Limited Song Mixing Without Genomic Gene Flow in a Contact Zone Between Two Songbird Species

Abstract

Song is considered to play an important role in the maintenance of prezygotic reproductive isolation between closely related songbird species. Therefore, song mixing in a contact zone between closely related species is often considered as evidence of hybridization. The Sichuan Leaf Warbler Phylloscopus forresti and the Gansu Leaf Warbler Phylloscopus kansuensis, which diverged 2 million years ago, have formed a contact zone in the south of the Gansu Province of China, where mixed songs have been observed. In this study, we investigated the potential causes and consequences of song mixing by integrating bioacoustic, morphological, mitochondrial, and genomic data with field ecological observations. We found that the two species display no apparent morphological differences, whereas their songs differ dramatically. We demonstrated that ∼11% of the males in the contact zone sang mixed songs. Two males singing mixed song were genotyped, and both were found to be P. kansuensis. Despite the presence of mixed singers, population genomic analyses detected no signs of recent gene flow between the two species, although two possible cases of mitochondrial introgression were identified. We conclude that the rather limited song mixing does not lead to, or result from, hybridization, and hence does not result in the breakdown of reproductive barriers between these cryptic species.

Keywords: contact zone; cryptic species; gene flow; hybridization; leaf warblers; song mixing.

Fig. 1.

Distribution, genomic sampling sites (n = 35), and morphological measurements of Phylloscopus forresti (n = 10), Phylloscopus kansuensis (n = 7), and 1 mixed singer. (A) Distribution areas of the two species according to BirdLife International (2016a, 2016b) and genomic sampling sites in this study. Note that the overlap zone according to our study is considerably smaller and placed further north than according to BirdLife International (2016a, 2016b). The nonbreeding range of P. kansuensis is unknown. (B) Principal component analysis (PCA) based on body shape indicators. (C) PCA based on plumage coloration.

Fig. 2.

Acoustic analysis of the two leaf warbler species. (A) Endless song (a) and verse song (b) of Phylloscopus forresti and verse song of Phylloscopus kansuensis (c). (B) Different verses of mixed singer Mix-HZ-1001 (a, c) and Mix-HZ-1101 (b, d) in the contact zone. (C) PCA based on 15 log-corrected bioacoustic parameters (n = 162 males). (D) Three slightly different verse types of P. kansuensis from different geographical areas shown in supplementary figure S8, Supplementary Material online.

Fig. 3.

Phylogenetic trees and population structure of Phylloscopus forresti and Phylloscopus kansuensis. (A) Cladogram inferred by Bayesian Inference from 11 mitochondrial genes (cytb, atp6, atp8, cox1, cox2, cox3, nad1, nad2, nad3, nad4, nad4l) in BEAST2. Note that two of the samples failed to assemble the mitochondrial genomes, thus 33 individuals were included in this analysis (vs. 35 in the analyses based on SNP data). (B) Cladogram inferred by ML from 2,484,080 informative genomic SNPs of 35 individuals in iqtree2 with best-fit model K3Pu + F + ASC + G4. (C) Mitochondrial tree inferred from partial cytb sequences (n = 48, cytb48 data set; target species: supplementary table S1, Supplementary Material online; close relatives: supplementary table S4, Supplementary Material online) with BEAST2. Posterior probabilities are labeled at the nodes. (D) Results of population structure analysis by FRAPPE based on 11,383,944 genomic SNPs inferred from the genomic data set (n = 35). The samples from the allopatric regions are located on both sides of the figure, while the samples from the contact zone are located in the middle. (E) Scatter plot from PCA of 11,383,944 genomic SNPs from the genomic data set (n = 35). Individuals from the contact zone are indicated by gray background in A, B, and D, and mixed singers are marked with asterisks in A, B, D, and E.

Fig. 4.

Potential ancient asymmetric gene flow between Phylloscopus forresti and Phylloscopus kansuensis inferred by fdM values of 100 kb windows in ABBA-BABA tests based on the genomic data set (n = 35). The models are shown by the cladograms on the right (F stands for P. forresti, K for P. kansuensis, C for contact region, A for allopatric region, and OU for outgroup) with red arrows showing the gene flow under test. Windows with high fdM are highlighted with red dots and corresponding scaffolds are labeled nearby.