Introduction

The process of ecological speciation has been increasingly brought to attention during the recent 2 decades (Rundle and Nosil 2005; Schluter 2001). A songbirds’ migratory direction is largely genetically determined (Helbig 1991). Because it is a heritable trait, differences in natural selection on migratory behaviors have the potential of mediating speciation. A striking example has been reported in Blackcaps Sylvia atricapilla (Berthold et al. 1992), in which a new migratory route, associated with both genetic and phenotypic differences (Van Doren et al. 2021), seems to have evolved in only about 30 years. So far, no consensus has been reached of which genes might underlie divergent migratory behaviors, though many examples exist (Delmore and Irwin 2014; Sanchez-Donoso et al. 2022; Toews et al. 2019).

Acting as natural laboratories, hybrid zones are essential for studying genes, determining divergent ecologically linked traits and how they contribute to creating reproductive isolation (Barton and Hewitt 1989). Hybrid zones between two differentially migrating taxa are called migratory divides and are important to locate for studying the genetics behind migration.

In Europe, there are two differentially migrating subspecies of Willow Warblers Phylloscopus trochilus (Sokolovskis et al. 2018). The nominate subspecies trochilus migrates toward SW, with wintering grounds in west Africa while the subspecies acredula migrates SSE and winter in east and south Africa. Ecologically and morphologically, the two subspecies are hard to tell apart. They differ on average in wing length and plumage coloration but overlap to such extent that individual identification is impossible in all but extreme cases (Bensch et al. 2009). The genetic differences between trochilus and acredula are also few and aggregate in inversion polymorphic regions on chromosome 1, 3, and 5 (InvP-Ch1, InvP-Ch3, and InvP-Ch5) (Lundberg et al. 2017, 2023). InvP-Chr1 and InvP-Chr5 are both associated with the direction of migration (Lundberg et al. 2017) while InvP-Chr3 correlates with a climatic gradient (Larson et al. 2014). A large repeat-rich block characterized by the expansion of a novel transposon in acredula (MARB-a) also segregates between migratory phenotypes (Caballero-López et al. 2022; Sokolovskis et al. 2023). Most recently, Sokolovskis et al. (2023) found that the MARB-a explains about 62% of the variation in migratory direction alone. While the trochilus allele for InvP-Ch1 has a dominant effect on migratory direction, so does the MARB-a but it also suppresses the effect of the former through an epistatic interaction (Sokolovskis et al. 2023).

In central Scandinavia and in eastern Poland, trochilus and acredula meet and form migratory divides (Bensch et al. 2009). The Polish hybrid zone has wide clines for the genetic markers linked to migration, consistent with it being the site of divergence (Bensch et al. 2009; Lundberg et al. 2023). In central Scandinavia on the other hand, the clines are steep, as expected for a tension zone (Barton and Hewitt 1985), and are likely the result of secondary contact between the two subspecies (Bensch et al. 2009; Lundberg et al. 2023). Located in between the distribution ranges of trochilus in southern Sweden and acredula in Finland, lie the Baltic islands of Åland, for which the genetic composition of Willow warblers is unknown. We examined the Willow warbler population structure on the Åland islands, to find out whether it consists of the subspecies trochilus, acredula or a mix between the two. Considering their intermediate location in between the main distribution ranges of ssp trochilus and acredula, we hypothesize that the islands represent a third hybrid zone, alongside the already know hybrid zones in Poland and central Scandinavia. We test this hypothesis by studying the genetic composition of Willow warblers on Åland and predict that birds from the islands comprise a genetic mix between the two parental subspecies.

Methods

We caught a total of 60 Willow warblers, equally divided between 3 sites spread across the Åland Islands in May 20–25th 2022. Our capture sites were located 19–21 km apart from each other on the main islands (Fig. S1), in the west (60°11'N, 19°32'E), central (60°15'N, 19°55'E), and in the east (60°13'N, 20°15'E). To capture singing males, we used mist nets with playback and thereafter took biometrics (e.g., wing length) and a 20 μl blood sample. As allopatric reference samples, we used previously published data collected with the same methods (Bensch et al 2009) from southern Sweden and mainland Finland, respectively. We also used samples from the Swedish hybrid zone from 2018 (Sokolovskis et al. 2023) to compare the distribution of transposon copy numbers and levels of heterozygosity with the Åland samples. From the blood samples, we successfully extracted genomic DNA from 58 individuals, following an ammonium acetate protocol. We used two biallelic SNP markers to genotype the birds for InvP-Chr 1 and 5 (SNP 65 and SNP 285 respectively), following a qPCR protocol (Zhao et al 2020). To estimate the presence of MARB-a, we analyzed transposable element (TE) copy numbers quantitatively following the qPCR protocol of Caballero-Lopez et al. (2022). A TE copy number over 7 indicates the presence of at least one chromosomal segment of the MARB-a.

We calculated pairwise FST values between the Åland population (n = 58) and allopatric reference samples from Sweden (n = 49) and mainland Finland (n = 35) following Weir and Cockerham’s (1984) method in the R package “hierfstat” (Goudet and Jombart 2022) To compare TE copy numbers between populations, we performed a Kruskal–Wallis one-way analysis of variance and a post hoc Wilcoxon rank sum test. Once again using the “hierfstat” package in R (Goudet 2005), we calculated the Hardy–Weinberg equilibrium for two chromosomal markers. The equilibria were computed for four sites: for birds from Åland, the two reference populations, and the Swedish hybrid zone (n = 105). We then tested the expected heterozygosity against the observed with a χ2 test to see if there were any signs of selection against heterozygotes. All statistical analyses were performed in R, version 4.1.2 (R Core Team 2021).

Results

Pairwise FST values between the Willow warbler population on Åland compared to southern Sweden and mainland Finland indicate that the birds on Åland are genetically intermediate for the two chromosomal markers included (see Fig. 1). Genetically, birds from Åland were slightly closer to the Swedish population (FST = 0.357) than to the population in mainland Finland (FST = 0.438) and resembled birds of the Swedish hybrid zone the most in genotype frequency (Fig. 1). Birds from Åland and the Swedish hybrid zone also showed minimal genetic distance between each other (FST = -0.002). The genetic distances between sites within populations were 87 small (Table S2).Wing lengths of birds from Åland were similar but slightly higher than our allopatric reference populations (Fig. S2).

Fig. 1
figure 1

Genotype frequencies in InvP-Chr 1 and 5 in Willow Warblers from Åland, the Swedish hybrid zone, allopatric trochilus from southern Sweden and allopatric acredula from Finland. Homozygote trochilus are shown in black bars, heterozygotes in grey, and homozygote acredula in white. Counts for each genotype and population combination are displayed above their respective bar

Full size image

The analysis of mean TE copy number between populations reported a difference in the mean for at least one of the groups (χ2 = 45.058, df = 2, p < 0.001). A post hoc test revealed the largest difference between the two allopatric populations (p < 0.001). The comparisons of Åland birds to the two allopatric populations also reported differences in rank sum, the largest of which were between Åland and Sweden (p < 0.001). The significance levels in the contrast between the Åland and the Finnish mainland population (p = 0.017) were smaller than between Åland and Sweden.

All Swedish birds had less than 6 TE copies while all Finnish mainland birds had more than 7 copies, in line with MARB-a being absent in the former and present in the latter (see Fig. 2). The frequency distributions of TE copy numbers on Åland and in the Swedish hybrid zone show a full spectrum overlap but the frequency of low copy numbers was higher in the Swedish hybrid zone (63%) than on Åland (43%).

Fig. 2
figure 2

Frequency distribution of TE copy numbers in birds from Åland, the Swedish hybrid zone, allopatric trochilus from southern Sweden, and allopatric acredula from Finland. Birds without the MARB-a (TE copy number < 7) are shown in black and birds with the MARB-a (TE copy number ≥ 7) are represented by grey bars

Full size image

The InvP-Chr 5 marker in Sweden had fewer heterozygotes than expected (χ2 = 4.117, p = 0.042). None of the chromosomal markers differed from Hardy–Weinberg equilibrium in the other populations (see Table S1).

Discussion

Åland Willow Warblers turned out to be genetically intermediate to the reference populations of allopatric trochilus and acredula, consistent with our hypothesis of the population being a mix between the two subspecies. The FST values for the two inversion polymorphisms suggest that Åland birds are more closely related to trochilus than to acredula with minimal difference to the Swedish hybrid zone. Intermediate FST values could, however, originate from samples with equal amounts of trochilus and acredula homozygotes. This can be excluded since all chromosomal markers were in Hardy–Weinberg equilibrium on Åland. We, therefore, find good support for our prediction of mixed genotypes. If the inversion polymorphisms alone determined migratory phenotype, the western migratory route would be dominating in Willow Warblers from Åland. When instead looking solely at the transposable element copy numbers, the distribution suggests quite the opposite migratory pattern. A majority of Åland birds have the MARB-a, which has a dominant epistatic effect on migration direction (Sokolovskis et al. 2023), and they would, therefore, be more likely to migrate SE. The combination of these opposing patterns render it probable that Willow Warblers on Åland will have mixed and intermediate migratory phenotypes, similar to birds from the central Scandinavian hybrid zone (Sokolovskis et al. 2023). Further supporting this hypothesis is the frequency distribution of transposable element copy numbers on Åland. The distribution resembles that of the Scandinavian hybrid zone and stands in contrast to the two allopatric populations. Where the allopatric populations are clearly separated, with all individuals having less than 7 copies in trochilus and more than 7 in acredula, Åland and the Scandinavian hybrid zone display a full spectrum of copy number frequencies (see Fig. 2). Hence, Åland Willow Warblers seem to constitute a blend of individuals homozygote for MARB-a, heterozygote for MARB-a and lacking the MARB-a.

There was no statistically significant difference in TE copy number rank sum between the three sites on Åland. If the islands represented a hybrid zone, with a gradual cline between trochilus and acredula, we would have expected to find significant FST values between the three sites on Åland and an increasing proportion of acredula alleles to the east. Instead, our results speak for the islands being genetically unstructured, and therefore better described as a hybrid swarm. To conclude whether Willow Warblers on the Åland islands express a diversity of migratory directions as in the Scandinavian migratory divide will however require tracking of their migration routes. Our findings of markedly higher heterozygosity levels in both Åland and the central Scandinavian hybrid zone fall well in line with our hypothesis of Willow Warblers on Åland consisting of hybrids between the two parental subspecies.

On the Åland islands, we discovered a previously unknown Willow Warbler hybrid swarm. Our findings fall in line with our prediction of mixed genotypes but the lack of evidence for a cline in allele frequencies makes us conclude that the population consists of a hybrid swarm rather than a hybrid zone. Overall, the population structure on the islands was similar to that of the Scandinavian hybrid zone which suggests that Åland Willow Warblers could represent a third migratory divide.