INTRODUCTION
The application of molecular methods combined with improved and more widespread recording of vocalizations, and increased sampling by ornithologists and bird banders across the breeding and wintering ranges of Asian birds, has greatly increased understanding of the taxonomy and distribution of sibling species among warbler genera. Among these, the large warbler genus Phylloscopus, with c. 60 species, is possibly the most taxing for identification.

We here document the addition of both Kamchatka Leaf Warbler P. examinandus and Sakhalin Leaf Warbler P. borealoides to the Thai avifauna, based on analysis of mitochondrial DNA recovered from tail feathers of birds trapped for banding and subsequently released.

The wide-ranging Arctic Warbler P. borealis was recently recognized as a complex of three species, including Japanese Leaf Warbler P. xanthodryas (breeding in Honshu to Kyushu), and Kamchatka Leaf Warbler P. examinandus which breeds in Kamchatka, Sakhalin and Hokkaido (Dickinson & Christidis 2014). Although P. examinandus was usually treated as an intergrade between P. borealis and P. xanthodryas (e.g., Ticehurst 1938) it has since been shown to lie in a separate clade from the other two species in the complex, differing in morphology and song (Alström et al. 2011, Saitoh et al. 2006, 2008, 2010).

Arctic Warbler (sens. lat.) is a widespread and common passage migrant and winter visitor in Thailand, mainly in the south and southern central part of the country, south of roughly 15 degrees N latitude, occurring in both lowland terrestrial forests and mangroves (Lekagul & Round 1991). Nominate P. b. borealis was the only taxon listed for Thailand by Deignan (1963) while P. xanthodryas is so far known from a single individual caught and banded on northwards migration in SE Thailand (Round et al. 2015). A small number of exceptionally long-winged "Arctic Warblers", including some specimens, from southern Thailand and Malaysia thought likely to be P. xanthodryas, under which name P. examinandus was previously often subsumed (Medway & Wells 1976, Wells 2007) are perhaps also likely to have been examinandus, but we have not examined these and their identification remains tentative.

The Sakhalin Leaf Warbler P. borealoides was described as a distinct species by Portenko (1950) on the basis of its song, which is distinctively different from that of Pale-legged Leaf Warbler P. tenellipes. It breeds in southern Sakhalin, the Kuril Islands and Japan south to Shikoku (Dickinson & Christidis 2014) and shows different habitat preferences from Ussuri and North Chinese-breeding Pale-legged Leaf Warbler (Weprincew et al. 1989). Understanding of its distribution and movements has been confounded, however, owing its great morphological similarity to the latter species. The Pale-legged Leaf Warbler has long been recognized as a widespread and common visitor in forested areas, including mangroves, throughout most of Thailand, though is commonest in the eastern part of the country and scarcer in the peninsula (Deignan 1963, Lekagul & Round 1991).

Arguably, separating these two species pairs, P. examinandus from P. borealis, and P. borealoides from P. tenellipes, on morphology are among the most difficult identification challenges among Asian Phylloscopus species.

METHODS
Biometrics and one or two tail feathers were collected from samples of Arctic Warblers and Pale-legged Leaf Warblers (sens. lat., both) banded mainly on passage at coastal localities in Thailand. Wing measurements were maximum chords and bill-length was measured to skull. All measurements were taken by either PDR or AJP.

DNA was extracted from tail feathers using DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany). We sequenced a partial region, the cytochrome c oxidase subunit I (COI) gene (c. 700 bp) of the mitochondrial genome, which is used as a standard DNA barcoding region for most animals (Heber t et al. 2003). The barcoding region was amplified using the following primers: L6697Bird (5'-TCAACYAACCACAAAGAYATCGGYAC-3') and H7390Thrush (5'-ACGTGGGARATRATTCCAAATCCTG- 3') following Saitoh et al. (2015). Neighbour-joining (NJ) trees with bootstrap values (1,000 replications) were constructed from the COI data using the Kimura 2-parameter model (K2P) model and Mega 5.2.2 (Tamura et al. 2011). A breeding season/location P. borealis VK57822 (Japan DNA Data Bank Accession no. LC087227) and a breeding season and location P. examinandus (specimen NSMT-A15083; DDBJ Accession no. AB843059) were used to provide sequences for comparison with Thai migrants of these two species, as was a P. borealoides specimen YIO-62722 (DDBJ Accession no.AB843689) and a P. tenellipes specimen 1998-5165 (DDBJ LC087226; Appendix 1 ). A sequence from P. xanthodryas specimen YIO-64546 (DDBJ accession number AB843687) was used as an outgroup for both NJ trees (Figs. 2 and 3 ).

Kamchatka Leaf Warbler P. examinandus
We were able to determine the DNA sequences of 23 "Arctic Warblers" sampled from 659 to 715 bp in the COI region (DDBJ accession Nos. LC087183, LC87186-88, LC87190-93, LC87195-87202, LC87209, LC87213-18; Appendix 1 ). Among these, five belonged in the P. examinandus clade, and 18 individuals belonged in the P. borealis clade (Fig. 2, Appendix 1 ). All but one of the sampled birds of both taxa, P. examinandus and P. borealis, were caught at locations that indicated they were almost certainly passage migrants: on northward (spring) migration 14 Apr. to 19 May, n = 13 both species combined; and southward (autumn) migration, 9 Sept. – 10 Oct., n = 9, P. borealis only). A presumed overwintering individual, another P. borealis, was sampled in lowland evergreen forest at Khao Soi Dao Wildlife Sanctuary, Chanthaburi Province, South-East Thailand, on 28 Dec. (Fig. 1, Appendix 1 ).

All five P. examinandus were caught at Laem Phak Bia, Phetchaburi Province, Central Thailand (Fig. 1 ), on northward migration: on 9 May 2010 (one); 14 May 2011 (one) and 19 May 2012 (three; Appendix 1 ).

The five P. examinandus were scarcely distinguishable in morphology from P. borealis (Table 1, Appendix 1 ), though tended to be slightly larger and longer-winged. One of these birds (band no. 1A04402) was markedly yellower on the supercilium and throat than typical borealis and brighter green above (Figs. 4a, 4b, and 5 ), apparently agreeing with descriptions of P. examinandus in Ticehurst (1938) and Alstrom et al. (2011). However, another individual, 1A04411 with virtually identical colouration to 1A04402, and caught on the same day (Fig. 6 ) proved to be a borealis on DNA (Appendix 1 ), illustrating the extreme difficulty of reliably separating these two species on morphology.