We present non-breeding season records of the recently named Alpine Leaf Warbler Phylloscopus occisinensis from Bangladesh (four individuals) and northern Thailand (one). Identification was based on mitochondrial DNA assay of feathers or blood from birds handled during ringing. Tickell's Leaf Warbler P. affinis (sensu lato) was abundant in scrub and scattered trees at the margins of wetlands in north-east Bangladesh, whilst the record from Thailand represents a significant eastwards extension of the previously recorded wintering range. Further sampling in South and South-East Asia will be necessary to resolve the winter ranges of the taxa affinis and occisinensis.
Within the ‘Tickell's Leaf Warbler Phylloscopus affinis species complex’, breeding around the margins of the Tibetan Plateau (which also encompasses the Sulphur-bellied Leaf Warbler P. griseolus), eastern populations of P. affinis were shown to constitute a distinct mitochondrial lineage that has been named Alpine Leaf Warbler P. occisinensis (Martens et al. 2008). Following this treatment, P. affinis breeds in the Himalayas, eastwards to Xizang, China, whereas P. occisinensis is distributed further east, beyond the Himalayan chain, in Yunnan, Sichuan, Qinghai and Kansu (China). Despite their deep mitochondrial divergence (15.8–16.0% in cytb, Martens et al. 2008; c.7.8% for the complete mitochondrial genome, Zhang et al. 2019) the two taxa are scarcely distinguishable using morphology and nuclear DNA, whilst any difference in song appears to be clinal, with ‘bioacoustic intermediacy' over a narrow contact zone in eastern Xizang (Zhang et al. 2019). The deep mitochondrial divergence has been attributed to ‘ghost introgression’ as an easterly distributed ancestor expanded west and interbred with another, unknown, and presumably extinct Phylloscopus species (Zhang et al. 2019). Treatment as two species is not universally followed, however (BirdLife International 2020).
Irrespective of whether P. affinis and P. occisinensis are treated as one or two species, it is of both faunal and possible future conservation interest that the winter (non-breeding) ranges of the two lineages are identified. The winter range of P. affinis (sensu lato) extends from northern and peninsular India (in the south-west Ghats) to Bangladesh and Myanmar (Rasmussen & Anderton 2005, Dickinson & Christidis 2014). Here we present records of overwintering birds sampled in north-east Bangladesh and northern Thailand.
Field collection.—Prior to their release, we collected feathers of three P. affinis (sensu lato) from a total of 28 individuals, mist-netted and ringed in low trees and scrub around two wetland sites in north-east Bangladesh during 2012–14 (Round et al. 2014), and from one additional individual of another 11 ringed at a third site during 2015 (Fig. 1).
Additionally, a single Phylloscopus with yellowish underparts and supercilium, lacking wingbars, in Hang Dong District, Chiang Mai province, northern Thailand (Fig. 1), first found on 17 January 2020 by Thammarat Kaosombat, was tentatively identified from digital photographs as either Tickell's or Alpine Leaf Warbler by Wich'yanan Limparungpatthanakij a day or two after its discovery. This individual attracted particular interest as there were no previous records of P. affinis or P. occisinensis in Thailand or Indochina. On 24 January 2020, it was caught in a mist-net, ringed, measured, photographed in the hand (Fig. 2), and a blood sample taken on filter paper before release. The same bird remained at the site and was repeatedly observed until 9 February 2020. It was not heard to call during this period.
Biometrics (mm) and mass (g) of five wintering Alpine Leaf Warblers Phylloscopus occisinensis sampled. The first four individuals were trapped in Bangladesh, the fifth in Thailand.
All individuals were caught in scrub in open marshy areas fringing wetlands on the plains, and mass and basic biometric data recorded (Table 1). The feathers were stored in labelled paper envelopes in a freezer at 0°C and the blood samples at -20°C.
Laboratory protocols.—Genomic DNA was extracted from feathers (the four Bangladesh individuals) and blood (the Thai bird) using NucleoSpin tissue kit (Macherey-Nagel), with 0.1% Dithiothreitol (DTT) added to increase the DNA yield (Olsson et al. 2005). Partial mitochondrial genes were amplified using primers: Bird F1 (5′-TTCTCCAACCACAAAGACATTGGCAC-3′) and Bird R1 (5′-ACGTGGGAGATAATTCCAAATCCTG-3′) for Cytochrome C oxidase subunit 1 (COI; Saitohetal.2015)andL14841(5′-AAAAAGCTTCCATCCAACATCTCAGCATGATGAAA-3′) and H15547 (5′-AATAGGAAGTATCATTCGGGTTTGATG-3′) for cytochrome b (Cytb; Helbig et al. 1995). Polymerase chain reaction (PCR) was performed using AccuStart II GelTrack PCR SuperMix (Quanta BioSciences) following the concentration of ingredients suggested by the manufacturer. The reactions for both genes were performed with the following steps: 94°C for three minutes followed by five cycles at 94°C for 30 seconds, 48°C for 30 seconds, 72°C for one minute, then 30 cycles at 94°C for 30 seconds, 51°C for 30 seconds, 72°C for one minute, and a final 72°C for five minutes (Saitoh et al. 2015) using an Eppendorf Mastercycler gradient thermocycler. PCR products were visualised with 1.5% agarose gel electrophoresis. The products successfully amplified were purified using NucleoSpin Gel and PCR Clean-up (Macherey-Nagel) and sent for sequence analyses using the Applied Biosystems BigDye Terminator v3.1 Cycle Sequencing Kit protocol.
Genetic analyses.—Sequences of collected samples were deposited in GenBank. These sequences and those of other species downloaded from GenBank (Table 2) were aligned using ClustalW in MEGA7 v.7.0.21 (Kumar et al. 2016) and used to conduct Maximum Likelihood (ML) and Bayesian Inference (BI) phylogenetic analyses, with Yellow-streaked Warbler P. armandii serving as the outgroup. Phylogenetic trees for cytb and COI were constructed separately because the sequences of both genes were not available for concatenative analyses. Kakusan4 (Tanabe 2007) was used to select the best-fit evolutionary model under Akaike information criterion (AIC; Akaike 1974) and Bayesian information criterion (BIC; Schwarz 1978) for ML and BI, respectively. The selected models for the ML trees were GTR Gamma (both COI and cytb) and for the BI trees HYK85 Gamma (both COI and cytb). The ML trees were constructed using RAxML v8.2.12 (Stamatakis 2014) on CIPRES Science Gateway V3.3 (Miller et al. 2010) with 1,000 bootstrapping iterations. A bootstrap value of 70% or higher was considered significant support. The BI trees were performed in MrBayes v3.2.6 (Huelsenbeck & Ronquist 2001) under a Metropolis-coupled, Markov chain Monte Carlo (MC-MCMC) approach, started from a random tree, run twice in parallel with a four-chain analysis for five million generations. The trees were sampled every 100 generations and 25% of these were discarded as ‘burn-in'. We ran Tracer v1.7.1 (Rambaut et al. 2018) to evaluate stationarity, and Effective Sample Size (ESS) values >200 were obtained for all outputs. Consensus topology, branch length, and bipartition posterior probability were estimated from the remaining trees, with posterior probabilities of 95% or higher considered as significant support. The trees for both ML and BI approaches were visualised and edited in FigTree v1.4.3 (Rambaut 2009).
Ring numbers, localities and accession numbers for collected samples, and the samples in GenBank used in this study.
The samples from all four Bangladesh-caught birds were successfully amplified for both COI and cytb genes, whereas only cytb was successfully amplified in the blood sample from the single northern Thai bird (Table 2). The alignments revealed 668 bp for COI (104 variable sites, 77 parsimony informative sites) and 673 bp for cytb (115 variable sites, 84 parsimony informative sites). As the phylogenetic analyses based on ML and BI approaches provided similar results, only the BI trees are shown (Fig. 3). Phylogenetic trees for COI and cytb were largely concordant in topology, and revealed the monophyly of all collected samples (AAA140, AAA504, AAA509, AAA631 and A35121) with P. occisinensis, with strong statistical support (COI: 91/100% and cytb: 100/100% for bootstrap values and posterior probabilities, respectively). The P. occisinensis clade was distinct from both affinis and griseolus. However, whilst the P. occisinensis clade was resolved as sister to P. affinis in the COI tree, occisinensis appeared as sister to P. griseolus in the cytb tree (Fig. 3). In the cytb tree indicated by Martens et al. (2008) affinis was sister to griseolus, while in a concatenated tree of tRNA, rRNA and cytb it was occisinensis and griseolus that appeared as sisters. Incongruities in mtDNA topologies among these taxa were discussed by Alström et al. (2018) and are not surprising as the mitochondrial divergence is almost identical among all three lineages (Zhang et al. 2019).
Four wintering P. affinis (sensu lato) caught and ringed at wetlands in north-east Bangladesh, and another individual, a presumed vagrant, in northern Thailand, the easternmost South-East Asian wintering record known (Clement 2020), proved to lie in the recently distinguished Alpine Leaf Warbler P. occisinensis lineage.
Even considering the greater proximity of Bangladesh to the Himalayan breeding areas of the more westerly distributed P. affinis (sensu stricto), the discovery of P. occisinensis there was not especially surprising considering that north-east India and Bangladesh also harbour wintering populations of other species that breed in south-west China such as Firethroat Calliope pectardens (Rasmussen & Anderton 2005, Round et al. 2014). Given the abundance of P. affinis (sensu lato) in trees and scrub around the margins of northern Bangladesh wetlands in winter, and considering that so few individuals were sampled, the question arises whether P. affinis (sensu stricto) and P. occisinensis might winter alongside each other in the same areas as yet undetected. Clearly more sampling is needed to improve understanding of the winter distribution of these taxa across South Asia. Because P. affinis (sensu lato) winters as far south as the Western Ghats (Rasmussen & Anderton 2005) it might be productive to first sample individuals there to discover whether there is any major disparity in the wintering ranges of the two, the more southerly and westerly distributed birds perhaps being exclusively P. affinis (sensu stricto). It also remains to be investigated whether names applied to specimens currently synonymized with P. affinis (e.g., Abrornis xanthogaster Hodgson, 1844, Phyllopneuste flaveolus G. R. Gray, 1846) might in fact be applicable to the newly described taxon.
In Bangladesh we thank the Chief Conservator of Forests and the IUCN Country Representative for facilitating our work, which was part-funded by the Wetland Trust, UK, in 2012 and 2014, and by the Bangladesh Bird Club and IUCN in 2015. Particular thanks are owed to Enam Ul Haque, Paul M. Thompson, Stephen Rumsey and Dr Jenjit Khudamrongsawat for invaluable logistic support and advice. In addition, we thank Per Alström and two referees, James Eaton and Jochen Martens, for their most helpful comments on drafts of this paper.