). For birds in our experiment that survived fall migration, learning could contribute to correct positioning on their return or on subsequent years. Considered together, our migratory behaviour results suggest that both genetic variation and the environment affect migratory behaviour; there is a genetic basis to this behaviour that is mediated by environmental variables they encounter on migration.

Conclusions

Overall, we did not find support for the hypothesis that divergent ecological adaptation to fall migration routes contributes to reproductive isolation between coastal and inland Swainson’s thrushes. While this does not eliminate a role for ecology in their speciation process, as ecological selection on hybrids may still contribute to isolation, these results suggest that Swainson’s thrushes may not conform to the classic model of ecological speciation. Orientation in the beginning of fall migration was determined by an interaction between release site and genomic ancestry, suggesting that both the environment and genetics influence migratory behaviour. Qualitatively, translocated birds ultimately followed routes consistent with expectations given their genomic ancestry, which may account for the limited effect of translocation on survival but ultimately supports the large body of research suggesting that variation in migratory behaviour is mediated by both environmental and genetic effects.

Supplementary Material

supplementary material

Acknowledgements

Funding for this research came from an NSF CAREER (IOS-2143004) and an NIH MIRA (1R35GM151012), both awarded to KED. We greatly appreciate assistance in the field from members of the Delmore lab (especially Hayley Madden, Scarlet Byron, Eilean McCutchon, and Himani Russell). We thank landowners who facilitate our fence of Motus stations (project #280) in British Columbia and all collaborators contributing to the Motus Wildlife Tracking system. We also thank two anonymous reviewers for suggestions that substantially improved the manuscript.

Footnotes

Conflict of Interest

The authors declare no competing interests.

Data Accessibility

Data included in this manuscript, including Motus detections and release/capture details for each individual, have been uploaded to Dryad (DOI: 10.5061/dryad.83bk3jb4d; link: datadryad.org/dataset/doi:10.5061/dryad.83bk3jb4d). Scripts for both bioinformatics and downstream analysis have additionally been uploaded to Dryad and are available on GitHub (bioinformatics: github.com/stephblain/thrush_hybrid_survival, analysis: github.com/stephblain/SWTH_translocation). All sequences have been uploaded to the NCBI Sequence Read Archive under BioProject number PRJNA979932.

References

  1. Åkesson S (2003). Avian Long-Distance Navigation: Experiments with Migratory Birds. In Berthold P, Gwinner E, & Sonnenschein E (Eds.), Avian Migration (pp. 471–492). Springer Berlin; Heidelberg. 10.1007/978-3-662-05957-9_33 [DOI] [Google Scholar]
  2. Alerstam T, Hedenström A, & Åkesson S (2003). Long-distance migration: Evolution and determinants. Oikos, 103(2), 247–260. 10.1034/j.1600-0706.2003.12559.x [DOI] [Google Scholar]
  3. Barton NH, & Hewitt GM (1985). Analysis of hybrid zones. Annual Review of Ecology and Systematics. Vol. 16, 16, 113–148. 10.1146/annurev.es.16.110185.000553 [DOI] [Google Scholar]
  4. Bates D, Mächler M, Bolker BM, & Walker SC (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48. 10.18637/jss.v067.i01 [DOI] [Google Scholar]
  5. Birds Canada. (2024). motus: Fetch and use data from the Motus Wildlife Tracking System (Version 6.1.1) [Computer software]. http://motusWTS.github.io/motus/ [Google Scholar]
  6. Blain SA, Justen HC, Easton WE, & Delmore KE (2024). Reduced hybrid survival in a migratory divide between songbirds. Ecology Letters, 27, e14420. [DOI] [PubMed] [Google Scholar]
  7. Breheny P, & Burchett W (2017). Visualization of Regression Models Using visreg. The R Journal, 9(2), 56. 10.32614/RJ-2017-046 [DOI] [Google Scholar]
  8. Calvert AM, Mackenzie SA, Flemming JM, Taylor PD, & Walde SJ (2012). Variation in songbird migratory behavior offers clues about adaptability to environmental change. Oecologia, 168(3), 849–861. 10.1007/s00442-011-2119-5 [DOI] [PubMed] [Google Scholar]
  9. Chernetsov N, Pakhomov A, Kobylkov D, Kishkinev D, Holland RA, & Mouritsen H (2017). Migratory Eurasian Reed Warblers Can Use Magnetic Declination to Solve the Longitude Problem. Current Biology, 27(17), 2647–2651.e2. 10.1016/j.cub.2017.07.024 [DOI] [PubMed] [Google Scholar]
  10. Cooper NW, Yanco SW, Rushing CS, Sillett TS, & Marra PP (2024). Non-breeding conditions induce carry-over effects on survival of migratory birds. Current Biology, 34(21), 5097–5103.e3. 10.1016/j.cub.2024.09.015 [DOI] [PubMed] [Google Scholar]
  11. Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA, Handsaker RE, Lunter G, Marth GT, Sherry ST, Mcvean G, Durbin R, & 1000 Genomes Project Analysis Group. (2011). The variant call format and VCFtools. Bioinformatics, 27(15), 2156–2158. 10.1093/bioinformatics/btr330 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Danecek P, Bonfield JK, Liddle J, Marshall J, Ohan V, Pollard MO, Whitwham A, Keane T, McCarthy SA, & Davies RM (2021). Twelve years of SAMtools and BCFtools. GigaScience, 10(2), 1–4. 10.1093/gigascience/giab008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Delmore KE, & Easton W (2019). BCInterior Thrushes (Project280). Motus Wildlife Tracking System, Birds Canada. https://motus.org/ [Google Scholar]
  14. Delmore KE, Fox JW, & Irwin DE (2012). Dramatic intraspecific differences in migratory routes, stopover sites and wintering areas, revealed using light-level geolocators. Proceedings of the Royal Society B: Biological Sciences, 279(1747), 4582–4589. 10.1098/rspb.2012.1229 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Delmore KE, & Irwin DE (2014). Hybrid songbirds employ intermediate routes in a migratory divide. Ecology Letters, 17(10), 1211–1218. 10.1111/ele.12326 [DOI] [PubMed] [Google Scholar]
  16. Fitzpatrick BM (2012). Estimating ancestry and heterozygosity of hybrids using molecular markers. BMC Evolutionary Biology, 12(131), 1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Funk DJ, Nosil P, & Etges WJ (2006). Ecological divergence exhibits consistently positive associations with reproductive isolation across disparate taxa. Proceedings of the National Academy of Sciences of the United States of America, 103(9), 3209–3213. 10.1073/pnas.0508653103 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Helbig AJ (1991). Inheritance of migratory direction in a bird species: A cross-breeding experiment with SE- and SW-migrating blackcaps (Sylvia atricapilla). Behavioral Ecology and Sociobiology, 28, 9–12. [Google Scholar]
  19. Hijmans RJ, Karney C, Williams E, & Vennes C (2024). geosphere: Spherical Trigonometry (Version R package version 1.5–20) [Computer software]. [Google Scholar]
  20. Holland RA, Thorup K, Gagliardo A, Bisson IA, Knecht E, Mizrahi D, & Wikelski M (2009). Testing the role of sensory systems in the migratory heading of a songbird. Journal of Experimental Biology, 212(24), 4065–4071. 10.1242/jeb.034504 [DOI] [PubMed] [Google Scholar]
  21. Irwin DE, & Irwin JH (2005). Siberian migratory divides: The role of seasonal migration in speciation. In Greenberg R & Marra PP (Eds.), Birds of Two Worlds: The Ecology and Evolution of Migration (pp. 27–40). Johns Hopkins University Press. [Google Scholar]
  22. Jenni-Eiermann S, Jenni L, Smith S, & Costantini D (2014). Oxidative Stress in Endurance Flight: An Unconsidered Factor in Bird Migration. PLoS ONE, 9(5), e97650. 10.1371/journal.pone.0097650 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Justen HC, Easton WE, & Delmore KE (2024). Mapping seasonal migration in a songbird hybrid zone—Heritability, genetic correlations, and genomic patterns linked to speciation. Proceedings of the National Academy of Sciences, 121(18), e2313442121. 10.1073/pnas. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Justen H, Lee-Yaw JA, & Delmore KE (2021). Reduced habitat suitability and landscape connectivity in a songbird migratory divide. Global Ecology and Biogeography, 30, 2043–2056. 10.1111/geb.13367 [DOI] [Google Scholar]
  25. Kulmuni J, & Westram AM (2017). Intrinsic incompatibilities evolving as a by-product of divergent ecological selection: Considering them in empirical studies on divergence with gene flow. Molecular Ecology, 26(12), 3093–3103. 10.1111/mec.14147 [DOI] [PubMed] [Google Scholar]
  26. Kuznetsova A, Brockhoff PB, & Christensen RHB (2017). lmerTest Package: Tests in linear mixed effects models. Journal of Statistical Software, 82(13). 10.18637/jss.v082.i13 [DOI] [Google Scholar]
  27. Laake JL, Johnson DS, & Conn PB (2013). marked: An R package for maximum likelihood and Markov Chain Monte Carlo analysis of capture-recapture data. Methods in Ecology and Evolution, 4(9), 885–890. 10.1111/2041-210X.12065 [DOI] [Google Scholar]
  28. Li H (2011). A statistical framework for SNP calling, mutation discovery, association mapping and population genetical parameter estimation from sequencing data. Bioinformatics, 27(21), 2987–2993. 10.1093/bioinformatics/btr509 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Li H, & Durbin R (2009). Fast and accurate short read alignment with Burrows – Wheeler transform. Bioinformatics, 25(14), 1754–1760. 10.1093/bioinformatics/btp324 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R, & 1000 Genome Project Data Processing Subgroup. (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics, 25(16), 2078–2079. 10.1093/bioinformatics/btp352 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Liedvogel M, Åkesson S, & Bensch S (2011). The genetics of migration on the move. Trends in Ecology and Evolution, 26(11), 561–569. 10.1016/jrtree.2011.07.009 [DOI] [PubMed] [Google Scholar]
  32. Louder MIM, Justen H, Kimmitt AA, Lawley KS, Turner LM, Dickman JD, & Delmore KE (2024). Gene regulation and speciation in a migratory divide between songbirds. Nature Communications, 15(98), 1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mack KL, & Nachman MW (2017). Gene Regulation and Speciation. Trends in Genetics, 33(1), 68–80. 10.1016/j.tig.2016.11.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. McKellar AE, Reudink MW, Marra PP, Ratcliffe LM, & Wilson S (2015). Climate and density influence annual survival and movement in a migratory songbird. Ecology and Evolution, 5(24), 5892–5904. 10.1002/ece3.1854 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. McWilliams SR, Guglielmo C, Pierce B, & Klaassen M (2004). Flying, fasting, and feeding in birds during migration: A nutritional and physiological ecology perspective. Journal of Avian Biology, 35(5), 377–393. 10.1111/j.0908-8857.2004.03378.x [DOI] [Google Scholar]
  36. Merlin C, & Liedvogel M (2019). The genetics and epigenetics of animal migration and orientation: Birds, butterflies and beyond. Journal of Experimental Biology, 222(Suppl_1), jeb191890. 10.1242/jeb.191890 [DOI] [PubMed] [Google Scholar]
  37. Miller CJJ, & Matute DR (2017). The Effect of Temperature on Drosophila Hybrid Fitness. G3 Genes\Genomes\Genetics, 7(2), 377–385. 10.1534/g3.116.034926 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Muheim R, Phillips JB, & Åkesson S (2006). Polarized Light Cues Underlie Compass Calibration in Migratory Songbirds. Science, 313(5788), 837–839. 10.1126/science.1129709 [DOI] [PubMed] [Google Scholar]
  39. Newton I (2024). Migration mortality in birds. Ibis, ibi.13316. 10.1111/ibi.13316 [DOI] [Google Scholar]
  40. Nilsson C, Klaassen RHG, & Alerstam T (2013). Differences in Speed and Duration of Bird Migration between Spring and Autumn. The American Naturalist, 181(6), 837–845. 10.1086/670335 [DOI] [PubMed] [Google Scholar]
  41. Perdeck AC (1958). Two types of orientation in migrating starlings, Sturnus vulgaris L., and Chaffinches, Fringilla coelebs L., as revealed by displacement experiments. Ardea, 46, 37. [Google Scholar]
  42. Piersma T, Loonstra AHJ, Verhoeven MA, & Oudman T (2020). Rethinking classic starling displacement experiments: Evidence for innate or for learned migratory directions? Journal of Avian Biology, 51(5), jav.02337. 10.1111/jav.02337 [DOI] [Google Scholar]
  43. Pulido F, & Berthold P (2003). Quantitative Genetic Analysis of Migratory Behaviour. In Berthold P, Gwinner E, & Sonnenschein E (Eds.), Avian Migration (pp. 53–77). Springer Berlin; Heidelberg. 10.1007/978-3-662-05957-9_4 [DOI] [Google Scholar]
  44. Rockwell SM, Wunderle JM, Sillett TS, Bocetti CI, Ewert DN, Currie D, White JD, & Marra PP (2017). Seasonal survival estimation for a long-distance migratory bird and the influence of winter precipitation. Oecologia, 183(3), 715–726. 10.1007/s00442-016-3788-x [DOI] [PubMed] [Google Scholar]
  45. Ruegg K (2008). Genetic, morphological, and ecological characterization of a hybrid zone that spans a migratory divide. Evolution, 62(2), 452–466. 10.1111/j.1558-5646.2007.00263.x [DOI] [PubMed] [Google Scholar]
  46. Ruegg KC, Hijmans RJ, & Moritz C (2006). Climate change and the origin of migratory pathways in the Swainson’s thrush, Catharsus ustulatus. Journal of Biogeography, 33, 1172–1182. 10.1111/j.1365-2699.2006.01517.x [DOI] [Google Scholar]
  47. Ruegg KC, & Smith TB (2002). Not as the crow flies: A historical explanation for circuitous migration in Swainson’s thrush ( Catharus ustulatus ). Proceedings of the Royal Society of London. Series B: Biological Sciences, 269(1498), 1375–1381. 10.1098/rspb.2002.2032 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Ruegg K, Slabbekoorn H, Clegg S, & Smith TB (2006). Divergence in mating signals correlates with ecological variation in the migratory songbird, Swainson’s thrush (Catharus ustulatus). Molecular Ecology, 15(11), 3147–3156. 10.1111/j.1365-294X.2006.03011.x [DOI] [PubMed] [Google Scholar]
  49. Rundle HD, & Nosil P (2005). Ecological speciation. Ecology Letters, 8, 336–352. 10.1111/j.1461-0248.2004.00715.x [DOI] [Google Scholar]
  50. Rundle HD, & Whitlock MC (2001). A genetic interpretation of ecologically dependent isolation. Evolution, 55(1), 198–201. [DOI] [PubMed] [Google Scholar]
  51. Rushing CS, Hostetler JA, Sillett TS, Marra PP, Rotenberg JA, & Ryder TB (2017). Spatial and temporal drivers of avian population dynamics across the annual cycle. Ecology, 98(11), 2837–2850. 10.1002/ecy.1967 [DOI] [PubMed] [Google Scholar]
  52. Satokangas I, Martin SH, Helanterä H, Saramäki J, & Kulmuni J (2020). Multi-locus interactions and the build-up of reproductive isolation. Philosophical Transactions of the Royal Society B, 375, 20190543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Schluter D (2009). Evidence for ecological speciation and its alternative. Science, 323(5915), 737–741. 10.1126/science.1160006 [DOI] [PubMed] [Google Scholar]
  54. Schmaljohann H, Fox JW, & Bairlein F (2012). Phenotypic response to environmental cues, orientation and migration costs in songbirds flying halfway around the world. Animal Behaviour, 84(3) 623–640. 10.1016/j.anbehav.2012.06.018 [DOI] [Google Scholar]
  55. Scordato ESC, Smith CCR, Semenov GA, Liu Y, Wilkins MR, Liang W, Rubtsov A, Sundev G, Koyama K, Turbek SP, Wunder MB, Stricker CA, & Safran RJ (2020). Migratory divides coincide with reproductive barriers across replicated avian hybrid zones above the Tibetan Plateau. Ecology Letters, 23, 231–241. 10.1111/ele.13420 [DOI] [PubMed] [Google Scholar]
  56. Smith AD, & McWilliams SR (2014). What to do when stopping over: Behavioral decisions of a migrating songbird during stopover are dictated by initial change in their body condition and mediated by key environmental conditions. Behavioral Ecology, 25(6), 1423–1435. 10.1093/beheco/aru148 [DOI] [Google Scholar]
  57. Taylor PD, Crewe TL, Mackenzie SA, Lepage D, Aubry Y, Crysler Z, Finney G, Francis CM, Guglielmo CG, Hamilton DJ, Holberton RL, Loring PH, Mitchell GW, Norris DR, Paquet J, Ronconi RA, Smetzer JR, Smith PA, Welch LJ, & Woodworth BK (2017). The Motus Wildlife Tracking System: A collaborative research network. Avian Conservation & Ecology, 12(1), 8. [Google Scholar]
  58. von Rönn JAC, Shafer ABA, & Wolf JBW (2016). Disruptive selection without genome-wide evolution across a migratory divide. Molecular Ecology, 25(11), 2529–2541. 10.1111/mec.13521 [DOI] [PubMed] [Google Scholar]
  59. Willemoes M, Blas J, Wikelski M, & Thorup K (2015). Flexible navigation response in common cuckoos Cuculus canorus displaced experimentally during migration. Scientific Reports, 5(1), 16402. 10.1038/srep16402 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplementary material
NIHMS2133757-supplement-supplementary_material.docx (503.5KB, docx)

Data Availability Statement

Data included in this manuscript, including Motus detections and release/capture details for each individual, have been uploaded to Dryad (DOI: 10.5061/dryad.83bk3jb4d; link: datadryad.org/dataset/doi:10.5061/dryad.83bk3jb4d). Scripts for both bioinformatics and downstream analysis have additionally been uploaded to Dryad and are available on GitHub (bioinformatics: github.com/stephblain/thrush_hybrid_survival, analysis: github.com/stephblain/SWTH_translocation). All sequences have been uploaded to the NCBI Sequence Read Archive under BioProject number PRJNA979932.

RESOURCES