'IqS 5-^0( & Journal of Asian Ornithology NATURAL HISTORY [ MUSEUM LIBRARY 26 NOV 2013 PURCHASED | m Forktail 29(2013) OBC Council David Buckingham (Chairman, Conservation Committee) Mike Edgecombe (Promotions) John Gregory (Treasurer) Tim Loseby (Art and Photographic Editor) Simon Roddis Steve Rowland Tony Sawbridge Graeme Spinks (Assistant Treasurer) Brian Sykes (Chairman) Margaret Sykes (Membership & General Secretary) Jo Thomas Richard Thomas (Internet) Publications Committee Stuart Butchart, Nigel Collar, John Eaton, Tim Loseby, Rene Pop, Nigel Redman, Simon Roddis, Brian Sykes Conservation Committee Nick Brickie, Stephen Browne, Dave Buckingham, Francis Buner, Mike Crosby, John Fellowes, Jim Wardill, Simon Wotton The Oriental Bird Club has been established for ornithologists throughout the world, both amateur and professional, who share a common interest in the region's birds and wish to assist in their conservation. The Club aims to: • Promote interest in the birds of the Oriental Region and their conservation • Liaise with, and promote the work of, existing regional organisations • Collate and publish material on Oriental birds OBC Representatives Susan Myers (Australia) Paul Thompson (Bangladesh) Filip Verbelen (Belgium) Keo Omaliss (Cambodia) Tony Gaston (Canada) Ding Chang-qing (China) Jiri Mlikovsky (Czech Republic) Klaus Mailing Olsen (Denmark) Hannu Jannes (Finland) Axel Braunlich (Austria and Germany) Paul Leader (Hong Kong) Janos Olah (Hungary) Asad Rahmani and Vishnu Singh (India) Ria Saryanthi (Indonesia) Chris Murphy (Ireland) Carlo Violani (Italy) Akira Hibi (Japan) Jin-Young Park (Korea) Tom Gray (Lao PDR) Mike Chong, Ooi Chin Hock and Anthony Wong (Malaysia) Charles Anderson (Maldives) Nyambayar Batbayar (Mongolia) Tony Htin Hla (Myanmar [Burma]) Yub Raj Basnet (Nepal) Jelle Scharringa and Bas van Balen (Netherlands) Jan 0ve Gjershaug (Norway) Aleem Ahmed Khan (Pakistan) Carmela Espanola and Arne Jensen (Philippines) Ray Tipper (Portugal) Fang Woei-horng (Taiwan) Lim Kim Seng (Singapore) Adam Riley (South Africa) Sarath Kotagama and Upali Ekanayake (Sri Lanka) Jonas Nordin (Sweden) Beat Wartmann (Switzerland) Philip Round and Pajaree Intravooth (Thailand) Robert Kennedy (USA) Jonathan Eames and Nguyen Cu (Vietnam) Membership of OBC Ordinary Member £15p.a. Family Member £20 p.a. Reduced Rate Member £10 p.a. For Oriental nationals resident in the region. We encourage all members to pay the full rate if they can afford it Supporting Member £25 p.a. Funding one Oriental member in addition to Ordinary membership Libraries and Academic Institutions £25 p.a. Business Supporter £45 p.a. For further information please write to: Oriental Bird Club, P.O. Box 324, Bedford MK42 OWG, UK OBC email address mail@orientalbirdclub.org OBC website http://www.orientalbirdclub.org/ The presentation of material in this publication and the geographical designations employed do not imply the expression of any opinion whatsoever on the part of the Oriental Bird Club concerning the legal status of any country, territory or area, or concerning the delimitation of its frontiers or boundaries. Cover picture: Cambodian Tailorbird Orthotomus chaktomuk, Highway 61, near Prek Kdam, Kandal province, Cambodia, 21 November 2012 by James A. Eaton ISSN 0950-1746 © Oriental Bird Club 201 3 The Oriental Bird Club is a Registered Charity No. 297242 FORKTAIL Number 29, 201 3 CONTENTS S. P. MAHOOD, A. J. I. JOHN, J. C. EAMES, C. H. OLIVEROS, R. G. MOYLE, HONG CHAMNAN, C. M. POOLE, H. NIELSEN & F. H. SHELDON A new species of lowland tailorbird (Passeriformes: Cisticolidae: Orthotomus) from the Mekong floodplain of Cambodia . 1 N.J. COLLAR &S. van BALEN Notes for the conservation of the Rufous-fronted Laughingthrush Garrulax rufifrons . 1 5 N. J. COLLAR, J. A. EATON & R. O. HUTCHINSON Species limits in the Golden Bulbul Alophoixus ( Thapsinillas ) affinis complex . 19 VLADIMIR YU. ARKHIPOV, TOM NOAH, STEFFEN KOSCHKAR & FYODOR A. KONDRASHOV Birds of Mys Shmidta, north Chukotka, Russia . . . 25 PAUL J. LEADER, GEOFF J. CAREY & PAUL I. HOLT Species limits within Rhopophilus pekinensis . 31 PAVEL KVARTALNOV, ABDULNAZAR ABDULNAZAROV, VERONIKA SAMOTSKAYA, JULIA POZNYAKOVA, IRINA ILYINA, ANNA BANNIKOVA & EUGENIA SOLOVYEVA Nesting of the Large-billed Reed Warbler Acrocephalus orinus: a preliminary report . 37 FRANCESCO GERMI, AGUS SALIM & ANDREA MINGANTI First records of Chinese Sparrowhawk Acc/p/ferso/oens/s wintering in Papua (Indonesian New Guinea) . 43 SHIAO-YU HONG, YUAN-HSUN SUN, HSIN-JU WU & CHAO-CHIEH CHEN Spatial distribution of the Tawny Fish Owl Ketupa flavipes shaped by natural and man-made factors in Taiwan . 48 M. MONIRUL H. KHAN Population, breeding and threats to the White-rumped Vulture Gyps bengalensis in Bangladesh . 52 JOHN D. FARRINGTON, ZHANG XIULEI & ZHANG MIN The birds of the Longbao National Nature Reserve and surrounding basin, Yushu county, Qinghai, China . 57 LI-HU XIONG & JIAN-JIAN LU Habitat specialisation in the Reed Parrotbill Paradoxornis heudei — evidence from its distribution and habitat use . 64 MIYUKI MASHIKO& YUKIHIKOTOQUENAGA Increasing variation in population size and species composition ratio in mixed-species heron colonies in Japan . 71 P. C. RASMUSSEN & N. J. COLLAR Phenotypic evidence for the specific and generic validity of Heteroglaux blewitti . 78 H. EDEN W. COTTEE-JONES, JOHN C. MITTERMEIER & DAVID W. REDDING The Moluccan Woodcock Scolopax rochussenii on Obi Island, North Moluccas, Indonesia: a 'lost' species is less endangered than expected . 88 PER ALSTROM, GANG SONG, RUIYING ZHANG, XUEBIN GAO, PAUL I. HOLT, URBAN OLSSON & FUMIN LEI Taxonomic status of Blackthroat Calliope obscura and Firethroat C. pectardens . 94 COLIN R.TRAINOR, STEPHEN J. S. DEBUS, JERRY OLSEN, JANETTE A. NORMAN & LES CHRISTIDIS Bonelli's Eagle Aquila fasciata renschi in the Lesser Sundas, Wallacea: distribution, taxonomic status, likely origins and conservation status . 100 EARL OF CRANBROOK, GOH WEI LIM, LIM CHAN KOON & MUSTAFA ABDUL RAHMAN The species of white-nest swiftlets (Apodidae, Collocaliini) of Malaysia and the origins of house-farm birds: morphometric and genetic evidence . 107 SOMSAK BUATIP, WANCHAMAI KARNTANUT & CORNELIS SWENNEN Nesting period and breeding success of the Little Egret Egretta garzetta in Pattani province, Thailand . 1 20 SIMON P. MAHOOD, SEBASTIEN DELONGLEE, FLORIAN KLINGEL, FALK WICKER & RICHARD CRAIK The status of Brown-chested Jungle Flycatcher Rhinomyias brunneatus in Vietnam . 124 JOHN C. MITTERMEIER, H. EDEN W. COTTEE-JONES, ENDANG CHRISTINE PURBA, NOVA MAULIDINA ASHURI, EKA HESDIANTI & JATNA SUPRIATNA A survey of the avifauna of Obi island, North Moluccas, Indonesia . 1 28 NATURAL HISTORY MUSEUM LIBRARY 26 NOV 2013 PURCHASED Forktail 29 (2013) Short Notes ASHOKA JAYARATHNA, SARATH W. KOTAGAMA & EBEN GOODALE The seasonality of mixed-species bird flocks in a Sri Lankan rainforest in relation to the breeding of the nuclear species. Orange-billed Babbler Turdoides rufescens . . . 138 T.M. BRAILE&K. WINKER New distributional records of Philippine birds from Bohol, Mactan, Olango, Busuanga and Luzon islands . 140 MERWYN FERNANDES & JAN WILLEM DEN BESTEN Some interesting breeding records for Pong Dam Wildlife Sanctuary, Himachal Pradesh, India . 141 TARIQ MAHMOOD, SYED M. USMAN-UL-HASSAN, MUHAMMAD S. NADEEM & AMJAD R. KAYANI Population and diet of migratory Common Starlings Sturnus vulgaris wintering in agricultural areas of Sialkot district, Pakistan . 143 EUN-MI KIM, CHANG-YONG CHOI & CHANG-WAN KANG Causes of injury and mortality of Fairy Pitta Pitta nympha on Jeju Island, Republic of Korea . 1 45 A. MOHAMED SAMSOOR ALI, S. RAMESH KUMAR & P. R. ARUN House Crow Corvus splendens nesting on pylons, Kutch district, Gujarat, India . 148 MARKR. BEZUIJEN New waterbird count data from the Heihe river in Gansu province, western China . . 1 50 SHAOBIN LI, WEIJUN PENG, CHENG GUO & XIN LU Breeding biology of the Small Snowfinch Pyrgilauda davidiana on the Tibetan plateau . 1 55 NATARAJAN EZHILARASI & LALITHA VIJAYAN Nest, eggs and nest sites of the Andaman Crake Rallina canningi . 1 58 TERESA M. PEGAN, JACK P. HRUSKA & JUSTIN M. HITE A newly described call and mechanical noise produced by the Black-and-crimson Pitta Pitta ussheri . 160 HUGH L. WRIGHT, SOK KO, NET NORIN & SUM PHEARUN White-shouldered Ibis Pseudibis davisoni population size and the impending threat of habitat conversion . 162 Errata . 165 Editorial notes . 166 Guidelines for contributors inside back cover FORKTAIL 29 (2013): 1-14 A new species of lowland tailorbird (Passeriformes: Cisticolidae: Orthotomus) from the Mekong floodplain of Cambodia S. P. MAHOOD, A. J. I. JOHN, J. C. EAMES, C. H. OUVEROS, R. G. MOYLE, HONG CHAMNAN, C. M. POOLE, H. NIELSEN &F. H. SHELDON Based on distinctive morphological and vocal characters we describe a new species of lowland tailorbird Orthotomus from dense humid lowland scrub in the floodplain of the Mekong, Tonle Sap and Bassac rivers of Cambodia. Genetic data place it in the 0. atrogularis-O. ruficeps-O. sepium clade. All data suggest that the new species is most closely related to O. atrogularis, from which genetic differences are apparently of a level usually associated with subspecies. However the two taxa behave as biological species, existing locally in sympatry and even exceptionally in syntopy, without apparent hybridisation. The species is known so far from a small area within which its habitat is declining in area and quality. However, although birds are found in a number of small habitat fragments (including within the city limits of Phnom Penh), most individuals probably occupy one large contiguous area of habitat in the Tonle Sap floodplain. We therefore recommend it is classified as Near Threatened on the IUCN Red List. The new species is abundant in suitable habitat within its small range. Further work is required to understand more clearly the distribution and ecology of this species and in particular its evolutionary relationship with O. atrogularis. INTRODUCTION After a hiatus of over half a century owing to the intense human conflicts in the area, the last two decades have witnessed the discovery of a flush of novel bird taxa in Indochina. These recent discoveries have been facilitated by better sampling of remote micro¬ habitats and to a much lesser degree the greater use of non- morphological characters in delimiting species. Most of these discoveries concerned babblers (Timaliidae) from isolated montane areas in Vietnam (Eames et al. 1994,Eames et al. 1999a, b, Eames& Eames 2001, Eames 2002). A smaller wave of discoveries involving a diverse range of taxa took place in forested limestone karst in Lao PDR, Vietnam and adjacent areas of China (Zhou Fang & Jiang Aiwu 2008, Woxvold et al. 2009, Alstrom et al. 2010). Only one new species, Mekong Wagtail Motacilla samveasnae, was named from Cambodian specimens, but it also occurs in Lao PDR, Thailand and Vietnam in ‘channel mosaic’ habitat on the Mekong and its major tributaries (Duckworth et al. 2001, Le TrongTrai & Craik 2008). Here we describe a new species of lowland tailorbird Orthotomus , confined to low elevation humid evergreen scrub in the floodplain of the Mekong and associated large rivers, in Cambodia. THE NEWTAILORBIRD During routine sampling of birds for avian influenza in 2009, four individual tailorbirds Orthotomus sp. were mist-netted and photographed in the hand: one on 28 and another on 29 January 2009 in a patch of scrub near a pond at Kraing Check, Kandal province (l l°4l'53.36"N 104°46'38.93,,E) (J. Reside per F. Goes in litt. 2012), one on 24 February 2009 and another on 12 March 2009 near to Phnom Tamao Zoo, Takeo province (lCH'Sh^N 104°50'22"E), in 3-5 m high scrub near rice fields (HN, A. Yang 103°30'Q"E 104°0'0"E 104°30'0"E 105°0,0,,E 1O5°30'O"E 106°0,0"E 106°30'0"E 107°0,0"E Figure 1. Distribution of records of Orthotomus chaktomuk and O. atrogularis within and close to the floodplain of the Mekong, Tonle Sap and Bassac rivers. 2 S. P. MAHOOD etal. Forktail 29 (2013) and P. Joyner in litt. 2012). Based on photographs and inferences drawn from incorrect location data (which misleadingly indicated that the birds had been caught close to the coast) these individuals were incorrectly identified as Ashy Tailorbird O. ruficeps-, the possibility of their representing aberrant Dark-necked Tailorbird O. atrogularis was considered and rejected based on general plumage similarity to O. ruficeps (F. Goes verbally 2012). On 29 January 2012, HN found a similar bird at PrekKsach, in a partially flooded construction site c.15 km from Phnom Penh (1 1°4T37.14"N 104°53'43.79"E) and, owing to similarity to the 2009 birds, assigned it to O. ruficeps. In earlyjune 20 1 2, photographs by AJIJ of an Orthotomus sp. from this site raised the interest of SPM. Subsequent field observations by SPM, AJIJ and T. D. Evans and discussion with J. W. Dqckworth, P. D. Round, CMP and C. Robson indicated to SPM that these birds might not be 0. ruficeps, but could perhaps represent an undescribed taxon. On 23 June 2012, SPM, HN and AJIJ searched for additional individuals at Prek Ksach and located five single males and two pairs. Between 23 June 2012 and 20 April 2013, intensive searches revealed at least 1 00 individuals at nine additional locations (Figure 1, Table SOM 1 [supplementary online material — see page 14]). Seven morphologically typical O. atrogularis comprising six males and one female were seen at five floodplain locations at or within 10 km of locations where birds of the new taxon were found (Figure 1, Table SOM 1). From photographs SPM re-identified all of the individuals mist- netted in 2009 as conforming to the new taxon, thus temporarily removing O. ruficeps from the list of birds recorded in Cambodia. Typical O. ruficeps has since been recorded in mangrove forest in coastal Cambodia close to the border with Vietnam (Mahood & Martin 2013). In August 2012 two adult males, one immature male and two immature females (aged by plumage, sexed internally) were collected for formal description (below). SPM was later able to compare these specimens directly with the Orthotomus material held at the Natural History Museum, Tring, UK (NHMUK) including a syntype (NHMUK 1 886. 10.1. 1830) of O. atrogularis nitidus (the subspecies in Cambodia) and the holotype of Olive-backed Tailorbird O. sepium (NHMUK1 880. 1.1. 4473), and also examined specimens at Naturalis Biodiversity Centre, Feiden, Netherlands (RMNH) including the holotype of O. ruficeps cineraceus (RMNH 137559). (The disjunct population of 0. ruficeps in coastal southern Vietnam and Cambodia has not been assigned to a subspecies; however, O. r. cineraceus is the subspecies recorded in mainland Asia: Madge 2006). All other Orthotomus taxa differed so extensively from the new form 01 23456789 10 Plate 1. Dorsal, ventral and lateral views of the holotype of 0. chaktomuk. that detailed comparison would be superfluous. A detailed list of all specimens examined is provided in Table SOM 2. The new taxon shows significant morphological differences from its close relatives and is sympatric with two lowland tailorbird species, O. atrogularis and Common Tailorbird (). sutorius , without signs of intergradation. We therefore consider that it represents a new species, which we name: Orthotomus chaktomuk, sp. nov. Cambodian Tailorbird http://zoobank.Org/urn:lsid:zoobank.org:act:23E9A09C-AD9C' 4346-A594-F 1 87DAFB60 1 3 Holotype and paratypes Study skins deposited in NHMUK (Table 1, Plate 1, Plate 2a-c) were collected by JCE and SPM at Bateay District, Kompong Cham Table 1 . Mensural and other relevant data of holotype and paratypes (in mm, except mass in g). Sex Age Culmen Tarsus Wing Tail Mass Testes length State of wing moult Collection date Holotype NHMUK reg.no. 2012.9.1 LSUMNS tissue accession no. B77286 M ad 13 19 47 42 8 4.5,2 PI-5 (R), P2-6 (L), tertials, greater and median coverts 8 August 201 2 Paratype NHMUK reg.no. 2012.9.2 LSUMNS tissue accession no. B77287 M ad 14 19.5 46 41 7 5,3.5 PI-6 (R), PI-6 (L), tertials, greater and median coverts r 9 August 2012 Paratype NHMUK reg.no. 2012.9.3 LSUMNS tissue accession no. B77288 M lyr 14.5 19 45 36 8 4,0.5 PI-6 (R), PI-6 (L), tertials, greater and median coverts 9 August 2012 Paratype NHMUKreg.no. 2012.9.4 LSUMNS tissue accession no. B77289 F lyr 13 18 41 35.5 6.5 n/a P2-3 (R), P2-3 (L), tertials and median coverts 9 August 2012 Paratype NHMUK reg.no. 2012.9.5 LSUMNS tissue accession no. B77290 F lyr 12 17 42 35.5 6 n/a PI -3 (R), PI -3 (L), tertials, greater and median coverts 8 August 201 2 M, Male; F, Female; ad, adult; lyr, first calendar year. State of wing moult lists feather tracts recently replaced; P, primaries; R, right wing; L, left wing. HARRY TAYLOR © NATURAL HISTORY MUSEUM, LONDON HARRY TAYLOR © NATURAL HISTORY MUSEUM, LONDON Forktail 29 (2013) A new species of lowland tailorbird ( Orthotomus ) from the Mekong floodplain of Cambodia 3 (b) V '* _ 2 ,KIT 3 H i V.lWotv TD ■- SfrAi-i Toe j 0123456789 10 Plate 2. Specimens of 0. chaktomuk: (a) dorsal, (b) ventral and (c) lateral views of the holotype and four paratypes (from left to right NHMUK 2012.9.1, 2012.9.2, 2012.9.3, 2012.9.4, 2012.9.5). province, Cambodia (1 1°56'53.94,'N 104°56'50.94"E), c.43 km north of Phnom Penh at c. 1 5 m elevation on 8 and 9 August 2012, and prepared by JCE. Tissue samples from the same individuals were deposited in Louisiana State University Museum of Natural Science, Baton Rouge, Louisiana, USA (Table 1). Holotype (NHMUK 2012.9.1): adult male; in active wing moult; one large testis (left testis 4.3 mm length, right testis 2 mm length). Paratypes aged by plumage: one adult male (NHMUK 2012.9.2); one immature male (NHMUK 2012.9.3) with unmoulted rectrices olive-green fringed; one immature female (NHMUK 2012.9.4) with some retained greater coverts fringed olive-green indicating immaturity, all other rectrices as adult; and one immature female (NHMUK 2012.9.5) with unmoulted rectrices as immature male. Diagnosis of species Head: in male entirely rich cinnamon-rufous crown and contrasting white cheeks, very similar to O. atrogularis , differing from 0. ruficeps and O. sepium in cheek colour (Table 2, Plate 3a- c). Rufous of crown less extensive in female. UpperpartS and wings: mid-grey in adult, superficially similar to those of 0. sepium but lacking olive tones, strikingly different from O. atrogularis which is yellowish-green (Table 2, Plate 3a, c); tail with dark grey subterminal band and whitish tips when fresh. UnderpartS: pale Table 2. Qualitative summary of plumage of Orthotomus chaktomuk and closely related species (all adult male). Species Crown Cheeks Malar stripe Mantle and rump Wing Chin Throat Breast Belly Vent Thighs Tail 0. chaktomuk Rich cinnamon- rufous Whitish White, black speckled Mid-grey Mid-grey Dark-grey with Dark-grey with Mid-grey with Light-grey, white speckling white speckling white speckling darker on flanks Greyish-white Whitish- cinnamon Mid-grey with blackish sub¬ terminal band and whitish tips O.a.nitidus Dull brick-red Whitish White, black speckled Bright yellow-olive Bright yellow-olive Black with Black with Whitish, black white speckling white speckling streaking on upper-breast White with greenish- yellow flanks Bright sulphur yellow Bright yellowish- orange Bright yellowish- green O.r.cineraceus Bright orange-rufous Orange-rufous Orange-rufous Dark Dull brownish-grey olive-brown Bright orange Mid-grey Pale-grey Whitish-grey White Bright orange-rufous Dull olive- brown, whitish terminal tips O.s. sepium Fore-crown rufous, central mid- and hind- crown olive-rufous Cinnamon- rufous Cinnamon- rufous Brownish- olive Dull brown with dull olive fringing Pale orange- rufous Dark olive-grey Greyish-olive Pale greyish- olive Whitish-grey Pale cinnamon- rufous Pale olive- brown with whitish terminal tips HARRY TAYLOR © NATURAL HISTORY MUSEUM, LONDON 4 S. P. MAHOODefa/. Forktail 29 (2013) grey ground colour with profuse blackish throat-streaking in males (largely absent in females) with white drop-shaped marks, extensive mid-grey on flanks, and white vent; underparts of both sexes superficially similar to those of respective sexes of O. atrogularis owing to throat-streaking, but greyer on flanks and vent white, O. ruficeps and O. sepium lacking throat-streaking in both sexes (Table 2, Plate 3b, c); further distinguished from other members of the genus by whitish-cinnamon thighs. Vocalisations : loud, lengthy, complex and highly varied. Very similar to O. atrogularis. Compared to O. atrogularis , phrases are given at a quicker pace and the gaps between phrases are shorter. Subjectively, these characteristics mean that the vocalisations of O. cbaktomuk sound faster and more complicated than those of O. atrogularis. Sexing and ageing Based on field observations (Plate 4a-k, Media Files SOM 1-3) and specimens (Plate 2a-c), female O. cbaktomuk can be distinguished from males by paler cinnamon-rufous on crown, which is restricted to forecrown and sides of mid-crown (in lateral view this appears as a short cinnamon-rufous supercilium), paler grey upperparts and wings and whitish underparts with usually faint dark streaking. The latter is usually evident only at the edges of the throat/upper breast, although some (possibly older birds) show stronger and more extensive streaking on throat and breast. Even in these extreme individuals, the degree of female streaking does not approach that in males (Plate 4a-g). All three immature paratypes show shorter tails than adults (Table 1). Immature birds possess bright yellowish-olive fringing to the wing-feathers (Plate 4h), which are moulted during August and replaced with grey adult- type feathers (Plate 4i-j). Immatures are browner (slightly olive) above and paler below, with reduced streaking (Plate 4h— j). Wing- feathers of subadults appear as in adults, except sometimes they retain yellowish-olive-fringed greater coverts (Plate 4k). Overall, subadults resemble adults, but are paler and less heavily marked below. In adults, there is individual variation in colour tone of grey feathering above and below, and intensity of throat-streaking (e.g. Media Files SOM 1-3). It is unknown if this is age-related. Description of species The detailed description below was completed in the NHMUK based primarily on the prepared specimens (Plates 1-2), supplemented by information from individuals observed and photographed in the field (Plate 4). It refers to the holotype unless otherwise stated. Although moult of body feathers was almost complete when specimens were collected, all adult specimens retained a few head, throat or breast feathers in pin. The holotype and paratypes were in wing moult. Moult of wing feathers is complete by late August and followed immediately by moult and replacement of tail feathers, which were very worn in all specimens. Subjective colour assessments of plumage are, where possible, followed by a formal colour classification taken from Smithe (1975). Head and face Crown from forehead to nape, lores, and feathers on orbital ring and just behind eye rich cinnimon-rufous (136 Raw Sienna) (slightly richer-coloured in the adult male paratype); hindcrown slightly darker and more brownish (23 Raw Umber). Crown feathers in moult with newer feathers slightly richer rufous. On the immature female paratypes the crown is less richly coloured than that of the holotype (240 Kingfisher Rufous) and the cinnamon-rufous lores and feathering on the orbital ring and immediately behind the eye are replaced by rufous-buff( 1 18 Warm Buff). The rufous crown feathering extends from the bill only as far back as the anterior of the mid-crown where dark-grey feathers predominate, imparting an overall greyish-brown colour (129 Dark Brownish Olive) to the hindcrown. Plate 3. Specimens of the haiotype of Orthotomus chaktomuk and closely related species (all males); (a) dorsal, (b) ventral and (c) lateral views (from left to right of 0. chaktomuk, O. sepium sepium, 0. ruficeps cineraceus and 0. atrogularis nitidus). PHOTOGRAPHS BY J. EATON (a,b,f), J. C. EAMES (c,d,j,k), A. J. I. JOHN (e,g,h,i) Forktail 29 (201 3) A new species of lowland tailorbird (Orthotomus) from the Mekong floodplain of Cambodia 5 Plate 4. Orthotomus chaktomuk (a-b) adult males in fresh plumage, 21 November 2012; (c-d) adult male in active moult (holotype), 8 August 2012 (NHMUK 2012.9.1); (e) adult male in worn plumage, 29 July 2012; (f) adult female in fresh plumage, 21 November 2012; (g) adult female in worn plumage, 29 July 2012; (h) immature male pre-moult, 16 July 2012; (i) immature male in active moult, 29 July 2012; (j) immature female in active moult, 8 August 2012 (NHMUK 201 2.9.5); (k) sub-adult female in active moult, 9 August 201 2 (NHMUK 2012.9.4;). 6 S. P. MAHOODeto/. Forktail 29 (2013) Five blackish rictal bristles per side, anterior two c.3 mm, twice the length of posterior three. Ear-coverts, cheeks and moustachial stripe almost white contrasting strongly with crown and underparts; however, feathers have buff (124 Buff) tips imparting an off-white wash. Feathers of submoustachial stripe and malar stripe white with very dark grey (82 Blackish Neutral Gray) bases and sometimes tips and fringes; white predominates, giving an impression of white speckling on a blackish base and contrasting strongly with the whitish cheeks. The malar stripe on the immature paratypes is quite different to that of the adult male specimens. It is made up of white feathers with pale grey central portions (85 Light Neutral Gray) and therefore contrasts little with the cheeks. Upperparts Boundary between hindcrown and upper neck abrupt. Upper neck, mantle and rump concolorous mid-grey (84 Medium Neutral Gray), slightly blue-toned approaching 78 Plumbeous (all feathers fresh and body moult apparently completed). Feathers on mantle and particularly rump relatively long and filamentous. Wings Wings of all prepared specimens in active moult (Table 1). On all specimens, fresh adult feathers are slightly darker grey (83 Dark Neutral Gray) than mantle, tinged very slightly brownish with mid¬ grey (84 Medium Neutral Gray) fringing (slightly broader on outer webs). Fresh primaries with off-white inner webs; worn adult rectrices buffy-brown (239 Ground Cinnamon) lacking fringing or pale webs. Underside of remiges dull silver-grey (84 Medium Neutral Gray). Underwing-coverts paler grey (85 Light Neutral Gray). Alula and axillaries contrastingwhite. Unmoulted rectrices of the immature male paratype and immature female paratypes differ strikingly from those of adult male specimens in being fringed bright olive-green (50 Yellowish Olive-Green). Tail Slightly rounded, outermost pair of rectrices 7 mm shorter than central pair. T ail of holotype very worn, buffy-brown (239 Ground Cinnamon), dorsal side slightly darker than ventral but heavily worn. Whitish-buff terminal tips just visible on all but central rectrices. Tail of immature female paratype (NHMUK 2012.9.4) less worn than that of the holotype (and other paratypes) and is dark greyish-brown (21 Fuscous) with broader whitish tips than those shown by other specimens. Field observations indicate that fresh tail feathers are mid-grey (similar in colouration to fresh wing feathers and therefore probably 83 Dark Neutral Gray or 84 Medium Neutral Gray) with a blackish-grey subterminal band (c. 1 cm wide) and whitish tips. Underparts The holotype shows white chin feathers with very dark grey (82 Blackish Neutral Gray) bases, tips and fringes, therefore darker overall than feathers on malar stripe, the latter overhanging those on throat. In the holotype, feathers of throat in an advanced stage of moult, some feathers in pin visible. Throat similar in colouration to chin although with much less white; feathers almost entirely solid dark grey (82 Blackish Neutral Gray) gradually becoming darker towards the breast (83 Dark Neutral Gray) with some white tips throughout. On the breast some dark grey feathers (83 Dark Neutral Gray) possess contrasting white rachis and base of barbs on distal two-thirds of feather, creating a pattern of whitish drops on a mid-grey background. On the edges of the breast, solid mid¬ grey (84 Medium Neutral Gray) feathers predominate. On the adult male paratype (NHMUK 20 12.9.2), the whitish drop-shaped marks on the breast are better developed than on the holotype and extend onto the throat, perhaps because the darker fringes are more worn. Field observations indicate that there is variation in the extent and intensity of dark throat-streaking in males (Media tiles SOM 1-2). The boundary between breast and belly is gradual; feathers tend towards lighter grey on belly (86 Pale Neutral Grey) and flanks (85 Light Neutral Gray). Flank feathers are relatively long. Vent greyish-white (paler than 86 Pale Neutral Gray). Thighs whitish-cinnamon (6 Salmon). The underparts of the immature male paratype (NHMUK 2012.9.3) differ from those of the adult male holotype in being paler with reduced dark grey on the throat and upper breast. There is an almost complete lack of dark tones on the throat, and the very dark grey (82 Blackish Neutral Gray) area on the throat is much smaller and barely extends onto the breast. On the throat and breast, feathers with white shafts and distal portions are more abundant than on the holotype, giving the throat a more speckled appearance. On the breast, solid white and pale grey feathers predominate such that the overall colour is whitish-grey (86 Pale Neutral Gray) rapidly grading to off-white on the belly. Flank feathers of the immature male are slightly whiter than those of the holotype. The underparts of the immature female paratypes are even paler than those of the immature male and almost completely lack dark tones. In those two specimens the chin and throat are white. Although there is a small area of mid-grey (84 Medium Neutral Gray) on the sides of the upper breast and the flank feathers are pale grey (86 Pale Neutral Gray or 85 Light Neutral Gray), the underparts are otherwise off-white. Bare parts Upper mandible dark horn, lower mandible pink horn, paler and pinker at base (more extensively pink on adult male paratype). Bill slender. Culmen decurved close to tip, not strongly carinated, tip very slightly hooked. Gonys convex. Tarsus and toes pinkish (slightly darker in adult male paratype, paler in immature male paratype); soles of the feet pale pink. Claws pale brownish pink, becoming paler towards tips. On female paratypes tarsi, feet, soles and lower mandibles are paler than those of the holotype. Iris orange-brown. Inside of mouth pale pink. Description of vocalisations For clarity we use the following terminology to describe vocalisations: note - a single song element; strophe - a continuous flow of notes, separated from other strophes by silent pauses; phrase - one or more strophes given in quick succession; and song - one or more phrases given in quick succession; strophe pace - number of notes per strophe/strophe length; phrase pace (for phrases with more than one strophe) - phrase duration/strophes per phrase. Note that recordings varied in length and quality, so only those with good quality strophes were analysed. Male O. chaktomuk songs are lengthy, often lasting more than one minute (Figure Slo-s, Slu, Media Files SOM 1-6). They consist of multiple phrases repeated at intervals of 0.42-4.30 seconds, typically much shorter than the maximum interval (mean: 1.7 seconds). Phrases are made up of 2-5 strophes, which are given at 0.12-0.95 second intervals. Males occasionally switch to a different strophe type mid-way through a song, although not within the same phrase. Strophes are also sometimes given singly. Strophes are trilled, consist of 3-18 notes and typically last 0. 17-0.49 seconds (Table 3). Twelve distinct male strophe types are known, ranging from up, down or ‘overslurred’ (the latter referring to sequences of notes that rise and then fall) trills (often with a louder initial or terminal note) to a mix of trilled notes and upslurs, downslurs or ‘overslurs’ (Table 3, Figure S la-1, Media Files SOM 1-6). Within strophe type, number of notes varies slightly (Table 4). Female O. chaktomuk vocalisations are typically emitted whilst the male is vocalising, but are sometimes given between male vocalisations (Figure Slo-s, Slu, Media File SI -6). Females give a stereotyped trill at a higher frequency than male vocalisations Forktail 29 (2013) A new species of lowland tailorbird ( Orthotomus ) from the Mekong floodplain of Cambodia 7 Table 3. Transliterations and univariate summary statistics of measurements of strophe characteristics of Orthotomus chaktomuk. Figure no. No. notes Length (s) Notes per second Max freq.(Hz) Min freq.(Hz) Bandwidth (Hz) Transliteration Male Sla 5-6 0.35-0.42 0.35-0.42 4,616-5,026 2,291-2,394 2,325-2,632 pi'pi'pi'Pm (5.5,07,2) (0.39,0.05,2) (0.07,0,2) (4,821,290,2) (2,343,73,2) (2,479,217,2) Sib 9-12 0.30-0.41 0.03-0.04 3,780-4,657 1,485-1,957 1,991-3,037 Chu'u'uVuVURH (10.6,1.0,25) (0.36,0.03,25) (0.03,0,25) (4,248,293,25) (1,766,136,25) (2,481,334,25) Sic 7-8 0.31-0.38 0.04-0.05 3,864-4,308 1,402-1,778 2,154-2,906 pu'ru'ru'RuWRU'RU (7.9,0.2,18) (0.37,0.02,18) (0.05,0,18) (4,101,143,18) (1,656,106,18) (2,444,201,18) Sid 12-14 0.38-0.47 0.03-0.03 3,658-4,206 1,504-2,667 1,368-2,668 TrWR'R'R'R'R'R'R'r (13.3,0.8,15) (0.43,0.03,15) (0.03,0,15) (3,946,156,15) (1,901,316,15) (2,045,294,15) Sle 10-18 0.28-0.40 0.02-0.03 4,306-4,545 1,547-1,886 2,602-2,998 EEeee\EEU (14.3,2.7,6) (0.32,0.04,6) (0.02,0,6) (4,393,102,6) (1,700,117,6) (2,694,154,6) Slf 9-12 0.27-0.35 0.03-0.03 4,836-5,609 1,499-1,838 3,192-4,110 Pleee/RRiEt (10.6,0.9,8) (0.31,0.03,8) (0.03,0,8) (5,271,112,8) (1,705,112,8) (3,566,297,8) Slg 10-16 0.33-0.49 0.03-0.04 2,839-4,286 1,282-1,890 1,183-3,004 PTeur'RWRW (12.1,1.7,54) (0.39,0.03,54) (0.03,0,54) (3,779,394,54) (1,608,119,54) (2,171,395,54) Slh 10-13 0.23-0.29 0.02-0.02 4,387-4,758 1,653-2,126 2,261-2,902 cheee\UP (11.1,0.9,7) (0.26,0.02,7) (0.02,0,7) (4,474,133,7) (1,866,196,7) (2,608,258,7) Sli 13-15 0.29-0.36 0.02-0.02 5,669-6,311 1,451-1,991 3,880-4,725 reeMuET! (14.0,0.9,8) (0.32,0.03,8) (0.02,0,8) (6,011,201,8) (1,687,178,8) (4,324,337,8) sij 3-4 0.17-0.24 0.05-0.06 3,830-4,821 1,128-2,120 2,394-3,522 P'p'Biu (3.5,0.5,33) (0.21,0.02,33) (0.06,0,33) (4,455,239,33) (1,610,259,33) (2,844,252,33) Slk 3-3 0.22-0.23 0.07-0.08 4,616-5,026 1,881-2,017 2,633-3,043 pi'pi'pui (3.0, 0,5) (0.23,0.01,5) (0.08,0,5) (4,828,153,5) (1,963,52,5) (2,865,152,5) S1I 3-4 0.19-0.27 0.06-0.07 3,680-4,240 1,573-1,949 1,731-2,633 Bi'Bi'BBit (3.1,0.3,18) (0.20,0.02,18) (0.07,0,18) (4,097,135,18) (1,742,115,18) (2,355,196,18) Female Sim 5-16 0.24-0.84 0.05-0.05 4,725-5,805 2,086-3,544 1,485-3,213 Tcri'i'i'i'i'iYi'iYi'i (9.4,3.4,25) (0.48,0.18,25) (0.05,0,25) (5,038,260,25) (2,673,398,25) (2,365,421,25) Sin 1-1 0.07-0.23 0.07-0.23 4,484-6,243 2,460-4.252 1,553-2,354 tew (1,0,17) (0.15,0.0717) (0.15,0.0717) (5,176,409,17) (3,189,376,17) (1,987,280,17) Analyses based on 12 recordings of male vocalisations from six pairs of Orthotomus chaktomuk in Kandal province, Cambodia, obtained as follows: (1) four pairs from c.40 km south-east of Phnom Penh (at or very close to 1 1°19'45.77"N 105°11'48.41"E);(2) one pair from c.15 km north of Phnom Penh (11°41'37.14"N 104°53'43.79"E); and (3) one pair from c.30 km north of Phnom Penh (1 1°50'18.24"N 104°49'26.55”E). Measurements taken in Raven Pro 1 .4 (Raven 2012). Values given are: minimum-maximum (mean; sd; sample size). In vocal transcriptions, notation follows Rasmussen & Anderton (2005). (typically 5-16 notes lasting 0.24-0.84 seconds; Table 3, Figure Sim, Media File S 1 — 6) . Females, and exceptionally males, sometimes produce a nasal squeak consisting of a single note with harmonics (Figure Sin, Media File S6). This vocalisation is usually given singly (Figure Sip), but occasionally more than one is repeated in quick succession; when many squeaks are given in sequence the first is usually longer than others (Figure Sit, Media File S6). Etymology The specific epithet ‘ chaktomuk ’ is a Khmer word meaning ‘four faces’. It is used in reference to the low-lying area at which the T onle Sap, Bassac and Mekong rivers come together to form an ‘X’ centred on Phnom Penh, itself historically known as ‘Krong Chaktomuk’ (literally ‘City of Four Faces’). Based on current knowledge, the global distribution of the new species is restricted to scrub within the dynamic floodplain created by the confluence of these waters. We use chaktomuk as a noun in apposition to the genus name, and it is thus invariable. Nomenclatural acts The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature (International Commission on Zoological Nomenclature 2012), and hence the new name contained herein is available under that Code from the electronic edition of this article. This published work and the nomenclatural act it contains have been registered in ZooBank, the online registration system for the International Commission of Zoological Nomenclature. The ZooBank Life Science Identifiers (LSIDs) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix “http://zoobank.org/”. The LSID for this publication is: urn:lsid:zoobank.org:pub:F177849T B6EE-4225-95B2-2843B32CBA08. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the digital repository BioTaxa (http://biotaxa.org). ECOLOGY AND BEHAVIOUR Habitat All observations of O. chaktomuk were made on level ground in very dense humid evergreen scrub (multi-stemmed woody plants, 2-6 m tall), sometimes admixed with long grasses or trees (Plate 5), at elevations of 3-25 m above sea level. Trees occur exceptionally; where present they are typically scarce, the scrub forming a dense layer with occasional tree canopies emerging from 8 S. P. MAHOODefa/. Forktail 29 (2013) ') Plate 5. Habitat at the type locality of Orthotomus chaktomuk. it. Orthotomus chaktomuk has not been seen in forest (defined as a habitat where trees predominate) and is therefore assumed to be absent from it. At all locations where birds have been found, the scrub is located within a floodplain and experiences seasonal or permanent (artificial) flooding. The presence of seasonally flooded scrub in any location is probably typically transitory, since in the absence of disturbance by people, large ungulates or hydrological processes it would presumably revert to seasonally flooded forest. Orthotomus atrogularis is sometimes found in seasonally flooded scrub occupied by 0. chaktomuk (Figure 1, Table SOM 1). Where the two species are syn topic, O. atrogularis is much the rarer. Typically O. atrogularis is found in the edge and canopy of taller forest habitats, showing a preference for disturbed and secondary forest because these offer an abundance of vines (Madge 2006, Wells 2007). In some parts of the Tonle Sap floodplain where 0. chaktomuk is absent, O. atrogularis is common in seasonally flooded forest and scrub, presumably because this habitat also offers an abundance of edge surfaces. At the other end of the habitat continuum, O. sutorius replaces O. chaktomuk in open scrub and gardens, although at some locations the two species are syntopic, even vocalising from the same individual plants (SPM pers. obs.). Orthotomus chaktomuk possibly occupies a habitat intermediate between those of O. atrogularis and 0. sutorius. However, as in other geographic areas where more than one lowland tailorbird species is present, habitat niches are difficult to define and distinguish. Ecological interactions and habitat associations of O. chaktomuk and other lowland tailorbirds are worthy of further research. Birds sharing the habitat of O. chaktomuk include widespread species often associated with gardens, e.g. Yellow-vented Bulbul Pycnonotus goiavier. Pied Fantail Rhipidura javanica, Oriental Magpie Robin Copsychus saularis, sometimes O. sutorius , and species usually associated with dense lowland humid evergreen scrub, including Striped Tit Babbler Macronous gularis. Yellow-bellied Prinia Prinia flaviventris, Plain Prinia Prinia inornata , Olive- backed Sunbird Cinnyris jugularis and O. atrogularis. From October to April, Palearctic migrants (e.g. Dusky Warbler Phylloscopus fuscatus and Siberian Rubythroat Luscinia calliope ) are abundant in this habitat. In locations where it occurs, O. chaktomuk often appears to be one of the most abundant bird species. Behaviour Owing to the structural characteristics of its habitat, O. chaktomuk is rarely seen without the aid of playback of vocalisations, and thus data on ‘normal’ behaviour are few. Almost all encounters have been with what appear to be adult male-female pairs, or adult male- female pairs with one subadult. Prior to moult, immature birds were seen singly, in male-female pairs of exclusively immature birds or in male-female pairs consisting of one immature and one adult bird. Birds usually stay within dense vegetation, where they glean and sally-glean from live and dead leaves of multi-stemmed bushes and occasionally vines, from ground-level to canopy. Orthotomus chaktomuk has not been observed foraging in trees. When vegetation is flooded, birds typically forage below the crown of the bush, on hanging branches just above the water. One individual that was lured out of dense scrub with the aid of playback foraged on long grass- stems, gleaning leaves of a vine that was growing amongst the grass. Individuals have been observed taking the following prey (once each): a small fly Diptera, a small spider Araneae, a small caterpillar Lepidoptera and a small katydid Tettigoniidae; all were consumed immediately. In response to playback, birds that have approached the observer have been seen to sing, usually in a duet, while perched (usually on or near the top of vegetation, including trees; Media F iles SOM 1 - 3) and occasionally in song flight. Singing is sometimes accompanied by rapid downwards tail-wagging. Sometimes, while singing in duet, perched birds droop and shiver their wings. Immature males gave a simpler, less developed song than adults. During March and April only males responded strongly to playback of vocalisations; females typically did not respond, or did so only briefly. Because this was in stark contrast to behaviour at other times of year it is thought to indicate that females were on the nest. Although there are no data on the timing of breeding of lowland tailorbirds in Cambodia, in Thailand O. atrogularis pairs with dependent young have been recorded from July to early September (Round 2008). The nest and eggs of O. chaktomuk remain to be described. Distribution The distribution of O. chaktomuk is incompletely known. It is apparently constrained by the distribution of seasonally flooded dense scrub within the floodplain of the Tonle Sap, Mekong and Bassac rivers in Cambodia (Figure 1). However, based on current data it is absent from part of this floodplain. Searches at various locations in apparently suitable habitat in the Tonle Sap floodplain have thus lar only found the species in the south-east (see Table SOM 1 for a list of all locations in the floodplain of the Mekong, Tonle Sap and Bassac rivers where searches for O. chaktomuk have been conducted). In the north of the Tonle Sap floodplain (where we have searched for and not found O. chaktomuk ), 0. atrogularis is abundant in habitat that is superficially structurally similar to habitat in the south-east, and it is unclear how far north and west along the lakeshore the distribution of O. chaktomuk extends. There is no biogeographic reason why O. chaktomuk should be absent from parts of the Tonle Sap floodplain, and the causes of its absence are unknown; O. atrogularis is scarce or absent at sites where O. chaktomuk was recorded (Table SOM 1). Orthotomus chaktomuk was not found in seemingly appropriate small seasonally flooded scrub patches at the northern limit of the Mekong floodplain (12°36'27.52"N 106° 01'36.06"E) in Kratie province (Table SOM 1, J. A. Eaton verbally 2012). Satellite data indicate that there is little, if any, suitable habitat for O. chaktomuk in the Mekong floodplain in Vietnam and it is currently unrecorded there (although no specific searches have been conducted). As might be expected, we have located only O. atrogularis in scrub habitats outside of the Mekong, Tonle Sap and Bassac floodplain (where these records were within 1 0 km of superficially suitable habitat for O. chaktomuk they are mapped on Figure 1). Based on current knowledge of its range, the distribution of O. chaktomuk covers less than c. 10,000 km2 (Figure 1); it therefore can be considered a restricted-range species (sensu Stattersfield etal. 1998). Conservation Orthotomus chaktomuk is restricted in distribution. Suitable habitat is patchy outside of the Tonle Sap floodplain and in the latter its Forktail 29 (2013) A new species of lowland tailorbird ( Orthotomus ) from the Mekong floodplain of Cambodia 9 distribution is poorly understood. Trends in loss, degradation and fragmentation of floodplain scrub are poorly documented and subject to considerable local variation (e.g. Packman et al. 2013). However, most floodplain scrub in Cambodia occupies land suitable for rice cultivation and could be further threatened by changes in ongoing burning, fuel-wood collection, cattle grazing (all of which potentially have a dual role because they also serve to slow succession) and the spread of the invasive plant Mimosa pigra. Ironically, O. chaktomuk might now be dependent on human activity to keep suitable scrubby habitat from becoming forest, since other anthropogenic impacts — eradication of wild ungulates, replacement of domestic animals by machines, water flow/level control, and changes in agricultural practices such as fallows and cyclical abandonment — have greatly curtailed processes that maintained the scrub. The species occurs in one protected area, Baray Bengal Florican Conservation Area, although at that site habitat is managed to maximise the area of grassland. It has already been lost from one site (Kraing Check) where birds were netted in 2009: visits in late 2012 found no birds and all suitable habitat had been converted to aquaculture ponds. We believe that 0. chaktomuk should be classified as Near Threatened on the IUCN Red List because it approaches the thresholds for Vulnerable under criteria Bla+bi,ii,iii,iv (IUCN 2001). Its Extent of Occurrence is 9,385 km2 and thus below the threshold for Vulnerable (<20,000 km2; criterion Bl). Although most locations where it occurs are small and isolated it has been found in the Tonle Sap floodplain where there is a large area of apparently suitable habitat (although it apparently does not occupy all of it). Because of this, its habitat cannot be considered severely fragmented (subcriterion a). Nonetheless it is inferred to be undergoing a continuing decline (subcriterion b) in (i) extent of occurrence, (ii) area of occupancy, (iii) area, extent and/or quality ofhabitat, and (iv) number of locations or subpopulations. Its Area of Occupancy has not yet been evaluated owing to uncertainty regarding both the distribution of suitable habitat and its distribution within apparently suitable habitat. Notwithstanding this assessment, if the species is found to be more widely distributed in the Tonle Sap floodplain, then it would warrant downlisting to Least Concern. Ongoing habitat loss is likely to be exacerbated by the impacts of hydropower development on the Mekong and its tributaries. Models of the effects of hydropower dams predict changes in the duration and size of the annual flood-pulse that will lead to a reduction in the extent of seasonally flooded habitats (Arias et al. 2012). Dam construction will also reduce fish populations (the primary protein source in rural Cambodia), cause changes in flood regime and lead to water shortages in the floodplain (Orr et al. 2012). These changes will probably lead to additional loss of floodplain scrub owing to expansion of agricultural land for rice production, fish ponds and grazing land for cattle. Construction has started on one mainstream lower Mekong dam (Xayaburi, in northern Lao PDR) and numerous tributary dams and ‘pre-construction’ works are thought to have begun on another (Don Sahong) in the far south of Lao PDR (International Rivers 2012); nine more mainstream dams are planned (Mekong River Commission 2011). TAXONOMIC CONSIDERATIONS Higher-level systematics Assignment of the new species to the genus Orthotomus is straightforward based on its overall structure, plumage and habits, which typify this genus. Genetic analysis provide additional support for this arrangement, as DNA sequence comparisons (detailed below) included specimens of the type species for the genus: O. sepium Horsfield, 1821. Orthotomus was previously placed in an expanded Sylviidae (Sibley & Ahlquist 1990) until that family was shown to be paraphyletic and the genus was transferred to the Cisticolidae, along with the cisticolas ( Cisticola ), prinias ( Prinia ) and a number of other genera (Alstrom et al. 2006). Within the Cisticolidae, Orthotomus occupies a clade with the genera Heliolais , Prinia and Urorhipis (Olsson et al. 2013). Orthotomus remained intact until molecular evidence led to the removal of four superficially similar species: the two African tailorbird species were transferred to the resurrected genus Scepomycter (Nguembock et al. 2007) and the Asian mountain tailorbirds, Mountain Tailorbird Phyllergates cucullatus and Rufous-headed Tailorbird P. heterolaemus, were shown to be not particularly closely related to ‘lowland tailorbirds’ — the most appropriate English group name for the remaining Orthotomus species (Alstrom et al. 2006, 2011). The Orthotomus comparisons by Sheldon et al. (2012), based on one mitochondrial and two nuclear DNA markers, suggested that O. sutorius is sister to the rest of the lowland tailorbirds, which comprise four relatively divergent clades: (l) Rufous-fronted Tailorbird O. frontalis, (2) Rufous-tailed Tailorbird 0. sericeus, (3) O. atrogularis-O. ruficeps-O. sepium, and (4) the rest of the Philippine endemics. Taxonomic implications from morphology Lowland tailorbirds appear to be relatively conservative in the evolution of distinctive plumage. For example O. atrogularis , the widespread O. sutorius and the Philippine endemics Philippine Tailorbird O. castaneiceps, 0. frontalis and Grey-backed Tailorbird O. derhianus share the same basic plumage pattern and colouration, but are not particularly closely related (Sheldon etal. 2012). Owing to this morphological congruence, various Philippine taxa were long considered part of O. atrogularis (Delacour & Mayr 1946). These were later split from that species and grouped under O. castaneiceps and 0. derhianus (Dickinson et al. 1991, Kennedy et al. 2000). Collar (2011) considered that morphological data alone were insufficient to afford O. casteneiceps chloronotus species status to resolve the geographically vexing situation created by the specific recognition of O. derhianus. Equally, treatment of O. frontalis as a species (proposed by Madge 2006) was deemed untenable using morphological information alone (Collar 2011). The plumage of 0. chaktomuk is typical of the O. atrogularis- 0. ruficeps-O. sepium clade. Within this grouping, all species are characterised by a rulous crown; white or rufous cheeks; grey, olive- grey or bright olive-green upperparts; and grey or whitish-grey underparts with or without heavy blackish throat-streaking (Table 2, Plate 3). Superficially, the head and underparts pattern and colouration of O. chaktomuk are similar to O. atrogularis, while colouration of the upperparts is more similar to 0. ruficeps. However, 0. chaktomuk shows a suite of plumage features that in combination are unique, and there are various additional subtle plumage differences between it and closely related species (Table 2, Plates 3,4). Immatures of all species within the clade possess yellow- olive fringing on the wing-feathers; adult O. atrogularis exhibit the same colouration on the wings, tail and much of the upperparts. Examination of specimens suggests that this colouration is not as vivid in immature O. ruficeps and 0. sepium as in 0. atrogularis and O. chaktomuk. In common with other lowland tailorbird species, those in the O. atrogularis-O. ruficeps-O. sepium clade show relatively minor geographic variation in morphology. All three species within the clade are polytypic. Although a detailed examination of morphological variation within these species was not completed, examples (including some type material) of multiple subspecies were examined (Table S2). Within each species, all examined specimens were superficially very similar. Morphological variation within 0. atrogularis is most marked along the Sabah-Sarawak border area, in common with other species that share a similar distribution (e.g. 10 S. P. MAHOODeta/. Forktail 29 (2013) White-rumped Shama Copsycbus malabaricus). The most morphologically divergent taxon within O. ruficeps , O. r. cagayanensis of Cagayan Sula, Philippines, which is apparently extensively brown or rufous-washed above with pale eyes (Madge 2006), was not examined. Within this context of limited geographic variation within species, any suggestion that O. chaktomuk should be considered a highly distinctive localised subspecies of O. atrogularis is untenable. Equally, there is no evidence that any lowland tailorbird species possesses regularly occurring colour morphs. Biometrics ol O. chaktomuk are similar to other species within the O. atrogularis- O. ruficeps-O. sepium clade (Table SOM 3). Using bill length/wing length as a proxy lor size, there is an indication that 0. chaktomuk is smaller than closely related species (Figure 2). This could not be confirmed statistically, because the sample size of O. chaktomuk is too small (Table SOM 3). In the field, O. chaktomuk appeared to have a shorter tail than local O. atrogularis. However, this could not be confirmed from the specimens collected of O. chaktomuk , because their tails are very worn. The description of a new species provides an opportunity to test the quantitative criteria for species delineation proposed by Tobias et al. (2010). These criteria use a scoring system for morphological, vocal and ecological features to assess taxonomic rank. Even when applied only to morphological features, O. chaktomuk exceeds the threshold score (7) for species status when compared with O. atrogularis, 0. ruficeps and 0. sepium. It scores 8 against O. ruficeps-. cheeks white rather than orange-rufous (3), throat and breast very dark grey with white speckling rather than unmarked pale grey (3), thighs whitish-cinnamon rather than orange-rufous (2). Its scores against O. sepium are similar. It scores 8 against O. atrogularis-. upperparts mid-grey rather than bright yellowish-olive (3), vent white rather than yellow (3), thighs whitish- cinnamon rather than yellowish-orange (2). Taxonomic implications from vocalisations Vocal data reaffirm the close relationship between O. chaktomuk and 0. ruficeps, O. sepium and O. atrogularis, in particular the last. Orthotomus vocalisations are difficult to define. Males and females often duet, or if three birds are present, all will vocalise simultaneously. Vocalisations of O. chaktomuk are typically lengthy, and those of males are extremely varied (Media File SOM 4-6). We do not think that we have documented the full vocal range of the species. In addition, comparisons with closely related species were hampered because it is doubtful that the full vocal repertoire of such species has been documented. The vocalisations of species in the O. chaktomuk-O. atrogularis-O. ruficeps-O. sepium clade fall into two distinct types. Those of 0. ruficeps and O. sepium are largely short, pure-tone whistles, while those of O. chaktomuk and O. atrogularis are restricted to short trills. To quantify the distinctiveness in songs of O. chaktomuk in relation to other species within the clade, we conducted a discriminant analysis (DA) on male vocalisations using XLStat (Addinsofc 2013). We randomly selected one song (defined as above) from each individual of O. chaktomuk from which we had obtained recordings (total five individuals) and randomly selected a similar number of songs from five individuals each of O. a. nitidus (the geographically closest subspecies of O. atrogularis ), 0. ruficeps and O. sepium. These included three recordings of O. atrogularis from Cambodia (including one from the floodplain of the Tonle Sap) and one recording from Vietnam made within 100 km of the border with Cambodia. (For a full list of recordings used in analyses see Table SOM 4.) Recordings were downloaded from www.xeno-canto.org and http://avocet.zoology.msu.edu/. From each recording we calculated mean values of the following variables: notes/strophe, length (in seconds) of longest note (one per 0.4 0.38 0.36 0.34 0.32 0.3 0 28 0.26 0.24 0.22 0.2 (a) O. chaktomuk O. a. nindus O. r onereceaus O. s. sepium 0.4 0.38 0.36 0.34 0.32 0.3 0.28 0.26 0.24 0.22 0.2 (b) O. chaktomuk O. a. nitidus O. r. dnereceaus O. s. sepium 0.4 0 38 036 0.34 0.32 03 0 28 0.26 024 0.22 0.2 (c) O chaktomuk O. a. nitidus O. r onereceaus O. s. sepium Figure 2. Box plots of bill-length (to skull) divided by wing-length of (a) males of Orthotomus chaktomuk (n = 3), 0. atrogularis nitidus (n = 1 5), 0. ruficeps cineraceus (n = 14) and 0. sepium sepium (n = 20); (b) females of 0. chaktomuk (n = 2), 0. a. nitidus (n = 10), 0. r. cineraceus (n = 6) and 0. s. sepium (n = 8); (c) males and females of 0. chaktomuk (n = 5), 0. o. nitidus (n = 25), 0. r. cineraceus (n = 20) and O. s. sepium (n = 28). Small squares represent the median; box indicates 50% of samples; bars indicate maximum and minimum. strophe), strophe length (seconds), maximum and minimum fundamental frequencies (one each per strophe), number ol strophes per phrase, length ol interval between phrases (in seconds) and, for phrases with more than one strophe, interval between strophes and length ol phrase (in seconds); we also calculated bandwidth (maximum minus minimum fundamental trequency within a given strophe), strophe pace (number ol notes per strophe/ strophe length) and, for phrases with more than one strophe, phrase pace (phrase length/strophes per phrase). Forktail 29 (2013) A new species of lowland tailorbird ( Orthotomus ) from the Mekong floodplain of Cambodia 11 -10 -SO S 10 FI (94.65 \) Figure 3. Multivariate vocal space of lowland tailorbirds in the 0. chaktomuk-O. atrogularis-O. ruficeps-O. sepium clade from discriminant analyses based on twelve song traits. Scatter plot of the first two canonical functions that discriminated songs of 0. chaktomuk, O. atrogularis, O. ruficeps and 0. sepium. • represent 0. chaktomuk ; @1, 0. atrogularis-, A, 0. ruficeps; +, 0. sepium. Table 4. Results of discriminant analyses of songs of species in the 0. chaktomuk-O. atrogularis-O. ruficeps-O. sepium clade, based on 1 2 acoustic variables, showing first three (of four) canonical functions. Canonical function Acoustic variable 1 2 3 Notes per strophe -0.742 -0.196 -0.155 Strophe length (seconds) 0.221 0.265 -0.524 Strophe pace 0.721 0.331 -0.270 Maximum frequency 0.227 -0.147 -0.009 Minimum frequency -0.021 -0.087 -0.183 Bandwidth 0.247 -0.105 0.112 Length of longest note (one per strophe) 0.873 0.187 -0.289 Strophes per phrase -0.015 0.391 0.457 Inter-strophe interval (seconds) -0.179 0.172 -0.006 Phrase length (seconds) -0.103 0.507 0.156 Phrase pace -0.131 0.405 -0.075 Inter-phrase interval (seconds) -0.330 -0.152 -0.117 Eigenvalue 43.01 1.362 1.071 % variance explained 94.6 3.00 2.36 Most of the analysed songs of O. chaktomuk and O. atrogularis clustered separately in multivariate vocal space and could be discriminated from each other and from the songs of O. ruficeps and O. sepium (Wilks’s X= 0.005, F = 2.222, P = 0.047; Figure 3). Overall, the analysis assigned songs to the correct species with 60% accuracy. Songs of O. chaktomuk were classified with 60% accuracy and songs of O. atrogularis were classified with 40% accuracy. Songs of O. ruficeps and O. sepium were classified with 40% and 100% accuracy, respectively. The discriminant analysis was mainly influenced by the length of the longest note, strophe pace and the number of notes per strophe (Table 4). In accordance with their acoustic similarities, O. chaktomuk and O. atrogularis respond to playback of each other’s vocalisations, indicating that inter-specific territoriality is common. There is individual variation in the magnitude of response, but this is poorly understood at present. Interspecific territoriality is a common trait in avian sister species whose ranges come into contact (e.g. Orians & Willson 1964, Murray 1971, Murray 1976). At locations where O. chaktomuk is sympatric with O. sutorius , the latter sometimes also responds to broadcast of vocalisations of the former by ascending the vegetation and singing. Taxonomic implications from ecology The apparently restricted distribution of O. chaktomuk differs from those of other Orthotomus species on the Asian mainland, which are typically wide (Madge 2006). Lowland tailorbirds are thought to have originated in southern Asia or possibly Sundaland. They rapidly spread widely in Sundaland and the Philippines and, more recently, additional species have evolved in both island groups (Sheldon et al. 2012). The O. chaktomuk-O. atrogularis-O. ruficeps-O. sepium clade emerged relatively recently, and, in contrast to the clade containing Philippine endemics, exhibits lower species richness (four versus six species), presumably owing to the lack of opportunity for speciation in lowland populations in a continental mainland versus an oceanic island setting. Lowland passerines in mainland Asia typically have large distributions. Those with smaller distributions are largely confined to successional habitats in the floodplains of large rivers, such as Black-breasted Parrotbill Paradoxornis flavirostris and Marsh Babbler Pellorneum palustre (Rasmussen & Anderton 2005). T ailorbird habitats are notoriously difficult to define (e.g. Mitra & Sheldon 1993), although lowland tailorbirds within the 0. chaktomuk-O. atrogularis-O. ruficeps-O. sepium clade have slightly clearer habitat preferences. In this clade, greater specialisation is thought to have helped these younger species avoid competition with the pre-existing generalist species O. sutorius and O. sericeus (Sheldon et al. 2012). The habitat preferences of O. chaktomuk are thought to be somewhat intermediate between O. atrogularis and O. sutorius. In addition, O. chaktomuk is apparently confined to successional floodplain habitat. Vegetational patterns in what is now its distribution have been shaped by cyclical ice-ages and interglacials over the past two million years. During glacial maxima, the sea-level was much lower than today and southern Vietnam and Cambodia were connected by land to what is now Peninsular Malaysia by the now submerged Sunda Shelf (Sathiamurthy & Voris 2005). It is thought that there were four large river basins on the Sunda Shelf and these great river systems connected the freshwater riverine faunas of many of today’s rivers (such as the Mekong) that are now restricted to Indochina, the Malay Peninsula or one of the greater Sunda Islands (Voris 2000). Floodplain habitats were therefore probably much more extensive during glacial maxima and it is possible that O. chaktomuk evolved on the Sunda Shelf. Orthotomus chaktomuk is now restricted to a much smaller area, constrained by the reduced availability of suitable habitat during an interglacial. The known distribution of O. chaktomuk lies within the distribution of O. atrogularis. No parts of the distribution of O. chaktomuk are more than a few tens of kilometres from locations where O. atrogularis is found. The two species are locally syntopic (Figure 1, Table SOM 1); for instance, c.200 m from the type locality a pair of O. atrogularis was observed in the same bush as a single O. chaktomuk (SPM, AJIJ, HC pers. obs.; photographed). Although nearly 50 O. chaktomuk-O. chaktomuk male-female associations have been observed, neither mixed pairs nor birds that are phenotypically identifiable as hybrids have been detected. Taxonomic implications from molecular analyses Within the genus Orthotomus , superficial morphological and vocal similarities between taxa have frequently clouded their taxonomic status. In this context, molecular techniques can provide a useful tool to infer relationships between taxa. T o determine the position of O. chaktomuk in the molecular phylogeny of Orthotomus we compared DNA sequences of mitochondrial ND2 and nuclear MUSK and TGF(32 genes of the type specimens with all other species of lowland tailorbird (Sheldon etal. 2012) (see Table SOM 5 for details of all specimens used in the genetic analyses). Tissues from the five O. chaktomuk specimens (Table 1) were preserved in 95% ethanol and stored in the University of Kansas Natural 12 S. P. MAHOODefo/. Forktail 29 (2013) History Museum (KUNHM) and Louisiana State University Museum of Natural Science (LSUMNS) tissue collections. DNA was extracted and sequenced following the protocol described in Sheldon et al. (2012), and the sequences deposited in GenBank: accession numbers KF015230-KF015247. The total number of DNA nucleotides was 1,041 of ND2, 614 of MUSK, and 613 of TGF(32. Separate and concatenated Bayesian phylogenetic analyses of these sequences using MrBayes ver. 3.2.1 (Ronquist et al. 2012) as in Sheldon et al. (2012) placed the novel taxon in a clade with 0. atrogiilaris , O. ruftceps and (). sepium ; all trees except TGF[32 placed 0. atrogiilaris and O. chaktomuk as sisters (Figure 4). This arrangement concurs with morphological and vocal analyses. The ND2 sequences of the four taxa had 90 variable and 78 parsimony informative sites. Between the one sample of O. atrogularis and the five samples of O. chaktomuk were 22 variable sites of which 12 consistently differed. The ND2 p-distance from O. chaktomuk to O. atrogularis averaged 1.3% (range 1 . 1 %— 1 .4%), to O. ruficeps 5.0% (4.6%-5.2%), and to O. sepium 6.5% (6.3%- 6.8%). Variation in the MUSK and TGFa2 sequences between O. chaktomuk and O. atrogularis was negligible (three and six sites, respectively). These genetic data are insufficient to resolve the relationship between O. chaktomuk and O. atrogularis owing to the small number of samples of 0. atrogularis compared (one) and because that sample was not of the subspecies sympatric with O. chaktomuk. Instead, the O. atrogularis sample was of the nominate subspecies collected in Sarawak, which is restricted to the Sundaic region (except Sabah, north Borneo). The genetic divergence between O. chaktomuk and the 0. atrogularis sample (1.3%) is small and broadly comparable to that between other lowland taxon-pairs on Borneo and mainland Asia that are considered subspecies, although there is considerable inter-species variation in genetic distances (e.g. Lim et al. 2010, Sheldon et al. 2012). A phylogeographic study including samples from all subspecies and biogeographically relevant populations of O. ruficeps , O. atrogularis and O. chaktomuk is required to clarify their evolutionary relationships. Relationship of Orthotomus chaktomuk to species within Orthotomus Orthotomus chaktomuk is locally syntopic (Figure 1, Table SOM 1) with the only species from which it shows apparently relatively low genetic divergence, O atrogularis. We have found no evidence of hybridisation, and the taxa satisfy the precepts of the biological species concept because they behave like separate species when they come into contact (e.g. Mayr 1963, 1999). Classification of the new taxon as a subspecies or highly localised colour morph of O. atrogularis is therefore untenable. Because O. chaktomuk is on a distinct evolutionary trajectory it also satisfies the phylogenetic species concept (Cracraft 1989). Although reported genetic divergence among sister species is typically lower in temperate regions than in the tropics this may be an artefact of incomplete sampling and incorrect taxonomy (Tobias etal. 2008, Sangster 2009). Recent studies are overturning the trend for lumping distinctive taxa into polytypic species and revealing cryptic diversity in widespread species (e.g. Collar 2006, 2011, Rheindt & Eaton 2010, Leader 2011, Moltesen et al. 2012, Rasmussen et al. 2012). By taking an integrated approach to taxonomy (as here), sister species are being recognised in tropical regions that differ in sampled regions of nuclear DNA by levels that are in line with those used in temperate regions (e.g. Irestedt etal. 2013). There are a number of plausible explanations for the apparently low genetic divergence between O. chaktomuk and O. atrogularis. The molecular phytogenies suggest that O. chaktomuk might be a relatively young lineage. Its diagnostic phenotypic traits that apparently prevent modern hybridisation might be encoded by a 100 100 100 J-L V: 100 100 q 100 - O. atrogularis 57037 O. chaktomuk 201293 O. chaktomuk 201294 O. chaktomuk 201295 O. chaktomuk 201291 O. chaktomuk 201292 O. ruficeps 47143 O. ruficeps 17284 O. ruficeps J 1 1 56 r O. sepium 219589 O. sepium 269052 O. sepium 56705 O. sepium 220247 O. sepium 703336 f 0.02 Figure 4. The Orthotomus chaktomuk, O. atrogularis, O. ruficeps, and 0. sepium clade extracted from the entire Orthotomus phylogeny, which was constructed from concatenated DNA sequences of ND2, MUSK, and TGFp2 via Bayesian phylogenetic inference as described in Sheldon et al. (2012). Numbers along branches indicate Bayesian posterior probabilities. The topology is the same as ND2 and MUSK trees by themselves. small number of genes that evolved rapidly under sexual selection (cf. Uy et al. 2009). This process might have occurred too rapidly for significant additional genetic differences to accumulate in parts of the genome not under intense selection or which are selected for other purposes (as are mitochondrial genes). If 0. chaktomuk is indeed a Sunda Shelf species, now confined to relict habitat in Indochina, then it might have been derived from Sundaic O. atrogularis rather than mainland populations. If this were the case then it might be expected to show greater genetic divergence from mainland O. atrogularis to which it is now locally syntopic than to the Sundaic nominate used here in the genetic analyses. An alternative explanation is that the apparent low genetic divergence between the species is a result of genetic introgression sometime during the last two million years (Rheindt & Edwards 2011). Periods of peak sea-level might be the most plausible time for genetic introgression to have occurred. A higher sea-level might have constrained suitable habitat for O. chaktomuk into a narrow band between the distributions of O. ruficeps and O. atrogularis. This process perhaps drove 0. chaktomuk through a population bottleneck and increased the chances of hybridisation with 0. atrogularis , leading to genetic introgression. If a comprehensive study of relationships within the O. chaktomuk-O. atrogularis-O. rufiiceps-O. sepium clade reveals that O. chaktomuk is more closely related to O. a. nitidus than to Sundaic taxa, then this is perhaps a more likely explanation for the low genetic divergence. Final remarks The modern discovery of an undescribed bird species close to sea- level within the limits of a large city in a populous country is extraordinary, but not unprecedented (cf. Stymphalornis sp. nov., an as-yet undescribed taxon restricted to marshes close to Sao Paulo, Brazil, discovered in 2005: Reinert et al. 2007). At least three interacting factors probably account for O. chaktomuk having gone unnoticed for so long. It inhabits a very small geographic range, and within this it is restricted to a very specific habitat type: dense floodplain scrub. This habitat is of little interest to birdwatchers and ornithologists because the other species that it supports are some of the most widespread and abundant birds in tropical South-East Asia. Even if its habitat were to attract more attention, the denseness of the habitat and the species’s skulking habits would more often than not render it invisible to the casual would-be observer. Vocalisations of O. chaktomuk are similar to those of O. atrogularis , which has a perplexing array of vocalisations with which birdwatchers rarely attempt to familiarise themselves fully. Orthotomus atrogularis is a common species within suitable habitat across a fairly broad range, and therefore there is little a priori reason Forktail 29 (2013) A new species of lowland tailorbird ( Orthotomus ) from the Mekong floodplain of Cambodia 13 for a birdwatcher or ornithologist to invest effort in trying to see a hidden, vocalising tailorbird in dense scrub in mainland South-East Asia. Moreover, collecting effort in Cambodia has been low: we have been able to trace only two O. atrogularis and one O. sutorius specimens (NHMUK 1928.6.26.1210, Eames & Ericson 1996) (the identification of all these specimens has been verified by the primary author, either first-hand or using photographs). Modern birdwatching effort in Cambodia is also limited and very localised. The factors discussed above also help explain why the first four individuals known were all mist-net captures. Their misidentification can be accounted for by the species’s superficial similarity to other species, observer inexperience and the sheer unlikelihood of alternative options (cf. Woxvold et al. 2009). The discovery of 0. chaktomuk indicates that new species of bird may still be found in familiar and unexpected locations. ACKNOWLEDGEMENTS We are most grateful to Keo Omaliss, Director of the Department of Wildlife and Biodiversity of the Forestry Administration of the Ministry of Agriculture, Forestry and Fisheries, Kingdom of Cambodia, for permission to collect and export scientific specimens; Mark Adams, Natural History Museum, Tring, and Steven van der Mije, Naturalis Biodiversity Centre, Leiden, for access to and high-quality images of scientific specimens; Will Duckworth for unwavering support and good advice; Philip Round and Craig Robson for their support for our initial identification; James Eaton for good advice and the use of his photographs and sound recordings; Robert Martin for the use his sound recordings; Martin Kennewell for his videos as supplementary online material; Tom Evans for his early role in the species’s discovery; Sarah Brook, Nigel Collar and John Pilgrim for their counsel; Phien Sayon for making the map; Harry Taylor for photographs of the type material at NHMUK; Angela Yang for information on the 2009 records; Andy Symes and Stuart Butchart at BirdLife International for advice on the conservation section; Normand David and Edward Dickinson for advice on nomenclature; Mauricio Arias for information on floodplain processes; Alexander Lees and Sidnei Dantas for help with sonagrams; Lucy Keatts, Chantha Yuthea and Tan Setha for help obtaining permits; Ulf Johansson for photographs of Orthotomus specimens from the Swedish Museum ofNatural History; Ann Mahood for assistance at NHMUK; Robert van Zalinge and Frederic Goes for field observations; Son Virak for assistance at Baray Bengal Florican Conservation Area; Tom Clements at WCS Cambodia for overlooking SPM’s absences in the field or museum, and for logistic support; and Frank Rheindt, Alexander Lees and five anonymous reviewers for comments that substantially improved the manuscript. REFERENCES Addinssoft (2013) XLStat. Available at: http://www.xlstat.com/en/. Alstrom, P., Ericson, P. G. P., Olsson, U. & Sundberg, P. (2006) Phylogeny and classification of the avian superfamily Sylvioidea. Mol. Phylogen. Evol. 38: 381-397. Alstrom, P., Davidson, P„ Duckworth, J. W., Eames, J. C„ Le Trong Trai, Nguyen Cu, Olsson, U„ Robson, C. &Timmins, R. J. (2010) Description of a new species of Phylloscopus warbler from Vietnam and Laos. Ibis 1 52: 145-168. Alstrom, P„ Hohna, S., Gelang, M., Ericson, P. G. P. & Olsson, U. (2011) Non- monophyly and intricate morphological evolution within the avian family Cettiidae revealed by multilocus analysis of a taxonomically densely sampled dataset. BMC Evol. Biol. 1 1 : 352. Arias, M. E., Cochrane, T. A., Piman.T., Kummu, M., Caruso, B. S. & Killeen, T. J. (2012) Quantifying changes in flooding and habitats in theTonle Sap Lake (Cambodia) caused by water infrastructure development and climate change in the Mekong Basin. J. Env. Mgmt. 1 12: 53-66. Cracraft, J. (1989) Speciation and its ontology: the empirical consequences of alternative species concepts for understanding patterns and processes of differentiation. Pp.28-59 in D. Otte & J. A. Endler, eds. Speciation and its consequences. Sunderland, MA: Sinauer Associates. Collar, N. J. (2006) A partial revision of the Asian babblers (Timaliidae). Forktail 22:85-112. Collar, N. J. (201 1) Species limits in some Philippine birds including the Greater Flameback Chrysocolaptes lucidus. Forktail 27: 29-38. Delacour, J. & Mayr, E. (1 946) Birds of the Philippines. New York: Macmillan. Dickinson, E. C., Kennedy, R. S. & Parkes, K. C. ( 1 991 ) The birds of the Philippines: an annotated check-list. Tring, UK: British Ornithologists' Union (Checklist 12). Duckworth, J. W., Alstrom, P., Davidson, P, Evans,T. D., Poole, C. M., Setha, T. & Timmins, R. J. (2001) A new species of wagtail from the lower Mekong basin. Bull. Brit. Orn. Club 1 21 : 1 1 2-1 42. Eames, J. C. (2002) Eleven new subspecies of babbler (Timaliinae) from Kon Turn province, Vietnam. Bull. Brit. Orn. Club 122: 109-141. Eames, J. C. & Eames, C. (2001) A new species of laughingthrush Garrulax Passeriformes: Sylviinae: (Garrulacinae) from the central highlands of Vietnam. Bull. Brit. Orn. Club 121:1 0-23. Eames, J. C. & Ericson, P. G. P. (1996) The Bjorkegren expeditions to French Indochina: a collection of birds form Vietnam and Cambodia. Nat. Hist. Bull. Siam Soc. 44: 75-1 1 1 . Eames, J. C., Robson, C. R. & Nguyen Cu (1994) A new subspecies of Spectacled Fulvetta Alcippe ruficapilla from Vietnam. Forktail 1 0: 1 41 -1 58. Eames, J. C., LeTrong Trai, Nguyen Cu & Eve, R. (1 999a) New species of barwing Actinodura (Passeriformes: Sylviinae: Timaliini) from the Western Highlands of Vietnam. Ibis 141 : 1-10. Eames, J. C., Le Trong Trai & Nguyen Cu (1999b) A new species of laughingthrush Garrulax Passeriformes: Sylviinae: (Garrulacinae) from the Western Highlands of Vietnam. Bull. Brit. Orn. Club 1 19: 4-1 5. International Commission on Zoological Nomenclature (2012) Amendment of Articles 8, 9, 10, 21 and 78 of the International Code of Zoological Nomenclature to expand and refine methods of publication. ZooKeys 219: 1-10. International Rivers (2012) Laos begins work on a second Mekong river dam. Available at: http://www.internationalrivers.org. Irestedt, M., Fabre, P.-H., Henrique, B.-F„ Jonsson, K. A., Roselaar, C. S., Sangster, G. & Ericson, P. G. P. (2013) The spatio-temporal colonization and diversification across the Indo-Pacific by a 'great speciator' (Aves, Erythropitta erythrogaster). Proc. Roy. Soc. Lond. B. 280: DOLdx.doi.org/ 1 0.1 098/rspb.201 3.0309. IUCN (2001) IUCN Red List Categories and Criteria: Version 3. 7. Gland, Switzerland & Cambridge, UK: IUCN Species Survival Commission. Kennedy, R. S„ Gonzales, P. C., Dickinson, E. C., Miranda, H. C. & Fisher, T. H. (2000) A guide to the birds of the Philippines. Oxford: Oxford University Press. Le Trong Trai & Craik, R. C. (2008) Mekong Wagtail Motacilla samveasnae — resident breeder in Vietnam? BirdingASIA 8: 68-69. Leader, P. J. (201 1) Taxonomy of the Pacific Swift Apus pacificus Latham, 1802, complex. Bull. Brit. Orn. Club 131: 81-93. Lim, H. C., Zou, F., Taylor, S. S., Marks, B. D., Moyle, R. G., Voelker, G. & Sheldon, F. H. (2010) Phylogeny of magpie-robins and shamas (Aves: Turdidae: Copsychus and Trichixos ): implications for island biogeography in Southeast Asia .J.Biogeog. 37: 1894-1906. Madge, S. C. (2006) Family Cisticolidae (Cisticolas and allies) [Orthotomus accounts], Pp.422-427 in J. del Hoyo, A. Elliott & D. A. Christie, eds., Handbook of the birds of the world, 1 1 . Barcelona: Lynx Edicions. Mahood, S. P. & Martin, R. W. (201 3) Ashy Tailorbird Orthotomus ruficeps: the first records for Cambodia. BirdingASIA 19: 121-122. Mayr, E. (1963) Animal species and evolution. Cambridge, MA: Belknap Press of Harvard University Press. Mayr, E. (1999) Systematics and the origin of species, from the viewpoint of a zoologist. Cambridge, MA: Harvard University Press. Mekong River Commission (201 1) Planning atlas of the lower Mekong River basin. Phnom Penh and Vientiane: Mekong River Commission. Mitra, S. & Sheldon, F. H. (1993) Use of an exotic tree plantation by Bornean lowland forest birds. Auk 1 10: 529-540. 14 S. P. MAHOOD etal. Forktail 29 (2013) Moltesen, M„ Irestedt, M., Fjeldsa, J., Ericson, P. G. P. & Jonsson, K. A. (2012) Molecular phylogeny of Chloropsidae and Irenidae - cryptic species and biogeography. Mol. Phylogen. Evol. 65: 903-914. Murray, B.J., Jr. (1 971) The ecological consequences of interspecific territorial behaviour in birds. Ecology 52: 414-423. Murray, B. J„ Jr. (1976) A critique of interspecific territoriality and character convergence. Condor 78:51 8-525. Nguembock, B„ Fjelsa, J„ Tillier, A. & Pasquet, E. (2007) A phylogeny for the Cisticolidae (Aves: Passeriformes) based on nuclear and mitochondrial DNA sequence data, and a re-interpretation of an unique nest-building specialisation. Mol. Phylogen. Evol. 42: 272-286. Olsson, U„ Irestedt, M., Sangster, G„ Ericson, P. G. P. & Alstrom, P. (2013) Systematic revision of the avian family Cisticolidae based on a multi¬ locus phylogeny of all genera. Mol. Phylogen. Evol. 66: 790-799. Orians, G. H. & Willson, M. F. (1964) Interspecific territories of birds. Ecology 45: 736-745. Orr, S., Pittock, J., Chapagain, A. & Dumaresq, D. (2012) Dams on the Mekong River: lost fish protein and the implications for land and water resources. Glob. Env. Change 22: 925-932. Packman, C. E., Gray, T. N. E., Collar, N. J., Evans, T. D„ van Zalinge, R. N., Son Virak, Lovett, A. A. & Dolman, P. M. (2013) Rapid loss of Cambodia's grasslands. Conserv. Biol. 27: 245-247. Rasmussen, P. C. & Anderton, J. C. (2005) Birds of South Asia: the Ripley guide. Washington D.C. and Barcelona: Smithsonian Institution and Lynx Edicions. Rasmussen, P. C„ Allen, D. N. S., Collar, N. J., DeMeulemeester, B., Hutchinson, R. O., Jakosalem, P. G. C., Kennedy, R. S., Lambert, F. R. & Paguntalan, L. M. (2012) Vocal divergence and new species in the Philippine Hawk Owl Ninox philippensis complex. Forktail 28: 1-20. Raven (2012) Raven interactive sound analysis software. Available at http:// www.birds.cornell.edu/brp/raven/RavenVersions.html#Raven14. Reinert, B. L., Bornschein, M. R. & Firkowski, C. (2007) Distribuigao, tamanho populacional, habitat e conserva^ao do bicudinho-do-brejo Stymphalornis acutirostris Bornschein, Reinert e Teixeira, 1995 (Thamnophilidae). Rev. Bras. Ecol. 1 5: 493-5 1 9. Rheindt, F. E. & Eaton, J. A. (2010) Biological species limits in the banded pitta Pitta guajana. Forktail 26: 86-91 . Rheindt, F. E. & Edwards, S. V. (201 1) Genetic introgression: an integral but neglected component of speciation in birds. Auk 1 28: 620-632. Ronquist, F.,Teslenko, M„ van der Mark, P., Ayres, D. L., Darling, A., Hohna, S., Larget, B., Liu, L., Suchard, M. A. & Huelsenbeck, J. P. (2012) MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst. Biol. 61 : 539-542. Round, P. D. (2008) Birds of the Bangkok area. Bangkok: White Lotus Company. Sangster, G. (2009) Increasing numbers of bird species result from taxonomic progress, not taxonomic inflation. Proc. Roy. Soc. Lond. B. 276: 31 85-31 91 . Sathiamurthy, E. & Voris, H. K. (2006) Maps of the Holocene sea level transgression and submerged lakes on the Sunda Shelf. Nat. Hist. J. Chulalonghorn Univ. Supplement 2: 1-44. Sheldon, F. H., Oliveros, C. H., Taylor, S.T., McKay, B., Lim, H. C., Rahman, M. A., Mays, H. & Moyle, R. G. (2012) Molecular phylogeny and insular biogeography of the lowland tailorbirds of Southeast Asia (Cisticolidae: Orthotomus ). Mol. Phylogen. Evol. 65: 54-63. Sibley, C. G. & Ahlquist, J. E. (1 990) Phylogeny and classification of birds: a study in molecular evolution. New Haven, CT: Yale University Press. Smithe, F. B. (1975) Naturalist's color guide. New York: American Museum of Natural History. Stattersfield, A. J., Crosby, M. J., Long, A. J. & Wege, D. C. (1998) Endemic Bird Areas of the world: priorities for biodiversity conservation. Cambridge UK: Bird Life International. Tobias, J. A„ Bates, J. M„ Hackett, S. J. & Seddon, N. (2008) Comment on the latitudinal gradient in recent speciation and extinction rate of birds and mammals. Science 31 9: 901. Tobias, J. A., Seddon, N., Spottiswoode, C. N., Pilgrim, J. D., Fishpool, L. D. C. & Collar, N. J. (2010) Quantitative criteria for species delineation. Ibis 152: 724-746. Uy, J. A. C., Moyle, R. G., Filardi, C. E. & Cheviron, Z. A. (2009) Difference in plumage color used in species recognition between incipient species is linked to a single amino acid substitution in the melanocortin-1 receptor. Amer. Nat. 1 74: 244-254. Voris, H. K. (2000) Maps of Pleistocene sea levels in Southeast Asia: shorelines, river systems and time durations. J. Biogeog. 27: 1 1 53-1 1 67. Wells, D. R. (2007) The birds of the Thai-Malay Peninsula, 2. London: Christopher Helm. Woxvold, I. A., Duckworth, J. W. & Timmins, R. J. (2009) An unusual new bulbul (Passeriformes: Pycnontidae) from the limestone karst of Lao PDR. Forktail 25: 1-12. Zhou Fang & Jiang Aiwu (2008) A new species of babbler (Timaliidae: Stachyris) from the Sino-Vietnamese border region of China. Auk 1 25: 420-424. SUPPLEMENTARY ONLINE MATERIAL Available on Oriental Bird Club website, links at http:// www.orientalbirdclub.org/ publications/ forktail29 Tables SOM 1 to SOM 5 Figure SOM 1 Media Files SOM 1 to SOM 6 Simon P. MAHOOD, Wildlife Conservation Society Cambodia Program, PO Box 1620, Phnom Penh, Cambodia. Email: smahood@wcs.org Ashish Joshia Ingty JOHN, Wildlife Conservation Society Cambodia Program, PO Box 1620, Phnom Penh, Cambodia. Email: ajohn@wcs.org Jonathan C. EAMES, BirdLife International Cambodia Programme, PO Box 2686, Phnom Penh, Cambodia. Email: Jonathan.Eames@birdlife.org Carl H. OLIVEROS, Biodiversity Institute and Department of Ecology and Evolutionary Biology, University of Kansas, Lawrence, KS 66045, USA. Email: oliveros@ku.edu Robert G. MOYLE, Biodiversity Institute and Department of Ecology and Evolutionary Biology, University of Kansas, Lawrence, KS 66045, USA. Email: moyle@ku.edu HONG CHAMNAN, Forestry Administration, Ministry of Agriculture, Forestry and Fisheries, Cambodia and Wildlife Conservation Society Cambodia Program, PO Box 1 620, Phnom Penh, Cambodia. Email: chong@wcs.org Colin M. POOLE, Wildlife Conservation Society Singapore, 352 Tanglin Road, Strathmore Block #01-08, Tanglin International Centre, Singapore 247671. Email: cpoole@wcs.org Howie NIELSEN, 193 Hollywood Blvd, Whitefield, Maine, USA. Email: birderhowie@gmail.com Frederick H. SHELDON, Museum of Natural Science and Department of Biological Sciences, Louisiana State University, Baton Rouge, LA 70803, USA. Email: fsheld@lsu.edu FORKTAIL 29 (2013): 15-18 Notes for the conservation of the Rufous-fronted Laughingthrush Garrulax rufifrons N. J. COLLAR &S. van BALEN The Rufous-fronted Laughingthrush Garrulax rufifrons, endemic to Java, has been recorded from a total of 15 montane sites, 14 in West Java (nominotypical rufifrons) and one in Central Java (subspecies slamatensis). It occupies montane forest generally in the range 1 ,000-2,000 m, although this may vary with site, and occurs in monospecific parties of birds but also in bird waves, and has or had an association with Javan Green Magpie Cissa thalassina. Breeding appears to be extended through the year, but lack of records in January-February and July- August may reflect real breaks in the cycle. A lack of recent records from bird markets and a recent hike in prices of captive birds supports other concerns that the Javan bird trade may have affected the species, which in the past 20 years appears only to have been observed at Gunung Gede-Pangrango. Surveys of known sites and of several montane forest reserves are needed before a heavy investment in captive breeding is made. INTRODUCTION Of all the species bearing the English name ‘laughingthrush’, now proposed as components of a large subfamily of babblers named Leothrichinae (Moyle etal. 2012), Rufous-fronted Laughingthrush Garrulax rufifrons — called Red-fronted in Andrew (1985) and MacKinnon (1988) and Plain-brown in Hellebrekers & Hoogerwerf (1967) — is the southernmost, being confined to the island of Java, Indonesia (Collar & Robson 2007). This fact, combined with its restriction to montane forest (Stattersfield etal. 1998), suggests a relictual distribution, and Berlioz (1930), in considering it ‘truly aberrant’, attributed this in part to its geographical isolation. The species has received virtually no attention from biologists, ecologists and scientific ornithologists, and there are no studies of it in the wild, but because, by contrast, it has received considerable attention from bird trappers, it has been treated all this century as Near Threatened (Stattersfield & Capper 2000, BirdLife International 2001). Recent anecdotal evidence suggests that its conservation status may have declined further since the 1990s (Collar et al. 2012 and below). This paper is therefore an attempt to assemble basic information relevant to its long-term conservation and make some appropriate preliminary recommendations. DISTRIBUTION Mees (1996) listed and mapped 1 1 localities for the species (treating Gn [=Gunung] Endut and Gn Salak separately) and on this basis remarked that it ‘may be assumed to occur throughout the highlands of West Java’. Our further collation of records (initials of museums are glossed in the Acknowledgements) suggests that this prediction was correct. Since its description in 1831 the Rufous-fronted Laughingthrush has been recorded at the following localities (listed as far as possible from west to east), increasing the total to 1 5 (but treating Gn Endut as part of Gn Salak): • Gn Karang above Ciomas and Ujungtebu, 1-13 April 1920 (Robinson & Kloss 1924: 285) and at an unspecified locality in April 1991 (D. A. Holmes in litt. 1991); • Gn Halimun, August and September 1922 (2 specimens in Naturalis), July 1982 (K. D. Bishop in litt. 2013); • Gn Salak (type locality designated by Deignan 1964), on the south-east slope, October 1882 (Vorderman 1886), at Gn Endut, 10 June 1897 (Bartels 1902, 1906, Mees 1996; 1 specimen in Naturalis), at Cianten, April-June 1932 (2 specimens inMZB),at Pasirreungit, 12-15 August 1981 (SvB), at Warungloa, heard once, 15 July 1981 (SvB), on the south¬ west slope at Awibengkok, 10 records of 1-3 birds, 3-9 September 1988 (SvB); • Gn Gede-Pangrango, May 1889 (1 specimen in Naturalis; Vorderman 1892), 1900-1926 (34 specimens and 3 clutches in Naturalis, 4 specimens in MZB), 1943-1947 (8 clutches in Naturalis, 2 specimens in MZB; also Hoogerwerf 1948), specifically at Puncak, 1970s (W. G. Harvey in a list supplied by thelateD. A. Holmes to SvB), Telaga Warna, 1979-1981 (SvB), Cibodas, October 1896 (1 specimen in Naturalis) and 21 September 1918 (Spennemann 1923), Cibodas and Kandangbadak, February-March 1916 (Robinson & Kloss 1924, Delsman 1927) and April 1941 (1 clutch in Naturalis), with many encounters 1979-1989 (Andrew 1985, SvB) including one at Cimungkat, July 1987 (SvB), sight records through the 1 990s and 2000s (J. Chance inlitt. 1991, J. A. Eaton, C. R. Robson in litt. 2013) and audio recordings in June-July 2009 (XC30475-76 by B. Cox, XC40473-74 by D. Edwards); • Cianjur, Cibeber, in the period 1946-1949 (G. F. Mees notebooks seen by SvB); • Gn Patuha, Koleberes, 1927-1929 (Bartels 1931: 336; hence Hoogerwerf 1948); • Situ Lembang, 15 March 1984, 8 birds (P. Andrew in litt. 2013); • Gn Tangkubanprahu (Mees 1996), July and October 1926, December 1955 and December 1957 (8 specimens in Naturalis); • Gn Malabar at Tirtasari, 12 May 1910 (Mees 1996; 1 specimen in Naturalis); • Gn Wajang, Cibitung (Mees 1996), April and May 1910 (3 specimens in Naturalis); • Gn Papandayan, late 1920s (Stresemann 1930), 1941-1942 (2 specimens in MZB), with subsequent records specifically at Kawahmanuk, 2 birds, 3 September 1987 (SvB); Gn Kendang, flock of 10-15 birds tape-recorded, 6 September 1987 (SvB); • Gn Rakutak, March 1900 (1 specimen in AMNH); • Gn Guntur, Garut, October 1900 (2 specimens in AMNH), including Kawah Kamojang, May 1923 (1 specimen in MZB), and ‘near Garut’ (Siebers 1929); • Gn Ciremay (Mees 1996), June 1930 (1 specimen in Naturalis); and • Gn Slamat (type and only locality for race slamatensis ) at Kaligua, 1916-1917 (Siebers 1929; type specimen in Naturalis, 3 paratypes in MZB), and at Purwokerto, March 1925 (Voous 1948, Mees 1996; 2 specimens in Naturalis). 16 N. J. COLLAR & S. van BALEN Forktail 29 (2013) ELEVATIONS, ECOLOGY AND NATURAL HISTORY The species is resident in and confined to ‘mixed original forest’ or ‘broadleaved evergreen forest’ at 900-2,500 m (Sody 1956, Collar & Robson 2007), this being a minor shift from elevations of 1,000- 2,400 m (Stattersiield et al. 1998, BirdLite International 2001). However, these limits represent extremes amalgamated from individual sites, and may vary considerably at each known site depending on ecological conditions, mountain height (the peaks of several mountains listed above lie below 2,400 m), and levels of deforestation. Moreover, nothing is known about the species’s relative abundance at different elevations, although Hoogerwerf (1950) indicated that on Gn Gede-Pangrango it was a common bird from Cibodas up to near the tops of the mountains. The site-specific elevations in Hoogerwerf (1948) — 1,500- 2,600 m on Gn Papandayan, 600-1,000 m at Ciomas on Gn Karang, 800-1,200 m at Cimungkat on Gn Cede, 600-1,000 m at Koleberes on Gn Patuha and 500-2,300 m on Gn Salak — are not intended to indicate the limits between which the species was certainly encountered; nor is there clear evidence to support Hoogerwerf ’s (1948) characterisation of the species as one ‘in certain areas probably living permanently between 2500 and (above) 3000 m’. On Gns Endut and Pangrango, Bartels (1902) gave its elevation as ‘3,000-3,500 feet’ (900-1,100 m), later changing this to ‘3,000-6,000 feet’ (900-1,800 m) (Bartels 1906); records from Cimungkat on Gn Gede were at 1,200 m (SvB). Some specimens on Gn Tangkubanprahu were at 1,500 m (Naturalis label data), as was the first record from Gn Papandayan (Stresemann 1930) , although subsequently birds were found in the latter locality at 1,900 m (Kawahmanuk) and at 2,525 m (Gn Kendang) (SvB). Records from Gn Salak are at 1,500 m (Vorderman 1886), and specifically at Pasirreungit at 1,350-1,900 m and Awibengkok at 1,000-1,150 m (SvB). It therefore appears that only one record, hitherto unpublished, pins the species to an elevation higher than 2,000 m; all other records traced come from below this altitude. The record from Situ Lembang was at 850 m (P. Andrew in litt. 2013), and those on Gn Karangwere at 600-900 m (Robinson & Kloss 1924), these apparently being the lowest elevations recorded for the species. The Rufous-fronted Laughingthrush occupies all strata of the forest but chiefly the undergrowth, and is ‘very agile’ (Hoogerwerf 1950). It occurs in loose, sometimes large monospecific groups but also participates in bird-waves (Hoogerwerf 1950, Andrew 1985), these latter sometimes comprising up to 1 5 different species on Gn Gede (van Balen 1992); in particular, it associates with the Javan Green Magpie Cissa thalassina (Koningsberger 1901, Bartels 1915— 1931) , such that on Gn Halimun in 1982 the two species were found together in a bundle of birds being carried by a poacher (K. D. Bishop in litt. 20 1 3). Its presence is best determined by its noisy, whinnying call, earning it the local name ‘horsebird’ (van Balen 1992) and placing it with the group oflaughingthrushes that possess a laughing call (Collar & Robson 2007). Various authors have given glancing accounts of the diet: ‘berries and insects, mostly beetles’ (MacKinnon 1988), beetles, snails and fruits of Melastoma malabathricum (Sody 1989), and these plus mantids and caterpillars (Collar & Robson 2007). Hoogerwerf ’s (1950) mention of small hard seeds and Sody’s (1989) of Melastoma may well both refer back to Vorderman’s (1886) account of stomachs ‘coloured black by fruit pulp, and filled with small hard seeds’ (our translation). The closest observer of the species described its diet as mainly and sometimes exclusively various forest fruits, supplemented with insects, mainly beetles including weevils, plus bugs, caterpillars, locusts, spiders, ants and small vertebrates such as frogs and lizards (Bartels 1915-1931; also Delsman 1927). Specimen labels in Naturalis mention Anomala beetles, small beetles, a large weevil, a phasmid, looper caterpillars and Ficus andLantana fruit as stomach contents. In captivity, birds caught wild mice in their enclosure (Pithart 2009). The nest is a sturdy, relatively small cup placed on a horizontal branch or in a fork usually fairly close (about 2 m) to the ground in smaller trees at the edge of forest (more details in Hoogerwerf 1 950, Hellebrekers & Hoogerwerf 1967). The usual clutch is three (blue- green) eggs, but sometimes two; nests have been found in March, April, May, June, September, November and December (Hoogerwerf 1949, 1950, Hellebrekers & Hoogerwerf 1967). Whether the gaps in breeding in January-February and July- August represent real seasonal differences, random variation or temporal patchiness in observer coverage is an open question. However, breeding in Prague Zoo followed a roughly similar schedule, with nests in April-June and August- October (Pithart 2009). Naturalis possesses birds marked as juveniles from January (1), May (2), June (1) and August (1), but these are full size and it is impossible to pin them to a likely month of birth; it also contains four specimens labelled as having full-sized gonads in April (male and female) and May (two males). In captivity the female was noted to do almost all incubation, which lasted 14-15 days, while the nestling period was 15-16 days; moult occurred slowly from autumn (occasionally July) through to December (Pithart 2009). Indeed, birds at the end and start of the year have been described as ‘gut im Gefieder’ (Bartels 1902), which presumably best translates as ‘in fresh plumage’. POPULATION TRENDS ANDTHREATS There has been no systematic monitoring of populations of this or any other forest bird species in Java, so a quantitative assessment of population trends is impossible. However, various items of qualitative information have accumulated to suggest that the Rufous-fronted Laughingthrush may now be in a more serious condition than has hitherto been realised, largely as a result of the singular Javanese tradition of bird keeping. ‘I am afraid that aviculture is a major source of bird destruction in Indonesia’, wrote Morrison (1980), having found Java to be ‘a singularly birdless island’. This was over 30 years ago. At that time, however, the Rufous-fronted Laughingthrush, being a bird of high, remote forests, may still have been common. On Gn Gede- Pangrango it was common in the 1940s (Hoogerwerf 1950) and in the 1980s (Andrew 1985), and there is no reason to imagine that it was less common at the other localities listed above in the 1980s, although as ajavan endemic it was protected under Indonesian law in 1979 (Noerdjito & Maryanto 2001). Only once in the early documentation was there an indication of relative rarity: it was scarce at Gn Patuha in the years 1927-1929 (Bartels 1931), presumably for natural reasons (‘only in the northern forests’). Extrapolation from experience at Gn Gede presumably lies behind MacKinnon’s (1988) general description of the species as ‘Locally not uncommon in montane forests’ and behind Mees’s (1996) remark that ‘Where this species occurs it is common, noisy, and conspicuous.’ Nevertheless, only two years after this comment, the species was said to be ‘fairly heavily exploited as a cagebird, which has rendered it uncommon in otherwise moderately secure habitat’ (D. A. Holmes in litt. 1998 in BirdLife International 2001), leading to its designation as a Near Threatened species, and in the mid-2000s it was described as ‘formerly common in Gede-Pangrango National Park... but now rare along main trail, reportedly owing to trapping’ (Collar & Robson 2007), although inquiries of leaders taking bird tours to Gn Gede do not suggest that numbers have obviously declined there (C. R. Robson in litt. 2013, J. A. Eaton in litt. 2013). Other evidence, however, certainly tends to support the notion that a real decline has been occurring for some years. Bird dealers Forktail 29(2013) Notes for the conservation of the Rufous-fronted Laughingthrush Garrulax rufifrons 17 in markets in Medan, Sumatra, recently reported that Rufous- fronted Laughingthrush is ‘becoming increasingly rare or difficult to find in economically viable numbers’ (Shepherd 201 1). Independently, it has been reported to have Vanished from the bird markets in Sumatra’ (P. Hospodarsky in Pithart 2009). Moreover, on Java at the start of the century the species ‘could be found in bird markets as a cheap local songster, selling for Rp 150,000 ($16)’, but in the past few years the price has increased tenfold and in 20 1 2 no birds could be found in bird markets (R. Sozer in Collar et al. 2012 and in litt. 2013). This latter testimony was independently supported by C. R. Shepherd [in litt. 2013): Dealers in the Barito Market and the Pramuka Market [Jakarta, Java] stated in 20 1 1 that this species was ‘difficult to find, or all gone’ ( susah or sudah habis). These kinds of statements do usually mean trappers are rarely bringing them in and are not finding them in their usual trapping areas. In 2012,1 only carried out one survey in Jakarta’s three largest bird markets ( June 2012) and did not see any. Moreover, there are parallels with declines and near-disappearances in other species that have been attributed to the demands of Javan bird-keeping, most notably that of the Javan Green Magpie (van Balen et al. 2013; also Collar et al. 2012). However, the current plight of the magpie, and the laughingthrush’s reported association with it, opens up the plausible if very remote possibility that the laughingthrush’s conservation status may not be so desperate. Since the magpie is a much more prized species in the Javan bird trade, it might conceivably be that when targeting the magpie trappers took many laughingthrushes simply as a ‘bycatch’, which could explain the latters’ low prices and wide availability a few years ago. Moreover, now that trade has reduced the numbers of magpies to near-zero (van Balen et al. 2013), trappers are perhaps no longer visiting areas where magpies once occurred, in which case the sudden disappearance of laughingthrushes from markets might simply reflect lack of trapping effort rather than lack of birds. Nevertheless, the rather high prices now commanded by the laughingthrush tend to suggest that its rarity is real and, as V. Nijman (in litt. 2013) has commented, there are high numbers of montane bird species still available for sale in Java’s markets, and ‘not all of them are expensive’. CONSERVATION NEEDS If protected area status improves the chances of long-term habitat conservation, then Gn Halimun-Salak, Gn Gede- Pangrango and perhaps Gn Guntur are likely to be the best- preserved of the sites at which the laughingthrush occurs (although at Gn Gede in the past 1 0 years there has been ‘shocking clearance’ for vegetable plots, ‘apparently inside the protected area, probably up to 2, 1 00 m in a c. 1 km belt above and east of the Cibodas Botanic Garden’: F. R. Lambert in litt. 2013). The reserve at Kawah Kamojan on Gn Guntur covers 8,000 ha at 1,400-2,250 m (MacKinnon et al. 1982), but other sites at which the species has been recorded have very small areas protected: the only reserves larger than 100 ha are at Telaga Patengan on Gn Patuha (150 ha), Gn Papandayan (844 ha) and Gn Tangkubanprahu (1,660 ha) (MacKinnon et al. 1982), but it is not known if they encompass laughingthrush habitat and viable populations. An area of 1 5,000 ha on Gn Slamat was long ago recommended for protection (see Stattersfield et al. 1998) but only two reserves, both less than 20 ha, exist there (MacKinnon etal. 1982); since it is the sole locality for the highly distinctive subspecies slamatensis (Siebers 1929, Voous 1948, Mees 1996) of Garrulax rufifrons , formal protection of the site is clearly highly desirable. Last records of the species from all known sites are: Gn Karang 1991, Gn Halimun 1984, Gn Salak (where on Endut it was ‘not rare’ at the start of the twentieth century: Bartels 1902, 1906) 1988, Gn Gede-Pangrango 2012, Cianjur at least 1949, Gn Patuha before 1931, Situ Lembang 1984, Gn Tangkubanprahu 1957, Gn Malabar 1910, Gn Wajang 1910, Gn Papandayan 1987, Gn Rakutak 1900, Gn Guntur 1923, Gn Ciremay 1 930 and Gn Slamat 1925. D. Liley (in litt. 2013) spent 5-6 weeks at Cikuya, on the southern slopes of Gn Halimun, mostly at 1,000-1,200 m, without seeing the species, and K. D. Bishop (in litt. 2013) visited Gn Halimun in August 2011 after a gap of 29 years and found no laughingthrushes; however, it is fair to note that the one location that most birdwatchers go to at Gn Halimun, Cikaniki, probably never had the species (between 1996 and 2009 field teams never recorded it: Prawiradilaga et al. 2003, Noske et al. 2011). Even so, it appears to be at least 20 years since there was a record of the species away from Gn Gede-Pangfango. Naturally, therefore, these sites need urgent surveying to determine the status of the forests and the continuing presence of the species (for which, given its noisiness, playback techniques would probably be highly effective). Other under-explored reserves which might hold the species are: GnBurangrang (2,700 ha; 1,000- 2,000 m); Gn Tampomas (1,250 ha; 1,000-1,700 m); Gn Sawal (5,400 ha; 600-1,764 m); Gn Simpang( 15,000 ha; 600-1,600 m) and Gn Tilu (8,000 ha; 1,200-2,177 m). Preferably, however, such a survey would involve line-transect or point-count work to establish baseline densities at the sites, and would target other rare species such as Javan Hawk Eagle Nisaetus bartelsi, Javan Trogon Apalharpactes reinwardtii, Javan Cochoa Cochoa azurea and Javan Green Magpie, along with (e.g.) certain primates. Study of the culture and economy of bird-keeping in Indonesia has led Jepson et al. (2011) to ‘argue that, in Indonesia at least, conservationists need to move beyond the moralistic, animal rights and protectionist logic that dominate \sic\ much wildlife trade discourse and embrace the development logic of pro-poor growth and more, better jobs’. Whatever one makes of this prescription it predicates a time-scale that completely mismatches the short-term needs of many species native to Java, and if acted on would merely vaporise their chances of survival. If conservationists do not focus on birds that are at greatest risk from trade activities on the island, the only logic they are likely to embrace is the logic of extinction. Captive breeding for conservation purposes (‘conservation breeding’) may therefore now be a lifeline for the Rufous-fronted Laughingthrush (Collar etal. 2012). However, the species has been bred only with some difficulty and only, apparently, in two European institutions, Tierpark Berlin (Kaiser 2006) and Prague Zoo (Pithart 2009). It has proved an aggressive and problematic species to keep, and endeavours to develop a significant captive stock may only be worth making once the evidence is clearer about its status in the wild. ACKNOWLEDGEMENTS We are most grateful to T. J. Trombone for locality data from specimens at the American Museum of Natural History (AMNH), and to the authorities at the Zoological Museum, Bogor (ZMB), and Naturalis, Leiden, for access to their collections. K. D. Bishop, N. Brickie, J. A. Eaton, F. R. Lambert, D. Liley, R. Sozer, C. R. Shepherd and A. Vaidl all greatly helped with our inquiries. R Andrew and V. Nijman were most helpful referees. REFERENCES Andrew, P. (1985) An annotated checklist of the birds of the Cibodas- Gunung Gede Nature Reserve. Kukila 2: 10-28. van Balen, B. (1 992) Birdwatching areas. Gunung Gede-Pangrango National Park, West Java, Indonesia. OBCBull. 15:27-29. 18 N. J. COLLAR &S. van BALEN Forktail 29(2013) van Balen, S. (B.), Eaton, J. A. & Rheindt, F. E. (201 3) Biology, taxonomy and conservation status of the Short-tailed Green Magpie Cissa [t.j thalassina from Java. Bird Conserv. Int. 23: 91-1 09. Bartels, M. E. G. (1902) Zur Ornis Javas. Natuurk. Tijdschr. Ned.-lndie6) : 1 29- 172. Bartels, M. (1906) Systematische Ubersicht meiner Java-Vogel. J. Orn. 54: 383-407, 497-519. Bartels, M. E. G. (1915-1931) Diaries and notebooks on Javan birds. Unpublished manuscripts and notebooks held at Naturalis, Leiden, Netherlands. Bartels, E. (1 931 ) Vogels van Kole Beres. Natuurk. Tijdschr. Ned.-lndie9 1 : 308- 348. Berlioz, J. (1930) Revision systematique du genre Garrulax Lesson. Oiseau 11: 1-27, 78-105, 129-159. BirdLife International (2001) Threatened birds of Asia: the BirdLife International Red Data Book. Cambridge UK: BirdLife International. Collar, N. J. & Robson, C. (2007) Family Timaliidae (babblers). Pp. 70-291 in J. del Floyo, A. Elliott & D. A. Christie, eds. Handbook of the birds of the world, 12. Barcelona: Lynx Edicions. Collar, N. J., Gardner, L„ Jeggo, D. F„ Marcordes, B., Owen, A., Pagel, T., Pes, T„ Vaidl, A., Wilkinson, R. & Wirth, R. (2012) Captive breeding and the most threatened birds in Asia. BirdingASIA 1 8: 50-57. Deignan, H.G. (1964) Subfamily Timaliinae, babblers. Pp. 240-427 in E. Mayr & R. A. Paynter, eds. Check-list of birds of the world, 10. Cambridge Mass.: Museum of Comparative Zoology. Delsman, H.C. (1927) Vogelleven in het oerbosch II. Tropische Natuur 16: 82-89. Hellebrekers, W. P. J. & Hoogerwerf, A. (1967) A further contribution to our oological knowledge of the island of Java (Indonesia). Zoo/. Verhandel. 88. Hoogerwerf, A. (1 948) Contribution to the knowledge of the distribution of birds on the island of Java, with remarks on some new birds. Treubia 19: 83-137. Hoogerwerf, A. (1949) Bijdrage tot de oologie van Java. Limosa 22: 1-279. Hoogerwerf, A. (1950) De avifauna vanTjibodas en omgeving, inclusief het natuurmonumentTjibodas-Gn. Gede (West-Java). Limosa 23: 1-1 58. Jepson, P„ Ladle, R. J. & Sujatnika (2011) Assessing market-based conservation governance approaches: a socio-economic profile of Indonesian markets for wild birds. Oryx 45: 482-491 . Kaiser, M. (2006) 50 Jahre Haltung und Zucht von Haherlingen ( Passeriformes : Garrulax, Liocichla, Babax) im Tierpark Berlin. Milu 1 1: 667-695. Koningsberger, J. C. (1901) De vogels van Java en hunne oeconomische beteekenis, 1 . Batavia: Kolff. MacKinnon, J. (1988) Field guide to the birds of Java and Bali. Yogyakarta, Indonesia: Gadjah Mada University Press. MacKinnon, J., Smiet, F. & Artha, M. B. (1982) A national conservation plan for Indonesia, III: Java and Bali. Bogor: Food and Agriculture Organization of the United Nations. Mees, G. F. (1946-1949) Unpublished field notebooks held in Naturalis, Leiden, Netherlands. Mees, G. F. (1996) Geographical variation in birds of Java. Publ. Nuttall Orn. Club 26. Morrison, A. (1980) A note on Javanese aviculture. Avicult. Mag. 86: 1 08— 109. Moyle, R. G., Andersen, M. J., Oliveros, C. H„ Steinheimer, F. D. & Reddy, S. (2012) Phylogeny and biogeography of the core babblers (Aves: Timaliidae). Sysf. Biol. 61 : 631-651 . Noerdjito, M. & Maryanto, I. (2001) Jenis-jenis Hayati yang Dilindungi Perundang-undangan Indonesia. Bogor: Museum Zoologi Bogoriense, Puslitbang Biologi-LIPI & The Nature Conservancy. Noske, R„ Prawiradilaga, D. M„ Drynan, D„ Leishman, A. & Rutherford, W. (2011) Understorey birds of Cikaniki Research Station, Gunung Halimun National Park, West Java: report of the Indonesian Bird Banding Scheme Training Programme. Kulika 1 5: 50-59. Pithart, K. (2009) Breeding of the Rufous-fronted Laughingthrush (Garrulax rufifrons) at Prague Zoo. Gazella 36: 151-176. (In Czech pp. 151-169 and English pp.170-176.) Prawiradilaga, D. M., Marakarmah, A. & Wijiamukti, S. (2003) A photographic guide to the birds of Javan montane forest: Gunung Halimun National Park. Jakarta: LIPI-JICA-PHKA. Robinson, H. C. & Kloss, C. B. (1924) A nominal list of the birds collected in Java. Treubia 3: 267-298. Shepherd, C. R. (201 1) Observations on trade in laughingthrushes ( Garrulax spp.) in North Sumatra, Indonesia. Bird Conserv. Int. 21 : 86-91 . Siebers, H. C. (1929) Neue Vogelrassen aus dem Indo-Australischen Gebiet. Treubia 11:1 49-1 53. Sody, H. J. V. (1956) De Javaanse bosvogels. Madjalah llmu Alam untuk Indonesia 1 12: 153-170. Sody, H. J. V. (1989) Diets of Javanese birds. Pp 164-221 in J. H. Becking, Henri Jacob Victor Sody (1 892- 1 959): his life and work. Leiden: Brill. Spennemann, A. (1923) Vogelleven in het oerbosch teTjibodas. Tropische Natuur) 2: 177-179. Stattersfield, A. J. & Capper, D. R., eds. (2000) Threatened birds of the world. Cambridge UK & Barcelona: BirdLife International & Lynx Edicions. Stattersfield, A. J., Crosby, M. J., Long, A. J. & Wege, D. C. (1998) Endemic bird areas of the world: priorities for biodiversity conservation. Cambridge UK: BirdLife International (Conservation Series 7). Stresemann, E. (1930) EineVogelsammlung vom Vulkan Papandajan (West- Java). Treubia 1 2: 425-430. Voous, K. H. (1948) Notes on a collection of Javanese birds. Limosa 21: 85- 100. Vorderman, A. G. (1886) Bijdrage tot de kennis der avifauna van den berg Salak. Natuurk. Tijdschr. Ned. -Indie 45: 304-414. Vorderman, A. G. (1 892) Java-vogels 1 . Natuurk. Tijdschr. Ned. -Indie 51 : 373- 416. N. J. COLLAR, BirdLife International, Girton Road, Cambridge CB3 0NA, UK. Email: nigel.collar@birdlife.org S. (BAS) van BALEN, Basilornis Consults, Muntendampad 15, 6835 BE Arnhem, Netherlands FORKTAIL 29 (2013): 19-24 Species limits in the Golden Bulbul Alophoixus (Thapsinillas) affinis complex N. J. COLLAR, J. A. EATON & R. 0. HUTCHINSON The Golden Bulbul Thapsinillas affinis of the Moluccan islands, Sula archipelago, Banggai islands, Togian islands and Sangihe, Indonesia, was until recently treated in Alophoixus before being placed in the resurrected genus Thapsinillas and shortly afterwards split into Northern and Southern Golden Bulbuls T. affinis and T. longirostris, but with a general consensus that a break-up into more species was required. We used plumage and morphometric analysis of museum specimens, supplemented by vocal samples, to determine where new species limits might be drawn. We found that the nine generally accepted subspecies break down into seven full species, five monotypic and two with two subspecies each: T. chloris on Morotai, Halmahera and Bacan (small, featureless; undifferentiated olive-green lores and ear-coverts, blackish base to submoustachial area; song reportedly a 'jumbled babbling'); T. lucasi on Obi (round yellow lores, yellow-tinged ear-coverts, seemingly simple often squeaky-toy-like vocalisations); T. affinis on Seram with race flavicaudus on Ambon (larger than previous two, with half-wedge yellow lores, broad yellow tips to tail, song a group of strong rich flat whistles); T. mysticalis on Buru (half-wedge yellow lores, partial yellow eye-ring, olive-green underparts, olive-grey tail, whistled phrases recalling domestic canary); T. longirostris on Sula with race harterti on Peleng and Banggai (longest-billed, large, undifferentiated olive-green lores, song a loud jumble); T. aurea on theTogian islands (golden-yellow underparts, vague half-wedge yellow lores, blackish frontal supercilial line, yellow-tinged rump, song seemingly more complex than in longirostris) and T. platenae on Sangihe (vivid yellow chin and submoustachial area to throat and breast, bright yellow triangular lores, almost-complete yellow eye-ring, song seemingly simple and nasal). Comprehensive vocal sampling and molecular work may shed light on the origins and colonisation routes of this geographically unusual cluster of species. INTRODUCTION The taxonomy of the Golden Bulbul Alophoixus ( Thapsinillas ) affinis complex of Wallacea, Indonesia, has long been considered problematic, owing to the considerable variation in plumage pattern and size shown by most of its subspecies (Hartert 1922, Delacour 1943, White & Bruce 1986). These subspecies possess an unusual and indeed unique distribution for a species in the region, in the geographic sequence given by White & Bruce (1986) as follows: chloris (North Moluccas: Morotai, Halmahera, Bacan); lucasi (Obi); affinis (Seram); flavicaudus (Ambon); mysticalis (Buru); longirostris (Sula); harterti (Peleng, Banggai); aurea (Togian Islands) and platenae (Sangihe). It is perhaps a measure of the uncertainty surrounding this complex that it has appeared in so many generic guises in the past hundred years. Until at least 1922 it was largely treated in Criniger (e.g. Wallace 1862a, b, 1863, Blasius 1888, Hartert 1903, 1922), but Delacour (1943) placed it in Microscelis (subgenus Iole ), Rand & Deignan (I960), Morony et al. (1975) and Andrew (1992) in Hypsipetes , White & Bruce (1986) and Coates & Bishop (1997) in Ixos, and Sibley & Monroe (1990) and Inskipp et al. (1996) in Alophoixus. Finally Dickinson & Gregory (2002) resurrected the genus Thapsinillas for the complex (a decision we follow hereafter), citing as diagnostic characters ‘typically dark oily green [plumage] , relieved by areas of yellow in some forms; crown not crested and feathers only slightly elongated ; bill much like Iole but perhaps more hooked and with lower mandible deeper; rictal bristles fewer and weaker’, but unaccountably omitting mention of the key criterion in the original description, namely that ‘from all the related genera with lengthened nostrils Thapsinillas may easily be distinguished... by its very short tarsus, this being considerably less than the exposed culmen’ (Oberholser 1905). Continuing this theme of taxonomic hesitancy, both Dickinson & Gregory (2002) and Dickinson & Dekker (2002) suspected that the variation between the subspecies in this resurrected genus ‘will justify subdivision into two to four species’. However, Delacour ( 1 943) bluntly cited ‘distribution’ as the reason to resist a split into two species based on ‘size and tail pattern’ (larger taxa with ‘particolored tail, dark olive and bright yellow’, smaller ones ‘strangely similar to M. ictericus (= Yellow-browed Bulbul Iole indica in Inskipp et al. [1996]). By contrast, Fishpool & Tobias (2005) took what they regarded as ‘a preliminary measure’ by separating the ‘Northern Golden Bulbul’ T. longirostris (with chloris, lucasi , harterti, aurea and platenae) from ‘Southern Golden Bulbul’ T. affinis (with flavicaudus and mysticalis) on account of reported vocal differences between these groups, thereby ‘drawing attention to the broadest rift in the complex, and paving the way for appropriate fieldwork and research into the song, morphology and genetics of all taxa involved’. These authors, like Dickinson & Dekker (2002), judged that ‘further subdivision’ would almost certainly be required, ‘in view of significant differences between the various island populations’. This was partially achieved by Rheindt & Hutchinson (2007), who, without going into detail, considered ‘Southern Golden Bulbul’ to comprise two morphologically and vocally distinct species, Buru Golden Bulbul T. mysticalis and Seram Golden Bulbul T. affinis (including flavicaudus). Steadily accumulating evidence on apparent differences in vocalisations of most of the taxa in the Thapsinillas affinis complex now prompts a more detailed review of their morphological and morphometric characters in order to attempt to reach a further stage in the revision of the Golden Bulbul complex. As Fishpool & Tobias (2005) observed, this is important not least because ‘some island races would prove to be very rare...’ such that ‘taxonomic review is vital for the compilation of a realistic conservation strategy for Wallacea, and must be made a priority’. METHODS We considered one line of hard evidence in this review, namely plumage and mensural characters from museum material, and supplemented it with morphological evidence from photographs as well as recordings and reports of vocalisations. Museum specimens of Golden Bulbuls were examined (NJC) in the Natural History Museum, Tring, UK (NHMUK), Naturalis, Leiden, Netherlands (Naturalis), Staatliches Museum fur Tierkunde, Dresden, Germany (SMTD), Staatliches Naturhistorisches Museum, Braunschweig, Germany (SNMB) and Zoologisches 20 N. J. COLLAR, J. A. EATON & R. O. HUTCHINSON Forktail 29 (2013) Museum (Museum fur Naturkunde), Berlin, Germany (ZMB). Each specimen was measured (by NJC) for length of bill (skull to tip), tarsus, wing (curved) and tail (tip to point of insertion), the characters of each taxon were logged in a matrix, and representative specimens were photographed. From these collections the numbers of specimens by taxon and island were: • chloris — North Moluccas: 39 specimens, 10 from Morotai, 16 from Halmahera, 13 from Bacan (11 males [m], 8 females [f], 20 unsexed M) • lucasi — Obi: 13 (7 m, 5 f, 1 u) • affinis — Seram: 12 (4 m, 3 1, 5 u) • flavicaudus — Ambon: 8 (6 m, 1 f, 1 u) • mysticalis — Buru: 21 (4 m, 10 1, 7 u) • longirostris — Sula (Taliabu & Mangoli): 23 (7 m, 2 f, 14 u) • barterti — Banggai (Banggai & Peleng): 13(1 m, 2 f, 10 u) • aurea — Togian: 2 (1 m, 1 f) • platenae — Sangihe: 3 (3 m) The large number of unsexed specimens and an occasional numerical bias in the sexed specimens prompted a comparison of males only (Table 2), but the full figures and standard deviations given in Table 1 are used in the analysis of character difference below. Photographs of live birds were assembled from our own collections (JAE, ROH), from those of colleagues, contacts and friends, and (with due care as to identification and provenance) from the internet (notably Oriental Bird Images). Sound recordings were likewise assembled from our own collections (JAE, ROH), Table 1. Means and standard deviation (in brackets) of four morphometric variables in all specimens of the Thapsinillas complex. Notes: a = sample size reduced by 1; b = sample size reduced by 2; c = sample size reduced by 6. These reductions were caused by damage to the parts being measured or (in the case of tarsi) their inaccessibility (being tucked tightly against the body). Taxon n Bill Wing Tarsus Tail chloris 39 22.3 (1.06)* 98 (3.44)* 18.6(0.6) 82.9(2.36)* lucasi 13 23.4(0.53)* 104.8(3.11) 18.6(0.69)" 86.3 (2.06) affinis 12 27.8(1.17) 109.2(4.35) 20(0.78)* 86.8(3.49) flavicaudus 8 28.4 (0.94) 111.4(3.25) 20(0.95) 93.5 (2.93) mysticalis 21 25.7(0.88) 104.4(4.48) 19.6(0.82)* 92.4(3.37) longirostris 23 29.8(1.31)* 115.6(4.62) 21.2(0.56)' 106.9 (4.56) harterti 13 28.9(1.24)* 120.7(5.11) 21.2(0.72)* 108.7(3.61)" aurea 2 27.1* 117 20.5 108 platenae 3 27.7 121 20 109.3 Table 2. Means of four morphometric variables in male specimens of the Thapsinillas complex. Note: 3 = sample size reduced by 1 . Taxon n Bill Wing Tarsus Tail chloris ii 22.7* 99.8 18.6 82.8 lucasi 7 23.6* 106 18.7 86.3 affinis 4 28.4 109.5 20* 87.3 flavicaudus 6 28.1 110.5 19.8 92.7 mysticalis 4 25.8 109.3 19.8 92.8 longirostris 2 28.7 112.5 21.5 105.5 harterti 1 29.5 123 22 113 aurea 1 27.1 124 21 111 platenae 3 27.7 121 20 109.3 those of others and the internet (AVoCet [AV],Xeno-Canto [XC] and the Internet Bird Collection [IBC]). They were compared qualitatively and informal descriptions and transcriptions of them prepared. Use of capitals in the transcriptions indicates emphasis (volume). We measured the degree of phenotypic differentiation between each taxon using a system in which an exceptional difference (a radically different coloration, pattern or vocalisation) scores 4; a major character (pronounced difference in body part colour or pattern, measurement or vocalisation) scores 3; a medium character (clear difference reflected, e.g. by a distinct hue rather than different colour) scores 2; and a minor character (weak difference, e.g. a change in shade) scores 1 ; a threshold score of 7 is set to allow species status; species status cannot be triggered by minor characters alone, and only three plumage characters, two vocal characters, two biometric characters (assessed for effect size using Cohen’s d where 0.2-2 is minor, 2-5 medium, 5-10 major and >10 exceptional), and one behavioural or ecological character may be counted (Tobias et al. 2010). Where additional characters are apparent but under these rules cannot be scored, the formula ‘ns [1]’ is used, signalling ‘not scored’ but giving in parenthesis the estimated value of the difference in question. RESULTS We review each taxon in turn for its diagnostic morphological, morphometric (Tables 1 and 2) and acoustic distinctiveness. However, the acoustic component of the analysis remains qualitative, because the vocalisations of each taxon appear to be variable and complex, so that only tentative and general comments on their diagnostic distinctiveness can be ventured from the limited and fragmentary material available. From this evidence a shared pattern of song nevertheless seems to exist between all taxa, which involves a hesitant series of staccato nasal or guttural notes that accelerate and switch abruptly either to a short jumble of babbled and fluty notes on often widely differing pitches or to a short series of fairly even whistles; but most taxa sound in varying degrees different, and if these findings are replicated widely by other recordings in future then they will add substantially to the case made below for the redrawing of species limits based on morphology. Photographs and museum label data indicate that there are no significant differences in the bare -part colours of any of the taxa: basically the bill is shiny black to plumbeous, reflecting light and looking whitish at some angles or in some photographs; the legs are brownish-grey; and the iris is reddish-brown to brown. There are slight variations in how museum labels report iris colour: for example, for the taxon mysticalis NHMUK 1969.29.203 gives ‘iris brown’, 1923.9.15.91 iris dark crimson’ and 1923.9.15.92 ‘eye red’, while the describer, Wallace (1863), also gives ‘iris red’, although photographs repeatedly show reddish-brown irides. Hombron & Jacquinot (1841) likewise gave 'iris rouge for their new species affinis , but in photographs it is reddish-brown. Two of the three known specimens of the very rare platenae are labelled by the collectors as having 'iris: rot-braun. Sample sizes of specimens of aurea and platenae were respectively two and three; and recordings of all taxa were inadequate in number, duration and representativeness. However, no clinching evidence depends on data stemming from these limited sources. In the following account, the size and shape of (yellow) lores are mentioned and require definition here. ‘Round’ (taxon lucasi ) lores means that the shape of the yellow patch is large and relatively circular, and comes into contact with the leading edge of the eye. ‘Half-wedge’ (taxa affinis, flavicaudus, mysticalis and aurea ) indicates that the patch of yellow is compressed into a flat triangular Forktail 29 (2013) Species limits in the Golden Bulbul Alophoixus (Thapsinillas) affinis complex 21 bar close to the line of the upper mandible and separated from the eye by an olive-green area. ‘Triangular’ (taxon platenae) describes a fuller area of yellow than the wedge, extending to the eye. Taxon chloris (Morotai, Halmahera, Bacan) This form is characterised by its small size (it is the smallest of the taxa in the complex) and its relatively featureless plumage; no differences were apparent between the three island populations. It differs from its geographically and morphologically closest relative, lucasi of Obi, by its olive-green vs yellow lores (3), olive-green vs olive -yellow ear-coverts (1), blackish base to submoustachial area vs all olive-green (2) and slightly smaller size and distinctly shorter wing (effect size -2.28) (2) — total score 8. Originally described by Wallace (1862a) under the pre¬ occupied name simplex, this form was renamed and further described by Finsch (1867), who pointed out that Wallace failed to mention the blackish submoustachial line. Finsch found this a very distinctive (‘ ganz besonders’) character, but in specimens examined for this review it proved to be constant but somewhat variable in strength. Fishpool & Tobias (2005) provided a description (‘a hurried, cheery, jumbled babbling’) that conforms closely with the general structure of Thapsinillas songs available to us. FFowever, brief recordings by ROH of two consecutive song strophes consist (after 2-3 brief staccato introductory twis notes) of three or so simple clear paired whistles, high-pitched at the start but each pair slightly lower than the preceding, morphing subtly into a slightly more drawn-out double-whistle with the stress on the first syllable, each again slightly lower than the last: pi-pi, pi-pi, pi-pi, wiwi, wiwi, wiwi, wiiwii, thus fairly closely resembling the falling-pitch song of T. affinis (below). Otherwise the only recording we have found is of a bird giving quiet thin sii calls in apparent mild alarm or for contact (IBC video under T. longirostris, A bird softly calling from a branch’). Taxon lucasi (Obi) Hartert (1922), while itemising Rothschild’s type specimens and therefore not reviewing the Golden Bulbul complex in any detail, remarked of lucasi , which he himself established as a full species (Fdartert 1903), that ‘though differing by itsyellow lores and larger size, [it] can hardly be anything but a subspecies of chloris', and lumped it accordingly (albeit keeping chloris separate from affinis). However, the morphological differences with chloris, as scored above, gainsay this judgement. The island of Obi is roughly equidistant from Seram, Buru and Taliabu, where three further relatives of lucasi occur, respectively affinis, mysticalis and longirostris. Of these, lucasi is closest in size and general structure to mysticalis and remotest from longirostris, but differs in turn from • mysticalis by its shorter bill, tarsus and tail (effect size for bill -3.22) (2); larger, much rounder yellow lores (2); lack of yellow partial eye-ring (2); largely yellow chin to vent vs largely (yellow- tinged) olive chin to vent (3); yellower ear-coverts (ns [1]) — total score 9; • affinis by its smaller size (effect size for bill -4.83) (2); larger, rounder yellow lores (2); yellower ear-coverts and submoustachial area (at least 1); paler and less extensive olive- green on breast and flanks (ns [ 1 ] ) ; lack of yellow tips to uppertail-coverts (ns [1]); olive-grey vs broadly yellow-tipped and -edged rectrices with entire undertail bright yellow (3) — total score 8; • longirostris by its smaller size (effect size for bill -6.03) (3); large round yellow vs olive-green lores (3); all-olive-grey vs bright yellow-fringed (on inner webs) rectrices (3); narrow whitish vs narrow yellow inner fringes to tertials ( 1); yellower ear-coverts (ns [1]) — total score 10. Recordings kindly sent by M. Thibault reveal only very simple calls: (a) a flat nasal penetrating tuuu-tuuu-tuuu-tuuu (3-4 notes separated by short pauses); (b) a high, thin, dropping-then-rising TSIIiuuuuii, starting like a squeaky toy but ending more richly whistled, this evidently the tweeuwip described by Linsley (1995) and mentioned in Coates & Bishop (1997); and (c) an equally high thin squeaky toy zu-WIIIT! zu-WIIIT! zu-iVIIIT! — these last sounds not dissimilar to those recorded from platenae (see below) but much thinner in tone, lacking the latter’s thrush-like richness. Linsley (1995) also mentioned groups giving ‘raucous calls reminiscent of Charmosyna placentis although without the harsh or scratchy quality of that species’. Taxon affinis (Seram) Morphological differences from lucasi (and by extension chloris ), aurea and platenae are scored above and below. It differs from • chloris by its greater size (effect size for bill length 4.68) (2); half- wedge yellow lores vs all olive-green lores (2); yellow tips to uppertail-coverts (1); rectrices broadly tipped and edged yellow (entire undertail bright yellow) vs olive-green (3) — total score 8; • mysticalis by its slightly larger size (effect size for bill length 1.99) (1); lack of partial yellow eye-ring (2); yellow vs olive- green belly to vent (3); rectrices broadly tipped and edged yellow (entire undertail bright yellow) vs olive-green (3); yellow tips to uppertail-coverts (ns[ 1 ] ) — total score 9; • longirostris by its rather smaller size and notably shorter tail (effect size for latter -4.82) (2); half-wedge yellow vs olive-green lores (2); darker and more extensive olive-green breast (2); different tail pattern, with broad yellow tips and all-yellow undersides vs broad yellow edges on both surfaces (3) — total score 9. A recording by F. R. Lambert (AV4805, XC67566) captures a single song strophe which starts with some scratchy clucking calls and then abruptly turns into a sequence of seven strong rich flat whistles, each longer and perhaps a shade lower in pitch than the previous, the last note most obviously lower: p’tupwupwud’p- p’TI-WI-WII- Will- Will I- Will II- WUUUUU. Another, by JAE, involves a very similar song but with the last two notes rolled throatily. Rheindt & Hutchinson (2007) also describe this song (‘a clean descending melodious whistle’) and present a sonogram of it. Isherwood et al. (1997) found that at one of their study sites (Wae Salas) ‘this species was found to possess a distinct variety of the usual call’, and Coates & Bishop (1997) independently mentioned two types of song (see ‘Conclusion and conservation’). Taxon flavicaudus (Ambon) Bonaparte (1850) gave a nugatory diagnosis of this taxon (translated from Latin: ‘olivaceous green, greenish-yellow below; throat, undertail mostly strong yellow’), but his scientific name nails the only discernible plumage difference from affinis-. in the rather small sample in NHMUK the specimens appear to have less olive markings in the rectrices than those of affinis and hence seem more fully yellow-tailed. White & Bruce (1986) suggested that flavicaudus males ‘tend to be lighter and yellower dorsally and on the breast, with a deeper yellow throat’, but admitted that ‘it is only a slightly differentiated form’. Measurements suggest that flavicaudus is also marginally larger than affinis (Tables 1 and 2). Consequently, always accepting that a larger sample of flavicaudus may show all these slight differences to be inconstant, flavicaudus is provisionally retained here as a valid taxon, but it is clearly conspecific with affinis. Given the proximity and biogeographical unity of Seram and Ambon, this is hardly surprising. Recordings of flavicaudus could not be found. 22 N. J. COLLAR, J. A. EATON & R. 0. HUTCHINSON Forktail 29 (2013) Taxon mysticalis (Burn) Differences from lucasi (and by extension cbloris ) and ajflinis (including flavicaudus) are scored above; chose from aurea and platenae are given below. Wallace (1863) gave this taxon the name mysticalis (not, incidentally, mystacalis ), meaning moustached (Jobling 2010), evidently because of its ‘remarkable half-yellow gape-bristles’. This character (rictal bristles yellow basally, black distally) is not particularly striking in specimens or photographs, nor is it unique within the complex, being shared with platenae and to a lesser degree with other taxa which show yellow lores; but olive-lored member taxa have all-black rictal bristles). Unique to mysticalis , however, is the extent of olive-green on the undersides, with only vague areas on the chin and vent being distinctly shaded yellow, the rest having the merest yellow tinge (score 3). It further differs from longirostris (including barterti) by its considerably smaller size and notably shorter tail (effect size -3.62) (2); half-wedge yellow vs olive-green lores (2); partial yellow eye-ring (ns [2]); dark olive- grey rv bright yellow-fringed rectrices (3); narrow whitish vs narrow yellow inner fringes to tertials (ns [2]) — total score 10. A recording by F. R. Lambert (AV4147, XC 67565) consists of single nervous low clucks, with occasional higher, very rapid chatters, and three times a drawn-out, flat whistle with a very curt downward inflection at the end, tweeeee(ub). These three calls also feature in recordings by JAE, but with the drawn-out whistle starting with a distinct short higher strangled tone, tswiUUUUUU(uh). However, other recordings by JAE also capture a series of song-phrases, starting with hesitant staccato accelerating notes before breaking into longer, musical whistles on (sometimes greatly) varying pitches and sometimes with glissandos, somewhat reminiscent of a domestic canary: pip up... pip-up... pipup-pipupipu WEE- WEE-WEE-puu -puu-puu- WEE-puii-PII- WEE- WEE- WEE. Jepson (1993) reported: ‘Call comprised a descending “si- si-seeow seeow seeow”, and typical bulbul chattering notes’. Taxon longirostris (Sula) As the name given it by Wallace (1862b) indicates, this form is the longest-billed taxon in the complex, although flavicaudus runs it close, and it is altogether the largest form, with the possible exception of aurea. It differs from lucasi (and by extension cbloris ), ajflnis (including_/7 avicaudus), mysticalis , aurea and platenae by the characters scored under those taxa. It differs little from barterti (see below). Recordings of longirostris by ROH all contain a song that consists of a throaty, rolling cb(a)rrrr, rapidly repeated several times and accelerating before breaking into a loud jumble of short whistled notes, some very clear: cbarrr... cbarrr... charrr-charrr-charrr- d idly!) 0 OdidlyD 0 O dully I) 0 0 ; or cbarrr... cbarrr... charrr-charrr- charrr-wididlyWAAbeDIbeDI, etc. However, the cbarrr component may nor be obligate, given the evidence under barterti below. Taxon harterti (Peleng, Banggai) Stresemann (1912) separated this form from longirostris on account of the darker olive coloration of the breast, less yellow upperparts and narrower yellow edges to the outertail. Specimens in SMTD, where 10 barterti are held alongside 7 longirostris , confirm this diagnosis; but as Eck (1976) observed, barterti is ‘only subtly differentiated’ (which is true also of its morphometries: see Table 1) and on morphological grounds it must remain a subspecies of longirostris , as biogeography might predict. Recordings by ROH reveal song-phrases similar or identical to those of longirostris ; however, two by P. Verbelen (AV3344, 3345) are of a singer that gives several clucks and only one very brief cbarrr before launching into its song, suggesting that the cbarrr component may be a separate call that is sometimes run together with the song. Taxon aurea (Togian) While noting the morphological proximity of this lorm to longirostris (which is indeed the closest taxon in plumage and size), Walden (1872) diagnosed it on its smaller size, ‘much shorter bill’ and ‘bright golden colouring of its plumage’. However, while a female specimen (ZMB 2000/26784) conforms in these respects, the type of this taxon, a male, actually has wing and tail longer and bill only 1.6 mm shorter than the mean for two male longirostris (Table 2). Both specimens are distinguished by their notably more golden-yellow underparts (2); much reduced yellow fringes to the tips and inner vanes of the rectrices (2); vague half-wedge yellow lores below a very narrow blackish-brown frontal supercilial line and notably darker olive-green crown (2); rump a shade yellower, less green (ns [1], well shown but perhaps a shade too obvious in Fishpool & Tobias 2005: 236); and presumed shorter bill (allow 1) — total score 7. Acoustically, aurea seems rather close to longirostris /barterti. However, multiple recordings by ROH on different dates suggest that (a) the homologous call in aurea to the 'cb(a)rrr call of longirostris lacks the latter’s rolling throaty quality, and (b) the short fluty babbling song is somewhat abrupt and simple in longirostris whereas in aurea it can be more protracted and typically ends with a set of very rich notes, slightly tailing off in pitch and volume, vaguely recalling the yaffling cadence of a Green Woodpecker Picus viridis. Taxon platenae (Sarsgihe) This is the most isolated, most threatened and in some ways most distinctive form in the Golden Bulbul complex. Blasius (1888), working with two syntypes (illustrated, with a photograph of one of them, in Hevers 2004), accurately characterised this bird as closest to aurea and longirostris but distinguished by its shorter bill (this is true for longirostris but not tor aurea). almost entirely uniform olive- green upperparts, and vivid yellow colour of the chin, throat, submoustachial area, eye-ring and inner vanes of all five outer rectrices. Our own examination of the only three specimens in existence (SNMB N 13945 and N43300, and RMNH [Naturalis] 84768) indicates that it is distinguished from all other taxa by its bright yellow triangular lores (much fuller and brighter than the yellow triangular lores of mysticalis against which it is here scored on this feature) extending to and contiguous with the eye-ring (2); bright yellow eye-ring, only broken by a narrow gap at the rear of the eye (much more obvious and complete than in mysticalis , in which it is confined to the ‘brow’ and a short arc on the lower rear edge) (3); yellowish ear-coverts and yellow submoustachial area, producing a broad yellow throat (ns [2]); and very broad yellow fringes to the inner vanes of the rectrices extending the length of the feathers, creating a different pattern from other taxa (2) — total score 7. A recording by P. Verbelen (AV3347) consists of a vigorously delivered series of fairly short, simple strophes composed of little groups of repeated thrush-like whistles. Recordings of this form by ROH reveal a consistent pattern of song, comprising two short abutting components, (a) four nasal but rich notes, each rising in pitch but each lower than the previous, the last cutting to (b) usually three high whistled notes, approximately: cui-cui-cui-cui-DEEP- pDEEP-pDEEP! (As noted above, in structure these sounds vaguely resemble those on a recording of T. lucasi , but are much richer and less strangled in tone.) CONCLUSION AND CONSERVATION Fishpool & Tobias (2005) separated the Golden Bulbul into Northern longirostris (with cbloris, lucasi , barterti, aurea and platenae as races) and Southern ajflnis (with flavicaudus and mysticalis as races) on account of their songs, the former lacking Forktail 29 (2013) Species limits in the Golden Bulbul Alophoixus ( Thapsinillas ) affinis complex 23 the ‘long sliding notes and descending cadence’ of the latter, affinis and flavicaudus possessing ‘a distinctive mournful series of sweet and minor-key notes, lasting 2-4 seconds, slightly erratic or meandering in pace and note length, but essentially slow and leisurely, sliding down scale almost throughout’, mysticalis ‘vaguely similar but much more complex’ — and hence a reason why Rheindt & Hutchinson (2007) recommended its separation Irom affinis. However, while Coates & Bishop (1997) support the account of the voice of affinis (‘main song... a lovely descending series of c. 15 short, clear, mellow whisdes... slightly slurred as the song dies away’) they also mention a second song type, ‘a rapidly swelling series of 20-30 pure, high-pitched whistled notes that climbs to a notably high pitch and ends abruptly’. Moreover, the clear resemblance of songs of chloris and affinis tends to confound the notion of a north- south divide in song types. This all suggests that the vocalisations of the taxa in this complex may be considerably more varied but also perhaps ultimately more homologous than we yet know, and that the sample used in descriptions above should not be considered anything more than partially representative. Even so, from the very limited material available to us we derive the impression that vocal differences largely support the seven-way split of the Golden Bulbul complex which the morphological evidence indicates, using the scoring system of Tobias et al. (20 1 0) : Halmahera Golden Bulbul Thapsinillas chloris Morotai, Halmahera, Bacan Obi Golden Bulbul Thapsinillas lucasi Obi Seram Golden Bulbul Thapsinillas affinis T. a. affinis Seram T. a. flavicaudus Ambon Buru Golden Bulbul Thapsinillas mysticalis Buru Sula Golden Bulbul Thapsinillas longirostris T. 1. longirostris Sula T. 1. harterti Peleng, Banggai Togian Golden Bulbul Thapsinillas aurea Togian Islands Sangihe Golden Bulbul Thapsinillas platenae Sangihe The conservation status of these seven species will require formal assessment against the IUCN Red List criteria, but a few preliminary remarks may be made here. From evidence in Fishpool & Tobias (2005), our own observations in the field (JAE and ROH) and material cited below, the first six species in the list above are relatively common in their various woodland/ forest habitats. Poulsen & Lambert (2000) tabulated records of chloris (Halmahera) indicating a high encounter rate, with birds found (albeit less commonly) even in mangrove. Linsley (1995) saw lucasi (Obi) in ‘small numbers (less than ten)... daily’, with two instances of breeding evidence ‘in scrub on the edge of disturbed forest’. Bowler & Taylor (1989) reported affinis (Seram) ‘common and widespread... in forested areas’ from sea-level up to c.900 m, while JAE saw them up to at least 1,300 m; Isherwood et al. (1997) also found the species common. Jepson (1993) called mysticalis (Buru) ‘common and widespread... in all types of forest’ (confirmed in Poulsen & Lambert 2000, and by JAE, ROH pers. obs.). Stones et al. (1997) found longirostris (Sula, specifically Taliabu) ‘abundant at each study site, in all habitat types surveyed, but most common in primary forest, both lowland and montane’ (confirmed by JAE, ROH pers. obs.), while Indrawan et al. (1997) reported harterti (Peleng) as ‘commonly seen’ in groups of three to four birds... in degraded forest at Monggias’ (confirmed by JAE, ROH pers. obs.). Coates & Bishop (1997) were concerned that aurea (Togian Islands) was ‘apparently rare and local’, but Indrawan et al. (2006) documented records from three of the seven larger islands in the group, finding it ‘relatively frequently’ on Togian itself and ‘relatively common’ on Walea Bahi (confirmed by ROH pers. obs., and J. Riley in litt. 2013). The status of platenae (Sangihe) is, however, worrying. Although Bishop (1992) observed it ‘commonly in secondary woodland and mixed tree crop plantations’ during a visit over lb- 19 May 1986, others have not been able to repeat this finding (Riley 1 997a,b). A year before, on 30 May 1 985, a male specimen (RMNH 84768) was collected on Gunung(Gn) Sahendaruman in ‘primary forest on eastern slope: 750 m: S of Liwung and SW of Kuma’ (Naturalis label data) by F. G. and C. M. Rozendaal, but it took until November 1996 before the species was seen again, with records of three birds twice and one bird once on three days, all evidently in the same area-on Gn Sahengbalira (Riley 1997b). These records were the only ones in four months’ fieldwork in 1995 and 1996, when the only local people to recognise photographs of the species (presumably from museum skins) were ‘in the village closest to the forest on Gunung Sahengbalira’ (Riley 1997b). Further fieldwork on Sangihe between August 1998 and March 1999 led Riley (2002) to suggest that platenae ‘is one of the island’s most endangered species’, being found only on Gn Sahendaruman with an estimated population of 50-230 birds. However, he noted that it was missed at one locality when not calling but found to be common there when it became vocal (Riley 2002), thereby confirming an earlier remark that ‘this can be a cryptic species, despite its bright coloration’ (Riley 1997b). Even so, visits to its small fragment of remaining habitat on Gn Sahengbalira in recent years have not produced any evidence to revise the view that this species is in trouble: JAE and ROH found four birds in August 2004, although a subsequent visit over two days in 2012 by ROH failed to record any. Of other observers visiting the area this century, R Verbelen saw several in November 2008 but B. Demeulemeester, R Gregory, J. Hornbuckle, C. Robson and M. Thibault ( in litt. or verbally to JAE, ROH) all failed to find it. Consequently, we judge that the Sangihe Golden Bulbul now requires urgent attention in order to secure its future. Clearly it would be valuable if this new arrangement of Thapsinillas were to be tested and corroborated by molecular study. Such work might also reveal the biogeographic history and colonisation routes of the taxa across this unusual range (which no other species or genus shares) . Moreover, a far more comprehensive sampling of vocalisations would also be of great interest, in part simply to determine the variation within individual taxa, in part to assess more confidently the degree of difference between taxa, and in part to test whether such differences correspond to the hoped- for molecular evidence. ACKNOWLEDGEMENTS We thank Robert Prys-Jones (NHMUK), Stephen van der Mije and Cees Roselaar (Naturalis), Michaela Forthuber (SNMB), Martin Packert (SMTD) and Sylke Frahnert (ZMB) for access to the specimens in their care, Alison Harding (NHMUK) for finding and assembling copies of the earlier literature on the complex for review, and those mentioned in the text for their recordings and records of members of the complex. We also thank L. D. C. Fishpool and J. A. Tobias for their helpful comments as referees. REFERENCES Andrew, P. (1992) The birds of Indonesia: a checklist (Peters' sequence). Jakarta: Indonesian Ornithological Society. Bishop, K. D. (1992) New and interesting records of birds in Wallacea.Ru/u7a 6: 8-34. 24 N. J. COLLAR, J. A. EATON & R. O. HUTCHINSON Forktail 29 (2013) Blasius, W. (1888) Criniger platenae nov. spec. Braunschweig. Anzeig. 9: 86. Bonaparte, C. L. (1850) Conspectum generum avium, 1. Leiden: E. J. Brill. Bowler, J. &Taylor, J. (1 989) An annotated checklist of the birds of Manusela National Park, Seram (birds recorded on the Operation Raleigh Expedition). Kukila 4: 3-29. Coates, B. J. & Bishop, K. D. (1997) A guide to the birds ofWaiiacea. Alderley, Queensland: Dove Publications. Delacour, J. (1943) A revision of the genera and species of the family Pycnonotidae (bulbuls). Zoologica 28: 17-28. Dickinson, E. C. & Dekker, R. W. R. J. (2002) Systematic notes on Asian birds. 25. A preliminary review of the Pycnonotidae. Zool. Verhand. 340: 93- 1 14. Dickinson, E. C. & Gregory, S. M. S. (2002) Systematic notes on Asian birds. 24. On the priority of the name Hypsipetes Vigors, 1831, and the division of the broad genus of that name. Zool. Verhand. 340: 75-91. Eck, S. (1 976) DieVogel der Banggai-lnseln, insbesondere Pelengs (Aves). Zool. Abhandi. Staatl. Mus. Tierk. Dresden 34: 53-100. Finsch, O. (1867) Ueber die Arten und das Genus Criniger. J. Orn. 1 5: 1 -36. Fishpool, L. D. C. &Tobias, J. A. (2005) Family Pycnonotidae (bulbuls). Pp.1 24- 251 in J. del Hoyo, A. Elliott & D. A. Christie, eds. Handbook of the birds of the world, 10. Barcelona: Lynx Edicions. Hartert, E. (1903) The birds of the Obi group, central Moluccas. Novit. Zool. 10: 1-17. Hartert, E. (1922) Types of birds in the Tring Museum. Novit. Zool. 29: 365- 412. Hevers, J. (2004) Die Typen des Staatlichen Naturhistorischen Museums in Braunschweig und ihre Autoren. Pp.335-397 in S. Ahrens, ed. 250Jahre Naturhistorisches Museum in Braunschweig. Braunschweig: Staatliches Naturhistorisches Museum. Hombron, M. & Jacquinot, M. (1 841 ) Description de plusiers oiseaux nouveaux ou peu connus, provenant de I'expedition autour du monde faite sur les corvettes I 'Astrolabe et la Zelee. Ann. Sci. Nat. (Paris) (2) Zool. 16: 312- 320. Indrawan, M., Masala, Y. & Pesik, L. (1 997) Recent bird observations from the Banggai islands. Kukila 9: 61-70. Indrawan, M., Somardikarta, S., Supriatna, J., Bruce, M. D., Sunarto & Djanubudiman, G. (2006) The birds of theTogian islands, Central Sulawesi, Indonesia. Forktail 22: 7-22. Inskipp, T., Lindsey, N. & Duckworth, W. (1996) An annotated checklist of the birds of the Oriental region. Sandy, UK: Oriental Bird Club. Isherwood, I. S., Willis, J. D. A., Edwards, T. R. K., Ekstrom, J. M. M„ Kuriake, S., Lubis, I. R., Notanubun, H„ Putnarubun, J., Robinson-Dean, J.C.&Tobias, J. A. (1997) Biological surveys and conservation priorities in north-east Seram, Maluku, Indonesia: final report of Wae Buta ‘96. Cambridge, UK: CBS Conservation Publications. Jepson, P. (1993) Recent ornithological observations from Buru. Kukila 6: 85- 109. Jobling, J. A. (2010) The Helm dictionary of scientific bird names. London: Christopher Helm. Linsley, M. D. (1995) Some bird records from Obi, Maluku. Kukila 7: 142-151. Morony, J. J„ Bock, W. J. & Farrand, J. (1975) Reference list of the birds of the world. New York: American Museum of Natural History (Department of Ornithology). Poulsen, M. K. & Lambert, F. R. (2000) Altitudinal distribution and habitat preferences of forest birds on Halmahera and Buru, Indonesia: implications for conservation of Moluccan avifaunas. Ibis 1 42: 566-586. Rand, A. L. & Deignan, H. G. (1960) Family Pycnonotidae. Pp.221-300 in E. Mayr & J. C. Greenway, eds. Check-list of birds of the world, 9. Cambridge, Mass.: Museum of Comparative Zoology. Rheindt, F. E. & Hutchinson, R. O. (2007) A photoshot odyssey through the confused avian taxonomy of Seram and Buru (southern Moluccas). BirdingASIA 7: 18-38. Riley, J. (1997a) The birds of Sangihe and Talaud, North Sulawesi. Kukila 9: 3-36. Riley, J. (1997b) Biological surveys and conservation priorities on the Sangihe and Talaud islands, Indonesia: the final report of Action Sampiri 1995 - 1997. Cambridge, UK: CSB Conservation Publications. Riley, J. (2002) Population sizes and the status of endemic and restricted- range bird species on Sangihe Island, Indonesia. Bird Conservation International 1 2: 53-78. Sibley, C. G. & Monroe, B. L. (1990) Distribution and taxonomy of birds of the world. New Haven: Yale University Press. Stones, A. J., Lucking, R. S., Davidson, P. J. & Raharjaningtrah, W. (1997) Checklist of the birds of the Sula Islands (1991 ), with particular reference to Taliabu Island. Kukila 9: 37-55. Stresemann, E. (191 2) Ornitologische Miszellen aus dem Indo-Australischen Gebiet. Novit. Zool. 19: 31 1-351. Tobias, J. A., Seddon, N„ Spottiswoode, C. N„ Pilgrim, J. D„ Fishpool, L. D. C. & Collar, N. J. (2010) Quantitative criteria for species delimitation. Ibis 152: 724-746. Walden, A. (1872) On some supposed new species of birds from Celebes and the Togian Islands. Ann. Mag. Nat. Hist. (4)9: 398-401 . Wallace, A. R. (1862a) On some new birds from the northern Moluccas. Ibis 4: 348-351. Wallace, A. R. (1862b) List of birds from the Sula Islands (east of Celebes), with descriptions of the new species. Proc. Zool. Soc. London 1 862: 333- 346. Wallace, A. R. (1863) List of birds collected in the island of Bouru (one of the Moluccas), with descriptions of the new species. Proc. Zool. Soc. London 1863: 18-32. White, C. M. N. & Bruce, M. D. (1986) The birds of Wallacea (Sulawesi, the Moluccas and Lesser Sunda Islands, Indonesia): an annotated check-list. London: British Ornithologists' Union (Check-list 7). N. J. COLLAR, BirdLife International, Wellbrook Court, Girton Road, Cambridge CB3 ON A, UK; and Bird Group, Department of Life Sciences, Natural History Museum, Akeman 5r, Tring, Herts HP23 6AP, UK. Email: nigel.collar@birdlife.org J. A. EATON, A-3A-5 Casa Indah I, Persiaran Sudan, Petaling Jaya, Selangor, 47410, Malaysia. Email: jameseaton@birdtourasia.com R. O. HUTCHINSON, 26 Sutton Avenue, Chellaston, Derby DE73 6RJ, UK. Email: robhutchinson@birdtourasia.com FORKTAIL 29 (2013): 25-30 Birds of Mys Shmidta, north Chukotka, Russia VLADIMIR YU. ARKHIPOV, TOM NOAH, STEFFEN KOSCHKAR & FYODOR A. KONDRASHOV A survey of avifauna was carried out in the Mys Shmidta area, north Chukotka, Russia from 8 June to 1 2 July 2011 . A total of 90 species was recorded in the area, which together with literature data made a final list of 1 04 species. For several species this area is beyond the northern, north-eastern or north-western limits of their known distribution. We collected new data for 19 globally or locally threatened species. Tundra Swan Cygnus columbianus, Emperor Goose Anser canagica, American Golden Plover Pluvialis dominica, Western Sandpiper Calidris mauri, Semipalmated Sandpiper C.pusilla, Northern House Martin Delichon urbica and Barn Swallow Hirundo rustica were all confirmed to be breeding. Breeding of Brent Goose Branta bernicla nigricans, Spectacled Eider Somateria fischeri and Steller's Eider Polysticta stelleri was judged to be 'very likely'. There was no evidence for breeding of Ross's Gull Rhodostethia rosea despite several records. Two Eurasian Dotterels Eudromias morinellus were recorded displaying for the first time in the area, but the status of the species is unclear. The area is important for Snowy Owl Nyctea scandiaca, and as moulting grounds for Emperor Goose. Canada Goose Branta canadensis, Baikal Teal Anas formosa, Bar-tailed Godwit Limosa lapponica, Slaty-backed Gull Larus schistisagus, Thayer's Gull L. thayeri, Black-headed Gull L. ridibundus, White-tailed Eagle Haliaeetus albicilla, Steller's Sea Eagle H. peiagicus, Osprey Pandion haliaetus, Arctic Warbler Phylloscopus borealis and House Sparrow Passer domesticus are more likely to be rare vagrants or migrants. An observation of a Pine Siskin Carduelis pinus is the first record for Eurasia. INTRODUCTION Mys Shmidta or the Cape of Shmidt is a prominent headland on the Arctic coast of Chukotka, Russia, ft is a remote place with a very harsh climate. Information on birds from this area is difficult to obtain but has inherent value because several globally or locally threatened species breed here and human impact on the habitat has been small. No systematic checklist of birds of the area has been made until now. Since Portenko (1972, 1973) worked on the birds of the Chukchi Peninsula, few ornithological papers have been published on the region and no avifauna surveys have been made. Existing publications include a short note by Tomkovich et al. (1991) on the area around Mys Shmidta airport, several notes by Stishov (1991, 1 992, 1 997) including one on the bird community of the Ekvyvatap river, two papers by Stishov & Maryuhnich ( 1 99 1 a, b) on particular species and short communications on brief visits by Andreev & Kondratyev (1996), Dorogoi (1996, 1997, 1998), Dorogoi & Beaman (1998) and Menyushina (2000). In this paper the results of an avifauna survey carried out in the Mys Shmidta area, north Chukotka, Russia, in summer 2011 are presented. Together with data from the literature the final list of birds for the area is 104 species. The purpose of the investigation was to survey this area for potential breeding grounds and suitable habitats for the Spoon-billed Sandpiper Eurynorhynchus pygmeus and to document the avifauna of this remote and hard-to-reach region, which had previously received little ornithological attention. STUDY AREA AND METHODS During an ornithological expedition to the Mys Shmidta area between 8 June and 12 July 2011 to survey breeding areas of the Spoon-billed Sandpiper, the other avifauna was also surveyed. Birds were identified mainly using binoculars and telescopes, and the species, number of individuals and habitat was noted in each case. Overall, research was focused on the Akatan lagoon and the Ekvyvatap river delta, plus as much of the Kosa Dvukh Pilotov Spit (Spit of Two Pilots) as could be accessed on foot or by kayak, the area immediately surrounding Mys Shmidta settlement and an area to the north-west of Mys Shmidta around the Erokynmanky lagoon (Table 1). There were four main habitat types in the area: gravel- sand, tussock tundra, grassy tundra and polygonal tundra. The spits running north-west to south-east were mainly gravel and sand, either pure gravel-sand or only sparsely covered with lichens and grass, at least on the main narrow sections running immediately parallel to the ocean. About 60% of the spits visited were covered by this type of vegetation. The higher ground on the spits and the areas around the lagoons were generally cotton-grass tussock tundra, and grassy tundra was found on lowlands surrounding the spits, which were still flooded until almost the end of June, being frequented by different species of geese. Dried polygonal tundra was found in patches around the Ekvyvatap river delta. Small hills around the settlement were covered by typical tussock tundra. The temperature in the region is generally cold; the highest temperature experienced during the day was 14°C and the typical daytime temperature was 5-7°C. Storms were f requent, with a week- long storm in the first week of July with strong winds and heavy snow. Table 1. Main survey sites in the Mys Shmidta, north Chukotka area, 2011. Site Coordinates Mys Shmidta, airport area Mys Shmidta, Kozhevikova cliff Mys Shmidta, sea coast Mys Shmidta, base of Kosa Dvukh Pilotov spit Mys Shmidta, settlement Tundra near Ryrkaypiy settlement Tundra near Mys Shmidta settlement 68.872°N 179.386°W 68.931°N 179.490°W 68.902°N 179.425°W 68.872°N 179.358°W 68.890°N 179.409°W 68.896°N 179.458°W 68.889°N 1 79.43 1°W Ekvyvatap River Delta & nearby tundra Akatan lagoon & Kosa Dvukh Pilotov spit Akatan lagoon mouth, Kosa Dvukh Pilotov spit 68.783°N 178.99EW 68.753°N 178.983°W 68.780°N 178.956°W Tynkergynpil'gyn lagoon, Kosa Dvukh Pilotov spit Tynkergynpil'gyn lagoon mouth, Kosa Dvukh Pilotov spit 68.683°N 178.689°W 68.573°N 178.449°W Erokynmanky lagoon, coast Erokynmanky lagoon, Odnobokiy stream Erokynmanky lagoon, Nutechikun spit Erokynmanky lagoon, tundra Erokynmanky lagoon, hills 68.951°N 179.918°E 68.969°N179.949°E 68.992°N 179.925°E 68.944°N 179.986°E 68.903°N 179.867°E RESULTS Ninety species were recorded in the area which together with literature data yielded a final list of 104 species (Appendix 1). Details of the most interesting and significant are given below, 26 VLADIMIR YU. ARKHIPOV etal. Forktail 29(2013) including new records, new breeding records, regional rarities and globally threatened species. Whooper Swan Cygnus cygnus Two records: one across the Akatan lagoon on 18 June was pursued by a Tundra Swan that appeared to strike it on the head and neck. Another Whooper Swan stayed near a tundra lake with a Tundra Swan on 23 and 24 June, in the Ekvyvatap delta. Images were obtained as these records are far to the north of known nesting areas (Krechmar & Kondrat’ev 2006). Tundra Swan Cygnus columbianus A common breeding species around Mys Shmidta: six nesting pairs within 25 km2 were recorded in the Ekvyvatap delta tundra. Some non-breeding pairs were also recorded, but no large groups were seen. A considerable number of birds had a mix of traits attributable to Cygnus columbianus bewickii or C. c. columbianus ; some individuals that could be reliably identified as either bewickii or columbianus belonged to mixed pairs. Of note, in all such mixed pairs the males appeared to possess columbianus traits. Possibly the study area is located in the area of overlap of the two forms (Rees 2006). Overall three nests with eggs were found: a nest in the Ekvyvatap River delta with one egg on 14 June and three eggs on 19 June; a nest in the same area found on 23 June but not checked for eggs; and a nest with three eggs found on 7 July on the Nutechikun spit. Snow Goose Anser caerulescens Breeding was not recorded. From 9 to 28 June Snow Geese were seen regularly with a maximum of 64 birds in several flocks on 13 June in the mouth of the Ekvyvatap River and 60 birds in four flocks on 14 June; all flocks were flying north-west. All observed individuals belonged to the white morph. Residents of Mys Shmidta reported that there is a pronounced autumn migration, occasionally reaching thousands of birds per day. Three Snow Geese rings were obtained from local hunters; two individuals, a male and a female, had been ringed as adults on Wrangel Island and one male as a juvenile in Alaska, 45 km east of Deadhorse. Emperor Goose Anser canagica Near Threatened. Common breeding species in the Ekvyvatap delta and surrounding area. W e observed several pairs holding territories, including one chasing away a Parasitic Skua Stercorarius parasiticus. A nest with three eggs was found in the Ekvyvatap delta on 14 June and a second nest was found nearby on 26 June. A pronounced migration of Emperor Geese to the north-west of Mys Shmidta was observed from 17 to 29 June. A maximum of 1,670 flew north¬ west between lOhOO and I6h00 on 27 June. It appears that this species is expanding its range westwards, as it was not observed in the vicinity of Mys Shmidta in the early twentieth century, but by 1970 it was breeding to the west ofUkouge lagoon (Portenko 1972) about 100 km south-east of the eastern part of the surveyed area. Hunters handed over a ring from a male bird collected at nearby Ryrkaypiy settlement; the bird had been ringed as an adult in Alaska 37 km south-east of Chevak. Canada Goose Branta canadensis A solitary Canada Goose was observed on 19 and 20 June in the Ekvyvatap delta. It was flying with two Greater White-fronted Geese Anser albifrons and was slightly larger than them. Judging by several traits, such as the relatively short bill and neck, it may be th eparvipes subspecies, common in northern Alaska. It appears that this is the first record of this form in Russia, apart from the introduced population in Eurasia (Koblik etal. 2006). Prior to this observation there was only one record of a wild Canada Goose in East Asia, seen on Hokkaido, Japan, in 2006, which was also presumed to be a parvipes form (Brazil 2009). The form minima was also observed in the vicinity of Anadyr airport in Chukotka. A solitary bird was seen in a flock of Brent Geese on a lagoon near the airport from 3 to 5 June 2011. Images of both birds were obtained. Brent Goose Branta bernicla Observed daily from 9 to 27 June in flocks and pairs in the Ekvyvatap delta and nearby on the Akatan lagoon with up to 143 birds on 17 June. On 1 1 June one pair was observed to be possibly breeding, but the birds were not seen on subsequent days. However, breeding is possible, as sporadic breeding pairs were observed in the first half of the twentieth century in the vicinity of Mys Shmidta (Portenko 1972). This species is commonly hunted by the local population; in the late 1990s one female shot in the vicinity of Polyarniy settlement had been ringed as a juvenile in Alaska, 20 km to the south of Chevak. Baikal Teal Anas formosa A single male was swimming on a lake near Mys Shmidta on 21 June. The area is outside the current breeding range (Krechmar & Kondrat’ev 2006), although in the first half of the twentieth century the species was observed in the area relatively frequently (Portenko 1972). Spectacled Eider Somateria fischeri Uncommon, possibly breeding in the tundra near the sea. Three pairs were observed on 26 June on lakes in the Ekvyvatap delta. Overall it was observed on 12 days with a maximum daily count of 32; all were flying to the north-west on 1 6 June, and 30 birds, mostly young males, were observed on the sea near Mys Shmidta on 10 July. Stishov (1992) found the species to be common on lakes and bogs around the lower reaches of the Ekvyvatap River. Steller's Eider Poiysticta stelleri Vulnerable. Common to uncommon at the time of migration. Possibly breeding in the tundra on coastal lakes. Four pairs were holding territories on 14 June in Ekvyvatap delta. The species was regularly seen in the vicinity of Mys Shmidta, with a maximum of 60 birds, including some pairs, on 9 June. Unfortunately the species, which is Vulnerable, is hunted in the area; eight rings were collected from individuals shot near Mys Shmidta; all had been ringed in Alaska. Bar-tailed Godwit Limosa lapponica Vagrant; only one record of a female feeding on tundra near the Ekvyvatap delta on 26 June. It is also a rare vagrant to Wrangel Island (Stishov etal. 1991). Red Knot Calidris canutus Seven observations were made, all on the sand spits or mudflats nearby: flocks of nine and five birds on 1 1 and 14 June on mudflats near the mouth of the Akatan lagoon; three flew north across the Akatan lagoon in the direction of Wrangel Island on 18 June; two were in tundra on the Spit ofTynkergynpil’gyn lagoon on 20 June with two more 1 .5 km away; in the same area on 21 June a bird was resting and also in the same area two birds were seen on 22 June; finally on 27 June three birds were in tundra on the spit of Akatan lagoon. Although most records were made on lichen tundra — suitable breeding habitat — displays were not seen or heard. Three individuals had colour flags and had been ringed in Australia — two with yellow flags in north-west Australia, and one with a green flag at Moreton Bay, near Brisbane, Queensland. Western Sandpiper Calidris mauri Late arrivals; during the first survey on 14 June in Ekvyvatap delta no more than five displaying males were seen, and on the following days, several other birds appeared in nearby areas. Two breeding Forktail 29 (2013) Birds of Mys Shmidta, north Chukotka, Russia 27 areas were found near Mys Shmidta. The first included 10 territories on the Ekvyvatap delta tundra, where a nest with four eggs was found on 25 June. The second area was to the north-west of Mys Shmidta on the Nutechikun spit, where a nest with four eggs was found on 7 July. Spoon-billed Sandpiper Eurynorhynchus pygmeus Critically Endangered. During the surveys the species was not recorded. However, the expedition found suitable habitat in the region, although limited in extent, and obtained invaluable information on the habitat in the area. The last record of this species in the area was in mid-June 1990 when several displaying males and at least two pairs were found in the Ekvyvatap delta (Stishov & Maryuhnich 1991a). In the 1970s Kishchinski (1988) found it to be a common breeder at the Ukouge lagoon about 100 km south¬ east of the eastern part of the area surveyed. ' Semipalmated Sandpiper Calidris pusilla Near Threatened. A rare breeding species in Eurasia, it was found nesting in areas with sparse vegetation on gravel and sandy patches of the tundra near lagoons. Breeding was recorded in two areas: Ekvyvatap delta (2-3 territories) and in the vicinity of Mys Shmidta, where on 9-1 1 June up to 10 males were displaying. A brood with four two-day-old chicks was observed on 25 June on the edge of the town. On 28 and 29 June in the vicinity of Mys Shmidta three more broods were observed and one pair was making alarm calls. The first nest of this species in this area was found on 8 June 1993 near the Mys Shmidta airport (Andreev & Kondratyev 1 996). In 1997 at least three pairs were breeding between the airport and the town; a nest at an advanced stage of incubation found on 1 July 1997, and on 4 July 1997 the same nest already had chicks. Two more nests with chicks were observed nearby on the same day, and one pair was observed between the airport and the town on 6 July 1997 (Dorogoi & Beaman 1998). Thus, the population of the Semipalmated Sandpiper in the vicinity of Mys Shmidta continues to thrive and, possibly, is growing slightly. American Golden Plover Pluvialis dominicus One territorial pair near a nest with a clutch of four eggs was found in the Ekvyvatap delta, with observations continuing between 23 June and 25 June. Images of adults and nest were obtained. This is the first documented breeding record in the Palearctic with both birds confirmed as American Golden Plovers. Prior to this observation an unsuccessful breeding attempt by an American Golden Ploverwith a Grey Plover Pluvialis squatarola was reported (Taldenkov 2006). In 1987-1990 on the lower reaches of the Ekvyvatap River about 30% of plovers had traits that resembled American Golden Plovers (Stishov 1991) and the same author mentioned breeding of the American Golden Plover in the same area in 1990-1993 without further details (Stishov 2004). Eurasian Dotterel Eudromias morinellus Three records in 2011. A solitary bird was displaying over the Ekvyvatap delta tundra on 14 June. On 15 June in the same area two displaying birds were seen about 1 km from each other, but during later surveys in this area the species was not seen. On a flat hilltop near the Odnobokiy stream a flock of nine was seen on 2 July. This species is regarded as rare since prior to our observations it was reported in Chukotka only once since 2000 (Tomkovich 2007a; P.S. Tomkovich in litt.). Records of Eurasian Dotterels near Mys Shmidta are contained in reports on breeding tundra birds in Russia (Dorogoi 1996, 1997). Thayer's Gull Larus thayeri On 17 June, one bird was resting on the ice of Akatan lagoon with Vega Gulls Larus vegae. The bird was about a third smaller than the Vega Gulls. This is the third record of Thayer’s Gull for Russia. Black-headed Gull Larus ridibundus Vagrant. On 18 June a bird in second-summer plumage was seen near the mouth of the Akatan lagoon and on 20 June two adult birds were swimming in a small bay near Kozhevnikova Cliff. Ross's Gull Rhodostethia rosea A summer visitor and migrant. Over 10 days between 10 June and 25 June juvenile and adult Ross’s Gull were observed at the confluence of the Akatan lagoon with up to eight birds on 1 1 June. Common Murre Uria aalge or Thick-billed Murre U. lomvia We observed seven murres flying north-west over the sea on 10 July, but unfortunately too far away for definitive identification to species level. The predominant species on Wrangel Island is Thick¬ billed Murre, but several hundred Common Murre also breed there (Stishov era/. 1991). Osprey Pandion haliaetus On 3 July an Osprey was seen attempting unsuccessfully to catch fish on Erokynmanky lagoon. Images were obtained. This is the most north-easterly record of the species in Eurasia and approximately 600 km from its known breeding range (Chereshnev 2008). White-tailed Eagle Haliaeetus albieilla The Mys Shmidta area is about 500 km from the known breeding range of the White-tailed Eagle (Chereshnev 2008). In June 2011 there were three records of White-tailed Eagles: a subadult was sitting on the sea ice in the mouth of Akatan lagoon on 13 June, an adult was seen in the mouth of Tynkergynpil’gyn lagoon on 23 June, and a first-year was flying along Two Pilot Spit on 27 June. Steller's Sea Eagle Haliaeetus pelagicus Vulnerable. An adult was observed on 27 June on the sea ice in the mouth of Akatan lagoon, and images were obtained; after about 10 minutes, the bird flew away to the north-west. This is the northernmost record of Steller’s Sea Eagle in Asia (BirdLife International 2001, Chereshnev 2008). Bam Swallow Hirundo rustica Two pairs of Barn Swallows were regularly observed in the settlements of Mys Shmidta and Ryrkaypiy. The pair in Mys Shmidta was nest building on 29 June inside an abandoned house. After the severe weather in early July, the nest was empty and birds were not seen again. One and two Barn Swallows were seen on 12 and 17 June respectively, on the spit near the mouth of the Atakan lagoon. Most of these records were of white-bellied birds. However, in Mys Shmidta settlement birds with bright red bellies were recorded alongside the white-bellied form: one bird on 8 June and two on 29 June, it is possible that these birds were the American subspecies H. r. erythrogaster. U nfortunately all observations of red- bellied birds were too brief for identification to subspecies to be confirmed. Northern House Martin Delichon urbica Breeding species in Mys Shmidta settlement. Nine nests were found on two four-storey buildings in June 2011. Following the severe storm at the beginning of July only two nests were still occupied. This record is the most north-easterly breeding record of this species in Eurasia. House Sparrow Passer domesticus Vagrant. On 17 June, an adult male was seen on Kosa Dvukh Pilotov Spit near an abandoned hut; an hour later it was seen again 28 VLADIMIR YU. ARKHIPOV etal. Forktail 29 (2013) near another abandoned hut about 3 km to the east. The closest breeding location is Pevek where breeding was reported in previous years (Tomkovich 2007b) and at least 10 lamilies with broods were seen there on 13 July 2011. Pine Siskin Carduelis pinus On 9 and 10 June 2011 we found a Pine Siskin on a sandy gravel spit near Mys Shmidta (68.872°N 179.358°W). The bird was not shy and it was possible to approach to a few metres and images were obtained. This is the first documented record for Russia and Eurasia (Arkhipov etal. in press). DISCUSSION During the 2011 expedition to northern Chukotka 90 species were recorded, 30 of which were first records for the wider region. Overall, this confirms the paucity of previous studies of this area. Including additional data Irom the literature, a total of 1 04 species has been recorded for the region indicating a relatively complete list in comparison with nearby areas. Near Pevek town 75 species have been recorded (Tomkovich 2007b) and 152 species have been found on Wrangel Island (Stishov et al. 1991). Nevertheless, the list may be far from exhaustive and be expanded by subsequent visitors, especially in respect of migrant birds. Northern House Martin and Barn Swallow were recorded attempting to breed around the settlements; these species had never previously been recorded breeding this far to the north-east. Breeding of several other species on the borders of their known western or north-western breeding range was recorded: Emperor Goose, American Golden Plover, Western Sandpiper and Semipalmated Sandpiper. Several other rare or endangered species were observed that were suspected to be breeding but without direct evidence, including two displaying Eurasian Dotterels and many Snowy Owls Nyctea scandiaca , with up to 13 individuals recorded per day. Emperor Geese moult in the area and a maximum day count of 1,670 was recorded. Finally, several species were observed which may be migrants, including Canada Goose, Baikal Teal, Bar¬ tailed Godwit, Slaty-backed Gull Larus schistisagus, Thayer’s Gull, Black-headed Gull, White-tailed Eagle, Steller’s Sea Eagle, Osprey, Arctic Warbler Phylloscopus borealis and House Sparrow. An observation of Pine Siskin is the first record for Eurasia. Disappointingly, in 2011 neither the Critically Endangered Spoon-billed Sandpiper nor the Near Threatened Yellow¬ billed Loon Gavia adamsii, both of which have been recorded in the area in previous years, were seen. ACKNOWLEDGEMENTS We thank Natalya Kveten and Oksana Makarova, heads of administrations of Mys Shmidta and Ryrkaypiy for hospitality and for help with organising our excursions. Warm thanks too to Pavel Tomkovich for useful comments on local birds and ornithological literature. We are very grateful to The David and Lucile Packard Foundation for the support to Birds Russia’s Spoon-billed Sandpiper conservation programme in 2011 and to Evgeny Syroechkovsky Jr, the leader of the Spoon-billed Sandpiper conservation team in Russia. REFERENCES Andreev, A. V. & Kondratyev, A. V. (1996) A new case of breeding Semipalmated Sandpiper in Chukotka. Information Materials of the Working Group on Waders 9: 34 (In Russian.) Arkhipov, V., Koshkar, S. & Noah, T. (in press) First record of Pine Siskin Carduelis pinus in Eurasia. BirdingASIA. B i rd Life International (2001) Threatened birds of Asia: the BirdLife International Red Data Book. Cambridge UK: BirdLife International. Brazil, M. (2009) Birds of East Asia. London: Helm. Chereshnev, I.A. (2008) Red data book of the Chukchi autonomous district Vol. 1 . Animals. Magadan: Dikiy Sever. (In Russian.) Dorogoi, I.V. (1996) Breeding conditions report for Ryrkarpiy settlement, Chukotka, Russia, 1996. In M. Soloviev & P.Tomkovich, eds .ARCTIC BIRDS: an international breeding conditions survey. (Online database), http:// www.arcticbirds.ru/info96/ru60ru11596r.html.Updated 11 December 2008. Accessed 23 August 2013. Dorogoi, I.V. (1 997) The fauna and distribution of waders in North-East Asia. Pp.53-87 in A.V. Andreev, ed. Species diversity and population status of waterside birds in North-East of Asia. (Series: Biological Problems of the North.) Magadan: North-East Science Centre, Far East Department of the Russian Academy of Sciences. (In Russian.) Dorogoi, I.V. (1998) Breeding conditions for waders in Russian tundras in 1 997: Mys Shmidta area. Information Materials of the Working Group on Waders 1 1 : 40. (In Russian.) Dorogoi, I.V. & Beaman, M. (1998) New data on breeding of Semipalmated Sandpiper in Eurasia. Information Materials of the Working Group on Waders 1 1 : 48-49. (In Russian.) Koblik, E. A., Redkin, Ya. A. & Arkhipov, V. Yu. (2006) Checklist of the birds of Russian Federation. Moscow: KMK Scientific Press. (In Russian with English introduction.) Kishchinski A. A. (1988) Ornithofauna of North-East Asia. History and modern state. Moscow: Nauka. (In Russian.) Krechmar, A.V. & Kondrat'ev, A.V. (2006) Waterfowl of North-East Asia. Magadan: North-East Science Centre, Far East Department of the Russian Academy of Sciences. (In Russian.) Menyushina, I.E. (2000). Breeding conditions report for Mys Shmidta settlement, Chukotka, Russia, 2000. In M. Soloviev & P.Tomkovich, eds. ARCTIC BIRDS: an international breeding conditions survey. ( Online database). http://www.arcticbirds.net/info00%5Cru36ru1 1600.html. Updated 1 1 December 2008. Accessed 23 August 2013. Inskipp, T„ Lindsey, N. & Duckworth, W. (1996) Checklist of the birds of the Oriental region. Sandy UK: Oriental Bird Club. Portenko, L.A. (1972) Birds of Chukotski peninsula and Wrangel island. Part I. Leningrad: Nauka. (In Russian.) Portenko, L.A. (1973) Birds of Chukotski peninsula and Wrangel Island. Part II. Leningrad: Nauka. (In Russian.) Sibley, D. (2000) The North American bird guide. New York: Chanticleer Press. Rees, E. (2006) Bewick's Swan. London: T & A. D. Poyser. Stishov, M.S. (1991) New data on distribution of some bird species on Chukotka Arctic coast. Ornithological problems of Siberia conference. Available: https://www.google.ru/#fp=5085c1 fl e640d3bd &newwindow=1 &psj = 1 &q=%220rnithological + problems+of +Siberia%22 +Conference+abstracts (In Russian.) Stishov, M.S. (1992) Bird community of the lower Ekvyvatap River (Vancarem Lowland, Chukotka). Russian Journal of Ornithology 1 (2):245— 25 1 . (In Russian.) Stishov, M. S. (1997) Breeding conditions report for Mys Shmidta settlement, Chukotka, Russia, 1997. In M. Soloviev & P.Tomkovich, eds .ARCTIC BIRDS: an international breeding conditions survey. (Online database), http:// www.arcticbirds.ru/info97/ru30ru11697r.html. Updated 11 december 2008. Accessed 23 August 201 3. Stishov, M. S. (2004) Wrangel Island - truly natural but a natural anomaly. Yoshkar-Ola: Izdatelstvo Mariyskogo Poligrafkombinata. (In Russian.) Stishov, M. S. & Maryuhnich, P. V. (1991a) Spoon-billed Sandpiper in the western Vankarem Lowlands. Pp.1 25-1 26 in V. Yu. Il'yashenko and L. N. Mazin, compilers The study of rare animals in the R.S.F.S.R. Moscow: Central Science Research Laboratory for Game Management and Nature Reserves. (In Russian.) Stishov, M. S. & Maryuhnich, P. V. (1991 b) Buff-breasted Sandpiper in the Cape Yakan area and Ekvyvatap River valley (Arctic coast of Chukotka). Pp.1 26-1 29 in V. Yu. Il'yashenko & L. N. Mazin, compilers The study of Forktail 29 (2013) Birds of Mys Shmidta, north Chukotka, Russia 29 rare animals in the R.S.F.S.R. Moscow: Central Science Research Laboratory for Game Management and Nature Reserves. (In Russian.) Stishov, M. S., Pridatko, V. I. & Baranyuk, V. V. (1991 ) Birds of Wrangel. Island. Novosibirsk: Nauka. (In Russian.) Taldenkov, I. A. (2006) Record of mixed breeding of American Plover and Gray Plover at northern Chukchi peninsula. Information Materials of the Working Group on Waders 1 9: 39-41 . (In Russian.) Tomkovich, P. S. (1998) Breeding conditions for waders in Russian tundras in 1994. International Wader Studies 10: 132-144. Tomkovich, P. S. (2007a) Population dynamics of the Dotterel Eudromias morinellus: alarming thoughts on the background of poor knowledge. Information Materials of the Working Group on Waders 20: 43-45. (In Russian.) Tomkovich, P. S. (2007b) Annotated bird list for Pevek town vicinity, Chukotka autonomous area, the Far East of Russia. Ornithologia 34 (2): 1 76-1 85. (In Russian.) Tomkovich, P. S., Masterov, V. B. & Soloviev M. Y. (1991) Avifauna of Mys Shmidta, Chukchi Sea. Ornithologia 25: 1 75-176. (In Russian.) Vladimir Yu. ARKHIPOV, Institute of Theoretical and Experimental Biophysics RAS, Pushchino, Moscow 142290, Russia; State Nature Reserve Rdeysky, Kholm, Novgorod 1 75270, Russia. Email: v.arkhipov@rambler.ru Tom NOAH, Bergstr.14 D - 15910 Schlepzig, Germany. Email: tomnoah@t-online.de Steffen KOSCHKAR, LiebigstraBe 91, D-35392, D-GieBen, Germany. Email: fasciolata@web.de Fyodor A. KONDRASHOV, Bioinformatics and Genomics Programme, Centre for Genomic Regulation, Dr. Aiguader 88, Barcelona, 08003 Spain; Universitat Pompeu Fabra (UPF), 08003 Barcelona, Spain; Institucio Catalano de Recerca i Estudis Avangats (ICREA), 23 Pg. Uuis Companys, 08010 Barcelona, Spain. Email: Fyodor. Kondrashov@crg.eu Appendix Annotated checklist of birds recorded in Mys Shmidta area Indicates a species not recorded during surveys and added from the literature. Status: R=Resident, P=Passage, S=Summer visitor, B=Breeder, B? = Possible breeder, V=Vagrant, U=Unclear, (Lit) indicates status according to the literature. Species Status Previous records Species Status Previous records ‘Willow Ptarmigan Lagopuslagopus R (Lit) 7,16 ‘Spotted Redshank Tringaerythropus V (Lit) 2,3 Rock Ptarmigan Lagopus mutus R 7 Wood Sandpiper Tringaglareola V 16 Whooper Swan Cygruscygnus U First record ‘Terek Sandpiper Xenus cinereus PB? (Lit) 9 Tundra Swan Cygnus columbianus B 9 ‘Grey-tailed Tattler Heteroscelus brevipes V (Lit) 7 Bean Goose Anserfabalis P 7 Ruddy Turnstone Arenaria interpres PB 3,4,10,16 Greater White-fronted Goose Anser albifrons PB 7,10,16 Long-billed Dowitcher Limnodromus scolopaceus PB 10 Snow Goose Anser caerulescens P 7,16 Great Knot Calidris tenuirostris V First record Em peror Goose Anser canagica PB First record Red Knot Calidris canutus PB? First record Canada Goose Branta canadensis V First record Western Sandpiper Calidris mauri PB First record Brent Goose Branta bernida PB? 7 ‘Spoon-billed Sandpiper Eurynorhynchus pygmeus PB? (Lit) 9,13 Northern Pintail Anas acuta PB 7,10 Little Stint Calidris minuta PB? 9 Baikal Teal Anas formosa P 7 Red-necked Stint Calidris ruficollis PB 2,3,4,7,9,15,16 Common Teal Anas crecca P First record Temminck's Stint Calidris temminckii PB 4,15,16 Greater Scaup Aythya marila PB? 7 Semipalmated Sandpiper Calidris pusilla PB 1,5 Common Eider Somateria mollissima PB 7,10,16 ‘Baird's Sandpiper Calidris bairdii PB? (Lit) 10 King Eider Somateria spectabilis PB? 7,10 Pectoral Sandpiper Calidris melanotos PB 3,4,7,10 Spectacled Eider Somateria flscheri PB? 10 ‘Sharp-tailed Sandpiper Calidris acuminata V (Lit) 7 Steller's Eider Polysticta stelleri P 7 Dunlin Calidris alpina PB 3,4,10,16 Harlequin Duck Histrlonicus histrionicus P First record ‘Curlew Sandpiper Calidris ferruginea P (Lit) 3,16 Long-tailed Duck Clangula hyemalis PB 7,10,16 Buff-breasted Sandpiper Tryngitessubruficollis PB 9,10,14 Black Scoter Melanltta nigra americana P First record Ruff Philomachus pugnax PB 2,3,10 White-winged Scoter Melanitta fusca steinegeri P First record Red-necked Phalarope Phalaropus lobatus PB 3,4,10,15 Red-breasted Merganser Mergus senator P First record Red Phalarope Phalaropus fulicaria PB 3,7,10 Snowy Owl Nyctea scandiaca R 4,6,8,11 Pacific Golden Plover Pluvialis fulva PB 7,9,10,16 Short-eared Owl Asio flammeus PB? 8 American Golden Plover Pluvialis dominicus PB 9,12 Sandhill Crane Grus canadensis PB 7,10,16 Grey Plover Pluvialis squatarola PB 10,16 Common Snipe Gallinago gallinago PB 3,4,10 Common Ringed Plover Charadriushiaticula PB 4,7,10,16 Bar-tailed Godwit Limosa lapponica V First record ‘Semipalmated Plover Charadriussemipalmatus PB? (Lit) 12 30 VLADIMIR YU. ARKHIPOV et al. Forktail 29(2013) Species Status Previous records Species Status Previous records Eurasian Dotterel Eudromias morinellus PB? 2,3 Red-throated Loon Gavia stellata PB 7,10 Pomarine Skua Stercorarius pomarinus PB 4,8,10,15 Black-throated Loon Gavia arctica PB 7,10 Parasitic Skua Stercorarius parasiticus PB 2,4,8,10 Pacific Loon Gavia pacifica PB 7,10 Long-tailed Skua Stercorarius longicaudus PB 2,4,8,10 *Yellow-billed Loon Gavia adamsii U (Lit) 7 Mew Gu\\ Larus canus PS First record Common Raven Corvus corax PB 8 Glaucous Gull Larus hyperboreus PB 8,10,16 *Grey-cheeked Thrush Catharus minimus V (Lit) 8 Slaty-backed Gull Laws schistisagus V First record *Dusky Thrush Turdusnaumanni V (Lit) 16 Vega Gull Larus vegae PB 8,10,16 Bluethroat Luscinia svecica P 16 Thayer's Gull Larus thayeri V First record Northern Wheatear Oenanthe oenanthe PB 10,16 Black-headed Gull Larus ridibundus V 16 Barn Swallow Hirundorustica PB First record Ross's Gull Rhodostethia rosea PS 8 Northern House Martin Delichon urbica PB First record *Sabine's Gull Xema sabini P (Lit) 8 Arctic Warbler Phylloscopus borealis P First record Black-legged Kittiwake Rissa tridactyla PB? 8 Horned Lark Eremophila alpestris U First record Arctic Tern Sterna paradisaea PB 8 House Sparrow Passer domesticus V First record Common Murre Uria aalge or Thick-billed Murre U. lomvia PS First record White Wagtail Motacilla alba PB 8,10 Black Guillemot Cepphus grylle PB 8 Yellow Wagtail Motacilla flava tschutschensis PB 8,16 Osprey Pandion haliaetus V First record Red-throated Pipit Anthus cervinus PB 10,16 White-tailed Eagle Haliaeetus albicilla V First record Pine Siskin Carduelis pinus V First record Steller's Sea Eagle Haliaeetus pelagicus V First record Hoary Redpoll Carduelis hornemanni PB 8,10,16 Rough-legged Buzzard Buteo lagopus PB 2 Common Redpoll Carduelis flammea V First record Merlin Falco columbarius U First record Little Bunting Emberizapusilla V First record *Gyrfalcon Falco rusticolus U (Lit) 7 Pallas's Bunting Emberiza pallasi V First record Peregrine Falcon Falco peregrinus PB? First record Lapland Longspur Calcarius lapponicus PB 8,10,16 Pelagic Cormorant Phalacrocorax pelagicus PB 7 Snow Bunting Plectrophenax nivalis PB 8,10,16 Previous records: 1 : Andreev & Kondratyev (1996), 2: Dorogoi (1996), 3: Dorogoi (1997), 4: Dorogoi (1998), 5: Dorogoi & Beaman (1998), 6: Menyushina (2000), 7: Portenko (1972), 8: Portenko (1973), 9: Stishov (1991), 10: Stishov (1992), 1 1 : Stishov (1997), 12: Stishov (2004), 13: Stishov & Maryuhnich (1991a), 14: Stishov & Maryuhnich (1991b), 15:Tomkovich (1998), 16: Tomkovich etal. (1991) FORKTAIL 29 (2013): 31-36 Species limits within Rhopophilus pekinensis PAUL J. LEADER, GEOFF J. CAREY & PAUL I. HOLT Rhopophilus pekinensis is a passerine endemic to north-east Asia occurring primarily in China; two or three subspecies are variously recognised. A review of museum material and fieldwork on the breeding grounds indicates that only two taxa ( R . p. pekinensis and R. p. albosuperciliaris ) are valid, and using criteria that grade morphological and vocal differences between allopatric taxa (Tobias etai. 2010), both achieve the threshold for species status. The English names Beijing Babbler and Tarim Babbler are proposed reflecting both the type location of each and the recently elucidated taxonomic affinities of Rhopophilus. INTRODUCTION The White -browed Chinese Warbler Rhopophilus pekinensis is a passerine endemic to north-east Asia, occurring from north-west China to north-east China and North (and previously also South) Korea (Cheng 1987, Dickinson 2003, Duckworth & Moores 2008, Brazil 2009, Moores et al. 2009, BirdLife International 2013a). Whilst placed in the family Cisticolidae (Dickinson 2003), it was included in the Timaliidae, in a clade with Sylvia and Paradoxornis, by Alstrom et al. (2006), based on myoglobin and cytochrome b sequence data. In light of this, it was placed within the Timaliidae by Collar & Robson (2007) and Gill & Donsker (2012), using the English name Chinese Bush-dweller to reflect the fact that it was no longer considered a warbler. Subsequently Gelang et al. (2009) and Moyle et al. (2012) proposed treatment of the Sylviidae (which includes Rhopophilus) as a family rather than a subfamily within the Timaliidae. Most authorities (Cheng 1987, Dickinson 2003, Zheng 2011, Gill & Donsker 2012) recognise three subspecies: pekinensis (eastern part of the range, type locality Beijing, China), leptorhynchus (central part of the range, type locality Gansu, China) and albosuperciliaris (western part of the range, type locality Xinjiang, China). The validity of leptorhynchus was questioned by Vaurie (1955), who suggested treatment as a synonym of pekinensis, and this is followed, albeit tentatively, by Collar &c Robson (2007). Two further taxa, ‘ beicki ’ (type locality north-west Nei Mongol, China) and ‘ major (type locality Qinghai, China) are not currently recognised and both have long been treated as synonyms of albosuperciliaris (Vaurie 1955, 1959). In this paper the relationship between pekinensis and albosuperciliaris and the validity of leptorhynchus are reviewed based upon an examination of museum material and fieldwork conducted in China; the taxa ‘ beicki ’ and ‘ major are also discussed. METHODS Museum specimens were examined at the Natural History Museum, Tring, UK (NHMUK) and the Museum fur Naturkunde, Berlin, Germany (ZMB). The type specimens of pekinensis, albosuperciliaris (NHMUK), leptorhynchus and ‘ beicki ’ (ZMB) were examined, as was material from the type locality of ‘ major (NHMUK). In total 55 specimens were examined comprising 29 albosuperciliaris (including one ‘ beicki ’ and three ‘ major ), 15 pekinensis and 1 1 leptorhynchus. The following biometrics were taken: wing (maximum chord), tail length (to base of tail measured under the undertail-coverts) and bill length (to skull); measurements taken accord with standard procedures (Redfern & Clark 2001). All measurements were taken by PJL. No plumage differences between males and females exist, but plumage differences attributable to age and especially feather wear were noted (juvenile birds were characterised by very fresh plumage and loose contour feathering). During fieldwork on the breeding grounds, sound recordings were obtained from Beijing, Hebei, Qinghai and Xinjiang. Recordings were made using Telinga Pro 5 or Pro 7 parabolic microphones with either a Sound Devices 722 or an HHB Portadisc MDP 500, and a Sony PCM-M10 with a Sennheiser ME66. Spectrograms were produced and analysis of various parameters carried out using Raven Pro 1.4 (Cornell Laboratory of Ornithology 2003-11). Contrast was adjusted for each recording to ensure all elements (defined as any continuous line on a sonogram) were retained, while minimising reverberation. Measurements were made using a spectrogram window size of 512. In all 122 strophes were analysed, comprising 67 from nine pekinensis and 55 from eight albosuperciliaris. Analysis of parameters of each strophe was based on those proposed by Tobias et al. (2010), and comprised: • start and finish times (from which duration was calculated); • lowest and highest frequency (from which frequency range was calculated); • peak frequency (the frequency at which peak power occurs); • pace (calculated by dividing strophe length by number of elements). For each individual, we calculated the mean of each parameter; we then used the mean and standard deviation of all individuals of each taxon to calculate Cohen’s d values (see below). Due to their regular occurrence in flocks, the exact number of different individuals recorded was not always certain, although the figures provided are considered conservative estimates. In order to review species limits between taxa we applied the quantitative scoring system proposed by Tobias et al. (2010) to assess the degree of phenotypic difference between allopatric taxa. These criteria were summarised by Collar (2011a, b) thus: an exceptional difference (a radically different colouration or pattern) scores 4; a major character (a pronounced and striking difference in the colour or pattern of a body part, or in measurement or vocalisation) 3; a medium character (clear difference reflected, e.g. by a distinct hue rather than a different colour) 2; and a minor character (a weak difference, e.g. a change in shade) 1. Tobias et al. (2010) set a threshold score of 7 to allow for species status; species status cannot be triggered by minor characters alone, and only three plumage characters, two vocal characters (one spectral and one temporal), two independent biometric characters and one behavioural or ecological character may be counted. Vocal and biometric characters were assessed for effect size using Cohen’s d computed via the online calculator at http://www.uccs.edu/ ~faculty/lbecker/, where 0.2-2 is minor, 2-5 medium, 5- 10 major and >10 exceptional. PAUL J. LEADER PAUL J. LEADER 32 PAUL J. LEADER, GEOFF J. CAREY & PAUL I. HOLT Forktail 29 (201 3) RESULTS Morphological differences between taxa As noted elsewhere ( Vaurie 1959, Collar & Robson 2007), there are pronounced plumage differences between pekinensis and albosuperciliaris. In general, albosuperciliaris is much paler and more uniform than pekinensis ; the key differences between the two are detailed in Table 1 and illustrated in Plates 1-4. During fieldwork it became apparent that there is a highly distinct difference in iris colour, with albosuperciliaris having a dark brown iris and pekinensis a glaring pale yellow iris. Table 1. Plumage and bare part differences between adult Rhopophilus pekinensis pekinensis and R. p. albosuperciliaris. pekinensis albosuperciliaris Head pattern Blackish lores, pale grey supercilium, grey-buff ear-coverts, bold blackish submoustachial stripe. Greyish lores, off-white to buff supercilium, buff ear-coverts, blackish submoustachial and mid-brown post-ocular stripe. Upperparts Crown and upperparts grey-brown with broad darker brown streaks, streaks longer and broader on mantle. Nape and crown flecked rufous. Crown and upperparts sandy-grey with narrow mid-brown streaks, streaks slightly longer and bolder on mantle. Nape uniform sandy-grey. Underparts Chin, throat and belly white, sides of breast and flanks boldly streaked rufous, lower flanks and undertail-coverts rich buff and contrasting strongly with upperparts. Chin, throat and belly white or off-white, sides of breast diffusely streaked apricot-buff, lower flanks and undertail-coverts pale buff. Tail Central rectrices pale brown, outer rectrices dark brownish-grey and with pale greyish tips. Central rectrices sandy-grey, outer rectrices mid brownish-grey and with whitish tips. Iris Glaring pale yellow, clearly paler than pupil. Very dark brown, similar in colour to pupil. Plate 1 . Adult male R. p. pekinensis, Shanxi, China, April 2012. Plate 2. Adult male R. p. albosuperciliaris, Xinjiang, China, June 2012. Plate 3. Adult male R. p. pekinensis, Shanxi, China, April 2012 Plate 4. Adult male R. p. albosuperciliaris, Xinjiang, China, June 201 2 PAUL J. LEADER PAUL J. LEADER Forktail 29 (2013) Species limits within Rhopophilus pekinensis 33 Consistent structural differences also exist with albosuperciliaris being larger than pekinensis in terms of wing, tail and bill length (Table 2), such that when wing and bill lengths are plotted there is no overlap between the two (Figure 1). Table 2. Average wing length, tail length and bill to skull (all measurements in mm) and standard deviation (SD) of pekinensis (n = 26) and albosuperciliaris (n = 29). pekinensis albosuperciliaris Mean SD Mean SD Wing 61.1 2.01 67.7 2.42 Tail 89.3 4.84 95.3 4.22 Bill (skull) 14.6 0.80 16.0 0.58 Figure 1. Bill to skull (mm) and wing length (mm) of pekinensis and albosuperciliaris. Wing length (mm) The validity of leptorhynchus and comments on 'major' and 'beicki' As noted above the treatment of leptorhynchus is inconsistent. Vaurie (1955) recognised leptorhynchus but noted that it was poorly differentiated from pekinensis and concluded that 'it is a matter of opinion whether or not it should be recognised in the nomenclature’. Specimens of leptorhynchus examined as part of this study were on average slightly smaller than pekinensis (0.8 mm shorter-winged, 0.9 mm shorter-tailed and 1.0 mm shorter-billed). There was, however, extensive overlap in biometrics (Figure 2). In addition there were no consistent plumage differences between the two, and plumage of the type specimen fell within the range of pekinensis sensu stricto. As such, we concur with Collar & Robson (2007) and consider leptorhynchus a synonym of pekinensis. Vaurie (1955) concluded that ‘ major was comparable to albosuperciliaris and not larger and more densely streaked as noted by Meise (1937) and that birds from the type locality of 'major fell within the range of plumage variation and size of albosuperciliaris from Xinjiang. An examination of specimens from the Qaidam Basin, Qinghai (the type locality of ‘major), and of birds in the field there provides nothing with which to contradict Vaurie’s conclusion. Meise (1937) described ‘beicki' from a single specimen collected in north-west Nei Mongol, China (note: Vaurie [1995] correctly mapped the type locality of ‘ beicki ’, but incorrectly labelled the province as Ningxia), and considered it similar in colouration to ‘major but smaller in size. Vaurie (1955) regarded any differences insufficient to establish the validity of' beicki and questioned the wisdom of recognising it based on just a single specimen. Examination of the type specimen as part of this study established that in terms of plumage it falls within the range of variation of albosuperciliaris. Differences in biometrics are limited to wing length (62.0 mm), with values for tail (90.5 mm) and bill to skull (16.2 mm) falling within the range of albosuperciliaris. Examination of the type also suggested nothing unusual regarding the condition or preparation of the specimen which may have resulted in the smaller wing measurement and, whilst further material may prove otherwise, there appears no reason at this stage to recognise ‘beicki'. Vocalisations Both pekinensis and albosuperciliaris are garrulous and gregarious, and are most often found in small foraging flocks, the members of which frequently utter contact and other vocalisations. Both taxa have a wide repertoire of vocalisations, comprehensive comparative analysis of which would require a very large dataset of recordings. Both taxa appear to have more than one territorial song, although we collected insufficient samples to clarify the situation. However, in the samples taken for this study, a single common vocalisation that appears to have the same territorial and/or advertising function was identified, and both taxa were seen perched prominently uttering it; based on this, we regard it as song. As a result, it has been possible to carry out the analysis described above. The relevant vocalisation is a short series of 2-5 very similar notes transcribed as pyoo, each descending in pitch; typical examples for each of the taxa are illustrated in Figures 3 and 4. The mean, standard deviation and Cohen’s d. values of the various Figure 2. Tail length (mm) and wing length (mm) of pekinensis and leptorhynchus. Wing length (mm) 34 PAUL J. LEADER, GEOFF J. CAREY & PAUL I. HOLT Forktail 29(2013) Figure 3. Typical pyoo vocalisation of pekinensis, Miyun Reservoir, Beijing, 4 November 2009. (Paul I. Holt) 4- Figure 4. Typical pyoo vocalisation of albosuperciliaris, Aksu, Xinjiang, 10 August 2005. (Paul I. Holt) measurements are presented in Table 3. Sample sizes of other vocalisations were insufficient to allow comparison in the absence of a thorough understanding of their function. Table 3. Mean, standard deviation (SD) and Cohen's d values of parameters (see text) selected for analysis of pekinensis and albosuperciliaris. pekinensis albosuperciliaris Cohen's d Mean SD Mean SD Low freq (Hz) 2093 145 2730 139 4.48 High freq (Hz) 2914 79 3630 202 4.67 Freq range (Hz) 821 108 900 172 0.55 Peak freq (Hz) 2670 108 3309 151 4.86 Duration (s) 1.34 0.17 1.32 0.32 0.08 No. of elements 3.51 0.83 2.96 0.52 0.79 Pace (elements/s) 0.39 0.06 0.44 0.05 0.91 From the summary statistics in Table 3, it can be seen that for low frequency, high frequency and peak frequency the mean values are higher in albosuperciliaris than pekinensis, with little overlap between the two taxa; these differences are clearly audible in recordings. Habitat differences A bird of dense secondary shrubland, pekinensis ranges from sea level (where generally rare) to at least 1,200 m and is found in degraded hill slopes, forest edge and forest clearings (Plate 5). Species regularly recorded in the same habitat include Vinous- throated Parrotbill Paradoxornis webbianus , Godlewski’s Bunting Ember iza godlewskii and Meadow Bunting A. cioides. On the other Plate 5. Typical habitat of R. p. pekinensis, Shanxi, China, April 201 2. Plate 6. Typical habitat of R. p. albosuperciliaris, Xinjiang, China, June 2012. hand, albosuperciliaris is a desert species occurring in areas of mature tamarisk and dense desert shrubland (Plate 6) particularly in areas where Phragmites are mixed with Chinese Date Ziziphus jujuba or ‘Shazhao’ — a central Asian xerophyte. It prefers low-lying, arid, sandy and often, but not always, well-drained areas and occurs from 780 to about 1,500 m in Xinjiang but up to 2,800 m in the Qaidam Basin, Qinghai. Lop Nur, Bayingol, is the lowest known site for this taxon but with the drying up of the lake and associated habitat changes there in recent years, it is quite possible that it is no longer present. It occurs alongside Eurasian Tree Sparrow Passer montanus. Saxaul Sparrow P ammodendri , Desert Whitethroat Sylvia minula, with which it shares a very similar breeding distribution (Olsson et al. 2013), Isabelline Shrike Lanius isabellinus and even Biddulph’s Ground Jay Podoces biddidphi. DISCUSSION Characters selected for comparison based on Tobias et al. (2010) were assessed (Table 4). Among biometric characters, only wing length was assessed because of the lack of clearly independent such characters (see Tobias et al. 2010). In terms of vocalisations, peak frequency and pace were selected; behavioural or ecological differences were represented by innate habitat. Geographical relationship (Tobias et al. 2010) is not applicable as the two taxa are allopatric, although Vaurie (1955) maps locations of both indicating that the two occur within approximately 300 km of each other. Overall, a score of 13 easily surpasses the threshold score of 7 for species status set by Tobias et al. (2010). Of the features listed above, the differences in iris colour is considered major and therefore ranks highly. Iris colour varies with PAULJ. LEADER PAUL J. LEADER Forktail 29(2013) Species limits within Rhopophilus pekinensis 35 Table 4. Characters selected for comparison of pekinensis and albosuperciliaris based on Tobias et al. (2010), with score (see text) in brackets. Character Score Plumage and bare parts Underparts Medium 2 Upperparts Medium 2 Iris colour Major 3 Vocal Peak frequency (Cohen's d) 4.48 2 Pace (Cohen's d) 0.91 1 Biometric Wing length (Cohen's d) 3.0 2 Behavioural or ecological differences Innate habitat 1 Total score 13 age in many passerine species, typically being duller in juveniles, so whilst it is possible that young pekinensis may show dull irides similar in colour to those of albosuperciliaris , the difference between pekinensis and albosuperciliaris appears to be consistent when breeding season adults are compared. The dark iris of albosuperciliaris was noted in the historical literature (Richmond 1896) but has been overlooked in recent times (Collar & Robson 2007). A comparable situation in two taxa closely related to Rhopophilus exists in V inous-throated Parrotbill and Ashy-throated Parrotbill P. alphonsianus , which have a dark brown and whitish iris respectively (Robson 2007). Whilst usually treated as separate species (Penhallurick & Robson 2009, Gill & Donsker 2012), recent genetic studies (e.g. Crottini et al. 2010) found these two taxa to be very closely related and suggested that alphonsianus may be a clinal morph of P webbianus. In addition, in Silver-eared Mesia Leiothrix argentauris, the subspecies laurinae from Sumatra is unlike other subspecies in that it has pale irides, and the subspecies orientalis (from south V ietnam and east Cambodia) of Blue-winged Minla Minla cyanouroptera can also be distinguished from other subspecies by its pale irides. However, species limits within Silver-eared Mesia and the taxonomic status of orientalis require further evaluation (Collar & Robson 2007). Other examples in which iris colour varies between subspecies include Masked Booby Sula dactylatra (O’Brien & Davies 1990), whilst Kemp & Delport (2002) described a new subspecies of Red-billed Hornbill Tockus erythrorhynchus largely on the basis of iris colour and their proposal that the Red-billed Hornbill complex is better treated as five separate species (based on consistent differences in the colour of signal areas between geographically discrete populations) has been adopted elsewhere (Gill & Donsker 2012). In this study, it is noteworthy that even without the score for iris colour a score of 10 would still readily achieve the threshold for species status. Based upon these results the following taxonomic treatment of two monotypic species is proposed: Beijing Babbler Rhopophilus pekinensis (Swinhoe, 1868) Tarim Babbler Rhopophilus albosuperciliaris (Hume, 1873) The English names reflect the geographical origin of the type specimens and the use of ‘Babbler’ reflects recent taxonomic studies which place Rhopophilus within the Timaliidae. ‘Bush-dweller’ (Collar & Robson 2007) is not adopted as we feel that ‘Babbler’ is more accurate and that ‘Bush-dweller’ gives little or no insight into the taxonomic relationships of the two species. We acknowledge that some authorities treat the Sylviidae as a separate family rather Figure 5. Map showing the approximate ranges of the two species Rhopophilus albosuperciliaris and R. pekinensis including the type localities. Irkutsk JTANA ULAAN BAATAR Harbin Changchun Urumqi BISHKEK KYRGYZSTAN ,Baotou Hamhung 'ONGYANG Dalian SEOUL Taejon .Qingdao Chonju Gwangju .Xuzhou ISLAMABAD Legend Huainan. Hefei nwala Xinyang # Nanjing Wuxi* .Sh Lahore Type localities ( T) pekinensis @ albosuperciliaris @ leptorhyrchus (4) "bieki" (5) "major" .Ludhiana Chengdu Wuhan Hangzf .Chongqin; Nanchang .Lichuan NEW DELHI KATHMANDU Jaipur Lucknow Guiyang Fuzhou ’Kanpur Allahabad • Patna Kunming Varanasi Rajshahi .Liuzhou - J DHAKA Geographic Range V//A albosuperciliaris V77A pekinensis Bhopal Jabalpur Guangzhou Hong Kong .Ahmadabad Khulna Kolkata • i Nanning Kaohsi Chittagong Indore Mandalay adodara •Surat .Nagpur HANOI- HainhnnP 36 PAUL J. LEADER, GEOFF J. CAREY & PAUL I. HOLT Forktail 29 (2013) than a subfamily within the Timaliidae (Gelang et al. 2009, Moyle etal. 2012), but refer to the use of the English name Sylviid Babblers for the Sylviidae (Gill & Donsker 2012) and note that the family includes a number of species which have ‘Babbler’ in their English name (e.g. African Hill Babbler Pseudoalcippe abyssinica). The BeijingBabbler occurs from North Korea, north to southern Jilin and then west across north China to Gansu and eastern Qinghai. According to BirdLife International (2013a), the range continues south through northern Sichuan, western Henan and north-eastern Hubei as far as south-western Anhui; however, we are unaware of any records from Sichuan, Hubei or Anhui and these provinces are omitted by Zheng (201 1 ), although it has been recorded from Henan since the 1930s (Fu 1937). The Tarim Babbler occurs in southern Xinjiang from the western part of the Tarim Basin (restricted to the rivers and oases around the margins of the Tarim Basin and avoiding the Taklamakan Desert proper) east to the Qaidam Basin, Qinghai. The ranges of the two species are shown in Figure 5. Beijing Babbler is a fairly common and widespread species found in shrubland and although its range has contracted and it is no longer recorded in South Korea (Moores et al. 2009) and has declined in North Korea (Duckworth 2006), it is probably not globally threatened. Tarim Babbler, whilst sometimes locally common, is probably facing similar threats to Biddulph’s Ground Jay and may be declining due to fragmentation and degradation of desert habitats caused by intensive grazing of livestock, extraction of fuelwood and conversion of suitable habitat to irrigated land (BirdLife International 2013b) and may qualify as Near Threatened. ACKNOWLEDGEMENTS We thank Lei Wei Dong, Gou Jun, Ma Ming and David Stanton for their help in the field in Xinjiang and Wang Qingyu for her help in Xinjiang and Beijing. Kadoorie Farm and Botanic Garden kindly funded the visit by PJL to the ZMB. Richard Lewthwaite assisted with distributional information and both Michael Leven and Per Alstrom provided useful comment on the English names of the two species. Hincent Ng prepared Figure 5 and Urban Olsson assisted with references. Mark Adams (NHMUK) and Sylke Frahnert (ZMB) kindly permitted access to specimens in their collections. Will Duckworth and Nial Moores commented on the status of pekinensis in Korea. Finally, Michael Leven and Nigel Collar provided very helpful comments on a draft of this paper. REFERENCES Alstrom, P„ Ericson, P. G. P„ Olsson, U. & Sundberg, P. (2006) Phylogeny and classification of the avian superfamily Sylvioidea. Mol. Phylogenet. Evol. 38: 381-397. BirdLife International (2013a) Species factsheet: Rhopophilus pekinensis. Downloaded from http://www.birdlife.org on 22/01/2013. BirdLife International (2013b) Species factsheet: Podoces biddulphi. Downloaded from http://www.birdlife.org on 22/01/2013. Brazil, M. A. (2009) Birds of East Asia. London: Christopher Helm. Cheng, T.-H. (1987) A synopsis of the avifauna of China. Beijing: Science Press. Collar, N. J. (2011a) Species limits in some Philippine birds including the Greater Flameback Chrysocolaptes lucidus. Forktail 27: 29-38. Collar, N. J. (201 1 b) Taxonomic notes on some Asian babblers (Timaliidae). Forktail 27: 100-102. Collar, N. J. & Robson, C. (2007) Family Timaliidae (babblers). Pp.70-291 in J. del Hoyo, A. Elliott & D. A. Christie, eds. Handbook of the birds of the world, 12. Barcelona: Lynx Edicions. Crottini, A., Galimberti, A., Boto, A., Serra, L., Liu, Y., Yeung, C., Yang, X. J., Barbuto, M. & Casiraghi, M. (2010) Towards a resolution of a taxonomic enigma: first genetic analyses of Paradoxornis webbianus and Paradoxornis alphonsianus (Aves: Paradoxornithidae) from China and Italy. Mol. Phylogenet. Evol. 57: 1312-1318. Dickinson, E. C. ed. (2003) The Howard and Moore complete checklist of the birds of the world (third edition). London: Christopher Helm. Duckworth, J. W. (2006) Records of some bird species hitherto rarely found in DPR Korea. Bull. Brit. Orn. Club 1 26: 252-290. Duckworth, J. W. & Moores, N. (2008). A re-evaluation of the pre-1 948 Korean breeding avifauna: correcting a 'founder-effect' in perceptions. Forktail 24: 25-47 Fu, T.-S. (1937) L'etude des oiseaux du Ho-Nan [Birds of Henan], Langres: Imprimerie Moderne. Gelang, M„ Cibois, A., Pasquet, E„ Olsson, U., Alstrom P. & Ericson, P. G. P. (2009) Phylogeny of babblers (Aves, Passeriformes): major lineages, family limits and classification. Zool. Scr. 35: 225-236. Gill, F. & Donsker, D„ eds. (201 2) IOC World Bird Names (v 3.1 ). Available at http://www.worldbirdnames.org [Accessed on 29/01/2013]. Kemp, A. C. & Delport, W. (2002) Comments on the status of subspecies in the red-billed hornbill ( Tockus erythrorhynchus ) complex (Aves: Bucerotidae), with the description of a new taxon endemic to Tanzania. Ann. Transvaal Museum 39: 2-8. Meise, W. (1937) in Stresemann, E., Meise, W. & Schonwetter, M. (1937) Aves Beickianae.L Ornithol. 85: 375-576. Moores, N., Park, J.-G. & Kim, A. (2009) The Birds Korea checklist: 2009. Available at http://www.birdskorea.org/Birds/Checklist/BK-CL-Checklist-Aug- 2009.shtml [Accessed 28/04/2013], Moyle, R. G„ Andersen, M. J., Oliveros, C. H., Steinheimer, F. D. & Reddy, S. (2012) Phylogeny and biogeography of the core babblers (Aves: Timaliidae). Sysf. Biol. 61: 631-651. O'Brien, R. M. & Davies, J. (1990) A new subspecies of Masked Booby Sula dactylatra from Lord Howe, Norfolk and Kermadec Islands. Mar. Ornithol. 18: 2-7. Olsson, U., Leader, P„ Carey, G„ Khan, A., Svensson, L. & Alstrom, P. (2013) New insights into the intricate taxonomy and phylogeny of the Sylvia curruca complex. Mol. Phylogenet. Evol. 67: 72-85. Penhallurick, J. & Robson, C. (2009)The generic taxonomy of parrotbills (Aves, Timaliidae). Forktail 25: 1 37-141 . Redfern, C. P. F. & Clark, J. A. (2001 ) Ringer's manual. Thetford: British Trust for Ornithology. Richmond, C. W. (1896) Catalogue of a collection of birds made by Doctor W. L. Abbott in eastern Turkestan, the Thian-Shan Mountains, and Tagdumbash Pamir, central Asia, with notes on some of the species. Proc. U.S. Natn. Museum 1 8: 569-591 . Robson, C. (2007) Family Paradoxornithidae (parrotbills). Pp.292-320 in J. del Hoyo, A. Elliott & D. A. Christie, eds. (2007) Handbook of the birds of the world, 12. Barcelona: Lynx Edicions. Tobias, J. A., Seddon, N., Spottiswoode, C. N„ Pilgrim, J. D., Fishpool, L. D. C. & Collar, N. J. (2010) Criteria for species delimitation based on phenotype. Ibis 152: 724-746. Vaurie, C. (1955) Systematic notes on Palearctic birds, No. 1 8. Supplementary notes on Corvidae, Timaliinae, Alaudidae, Sylviinae, Hirundinidae and Turdinae. Amer. Museum Novit. 1 753. Vaurie, C. (1959) The birds of the Palearctic fauna. Passeriformes. London: H. F. & G. Witherby. Zheng, G., ed. (201 1 ) A checklist on the classification and distribution of the birds of China. Beijing: Science Press. (In Chinese.) PaulJ. LEADER, AEC Ltd., 127 Commercial Centre, Palm Springs, New Territories, Hong Kong. Email: pjl@aechk.hk Geoff J. CAREY, AEC Ltd., 127 Commercial Centre, Palm Springs, New Territories, Hong Kong. Email: gjc@aechk.hk Paul I. HOLT, Bracken Dean, Pendleton, Clitheroe, Lancashire, UK. Email: piholt@hotmail.com FORKTAIL 29 (201 3): 37-42 Nesting of the Large-billed Reed Warbler Acrocephalus orinus: a preliminary report PAVEL KVARTALNOV, ABDULNAZAR ABDULNAZAROV, VERONIKA SAMOTSKAYA, JULIA POZNYAKOVA, IRINA ILYINA, ANNA BANNIKOVA & EUGENIA SOLOVYEVA Large-billed Reed Warbler Acrocephalus orinus has a limited breeding distribution. It is known to inhabit valleys of the Panj river and its tributaries in Gorny Barakhshan Autonomous Republic (Tajikistan) and Badakhshan province (Afghanistan). Here we give descriptions of nests and eggs of this species based on 18 fresh nests found in Panj and Ghund valleys (Tajikistan) in 201 1. Unlike the closely related species A. dumetorum and A. scirpaceus, Large-billed Reed Warbler has nests built with a layer of wool and seed tufts. Nests are placed on twigs of sea-buckthorn, willow and other bushes, herbs and reed stems over dry soil. Large-billed Reed Warbler clutch size is relatively small (on average, 3.77 ± 0.83 eggs (n = 1 3)). The ground colour of eggs is usually white, not bluish, greenish or rosy as in the related species. INTRODUCTION Until recently Large-billed Reed Warbler Acrocephalus orinus remained one of the least studied bird species of the Palaearctic fauna. A. O. Hume discovered the first specimen on 1 1 November 1867 in the Sutlej valley, Himachal Pradesh, India (Hume 1869, 1870). He described the bird as Phyllopneuste macrorhynchus, and later referred to it as A. macrorynchus. Oberholser (1905) changed the name to A. orinus. The taxonomic status of this form remained uncertain until the beginning of the twenty-first century when Bensch & Pearson (2002) studied the type specimen in detail, including sequencing of mitochondrial and nuclear DNA. This study confirmed the specific status of Large-billed Reed Warbler, although some doubts remained (McCarthy 2006) until the moment when P. D. Round caught a live bird near Bangkok, Thailand (Round et al. 2007). Further studies of museum collections and searches for living individuals helped to elucidate possible breeding, moulting and wintering areas (Svensson et al. 2008, 2010, Timmins et al. 2009, Koblik et al. 2010, 2011). In 2009, a bird feeding fledglings was caught in south-east Tajikistan, not far from the border with Afghanistan, in the Shakhdara river valley (Aye et al. 2010). Museum specimens in the Zoological Museum of Moscow University (Moscow, Russia) and the Institute of Zoology and Parasitology (Dushanbe, Tajikistan) reveal that previous records of Blyth’s Reed Warbler Acrocephalus dumetorum breeding in the Vanj and Ghund river valleys in fact refer to Large-billed Reed Warbler (Kvartalnov et al. 201 la, b, Kvartalnov & Garibmamadov 2012). Although fledglings were recorded in 1961 by A. V. Popov in the Vanj valley, near the village ofGhijovast (Abdusalyamov 1973, Kvartalnov et al. 2011b) and in 2009 by R. Aye and colleagues (Aye etal. 2010), and birds collecting nest material were observed in 1976 by V. V. Kashinin in the lower Ghund valley, near the village of Barsem (unpublished manuscript - see Kvartalnov et al. 2011a), no nests of this species have ever been described. The mystery of the Large-billed Reed Warbler could have been solved in the mid-twentieth century. When in 1937 A. B. Kistyakovsky took part in a Pamir expedition, he found that reed warblers in the south-western Tajik Pamir mountains (in the environs of the town of Khorog) were not typical Blyth’s Reed Warblers. He therefore prepared a description of a new A. dumetorum subspecies, but his manuscript and the intended type series were destroyed in a fire together with all zoological collections in Kiev University during the German occupation in the Second World War (Nowak 200 1 ). Kistyakovsky ( 1 950) wrote his opinion of the systematic position of this form, which he thought to be endemic to Gorny Badakhshan, Tajikistan. Other naturalists who had visited the Pamir mountains in the twentieth century did not distinguish those birds from typical Blyth’s Reed Warblers. In 2010-2011 we studied spring migration of Blyth’s Reed Warbler and breeding biology of Large-billed Reed Warbler in Tajikistan. Blyth’s Reed Warbler was found to be common during spring passage in the vicinity of Dushanbe and in the south-west part of the republic, but we failed to prove its breeding in Taj ikistan, although this was suspected by Abdusalyamov (1973), Portenko & Stubs (1976) and other ornithologists. There is no doubt that Blyth’s Reed Warbler is a transient in all regions to the south of Almaty in south-east Kazakhstan, and that all nests found there that had been attributed to that species belong to others, including Paddyfield Warbler Acrocephalus agricola and Sykes’s Warbler Hippolais rama (Ivanitskii et al. 2012). Data about phenology, breeding biology, social behaviour, acoustics and morphometry of the Large-billed Reed Warbler collected in 2011 are presented in this article with additional data from 2012. This information is to help other ornithologists to search for and distinguish nests of Large-billed Reed Warbler. MATERIALS AND METHODS The main field observations were conducted in the Panj valley near the village of Zumudg, Ishkashim region, Gorny Badakhshan Autonomous Region, Tajikistan (36.9 17°N 72.183°E) between 10 June and 11 July 2011. Additional data were collected in the Apharv forest area in the Panj valley (36. 800°N 71. 550°E) and near the village of Langar in the lower Pamir river valley (37.033°N 72.667°E). AA inspected riverside forests in the Ghund valley near the villages of Charthem (37.717°N 72.167°E), Vuzh (37.717°N 71.933°E) and Dehmiyona (37.700°N 71.917°E) (Figure 1). From 23 May to 23 July 2012 PK and colleagues studied breeding biology and social behaviour of Large-billed Reed Warbler near the village of Dehmiyona; the resulting data are not included here. Nine adult Large-billed Reed Warblers were caught in mist- nets and traps at nests near Zumudg (Plate 1). The birds were identified by measurements of bills, wings, tails and legs according to Svensson et al. (2010), Koblik et al. (2010) and from our experience of working with series of Large-billed and Blyth’s Reed Warblers in collections of the Zoological Museum of Moscow University (Ivanitskii et al. 2012). Adults were marked with metal and colour rings and by grease paint colouring on breast and head for individual identification (a harmless method that we used previously with other warbler species). Blood samples were taken 38 PAVEL KVARTALNOV etol. Forktail 29 (2013) Figure 1. A map of localities of some historical and recent observations of Large-billed Reed Warbler A. orinus in Tajikistan. A. orinus JQ651 381.1 Zumudg A. orinus GU247955 Burma A. orinus GU247957 India A. orinus HM352785 Khorog. Tajikistan A. orinus GU247958 India A. orinus GU247952 Pakistan A. orinus GU247953 Afghanistan A. orinus GU247950 Afghanistan A. orinus GU247954 Kazakhstan (?) 43 71 70 90 84 99 98 98 97 98 A. A. orinus HM352786 Khorog, Tajikistan ~~ A. orinus GU247949 Afghanistan A. orinus DQ681065 Thailand A. orinus GU247956 Burma A. orinus JQ651 384.1 Zumudg A. orinus GU247951 Afghanistan A. orinus JQ651380.1 Zumudg A. orinus JQ651 382.1 Zumudg A. orinus JQ651 383.1 Zumudg “A. orinus HM352789 Shakhdara r., T ajik istan scirpaceus 97 97 97 99 98 I A. concinens ^ A. dumetorum \A palustris 97 97 j ^ A. agricola ' S. borin AY329474 0.01 Figure 2. The relationships between Large-billed Reed Warbler and closely related species of the genus Acrocephalus based on NJ and MP analyses. Genbank accession numbers for Eurasian Reed Warbler A. scirpaceus, Marsh Warbler A. palustris etc. are given in the text. from all caught adult birds and eight nestlings (also marked with metal rings). Specific identification was supported by analysis of mt DNA (Figure 2). Total DNA was extracted from dried blood samples using the standard protocol of proteinase K digestion, phenol-chloroform deproteinisation and isopropanol precipitation (Sambrook et al. 1989). Mitochondrial DNA sequences were obtained from five Large-billed Reed Warblers caught in 201 1 near Zumudg. The partial cytochrome b gene (207 bp) was amplified in one polymerase chain reaction (PCR) with the forward/reverse primer combination L14841/H 1 5 149 (Kocher et al. 1989). Typical conditions for cyt b amplification included initial denaturation at PAVEL KVARTALNOV Forktail 29 (2013) Nesting of the Large-billed Reed Warbler Acrocephalus orinus: a preliminary report 39 Plate 1. Panj river valley near Zumudg village, 25 June 201 1 . 94°C for 3 min, 35 cycles of 94°C for 30 s, annealing at 51°C for 1 min, and extension at 72°C for 1 min, followed by a final extension at 72°C for 10 min and an indefinite hold at 4°C. PCR products were visualised on 1 % agarose gel and then purified using DEAE (Whatman) or NHJitOH. Approximately 10-50 ng of the purified PCR product were used for sequencing with each primer by the autosequencing system ABI 3 1 00-Avant in conjunction with ABI PRISM’BigDyeTM Terminator, version 3.1. Cytb sequences were aligned by eye using BioEdit 7.0. The final alignment of the mitochondrial region included 207 bp, of which 54 sites were variable and 40 sites were parsimony-informative. For the analysis we also used GenBank data (A. dumetorum,A. orinus , A. agricola , Eurasian Reed Warbler U. scirpaceus. Marsh Warbler A. palustris and Blunt-winged Warbler H. concinens). Phylogenetic neighbour-joining (NJ) and maximum parsimony (MP) analyses were performed using MEGA 4.0.0.4083. The NJ tree was reconstructed using the uncorrected p-distance. Unweighted MP analysis was performed using heuristic search starting with stepwise addition trees (random addition sequence, 10,000 replicates). To assess clade stability in the MP and NJ trees, 1,000 bootstrap pseudoreplicates were analysed. GenBank accession numbers of obtained sequences are JQ65 1 380-JQ65 1384. Genbank accession numbers of other sequences used in this work comprise: A. agricola AJ004245- AJ004248, AJ0043.30, AJ004331, AJ004775, AJ004776, FJ883021, Y15694; A. concinens AJ004260-AJ004262, FJ883027; A. dumetorum : AJ004263, AJ004264, AJ004336-AJ004340, AJ004773, FJ883028; A. orinus -. DQ681065, GU247949- GU247958, HM352785, HM352786, HM352789; A. palustris- AJ004293, AJ004294, AJ004344, AJ004345, AJ004774, EU86 1 03 1 , FJ883036; A. scirpaceus- AJ004301-AJ004304, AJ004771, AJ004772, AM889139, FJ883039, NC 010227, Z73483. The identification of uncaught birds was based on characteristic songs (Timmins et al. 2010, Ivanitskii et al. 2012). Recordings of songs of five marked males proved that the song described by Timmins et al. (2010) belongs to the Large-billed Reed Warbler (Ivanitskii et al. 2012). Nests that AA found in the Ghund valley were identified by comparison with known Large-billed Reed Warbler nests from the Panj valley. For comparison we also used unpublished data from 51 Blyth’s Reed Warbler nests found and described by PK in 2007-2009 in the Kostroma region, Russia. RESULTS AND DISCUSSION The Large-billed Reed Warbler is a common species in suitable habitat in the Panj, Ghund and lower Pamir valleys. We found nests near the villages of Zumudg, Charthem, Vuzh and Dehmiyona, observed actively singing males in the Apharv forest area, and saw singing males and territorial pairs at the village of Langar (Figure 1). The birds inhabited thickets of sea-buckthorn Hippopbae ramnoides and willow Salix turanica, S. shugnanica and S. wilhemsiana intertwined with clematis Clematis hilariae , with sparse ground cover of liquorice Glycyrrhiza uralensis , reed Pbragmites australis and other species. Other bird species observed in the same habitat of Panj valley include Hume’s Lesser Whitethroat Sylvia althaea. Mountain Chiffchaff Pbylloscopus sindianus. Common Rosefinch Carpodacus erythrinus, Cetti’s Bush Warbler Cettia cetti , Common Nightingale Luscinia megarhyncbos , Bluethroat Luscinia svecica. Black-billed Magpie Pica pica, Isabelline Shrike Lanius isabellinus phoenicuroides and Citrine Wagtail Motacilla citreola calcar ata. The only other Acrocephalus warbler recorded around Zumudg during our observations was Clamorous Reed Warbler A. stentoreus (a single transient bird). According to observations of PK in 20 1 2, in the Ghund valley Large-billed Reed Warblers also breed in S. turanica, wild rose Rosa beggerana and honeysuckle Lonicera stenantha thickets with Astragalus longistipitatus, Potamogeton cariatum and other herbs along canals among crop fields on alluvial fans. Most Large-billed Reed Warblers were found near river banks, canals or other wet localities. The birds breed in monogamous pairs, although attempted extra-pair copulations by at least three paired and two unmated territorial males were observed. We described 1 5 nests and one abandoned construction built in June andjuly 2011 in the Panj valley (Plates 2, 3), plus seven remains of nests built in 2009-2010 in the same area, and three recent nests in the Ghund valley in 20 1 1 . Thirteen nests had complete clutches. Nests are built by females (based on observations of building of nine nests in 2011, including two nests observed from the first day of construction). Most were in sea-buckthorn thickets, but one was in a willow bush. Nearly all nests were placed over dry soil, except one that was built on a branch over a canal temporarily filled by water. Large-billed Reed Warblers attached nests to sea-buckthorn twigs at forks (six nests), stems of Artemisia (three nests), liquorice PAVEL KVARTALNOV PAVEL KVARTALNOV 40 PAVEL KVARTALNOV etal. Forktail 29 (2013) stems (three nests), thin willow stems (two nests), willow twigs at forks (two nests), twigs of clematis at forks (two nests), reed stems (one nest), reed stems and willow twigs (one nest), reed and liquorice stems (one nest), liquorice stems and willow twigs (one nest), sea-buckthorn and willow twigs (one nest), reed, liquorice stems and a sea-buckthorn twig (one nest). Females began nest construction with a platform of dry plant debris, but from the first day they braided vertical supports (stems, etc.) with plant fibres. Nests were fastened to stems more firmly than the nests of Blyth’s Reed Warbler. The principal material consists of bast and bark fibres of clematis, willow, liquorice and Artemisia , fibres of reed sheath, dry leaves, stems and ears of grasses, goat wool and clematis seed tufts. Bast strips also form an outer covering that disguises the nest in thickets. An inner part of the structural layer is made with wool and seed tufts, and rarely includes bird feathers. Nests of Blyth’s Reed Warbler usually lack such soft materials (Plate 5). The upper edges of Large-billed Reed Warbler nest cups are made usually with ears of grasses. Nests are lined with clematis bast fibres or (rarely) with thin dry grass stems, with the addition of mammal hairs. Nests found near Zumudg (n = 15) had the following measurements (average and standard deviation): outer diameter 81.7 ± 8.0 mm; height 68.6 ± 9.4 mm; inner diameter 54.8 ± 2.0 mm; depth 45.1 ± 3.0 mm. The height of nests above ground or water was 30-168 cm (average 82.1 cm). Nests inspected in the Ghund valley in 201 1 were placed at 110-210 cm above ground. Most full clutches in the Panj valley were of four eggs (n = 7), rarely two (n = 1), three (n = 3) or five eggs (n = 2); on average, 3.77 ± 0.83 (n = 13), including a replacement clutch that contained three eggs (the first clutch consisted of four eggs). We also found a nest with three nestlings. Nests found in Ghund valley had three (two nests) and four (one nest) eggs. The usual clutch size of the Blyth’s Reed Warbler in Kostroma region was 5-6 eggs, rarely four eggs, on average (n = 37) 5.53 ± 0.56 eggs. Plate 2. Nest 5-1 1 of the Large-billed Reed Warbler A. orinus with full clutch (two unhatched eggs from this nest are now held in the Natural History Museum, Tring, UK), 21 June 201 1. Plate 3. Nest 11-11 of the Large-billed Reed Warbler with full clutch (now held in the Zoological Museum of Moscow University), 9 July 2011. Plate 4. A clutch from nest 13-11 of the Large-billed Reed Warbler showing markings on eggs, 1 1 July 201 1 . Plate 5. A nest of Blyth's Reed Warbler A. dumetorum with full clutch, Kostroma region, Russia, 14 June 2007. PAVEL KVARTALNOV PAVEL KVARTALNOV ABDULNAZAR ABDULNAZAROV PAVEL KVARTALNOV Forktail 29 (2013) Nesting of the Large-billed Reed Warbler Acrocephalus orinus: a preliminary report 41 Clutches of Blyth’s Reed Warbler from South Siberia had on average 4. 8-5. 2 eggs (Totunov 1981) or 5.1-5.76 eggs (Kuranov 2008) in different years. Eggs of Large-billed Reed Warbler have a dirty-white (rarely creamy white or pure white) shell covered with small olive-brown superficial spots that usually (but not always) form a sparse cap or a ring at the larger end, and with more sparse dark-brown superficial specks which rarely (as opposed to eggs of Blyth’s Reed Warbler) lay over larger spots (Plates 4 & 5). Sparse and small deep bluish spots also form a cap or a ring at the larger end. Some eggs are covered not only with sparse spots but also with dense small olive brown specks that can hide the basic shell colour. Fresh eggs are dull or with a weak gloss. Eggs of Blyth’s Reed Warbler are greenish or rosy, usually with no white background (Chernyshov 1998; our data). Measurements of eggs (in mm; n = 58): 17.72 ± 0.55 (16.8- 19.1) x 13.14 ± 0.31 (12.3-13.8). Egg weight (n = 48): 1.51 ± 0.15 (1.13-1.76) g. One Large-billed Reed Warbler nest with a Plate 6. A newly-hatched nestling of Large-billed Reed Warbler from nest 5-1 1,3 July 2011. I Plate 7. Female Large-billed Reed Warbler feeding nestlings, nest 1- 11, 10 July 2011. complete clutch (three eggs) is now deposited in the Zoological Museum ofMoscow University (ZMMU Q-8036); two unhatched eggs were sent to the Natural Elistory Museum, Tring, UK (NHMUK E/2012.5.1). Eurasian Reed Warbler has been found in Gorny Badakhshan and nearby Afghanistan, although it is not known to breed there (Abdusalyamov 1973, Timmins et al. 2010). It has nests without the layer built with wool and seed tufts (Kvartalnov et al. 2006), and eggs with bluish or olive-greenish basic shell colour, not white (Nikiforov et al. 1989). Only female Large-billed Reed Warblers have brood-patches, but both partners incubate eggs and provide food to nestlings and fledglings (Plates 6, 7, 8). Large-billed Reed Warblers in the Panj valley have a relatively wide range in dates of arrival on the breeding grounds (compared with Mountain Chiffchaff and other passerines inhabiting the Pamir Mountains: our observations). The nest found on 5 July with young birds ready to fledge must have been built during the first days of June. Most other nests near Zumudgwere built after 10 June. Several new birds reached the breeding grounds at the end of June and the beginning of July. The latest of the first clutches found near Zumudg was finished on 7 July. On the last day of investigations (11 July) we observed a female in the territory of a male which was singing from 9 July. Of two nests built for replacement clutches, one was still empty on 1 1 July. In the lower Pamir river valley on 2 July we found several actively singing bachelor males, newly formed pairs and apparently non-territorial birds. AA found nests with clutches on 23 June, 16 July (nestlings hatched 17 July) and 17 July in the Ghund valley. V. V. Kashinin in the lower Ghund valley (37.550°N 71.733°E) observed the peak of nest building to be in mid-June (Kvartalnov et al. 201 lb). A. V. Popov saw fledglings on 22 June 1961 in the Vanj valley (38.550°N 71.733°E) (Abdusalyamov 1973, Kvartalnov et al. 2011b); fledglings possibly of this species were recorded by A. V. Popov in the Shakhdara valley on 26 July (birds were not collected, and the year is unknown) (Abdusalyamov 1973). Aye et al. (2010) saw fledglings near the Shakhdara River on 19 July 2009. According to museum collections (Koblik et al. 2010, 2011) and our observations in 2012, Large-billed Reed Warblers reach breeding grounds in Gorny Badakhshan in the final third of May. AA saw the last birds that he thought to be this species in the first ten days of September in the Panj valley, but this needs to be confirmed by mist-netted birds. Plate 8. A young Large-billed Reed Warbler ready to fledge, nest 1 2- 11,5 July 2011. PAVEL KVARTALNOV 42 PAVEL KVARTALNOV etal. Forktail 29 (2013) ACKNOWLEDGEMENTS The authors are grateful to the Director of the Zoology and Parasitology Institute (Dushanbe) Abdusattor Saidov for help in organising our expeditions, to Shamsher Mirzobekov (Zumudg) and his family for their hospitality, to Tamara Krutenko (Lomonosov MSU) for plant identification and to Nigel Collar, Brian Sykes and two anonymous referees for their careful work with the text of our paper. This work was supported by the Rufford Small Grants Foundation and by Russian Fund for Basic Research (##10-04- 00483, 13-04-01771). The publication is dedicated to the memory of Alexander Kistyakovsky (1904-1983). REFERENCES Abdusalyamov, I. A. (1973) Fauna of Tadjik SSR. Vol.XIX. Birds. Part 2. Dushanbe: Donish. Aye, R., Hertwig, S.T. & Schweizer, M. (2010) Discovery of a breeding area of the enigmatic Large-billed Reed Warbler Acrocephalus orinus. J. Avian Biol. 41: 452-459. Bensch, S. & Pearson, D. (2002) The Large-billed Reed Warbler revisited. Ibis 144:259-267. Chernyshov, V. M. (1998) [Character of inheritance of egg shell colour in Blyth's Reed Warbler (Acrocephalus dumetorum)]. Pp. 18-20 in Actual' nye problemyoologii. Lipetsk: Lipetsk State Pedagogical University Press. (In Russian.) Hume, A. O. (1869) To the Editor of The Ibis'. Ibis (2) 5: 355-357. Hume, A. O. (1871) Stray notes on ornithology in India. No. VI. On certain new or unrecorded birds. Ibis (3) 1 : 23-38. Ivanitskii, V. V., Kvartalnov, P. V., Marova, I. M., Samotskaya, V. V. & Bannikova, A. A. (2012) Modes of stay, vocalization patterns and morphological variability of Blyth's Reed Warblers (Acrocephalus dumetorum) in Central Asia. Zoologicheskii Zhurnal 91: 1459-1509. (In Russian with English summary.) Kistyakovsky, A. B. (1950) [Materials on the zoogeography of Pamir (based on the distribution of land vertebrates)]. Proceedings of the Zoological Museum of Shevchenko Kiev State University 2: 5-58. (In Russian.) Koblik, E. A., Arkhipov, V. Yu. & Redkin, Ya. A. (2010) New data on distribution of the Large-billed Reed Warbler Acrocephalus orinus Oberholser, 1905. Russian Journal of Ornithology 19 (Express-issue 596): 1619-1633. (In Russian.) Koblik, E. A., Redkin, Ya. A., Meer, M. 5., Derelle, R., Golenkina, S. A., Kondrashov, F. A. & Arkhipov, V. Yu. (201 1 ). Acrocephalus orinus: A case of mistaken identity. PLoS ONE 6: No. 4. - e17716. doi:10.1 371/ journal. pone.001 7716. Kocher, T. D„ Thomas, W. K„ Meyer, A., Edwards, S. V., Paabo, S„ Villablanca, F. X. & Wilson, A. C. (1 989) Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proc. Natn. Acad. Sci. USA 86: 6196-6200. Kuranov, B. D. (2008) Peculiarities of nesting biology in the Blyth Reed Warbler (. Acrocephalus dumetorum, Passeriformes, Sylviidae) in urban habitats. Zoologicheskii Zhurnal 87: 466-475. (In Russian with English summary.) Kvartalnov, P. V. & Garibmamadov, G. D. (201 2) Ornithological and oological collections of E.N. Pavlovsky Zoology and Parasitology Institute, Tajikistan: their history and present significance. Pp.1 62-1 65 in Nazemnye Pozvonochnye Zhivotnye Aridnykh Ekosistem. Tashkent: Chinor Enk. (In Russian with English summary.) Kvartalnov, P. V., Ivanitskii, V. V. & Marova, I. M. (2004) Ecology and social behaviour of the Reed Warbler ( Acrocephalus scirpaceus) at the eastern Azov sea coast. Ornitologia 33: 100-108. (In Russian with English summary.) Kvartalnov, P. V., Ivanitskii, V. V., Marova, I. M. & Samotskaya, V. V. (2011a) A bewitched bird: history of Large-billed Reed Warbler. Priroda 6: 35-40. (In Russian with English summary.) Kvartalnov, P. V., Samotskaya, V. V. & Abdulnazarov, A. G. (2011b) From museum collections to live birds. Priroda 12: 54-56. (In Russian.) McCarthy, E. M. (2006) Handbookof avian hybrids of the world. Oxford: Oxford University Press. Nikiforov, M. E., Yaminskiy, B.V.& Shklyarov, L. P. (1 989) Ptitsy Belorussii [Birds of Belorussia: a guide for nests and eggs], Minsk: Visheyshaya shkola. (In Russian.) Nowak, E. (2001) Erinnerungen an Ornithologen, die ich kannte (Teil 6). Professor Alexander Bogdanowitsch Kistiakowskij (1904-1983). Berkut 10: 234-242. (In German and Russian.) Oberholser, H. C. (1905) Birds collected by Dr. W.L. Abbott in the Kilimanjaro region, East Africa. Proc. US Natn. Mus. 28: 823-936. Portenko, L. A. & Stubs, J. (1976) Acrocephalus dumetorum Blyth. Atlas der Verbreitung Palaearkischer Vogel 5: 1 -5. Round, P. D„ Hansson, B., Pearson, D. J., Kennerley, P. R. & Bensch, S. (2007) Lost and found: the enigmatic Large-billed Reed Warbler Acrocephalus orinus rediscovered after 1 39 years. J. Avian Biol. 38: 1 33-1 38. Sambrook, J., Fritch, E. F. & Maniatis, T. (1989) Molecular cloning: a laboratory manual. New York: Cold Spring Harbor Laboratory Press. Svensson, L„ Prys-Jones, R., Rasmussen, P.C. & Olsson, U. (2008) Discovery of ten new specimens of Large-billed Reed Warbler Acrocephalus orinus, and new insights into its distributional range. J. Avian Biol. 39: 605-610. Svensson, L„ Prys-Jones, R., Rasmussen, P. C. & Olsson, U. (2010) The identification and distribution of the enigmatic Large-billed Reed Warbler Acrocephalus orinus. Ibis 1 52: 323-334. Timmins, R. J., Mostafawi, N., Rajabi, A. M., Noori, H., Ostrowski, S., Olsson, U., Svensson, L. & Poole, C. M. (2009) The discovery of Large-billed Reed Warblers Acrocephalus orinus in north-eastern Afghanistan. BirdingASIA 12:42-45. Timmins, R . J., Ostrowski, S., Mostafawi, N., Rajabi, A. M„ Svensson, L. & Olsson, U. (201 0) New information on the Large-billed Reed Warbler Acrocephalus orinus, including its song and breeding habitat in north-eastern Afghanistan. Forktail 26: 9-23. Totunov, V. M. (1981) [Breeding of the Blyth's Reed Warbler ( Acrocephalus dumetorum Blyth.) in vicinities of Chany lake.] Pp. 160-165 in Ekologiyai biotsenoticheskiye svyazi perelyotnykh ptits Zapadnoy Sibiri. Novosibirsk: 'Nauka' Press, Siberian Branch. (In Russian.) Pavel KVARTALNOV, Vertebrate Zoology Department, Faculty of Biology, Lomonosov Moscow State University, Leninskiye Gory, Moscow 1 1 999 1, Russian Federation. Email : cettia@mail.ru Abduinazar ABDULNAZAROV, Pamir Biological Institute, Khorog 736002, Tajikistan. Email: abdu_70@mail.ru Veronika SAMOTSKAYA, Vertebrate Zoology Department, Faculty of Biology, Lomonosov Moscow State University, Leninskiye Gory, Moscow 1 19991, Russian Federation Julia POZNYAKOVA, Vertebrate Zoology Department, Faculty of Biology, Lomonosov Moscow State University, Leninskiye Gory, Moscow 1 19991, Russian Federation Irina ILYINA, Vertebrate Zoology Department, Faculty of Biology, Lomonosov Moscow State University, Leninskiye Gory, Moscow 1 19991, Russian Federation Anna BANNIKOVA, Vertebrate Zoology Department, Faculty of Biology, Lomonosov Moscow State University, Leninskiye Gory, Moscow 1 19991, Russian Federation Eugenia SOLOVYEVA, Vertebrate Zoology Department, Faculty of Biology, Lomonosov Moscow State University, Leninskiye Gory, Moscow 1 1 999 1, Russian Federation F. GERMI © NATURAL HISTORY MUSEUM, LONDON FORKTAIL 29 (2013): 43-47 First records of Chinese Sparrowhawk Accipiter soloensis wintering in Papua (Indonesian New Guinea) FRANCESCO GERMI, AGUS SALIM & ANDREA MINGANTI Chinese Sparrowhawks Accipiter soloensis were found for the first time wintering in mainland Papua (Indonesian New Guinea) during a field survey carried out between December 2010 and March 201 1. A combination of 39 road, boat and foot transects were completed in the provinces of Papua and West Papua, covering a total of 2,303 km, of which 1,948 km were on roads or footpaths and 355 km on rivers. Transects were supplemented by frequent spot counts and stops to broadcast recordings of Chinese Sparrowhawk vocalisations. Routes covered eight sample areas in the most representative habitats of the region. A total of 1 0 Chinese Sparrowhawks were recorded at four locations, all close to the coast. The new records are up to 1,200 km east of the easternmost extent of the previously known wintering range, thus proving that this species does winter in Indonesian New Guinea, although most likely at low density. Seventeen other raptor species were recorded on the transects. In addition, 12 days were spent between 6 and 17 March 2011 at a suitable coastal watch site at the westernmost point of West Papua, but no visible migration of Chinese Sparrowhawk was observed. INTRODUCTION The Chinese Sparrowhawk Accipiter soloensis is a small-sized accipiter, whose breeding grounds, although imprecisely known, are in China, Korea and Ussuriland, and whose wintering grounds are thought to be mainly in the Philippines and eastern Indonesia. Despite this being the most numerous migratory raptor in East Asia, with an estimated global population running well into six figures (Ferguson-Lees & Christie 2001), very few records exist from the presumed wintering range, suggesting that the winter distribution remains largely unknown (Wattel 1973, White 1976, Mees 1982, White & Bruce 1986, Beehler et al. 1986). Recent migration research (Germi 2005, Germi & Waluyo 200 6, Nijman et al. 2006, Germi et al. 2009) showed that at least 350,000 individuals of this species are streaming into eastern Indonesia each autumn, through both the Sangihe-Talaud Plate 1 . The five Chinese Sparrowhawks collected in New Guinea in the nineteenth century at unspecified localities, held at the Natural History Museum, Tring, UK. From left to right: NHMUK 1955.6.N.20.2635; 1873.5.12.1638; 1955.6.N.20.2636; 1 955.6.N.20.2633 and 1955.6.N.20.2632. 44 FRANCESCO GERMI, AGUS SALIM & ANDREA MINGANTI Forktail 29 (2013) Archipelago in the north and Bali in the west. Nevertheless, other than scattered records from Sulawesi’s northern peninsula, winter observations of Chinese Sparrowhawk elsewhere on this island are negligible, despite some areas having been relatively well surveyed (Hartert 1896, Meyer & Wiglesworth 1898, Rozendaal & Dekker 1989). Likewise, there are very few winter records of this species from the Moluccas and the Lesser Sundas. A handful of specimens and records provide solid, albeit scarce, evidence that this species occurs further to the east in Wallacea and at least in the extreme western tip of New Guinea (Coates 1985, Beehler et al. 1986, Coates & Bishop 1997, Gjershaug& Rov 2000). Only four records from the islands off western Papua are known in the literature: Meyer & Wiglesworth (1898) quoted Salvadori (1880-1882) with two specimens collected at Gagi (by Bernstein) and Waigeo (by Platen) islands and one from an unspecified locality in ‘ Nova Guinea (by Wallace). Ripley (1964) reported one female collected at Efman (Jefman) Island near Sorong. Additionally, we located five specimens held in the ornithological collections at the Natural History Museum (formerly British Museum of Natural History), Tring, collected in New Guinea in the nineteenth century at unspecified localities (Plate 1). These specimens appear to have been overlooked in the relevant literature. As the islands of eastern Indonesia are regularly visited by birding tours, it is remarkable that such a large number of Chinese Sparrowhawks could have been so consistently unrecorded during the wintering months. Moreover, our observations on Sangihe Island (Germi et al. 2009) showed that when present the Chinese Sparrowhawk is easy to locate, both because of its strong diurnal vocal activity and when perched or hunting from prominent tree branches. The possible explanation for the fact that the species has been so widely ‘overlooked’ in Wallacea, may be: (1) it has been genuinely unrecorded due to an absence of observers present at the right time of year; or (2) the lack of suitable wintering areas in parts of Wallacea, so that a large proportion of the migrants disperse into remote and poorly surveyed areas as far as Papua (Indonesian New Guinea), or both. In order to address the paucity of winter records in Wallacea, we undertook this survey in Papua to establish if Chinese Sparrowhawk overwinter on the island, simultaneously carrying out the first raptor road counts in New Guinea. METHODS Study area Papua, the Indonesian half of the island of New Guinea, covers an area of 416,129 km2. It is still largely covered in relatively undisturbed primary forests, the largest tropical forest wilderness remaining in the Asia-Pacific region. For the past 50 years it was essentially inaccessible to all but a few field researchers, and thus a terra incognita (Marshall & Beehler 2007). Dominated by the massive Central Cordillera (more than 3,000 m high) that generates abundant rainfall, rivers drain into vast forested interior basins and alluvial floodplains. In the far south-east, near the Papua New Guinea border, is a swath of savannah known as the Trans-Fly, ecologically resembling northern Australia rather than New Guinea. To the west, the heavily logged Vogelkop and Bomberai peninsulas are dominated by small mountain ranges and island groups. Papua’s equatorial climate is dominated by the North-west Figure 1. Map of Papua, Indonesia. Black lines show all transects carried out during December 2010 - March 201 1 in the numbered sample areas: (1) Merauke; (2) Kimaam; (3) Asmat; (4) Timika; (5) Wamena; (6) Manokwari; (7) Sorong; (8) Biak. Black stars: new Chinese Sparrowhawk records January- March 201 1. White stars: old Chinese Sparrowhawk records 1880-1964. Black asterisk: migration watch site at Tanjung Kasuari, March 201 1. 130°E 132°E 134° E 136'E 138“E 140°E 142°E 130°E 132*E 134°E 136°E 138*E 140°E 142*E SoZ Sotr So9 So8 Forktail 29 (2013) First records of Chinese Sparrowhawk wintering in Papua (Indonesian New Guinea) 45 Table 1 . Sample areas, transects and Chinese Sparrowhawk records, Papua, December 201 0-March 2011. Sample area Coordinates Habitat types Transects Chinese Sparrowhawk records Merauke 8.479°S 140.580°E 1,3,4,5,13,14,15 road 454 km; foot 22 km Kimaam 7.957°S 138.862°E 1,5,6,14 foot 48 km; river 28 km 7.980°S 1 38.853°E, one on 10 January roosting at Kimaam airstrip Asmat 5.518°S 138.1 17°E 4,6,7,14 foot 17 km; river 327 km Timika 4.755°S 136.865°E 4,6,7,10,14, 15 road 708 km; foot 13 km 4.691°S 136.887°E, a flock of six in westward migration on 3 March over coastal swampland Wamena 3.964°S 138.946°E 2,8,13,14 road 160 km; foot 11 km Manokwari 0.791 °S 134.023°E 9,10,13,14,15 road 175 km 0.808°S 134.053°E, one on 12 February roosting in secondary forest Sorong 0.823°S 131 .231 °E 10,13,14,15 road 328 km; foot 12 km Biak 1.048°S 135.959°E 11,12,14,15 point counts 1.094°S 136.330°E, two singles at the same locality, 14 and 15 February, roosting in degraded woodland Habitat types: 1 Trans-Fly savannah, 2 Grassland, 3 Inundated grassland , 4 Swampland, 5 Monsoon forest , 6 Mangrove forest , 7 Lowland rainforest, 8 Montane forest, 9 Coastal forest, 10 Secondary forest , 1 1 Degraded woodland, 12 Semi-arid scrub, 13 Agricultural landscapes, 14 Villages, 15 Transmigration settlements. Monsoon and the South-east T rade Winds. The main rainy season occurs from November to March; however in the wetter areas the seasons are reversed, and most rain falls in the April-October period. The highest rainfall is recorded in the southern scarp of the Central Cordillera, with more than 5,000 mm/year. Temperature varies little, with elevation being the key variable. At sea level, where most of our survey areas were located, the average temperature is 25-27°C (Marshall & Beehler 2007). Survey techniques Our general methodology followed a combination of classic survey techniques described in the literature (Fuller & Mosher 1987, Bibby et al. 1998, Bibby etal. 2000, Bird & Bildstein 2007, Malan 2009). We also reviewed specific literature on previous raptor road surveys from other parts of the world in order to adapt our own protocols to local conditions (Millsap & LeFranc 1988, Ellis et al. 1990, Hanowsky & Niemi 1995, Vergara 2010). Thirty-nine road, boat and foot transects were carried out in eight sample areas in the Indonesian administrative provinces of Papua and West Papua (Figure 1), chosen by habitat type and for their accessibility. The following main habitats were surveyed: Trans-Fly savannah and grassland, swamp forest, mangrove forest, lowland rainforest, secondary forest and agricultural landscapes (Table 1). Logistical difficulties such as lack of roads and suitable transport, high cost of transport, access restrictions due to oil and mining exploration and floods due to the rainy season hindered the fieldwork. The resulting poor accessibility and the insecurity of some areas restricted the number of study sites and the time spent in them. All 39 transects were carried out in the middle of the wet season (December-late February) and recorded as GPS tracks, and all raptor sightings as GPS waypoints. The length of routes was measured by using a GPS, the vehicle odometer and topographic maps. As habitat structure, detectability of raptors and driving conditions were too variable, we did not follow a strict methodology protocol (i.e. strip, line or point transects), but a combination of the three. To facilitate relocation of the routes for future surveys, we downloaded all recorded tracks in a GPS navigation software (Touratech QV). Sample segments were arranged by habitat type. To minimise differences in detectability among transects, we standardised the time of day and weather conditions. Driving speed was 40-60 km/hr on paved road and 20-40 km/hr or less on dirt roads, although road conditions and habitat type were too variable to permit a uniform driving speed. Roadside counts were conducted by 1-2 observers assisted by one driver. Observers looked carefully for accipiters while driving slowly (<50 km/hr), and during frequent stops to identify distant birds or scan the landscape, from after sunrise to before sunset (07h00- 18h00), using 10x binoculars and a 20-60x telescope. All accipiters seen perched or flying were counted and identified if possible. As habitat structure in the proximity of the roads affected detectability of birds, we stopped at points spaced about 1-3 km in suitable habitats to look and listen for accipiters during a 5-10 minute period, or to conduct foot transects when necessary. During these stops we broadcast three different recordings of Chinese Sparrowhawk vocalisations in an effort to elicit vocal responses or approaches, using a standardised protocol (Parker 1991). Twelve days from 6 to 17 March 2011 were spent at a peninsular migration watch site, Tanjung Kasuari (0.823°S 131.231°E), near the westernmost point of West Papua, in the proximity of the town of Sorong, to assess whether migratory movements were visible at this site. Methods were similar to those used in previous migration studies in Indonesia (Germi etal. 2009). RESULTS Transects covered a total length of 2,303 km, of which 1,948 km were on roads or footpaths and 355 km on rivers. T ransects in forest habitats resulted in very few raptor sightings, probably due to poor detectability in dense forest, thus suggesting that this habitat type requires a different methodological approach to carry out raptor counts (Thiollay 1989, Whitacre 1991, Whitacre & Turley 1991). Eighteen raptor species were recorded during transects, all at low density (Table 2), and including three species (Black-shouldered Kite Elanus caeruleus, Bat FFawk Macheiramphus alcinus and Brown Goshawk Accipiter fasciatus ) from areas outside their known distribution, thus extending their range within New Guinea. Chinese Sparrowhawk was observed five times at four sites (Kimaam, Timika, Manokwari and Biak), 10 individuals in all (Table 1), up to 1,200 km east of the previously known limit of the wintering range, thus proving that this species does winter in Papua. These are the first records for the species in mainland New Guinea. All five records were at sea level, near coasts, in degraded habitats and swampland. Individuals were observed at sufficiently close range to permit positive identification using diagnostic field characters. The authors are very familiar with the identification of this species from previous field studies in the region (Germi 2005, Germi & Waluyo 2006, Germi et al. 2009). New records Kimaam, Dolok island: 1 0 January 2011, one adult male observed and photographed at short distance, roosting on a light pole at Kimaam airstrip (7.980°S 138.853°E) in Trans-Fly savannah. Manokwari: 12 February 2011, one adult male roosting on a tree branch in secondary coastal forest in the proximity of the town of Manokwari (0.808°S 134.053°E). 46 FRANCESCO GERMI, AGUS SALIM & ANDREA MINGANTI Forktail 29 (2013) Tabie 2. Raptor species recorded during transects, Papua, December 201 0-March 2011. Species Merauke Kimaam Asmat Timika Wamena Manokwari Sorong Biak Total Black-shouldered Kite Elanus caeruleus 12 12 Bat Hawk Macheiramphus alcinus 1 1 Pacific Baza Aviceda subcristata 4 3 7 Long-tailed Honey-buzzard Henicopernislongicauda 2 6 4 12 Whistling Kite Haliastur sphenurus 67 8 11 86 Brahminy Kite Haliastur indus 15 9 6 11 1 8 6 4 60 White-bellied Fish-eagle Haliaeetus leucogaster 2 3 4 5 1 15 Papuan Harrier Circus spilothorax 4 4 3 10 21 Chinese Sparrowhawk Accipiter soloensis 1 6 1 2 10 Grey Goshawk Accipiter novaehollandiae 1 4 5 Varied Goshawk Accipiter hiogaster 1 1 3 1 6 Collared Sparrowhawk Accipiter cirrhocephalus 3 2 2 5 1 3 1 17 Brown Goshawk Accipiter fasciatus 3 2 5 Gurney's Eagle Aquilagurneyi 1 1 2 Wedge-tailed Eagle Aquila audax 3 3 Oriental Hobby Falco severus 1 1 Brown Falcon Falco subniger 2 2 Peregrine Falcon Falco peregrinus calidus 1 8 2 11 Note: The taxonomy of Accipiter novaehollandiae and A. hiogaster remains controversial among authors. Here they are treated as separate species. Biak island: 14 and 15 February 2011, two single females (possibly the same bird) observed at the same site, roosting on a tree branch in degraded woodland (1.094°S 136.330°E). Timika: 3 March 2011, a flock of six individuals soaring and gliding at low altitude, apparently in westward migration, observed and photographed over swampland (4.69 1°S 136.887°E) along the mining road between the town of Timika and the port of Amamapare. From previous observation in Sangihe (Germi et al. 2009), we found that Chinese Sparrowhawk can be conspicuously vocal when perched on exposed tree branches, especially in the early morning. Although we broadcast Chinese Sparrowhawk vocalisations on numerous occasions in seven sample areas (excluding Biak) and towards the individual observed in Manokwari, we never obtained any response or approach. During 1 2 days spent at a suitable watch site at the westernmost point of Papua during the migration season, no movements of Chinese Sparrowhawk were observed, although dates were chosen to precede by a few days the known passage period in North Sulawesi. DISCUSSION W e plotted the known records of migrating Chinese Sparrowhawks in Wallacea and the handful of historical winter records from the islands off the westernmost tip of Papua. This hinted at a probable winter distribution in poorly surveyed but ecologically suitable areas in Papua, particularly in the open habitats of the south-east. As this territory is subject to heavy flooding during the rainy season, its wintering avifauna is poorly known, partly because of the inaccessibility of much of the region. Moreover, very few formal surveys have been published from Papua in the past 20 years and only a handful of informal accounts from short visits by ornithologists and birdwatchers have appeared (Marshall & Beehler 2007). Records of wintering Chinese Sparrowhawks in other parts of Indonesia show that this accipiter is not a closed-canopy forest species; instead it occupies a variety of open habitats ranging from forest edge, secondary forest and scrub to agricultural landscapes. Previous observations in central Flores (Germi et al. 2009) showed that wintering Chinese Sparrowhawks feed primarily on cicadas, and the majority ot birds leave the island in late November once cicada emergences cease, and only small numbers remain through the winter. Interestingly, large emergences of cicadas are known to occur at the beginning ol the rains in October-November in Papua and in October-February in the Moluccas, suggesting that Chinese Sparrowhawks move east when insect abundance shifts with the advancing rainy season (Germi et al. 2009). The south-east ol Papua, part of the Greater Trans-Fly, is characterised by an extensive mosaic of monsoon forest, secondary forest, swampland and swathes of savannah, thus potentially ideal wintering habitat. Although this region was initially the focus of our study, lack of accessibility and the subsequent paucity of records (only one individual at Kimaam, Dolok island) prompted us to extend the survey into different parts of Papua. Most areas surveyed presented several logistical constraints and difficult access. Suitable habitat covered large inaccessible areas, and the scarcity of records during this study suggests that the species disperses at low density within the vast region. However our records, although small in number, clearly indicate that an unknown proportion of this accipiter population overwinters in Papua, apparently in coastal areas. The absence of migrating Chinese Sparrowhawks in March (migration season) at the watch site of Tanjung Kasuari (Figure 1) might reflect the use of a different route further to the south. We assumed that migrants would travel from the Sorong area to Waigeo and Halmahera islands, but the lack of migrants at Tanjung Kasuari could indicates that migrants leave Papua via Salawati Island, where shorter water crossings are available, and continue east through Kofiau, Obi and the Sula islands before reaching Forktail 29(2013) First records of Chinese Sparrowhawk wintering in Papua (Indonesian New Guinea) 47 Sulawesi. Access restrictions on the Papua coast in the proximity of Salawati prevented us from testing this hypothesis. ACKNOWLEDGEMENTS This survey was funded by The Peregrine Fund (USA) and Hawk Mountain Sanctuary (USA). We wish to thank Cheryl Tipp (British Library of Wildlife Sounds) and Yasunori Nitani (Asian Raptor Research and Conservation Network) for providing Chinese Sparrowhawk recordings; Mark Adams at the Natural History Museum, Tring, for access to specimens; Daniel Natush and Jessica Lyons for their help in Timika and for sharing raptor records; Benja Mambai and Marco Wattimena (WWF Papua) for logistical assistance; Rick Watson and Keith Bildstein for reviewing an earlier draft of the manuscript and two anonymous referees for helpful comments. We are also grateful to the following people for their kind cooperation at various stages of the study: Stephen Bird, Giuseppe Carpaneto (Department of Environmental Biology, Rome University), Tom Edwards, James Ferguson- Lees, Lloyd Kiff (Global Raptor Information Network/The Peregrine Fund), Marco Panella (Italian Ministry of Forests) and Michael Schneider. REFERENCES Beehler, B. M., Pratt, T. K. & Zimmerman, D. A. (1986) Birds of New Guinea. Princeton USA: Princeton University Press. Bibby, C„ Jones, M. & Marsden, S. (1998) Bird surveys. London: Expedition Advisory Centre. Bibby. C. J., Burgess, N. D., Hill, D .A. & Mustoe, S. FI. (2000) Bird census techniques. London: Academic Press. Bird, D. M. & Bildstein, K. L. (2007) Raptor research and management techniques. Blaine USA: Plancock blouse Publishers. Coates, B. J. (1 985) The birds of Papua New Guinea. Vol. 1 . Alderley, Queensland: Dove Publications. Coates, B. J. & Bishop, K. D. (1997) A guide to the birds of Waiiacea: Sulawesi, the Moluccas and Lesser Sundas islands, Indonesia. Alderley, Queensland: Dove Publications. Ellis, D. H„ Glinsky, R. L. & Smith, D. G. (1990) Raptor road surveys in South America. J. Rap. Res. 24 (4): 98-1 06. Ferguson-Lees, J. & Christie, D. A. (2001) Raptors of the world. London: Christopher Helm. Fuller, M. R. & Mosher, J. A. (1987) Raptor survey techniques. Pp. 37-65 in B. A. Giron Pendleton, B. A. Millsap, K. W. Cline & D. M. Bird, eds. Raptor management techniques manual. Washington DC: Institute for Wildlife Research, National Wildlife Federation, Scientific and Technical Series No. 10. Germi, F. (2005) Raptor migration in East Bali, Indonesia: observations from a bottleneck watch site. Forktail 21 : 93-98. Germi, F. & Waluyo, D. (2006) Additional information on the autumn migration of raptors in East Bali, Indonesia. Forktail 22: 71-76. Germi, F„ Young, G. S., Salim, A., Pangimangen, W. & Schellekens, M. (2009) Over-ocean raptor migration in a monsoon regime: spring and autumn 2007 on Sangihe, North Sulawesi, Indonesia. Forktail 25: 104-1 16. Gjershaug, J. O. & Rov, N. (2000) The raptor fauna in North Moluccas, Indonesia. Pp. 240-247 in H. Ichinose, T. Inoue & T. Yamazaki, eds. Proceedings of the first symposium on raptors of Asia. Shiga Japan: ARRCN. Hanowsky, J. M. & Niemi, G. J. (1995) A comparison of on- and off-road bird counts: do you need to go off road to count birds accurately? J. Field Ornithol. 66 (4): 469-483. Hartert, E. (1896) On ornithological collections made by Mr. Alfred Everett in Celebes and on the islands south of it. Novit. Zook 3: 1 48-1 83. Malan, G. (2009) Raptor survey and monitoring. Pretoria, South Africa: Briza Publications. Marshall, AJ. & Beehler, B.M. (2007) The ecology of Papua. 2 vol. Singapore: Periplus. Mees, G. F. (1 982) Birds from the lowlands of southern New Guinea (Merauke and Koembe). Zook Verhand. 191. Meyer, A. B. & Wiglesworth, L. M. (1898) The birds of Celebes and neighbouring islands. Berlin: Friedlander. Millsap, B. A. & LeFranc, Jr., M .N. (1988) Road transect counts for raptors: how reliable are they? J. Rap. Res. 22 (1 ): 8-1 6. Nijman, V., Germi, F. & van Balen, S. (2006) Relative status of two species of migrant sparrowhawks on Java and Bali, Indonesia. Emu 106: 157-162. Parker, T. A. (1991 ) On the use of tape recorders in avifaunal surveys. Auk 1 08: 443-444. Ripley, S. D. (1964) A systematic and ecological study of birds of New Guinea. Bull. Peabody Mus. Nat. Hist. 19: 1-87. Rozendaal, F. G. & Dekker, R. W. R. J. (1989) Annotated checklist of the birds of the Dumoga-Bone National Park, North Sulawesi. Kukila 4 (3-4): 85- 109. Salvadori, T. (1880-1882) Ornitologia della Papuasia e delle Molucche. 3 vol. Turin: G. B. Paravia. Thiollay, J. M. (1989) Censusing of diurnal raptors in a primary rain forest: comparative methods and species detectabilty. J. Rap. Res. 23 (3): 72- 84. Thiollay, J. M. & Rahman, Z. (2002) The raptor community of Central Sulawesi: habitat selection and conservation status. Biol. Conserv. 1 07: 111-122. Vergara, P. (2010). Time-of-day bias in diurnal raptor abundance and richness estimated by road surveys. Catalan Journal of Ornithology 26: 22-30. Wattel, J. (1973) Geographical differentiation in the genus Accipiter. Cambridge USA: Nuttall Ornithological Club. Whitacre, D. F. (1991) Censusing raptors and other birds in tropical forest: further refinements of methodology. Pp.31-42 in D. F. Whitacre, W. A. Burnham & J. P. Jenny, eds. Maya Project: Use of raptors and other fauna as environmental indicators for design and management of protected areas and for building local capacity for conservation in Latin America. Progress report IV. Boise USA: The Peregrine Fund. Whitacre, D. F. & Turley, C. V. (1991) Further comparisons of tropical forest raptor census techniques. Pp.71-92 in D. F. Whitacre, W. A. Burnham & J. P. Jenny, eds. Maya Project: Use of raptors and other fauna as environmental indicators for design and management of protected areas and for building local capacity for conservation in Latin America. Progress report IV. Boise USA: The Peregrine Fund. White, C. M. N. (1976) Migration of Palearctic non-passerine birds in Waiiacea. Emu 76:79-82. White, C. M. N. & Bruce, M. D. (1986) The birds of Waiiacea (Sulawesi, the Moluccas and Lesser Sunda Islands): an annotated check-list. London: British Ornithologists' Union (Checklist No. 7). Francesco GERMI, 46 (2F2) Elm Row, Edinburgh EH7 4AH, UK. Email: fgermi@yahoo.co.uk Agus SALIM, Jl. Palem Putri VIII No. 5-7, Taman Jasmin 5, Bogor, 16113, West Java, Indonesia. Email: salim@asalim.org Andrea MING ANTI, Via Monti di Primavalle 96/a, 00 1 68 Rome, Italy. Email: andming@alice.it FORKTAIL 29 (201 3): 48-51 Spatial distribution of the Tawny Fish Owl Ketupa flavipes shaped by natural and man-made factors in Taiwan SHIAO-YU HONG, YUAN-H5UN SUN, HSIN-JU WU & CHAO-CHIEH CHEN This study investigated the distribution of the Tawny Fish Owl Ketupa flavipes, a rare top predator in Taiwan, and examined natural and man-made factors that affect it. Records of Tawny Fish Owls from 1 993 to 2006 were compiled from field studies, literature surveys, museum notes and specimens, and interviews with researchers, birdwatchers and indigenous hunters. In total, 91 Tawny Fish Owl territories were identified, widely distributed along mountain streams in the Central Mountain Range between 48 and 2,407 m — more than half of them were below 700 m. The upper altitudinal range of the owls is probably limited by food availability and stream size. Territories were on average 431 m higher on the west side of the Central Mountain Range than on the east. Habitat selection analysis further indicated that, in proportion to the land area available, the Tawny Fish Owl was absent from areas below 500 m.This is apparently due to extensive deforestation of lowlands for agriculture and urbanisation on the west side of the island. It is recommended that a protected area be established in the north-east part of Taiwan, to preserve the remaining lowland streams and riparian forests still inhabited by the species. INTRODUCTION Fish owls are top predators in freshwater food chains and are important indicator species for healthy stream ecosystems (Duncan 2003, Wu et dl. 2006). The Tawny Fish Owl Ketupaflavipes is widely distributed in the Himalayas, eastern Indochina, south China and Taiwan (Voous 1988, Marks et al. 1999). It reaches 58 cm in body length (Sun 1996), making it one of the largest raptors within its range. However, it is so rarely observed in the wild that it is considered to be rare over most of its range (Marks et al. 1999). It is currently listed in CITES Appendix II (UNEP-WCMC 2009). The species was first reported from Taiwan by Kuroda (1916). Since then its natural history has remained poorly known (Voous 1988). In 1989, when the Wildlife Conservation Law was first implemented in Taiwan, the Tawny Fish Owl was listed as an Endangered Species (Class I) due to poaching (Sun 1996) and destruction of riparian forests (Severinghaus 1987). Even though its population was still considered very small (Fang 2005), Tawny Fish Owl was down-listed to Rare and Precious Species (Class II) in 2008, due to better knowledge of its population size (Sun 1996, Sun etal. 2000,2004). As the distribution and population size of the Tawny Fish Owl in Taiwan remained unclear, further investigation was needed to correctly assess its protection status. The aims of this study were to determine its spatial distribution and to evaluate factors affecting its distribution by identifying locations throughout Taiwan where Tawny Fish Owl had been recorded recently. METHODS Study area Taiwan is a mountainous island lying between 21 and 25°N and 120 and 122°E in subtropical East Asia, about 150 km off the coast of south-east China. It is 394 km long and 144 km wide with a total area of 36,000 km2. Hills and mountains with elevations higher than 100 m cover about 70% of the island (Taiwan Forestry Bureau 1995). The Central Mountain Range, with more than 200 peaks over 3,000 m in height, runs from north to south. The rivers that originate from the mountain range are short and steep with rapid currents, particularly on the east side. A lower Coastal Mountain Range is present between the east coast and the Central Mountain Range. Forests cover 59% of the island, of which 73% is natural forest (44% of the island) (Taiwan Forestry Bureau 1995). Broad-leaved forests predominate below 1,500 m, coniferous forests above 2,500 m, with mixed forests in between. The coastal plains are much wider on the west side than on the east, and most of them have been converted to urban or agricultural areas. Only small remnants of the original riparian forests remain on the peripheral hilly areas on the west side of the mountain range, whereas most forests on the east side remain little affected because of access difficulties (Taiwan Forestry Bureau 1995). Annual precipitation is between 1,000 mm and 4,800 mm, and decreases gradually from north-east to south¬ west because of monsoon rainfalls and incidence of typhoons (Chiu 2006). Annual average temperature is about 25°C at sea level and 4°C at about 3,800 m (Chiu 2006). Data collection and analysis Tawny Fish Owl records between 1993 and 2006 were collected using the following approaches to identify locations where it occurred. First, faunal reports and the bird-sighting databases of wild bird societies were searched for records, and the collection data for Tawny Fish Owl specimens preserved in the Academia Sinica (AS), the National Taiwan Museum (NTM), National Museum of Natural Science (NMNS), and Taiwan Endemic Species Research Institute (TESRI) were examined. A total of 42 records were obtained: nine records from faunal reports, 15 from wild bird societies, two from AS, five from NTM, six from NMNS, and five from TESRI. Second, 19 birdwatchers and 1 19 hunters from 27 indigenous villages located near rivers where there were no or very few records of Tawny Fish Owls were interviewed. All information indicating the presence of owls was collected, including hunting captures, sightings, calling birds, pellets, feathers, droppings and food remains. Tawny Fish Owl calls are easily distinguished from those of other owls in Taiwan (Sun 1996). Interviewees had to describe the sound of the owl clearly before each record was considered valid. As a fish-eater, the droppings and food remains of Tawny Fish Owls were easily identifiable (Wu et al. 2006). We obtained 52 records from the interviewees. Third, field surveys were conducted to identify Tawny Fish Owl sites, using the same clues as above to detect the presence of owls. The sites surveyed included 15 coldwater fish farms (Rainbow Trout Oncorhynchus mykiss and smelt [Osmeridae spp. ] ) near mountain streams, mostly in northern Taiwan. The owls preyed on fish in the farms and often left remains, such as fish scales, bloodstains and internal fish organs, along with their own feathers, near the predation sites (Sun et al. 2004). The Tawny Fish Owl has a strong territorial habit (Fogden 1973, Sun 1996, Sun et al. 2000) . A territory of a pair of owls was Forktail 29(2013) Spatial distribution of the Tawny Fish Owl Ketupa flavipes in Taiwan 49 estimated to be about 6.2 km along a river and less than 550 m from the bank, derived from two territories measured by radio¬ tracking (Sun et al. 2000). In this study any owl records obtained within a 6.2 km length of river were considered to refer to a single pair, and the midpoint of the two outermost records was used to represent the territorial site. To construct the Tawny Fish Owl’s altitudinal distribution chart, the elevation of each of the owl territories was obtained by using coordinates under Identify (ArcGIS) on a Digital Terrain Model (DTM) of Taiwan (precision 40 m x 40 m). The perpendicular distance of each owl territory from the crest line of the Central Mountain Range was calculated, to illustrate how altitudinal distribution varied across the island. A 50 m interval contour map was created from the DTM by 3D Analyst, and then transformed to a 3D topographic chart in ArcScene. The 3D topographic chart was rotated to calculate the altitudinal profile of the island (between 25 and 375 km south of the northern tip of the island), as viewed by an observer standing on the south end of the island. Elevations of territories to the east and west sides of the Central Mountain Range were compared with an independent r-test. The altitudinal range of the island from sea level to 3,950 m was divided into four bands: <500 m, 500-1,000 m, 1,000- 1,500 m and >1,500 m. The proportion of owl territories in each altitude band was compared with the proportion of the land area in each altitude band. A use and availability analysis (Litvaitis etal. 1994) was conducted using a chi-squared test and multiple simultaneous comparisons between Bonferroni confidence intervals of observed use and the proportion of land area in each altitude band. All statistical analyses were conducted with SPSS 10.0.7C for Windows with an a-level set at 0.05. RESULTS A total of 153 owl records were obtained. Of these, 15 museum specimens and 1 1 records from wild bird societies had no clear location and were excluded from the study. The 127 valid records consisted of 37 from birdwatchers, 28 from indigenous hunters, 35 from field surveys along rivers, 1 5 from fish farms, 9 from fauna reports and 3 from museum collections. Of these, 39 records were direct sightings of owls, 35 hunting captures or owls found dead, 28 sites with owl pellets, 17 calling birds and 8 records of feathers, food remains, droppings or other signs (Table 1). Taking owl records found within a 6.2 km section of a river to represent single pairs, the 127 valid records indicate 91 owl territories. Two of these territories had four records and 70 Table 1. Sources and categories of 127 Tawny Fish Owl records obtained in 1993-2006 for this study. Captured birds included those caught by hunters or found dead by birdwatchers and those preserved as specimens in museums and research institutions. Sources Categories Birdwatchers Field survey Hunters Fish farm Reports Museums Total Sightings 13 13 11 2 39 Captures 2 15 15 3 35 Pellets 10 14 4 28 Calling birds 8 5 2 2 17 Feathers/ food remains 3 1 4 Droppings 2 2 Others 1 1 2 Total 37 35 28 15 9 3 127 Figure 1. Locations of 91 verified Tawny Fish Owl territories in Taiwan, 1993 to 2006. 120°0'0"E DIWE 122 0'0"E 25 0’0"N 24 0'0"N 23 0'0"N 22“0'0"N Elevation (m) 0-500 501 - 1,500 1 ,50 1 - 2,500 01020 40 60 80 territories contained only one. The territories were distributed along mountain streams in the Central Mountain Range (Figure 1 ). There were no records from either of the coastal plains or the Coastal Mountain Range in the east. The altitudinal distribution of the owl territories ranged between 48 m and 2,407 m with mean 687 m and the middle 50% range between 300 m and 1,100 m (Figure 2). The distribution was skewed toward low elevations (skewness = 0.97), implying that most of the owl territories were located at lower altitudes. The transverse altitudinal distribution of the owl territories in relation to the Central Mountain Range is shown in Figure 3. There was significant difference {t = 3.04, df = 55 ,P < 0.01) between the Figure 2. Frequency distribution of the elevations of 91 verified Tawny Fish Owl territories in Taiwan, 1 993-2006. in CD > o c CD Z3 cr CD LL 14 ] 12 10 8 0 1,000 1,500 2,000 2,500 Elevation (m) 50 SHIAO-YU HONG etal. Forktail 29(2013) Figure 3. Altitudinal distribution of Tawny Fish Owl territories relative to the crest line of the Central Mountain Range. The grey outline shows the altitudinal profile of the island, viewed from the south end. (The vertical scale is magnified 30-fold as the elevation of Mount Yushan, the highest peak of Taiwan, is relatively small [4 km] compared with the width [144 km] of the island.) Crest line of Central Mountain Range 0 1 — - - 1 - 1 - 1 - 1 — - - 1 - 1 — - - 1 — 1 - 1 0 20 40 60 80 100 120 140 160 Transverse distance (km) east and west side of the mountain range. Most of the territories on the west side were above 500 m, whereas more than half of the territories on the east side were below 500 m. The former averaged 431 m higher than the latter. Territories were generally situated in deep valleys, well below the surrounding mountain ridges (Figure 3). Tawny Fish Owls did not use the four altitude bands in proportion to the land area of the island (%2 = 41.57, df = 3, P < 0.0001). Multiple comparisons showed that the owls appeared in land below 500 m and above 1,500 m in a significantly lower proportion compared with the land area of these altitude ranges (Table 2). In contrast, the proportion of owl territories was twice the proportion of the land area between 500 and 1,500 m, indicating that the owls selected this altitude range. DISCUSSION The Tawny Fish Owl mainly takes prey from streams (Sun 1996, Wu et al. 2006). Its upper altitudinal range may be constrained by the distribution of stream fishes (Voous 1988, Marks et al. 1999). In Taiwan, the highest reported elevation for stream fish is 2,400 m (Tzeng 1986, Wang 20 10), which coincides with the upper limit of the owl’s altitudinal range (Figure 2). Fish abundance and diversity increase at lower elevations, for example 10 fish species occur at 1,500 m (Wang 2010) and only six owl territories were recorded above 1,500 m suggesting that fish abundance and diversity is important to the species. However, amphibians and crabs were found to be the most important prey items for Tawny Fish Owls at Sakatang Stream, eastern Taiwan (Wu et al. 2006), and the abundance and diversity of these taxa also increase at lower altitudes. The occurrence of the Tawny Fish Owl was also related to the distribution of mountain streams. Streams with stable flow above 1,500 m are only found in central regions where large mountains occur. Tawny Fish Owls seldom forage in small creeks less than 5 m wide, perhaps due to food scarcity and the poor manoeuvrability of a bird with a wingspan of 1.5 m (Sun et al. 2000). As a result, the highest parts of the range mainly occurred around the largest mountains. The latter phenomenon was possibly enhanced by the Massenerhebung effect, which predicts that locations at the same elevation are warmer on large mountains compared to smaller ones, because the large mountains release heat more slowly (Flenley 1994). Conversely, local differences between the altitudinal range of the owls and the surrounding mountain ridges were larger in small mountains and mountains with steep slopes. For example, the owl’s upper altitudinal limit on the east side of the Central Mountain Range, where rivers are much smaller, steeper and often dry up, was lower than that on the west side. This may partly explain why owl territories on the east side of the island were lower than those on the west side. In addition, no Tawny Fish Owls were recorded in the Coastal Mountain Range where rivers are short and small. Similarly, the owls were seldom found in streams on low mountains in the south of Taiwan. In contrast, streams in the north of Taiwan have regular water flow all year round, due to the north-east monsoon rainfalls (4,000 mm/year) (Chiu 2006), so the owls occurred further upstream, closer to the mountain ridge. The distribution of Tawny Fish Owls in Taiwan was also influenced greatly by habitat loss due to human activities. Destruction of lowland habitat is widespread, especially on the west side of the island, due to the conversion of natural forests to farmland and urbanisation over the last 200 years. The habitat selection analysis indicated that there were fewer owls in lowland areas, even though lowlands should have more suitable streams than the hills — wider, with more reliable flow, and greater prey abundance and diversity. Most Tawny Fish Owl territories to the east of the Central Mountain Range were below 500 m, but only two owl territories were found below that elevation on the west side. The lower altitudinal limit on the west side is likely to have retreated into higher, mountainous areas, as is the case for many forest birds and mammals on the island (Liu et al. 2003, Shiu 2003). It is likely that the loss of natural forests below 500 m has eradicated Tawny Fish Owls from the western lowlands, since riparian natural forests are the main habitat type used by fish owls (Hayashi 1997, Sun et al. 2000). Poaching is one of the most serious threats to the Tawny Fish Owl population (Severinghaus 1987) and at least 61 records of hunting have been reported since the species was legally protected in Taiwan in 1989 (Wang et al. 1995). Of the 28 owl records obtained from indigenous hunters, 54% were the result of poaching. Only a small proportion of indigenous hunters were interviewed, suggesting that poaching is likely to be widespread across the island. More than 60% of the owl territories found in this study were situated outside areas protected for wildlife (Hong 2007). The lowlands are largely unprotected and those on the developed western part of the island are already deforested, therefore we recommend the establishment of a protected area for the Tawny Fish Owl along lowland streams in north-east Taiwan, where the original riparian forests still remain nearly intact (Taiwan Forestry Table 2. The effects of altitude on territory selection by Tawny Fish Owls in Taiwan (%2= 41 .57, P < 0.0001). The Bonferroni confidence intervals show that the % use of each altitude band is higher (selected) or lower (avoided) than the % availability. Altitude band (m) N Use % Availability % Selection index Standardised index Bonferroni confidence intervals <500 35 0.3846 0.5276 0.7290 0.1307 0.2571 < Pi <0.51 21 500-1,000 30 0.3297 0.1542 2.1379 0.3835 0.2065< Pi <0.4529 1,000-1,500 20 0.2198 0.1071 2.0521 0.3681 0.11131,500 6 0.0659 0.2112 0.3122 0.0560 0.0009< ft' <0.1310 Forktail 29(2013) Spatial distribution of the Tawny Fish Owl Ketupa flavipes in Taiwan 51 Bureau 1995). This is possibly the only place in Taiwan where the Tawny Fish Owl range still reaches the coastline. This recommendation is in accordance with other studies that have recommended low-elevation protected areas for threatened species including Clouded Leopard Neofelis nebulosa , Leopard Cat Prionailurus bengalensis , Small Indian Civet Viverricula indica, Formosan Pangolin Manis pentadactyla and Fairy Pitta Pitta nympha (Liu et al. 2003, Chiang & Pei 2004, Lee et al. 2006). ACKNOWLEDGEMENTS We thank C. L. Bridgman and L. C. Lo for providing useful suggestions, and C.-F. Tsai for editorial help during the preparation of this paper. C. T. Yao and Y. J. Chen helped locate specimen records. We are also indebted to numerous birdwatchers and hunters who were willing to share with us their information on the Tawny Fish Owl. REFERENCES Chiang, P.-J.& Pei, K. J.-C. (2004) Present status and conservation of Formosan Clouded Leopard and other medium-to-large mammals at Tawu Nature Reserve and vicinities (III). Taipei: Taiwan Forestry Bureau. (In Chinese with English abstract.) Chiu, C.-A. (2006) Applying the ecoclimatic indices to predict the potential natural vegetation of Taiwan. PhD thesis, National Chung-Hsing University, Taichung, Taiwan. (In Chinese with English abstract.) Duncan, J. R. (2003) Owls of the world: their lives, behavior and survival. New York: Firefly Books. Fang, W.-H. (2005) A guide to threatened birds of Taiwan. Taipei: Owl Publishing House. (In Chinese with English abstract.) Flenley, J. R. (1994) Cloud forest, Massenerhebung effect and ultraviolet insolation. Pp.1 50-55 in L. S. Hamilton, J. O. Juvik & F. N. Scatena, eds. Tropical montane cloud forests. New York: Springer-Verlag. Fogden, M. (1973) Fish-owls, eagle owls and the Snowy Owl. Pp. 53-85 in J. A. Burton, ed. Owls of the world: their evolution, structure, and ecology. New York: A. W. Visual Library. Hayashi, Y. (1997) Home range, habitat use and natal dispersal of Blakiston's Fish-Owls. J. Raptor Research 31: 283-285. Hong, S.-Y. (2007) Distribution pattern of Tawny Fish Owls ( Ketupa flavipes) in Taiwan. MS thesis, National Pingtung University of Science and Technology, Pingtung, Taiwan. (In Chinese with English abstract.) Kuroda, N. (1916) Rare species of avifauna of Formosa . Zoological Magazine Tokyo 28: 263-264. (In Japanese.) Lee, P.-F., Bai, M.-L. & Lin, R.-S. (2006) Habitat preference and distribution prediction of vulnerable Fairy Pitta (Pitta nympha) in Taiwan by remote sensing and GIS. Taipei: Council of Agriculture. (In Chinese with English abstract.) Litvaitis, J. A., Titus, K. & Anderson, E. M. (1994) Measuring vertebrate use of terrestrial habitats and food. Pp. 254-274 inT. A. Bookhout, ed. Research and management techniques for wildlife and habitats. Fifth edition. Bethesda, MD: The Wildlife Society. Liu, C.-N., Liu, C.-H. & Chang, C.-H. (2003) Current condition and conservation of medium-to-large mammals at low altitude. Natural Conservation Quarterly 43: 61-66. (In Chinese.) Marks, J. S., Canning, R. J. & Mikkola, H. (1999) Family Strigidae (typical owls). Pp. 76-242 in J. del Hoyo, A. Elliott & J. Sargatal, eds. Handbook of the birds of the world, 5. Barcelona: Lynx Edicions. Severinghaus, L. L. (1987) The Tawny Fish Owl. Pp.354-355 in A.W. Diamond, L. L. Severinghaus & C. Chen, eds.Sove the birds. Frankfurt, Germany: Pro Nature. Shiu, H.-J. (2003) Spatial and seasonal variations in avian assemblages in Taiwan. PhD thesis, National Taiwan University, Taipei, Taiwan. (In Chinese with English Abstract.) Sun, Y.-H. (1996) Ecology and conservation of Tawny Fish Owl in Taiwan. PhD thesis, Texas A & M University, Texas, USA. Sun, Y.-H., Wang, Y. & Lee, C.-F. (2000) Habitat selection by Tawny Fish-Owls ( Ketupa flavipes) in Taiwan. J. Raptor Research 34: 102-107. Sun, Y.-H., Wu, H.-J. & Wang, Y. (2004) Tawny Fish-Owl predation at fish farms in Taiwan. J. Raptor Research 38: 326-333. Taiwan Forestry Bureau (1995) Thethird forest resource and land use inventory in Taiwan. Taipei: Taiwan Forest Bureau. (In Chinese.) Tzeng, C.-S. (1986) Distribution of the freshwater fishes of Taiwan. J. Taiwan Museum 39: 127-146. UNEP-WCMC (2009) UNEP-WCMC species database: CITES-listed species. Available online: http://www.unep-wcmc.org. Accessed 22 November, 2010. Voous, K. H. (1988) Owls of the Northern Hemisphere. London: Collins. Wang, H.-W. (2010) Ecoregion classification by using drainage fish community in Taiwan. MS thesis, National Kaohsiung Normal University, Kaohsiung, Taiwan. (In Chinese with English abstract.) Wang, Y., Sun, Y.-H. & Wu, H.-J. (1995) The distribution and use of the listed endangered birds by indigenous peoples and the ecology of Tawny Fish- Owl in Taiwan. Taipei: Taiwan Normal University. (In Chinese with English abstract.) Wu, H.-J., Sun, Y.-H., Wang, Y. &Tseng, Y.-S. (2006) Food habits ofTawny Fish- Owls in Sakatang Stream, Taiwan. J. Raptor Research 40: 111-119. Shiao-Yu HONG, Institute of Wildlife Conservation and Graduate Institute of Bioresources, National Pingtung University of Science and Technology, Pingtung 9(2, Taiwan Yuan-Hsun SUN, Institute of Wildlife Conservation, National Pingtung University of Science and Technology, Pingtung 912, Taiwan Hsin-Ju WU, Department of Life Science, National Taiwan Normal University, Taipei 1 17, Taiwan Chao-Chieh CHEN (Corresponding author) Department of Biomedical Science and Environmental Biology, Kaohsiung Medical University, Kaohsiung 807, Taiwan. Email: chen51 23@kmu.edu.tw or chenkmu@gmail.com FORKTAIL 29 (2013): 52-56 Population, breeding and threats to the White-rumped Vulture Gyps bengalensis in Bangladesh M.MONIRULH. KHAN The population of the White-rumped Vulture Gyps bengalensis in Bangladesh has declined very rapidly in recent years, so a research-cum- conservation project was launched in July 2008 that continued until June 201 2. Three species of vultures were found during the survey — White-rumped Vulture, Himalayan Vulture Gyps himalayensis and Cinereous Vulture Aegypius monachus. Based on nesting sites and frequent sightings of vultures, a total of six 'hotspots' were identified in the areas of Moulvibazar, Habiganj, Haor Basin, Mymensingh, Sundarbans (northern end) and Barisal.The total population of the White-rumped Vulture in suitable habitats across the country shows that numbers have drastically declined from 1,972 to 816 (nearly 60% drop) in four years. In two consecutive breeding seasons only 5 out of 32 and 8 out of 31 nests were successful in producing fledglings (one from each nest). The overall breeding success was very low (1 5.6-25.8%). The reason for such poor breeding success was sudden death or disappearance of parent birds, apparently due to poisoning by diclofenac, a veterinary drug used to treat livestock ailments. The project identified poisoning as the principal cause of vulture decline. Although the Government of Bangladesh banned use of veterinary diclofenac from 25 October 2010, 53% of the veterinary drug stores still sell it illegally. Awareness campaigns have made people aware of vulture conservation and the adverse effects of diclofenac. INTRODUCTION Historically, seven species of vultures — White-rumped Gyps bengalensis , Himalayan G. himalayensis , Griffon G.fulvus , Slender- billed G. tenuirostris , Cinereous Aegypius monachus, Red-headed Sarcogyps calvus and Egyptian Neophron percnopterus — have been recorded in Bangladesh (Khan 2008, Siddiqui et al. 2008). However, only the White-rumped Vulture, Himalayan Vulture and Cinereous Vulture have been seen in the last four years and only the White-rumped Vulture is now known to breed in Bangladesh. Populations of the White-rumped Vulture and other resident Gyps vulture species have declined very rapidly since the mid- 1 990s across the Indian subcontinent (Prakash 1999, Gilbert et al. 2006, Prakash et al. 2007, Chaudhary et al. 2012). Declines in numbers of the White-rumped Vulture have exceeded 99.9% in India (Prakash et al. 2007) and the species is classified as Critically Endangered (BirdLife International 2001, BirdLife International 2012). If the rate of decline cannot be arrested, the species will disappear from the Indian subcontinent in the next few years. This species not only plays a key role as a scavenger but is also part of the heritage of the Bengal region. The people of Bangladesh are not hostile to vultures, but they have been almost totally unaware of the threats and dire situation that vultures are facing. A decade ago the White-rumped Vulture was a common and widely distributed bird in Bangladesh (Harvey 1990, Thompson &Johnson 1996). Recent studies in India, Nepal and Pakistan confirm that vultures are poisoned when they feed on the carcasses of cattle treated with the veterinary drug diclofenac shortly before their death (Green et al. 2004, Oaks et al. 2004, Shultz et al. 2004). This is also likely to be the main cause of the decline in the vulture population in Bangladesh, but there may be other factors contributing to the decline. Because the White-rumped Vulture is a globally and nationally threatened species (BirdLife International 2001, BirdLife International 2012) it was necessary to take measures to save it from local extinction. Here the findings of a research-cum-conservation project focusing on this species, started in 2008, are reported. The aim was to understand the conservation status of the White- rumped Vulture in Bangladesh and to implement ways of reducing population decline. The specific objectives were to estimate the relative abundance of populations in different parts of the country, assess the population trend, identify important vulture ‘hotspots’ (where they nest and are frequently sighted), record breeding success and assess the threats (focusing mainly on the availability and use of veterinary diclofenac). Moreover, vulture conservation awareness programmes were conducted to spread knowledge of the dire situation faced by the vulture population and discourage the use of diclofenac to treat cattle. METHODS Study area The project was implemented in different parts of the country, but focused on Greater Sylhet (north-east), Greater Khulna (south¬ west), Greater Mymensingh (north) and Greater Barisal (south) Figure 1. Bangladesh showing hotspots (shaded black) for the White- rumped Vulture Gyps bengalensis. Forktail 29 (2013) Population, breeding and threats to the White-rumped Vulture Gyps bengalensis in Bangladesh 53 where the White-rumped Vulture was known to nest and roost regularly (Figure 1 ). Geographically, Bangladesh is located between 20.567°-26.550°N and 88.017°-92.683°E. The total area of the country is 147,570 km2, with a population of around 160 million people. The climate is tropical monsoon, characterised by marked seasonal variations. Abundant rainfall during the monsoon (July- October) is followed by a cool winter period (November-February), and then a hot and dry summer (March-June). Bangladesh can be divided into three main physiographic divisions — Tertiary hills, Pleistocene terraces and recent plains (Khan 2008). Field methods A small team of researchers carried out the surveys, but local people were also involved in most areas. Five members of the research team had been trained in different aspects of wildlife biology. The project team worked closely with villagers and labourers on tea plantations, since the vultures mainly occur and breed in and around villages and tea estates. Journalists, veterinarians and other professional people were also involved with various activities. The project team liaised closely with the Bangladesh Forest Department. Work was carried out from July 2008 to June 2012 (hence 2008-2009 is the period between July 2008 and June 2009, so that each breeding season falls in one slot), but not all the activities were carried out each year of the project. Some data were collected during 2005- 2008 to confirm the relative abundance of vultures in different areas and find the hotspots. Surveys to assess relative abundance and identify vulture hotspots Between 2005 and 2008, vulture sightings by 50 local people were recorded every year in each of the seven administrative divisions (Dhaka, Chittagong, Sylhet, Rajshahi, Rangpur, Khulna and Barisal) so that division-wise the relative abundance of vulture populations could be estimated. In order to get the best output from the limited resources available for awareness and conservation activities, hotspots were identified on the basis of the occurrence of nesting sites and frequent sightings (where interviewees have seen vultures at least once in every two-month period). Population surveys To determine the population trend, the team selected potential survey areas throughout the country and systematically visited the known roosting sites in the morning and in the afternoon to count White-rumped Vultures in roosting colonies. Sightings of other vulture species were recorded during these surveys. The survey team interviewed local people about their vulture sightings and, in areas where interviewees claimed recent sightings of roosting vultures, the survey team stayed and counted the vultures when they returned to roost. In areas where the interviewees said that there were no recent sightings at roosting colonies, the survey team quickly moved on to new areas. The same roosts visited during 2008-2009 were visited again in 2009-20 10,2010-2011 and 2011-2012. Since the count was conducted in all the potential sites for vultures, although not in each and every part of the country, the annual counts can be considered as total counts for the country or a very close representation of it. More importantly, however, the count has been repeated in a standardised way so that the data are comparable year on year. Monitoring breeding success The breeding success of vultures was documented by periodically (at least once every two weeks) observing every known nest during the breeding season (dry season: October to March). Telescopes and binoculars were used for these observations. A commercially made camouflaged hide was often used so that the nesting vultures were not disturbed by the presence of observers. The main information recorded was the fate of nests — if the nestling from a particular nest flew (fledged), the nest was treated as successful. Information on nesting trees (species, nest height from the ground) was also recorded. Surveys for diclofenac The availability of veterinary diclofenac and other non-steroidal anti¬ inflammatory drugs (NSAIDs) was assessed in two ways. First, undercover surveys (posing as buyers) of the veterinary drug stores throughout the country were undertaken (70 drugstores every year), from 2008-2009 to 20 1 1 -20 1 2. Second, the use of diclofenac was recorded by interviewing local cattle-owners (86 individuals in total) in different areas of the country. This was done openly (since the cattle-owners were not as secretive as the veterinary drug salesmen) by using. a standard questionnaire that also included a question on what the cattle-owners do with dead cattle. RESULTS Based on sightings of the White-rumped Vulture during 2005- 2008 surveys by local people in suitable habitats of the seven administrative divisions, the highest relative abundance was found to be in Sylhet (5.1 sightings/interviewee/year) and the lowest in Chittagong (0.7 sightings/interviewee/year) (Table 1). This gives an average for the country of 2. 8 sightings/interviewee/year. Based on nesting sites and frequency of sightings, a total of six hotspots were identified: the areas of Moulvibazar, Habiganj, Haor Basin, Mymensingh, Sundarbans (northern end) and Barisal (Figure 1). Table 1. Relative abundance of the White-rumped Vulture in suitable areas of seven divisions of Bangladesh. Division Total no. of sightings during 2005-2008 by 50 interviewees Average no. of sightings/ interviewee/year Dhaka 645 3.2 Chittagong 141 0.7 Sylhet 1,022 5.1 Rajshahi 443 2.2 Rangpur 420 2.1 Khulna 640 3.2 Barisal 466 2.3 Survey data from 2008-2009 to 201 1-2012 demonstrate that the White-rumped Vulture declined by nearly 60% over the four- year study period throughout the country, and that other species of vultures are extremely rare. The total population of the White- rumped Vulture in suitable habitats across the country (which can be treated as the total Bangladesh population) shows that the population drastically declined from an estimated 1,972 in 2008- 2009, 1456 in 2009-2010, 991 in 2010-2011 and 816 in 2011- 2012. Other than the White-rumped Vulture there were 16 sightings of Himalayan Vulture from Moulvibazar (north-east), Sunamganj (north-east), Habiganj (north-east), Jamalpur (north), Sirajganj (west) and Rangpur (north-west). All sightings were between October and March indicating that they were winter visitors. Most of the sightings were small flocks, but the largest roosting flock of 21 birds was sighted in a tall Silk Cotton Salmalia sp. tree at Dawrachara Tea Estate, Moulvibazar, on 3 March 20 1 2. Himalayan Vultures were seen roosting, soaring, or feeding with White-rumped Vultures six times. There were only four sightings of Cinereous Vulture from Rajshahi (north-west), Madhupur Tract (central), Narshingdi (north-east) and Moulvibazar (north-east). All were solitary birds, including two juveniles, sighted at different seasons, indicating that they were probably vagrant individuals. 54 M. MONIRUL H. KHAN Forktail 29 (2013) Table 2. Breeding status of the White-rumped Vulture in seven divisions of Bangladesh Division Number of nests observed Season-1 Season-2 Unsuccessful (nesting birds died/ vanished) Season-1 Season-2 Status of breeding Successful (nestling left nest) Season-1 Season-2 Breeding success (%) Season-1 Season-2 Dhaka 8 6 7 5 1 1 12.5 16.7 Chittagong 0 0 - - - - - - Sylhet 14 13 12 10 2 3 14.3 23.1 Rajshahi 0 0 - - - - - - Rangpur 0 0 - - - - - - Khulna 10 9 8 7 2 2 20.0 22.2 Barisal 0 3 - 1 - 2 - 66.7 Total/Overall for Bangladesh 32 31 27 23 5 8 15.6 25.8 NB. Season 1 (October 2009-March 2010) and Season 2 (October 2010-March 2011) were the two consecutive breeding seasons of vultures. Breeding success of the White-rumped Vulture for all nests in all the known breeding areas in Bangladesh was documented. In Season 1 (October 2009 to March 2010), a total of 32 nests were observed of which only five birds from five nests successfully fledged (clutch size is one), giving an overall breeding success of 1 5.6% (Table 2). In Season 2 (October 2010 to March 2011), a total of 31 nests were observed from which eight birds fledged (25.8% success rate). At unsuccessful nests the parent birds were found dead, either on the ground near the nest (n = 1 1) or on/beside the nest (n = 4), or the parent birds just vanished suddenly (n = 35), indicating that they had probably died elsewhere. Although no post-mortems were undertaken, the dead vultures were apparently in good health, indicating sudden death that can be caused by poisoning. It was observed that, whether the nest had an egg or nestling, one parent almost always attended the nest to guard against crows (House Crow Corvus splendens , Large-billed Crow C. macrorbynchos ) and other raptors (Pallas’s Fish Eagle Haliaeetus leucoryphus , Steppe Eagle Aquila nipalensis). The parents take shifts so that both can feed and bring food for the nestling. Vulture nests were found in the following trees -.Albizia lebbeck , Albizia procera, Antbocephalus chinensis , Bombax ceiba, Borassus flabellifer. Cocos nucifera , Ficus benghalensis , Ficus religiosa, Mangifera indica, Sivietenia mabagoni. Vultures had no preference for nesting tree species, but all nesting trees were large. Nests were constructed at heights ranging from 7 to 17 m. The veterinary drug stores were surveyed for four years, from 2008-2009 to 201 1-2012, and it was found that the availability of diclofenac in stores has decreased from 100% to 53% and the availability of three other NSAIDs (meloxicam, ketoprofen and sodium salicylate) had increased (Figure 2). This is in response to the ban on production of veterinary diclofenac from 25 October 2010 and the banning of the use of veterinary diclofenac from six Figure 2. Availability of different NSAIDs in veterinary drug stores in Bangladesh. N.B. Otherthan NSAIDs, paracetamol and dexamethasone were available, which are used for the same purposes. 120 a 00 3 $5 100 ■a >■ )— fZ c | 80 60 QJ HQ rz C <3J 0.05). In RDT, the Reed Parrotbill was not recorded after May 2007 when the reeds died back, but it was recorded in almost all months in RST (Figure 3), indicating that presence of the Common Reed is a necessary precondition for survival of the Reed Parrotbill. DISCUSSION The use of reed vegetation by Reed Parrotbills is reported frequently in published literature. These reports, combined with the observations from Chongming Island, indicate that the species uses reedbeds or reed-dominated vegetation as habitat throughout its Chinese distribution. Detailed local observations at Chongming Island and Chongxi Wetland Research Centre also indicated that the species was almost exclusively associated with reeds and that birds were not found in nearby areas without reeds. It was concluded that the Reed Parrotbill is a reed-dominated habitat specialist. The species’s distribution map for China (Figure 2) includes many new records. In the south, the range extends to the southern shore of Hangzhou Bay but the northern extent cannot be determined yet, as there are records of Reed Parrotbills in south¬ east Russia. This study revealed that the species has a larger range than previously thought. Further changes to the known range are anticipated, as research and hireling activities increase. At Chongxi Wetland Research Centre, when reed shoots disappeared from mixed vegetation, Reed Parrotbills were rarely recorded in the reed-free vegetation. They were unable to persist in these areas by utilising more distant patches oi reeds and disappeared along with the reeds. This implies that the species is dependent on reed vegetation and this dependence on reeds might constrain their ability to use other vegetation in the absence oi reeds. There are published accounts of Reed Parrotbills using non-reed vegetation close to large areas of reeds (Su et al. 1987, Hou et al. 1997a, Zhao et al. 2004, Han et al. 2007). The species has been observed using bushes and woodland close to reeds when the reed-vegetation was disturbed by, for example, reed harvesting (La Touche 1906, Wang & Tian 1988). In the absence oi disturbance, birds occasionally visit nearby non-reed vegetation, for example Spartina, but the number oi individuals and their density were much lower than in reeds (Dong etal. 2010, Gan et al. 2010). Birds evidently disperse readily into non-reed vegetation near reeds, but it is not known whether they use resources within the non-reed vegetation or make only transitory visits. Birds visiting Spartina close to reeds only spent very short periods of time there (Dong et al. 2010). Records of the species in non-reed vegetation indicate that the degree of habitat specialisation is not fully understood. It is Figure 3. Monthly variation of Reed Parrotbill density (±SE) in reedbeds, reed patches, reeds with dense trees and reeds with sparse trees during 2006 to 2008, at Chongxi Wetland Research Centre. Forktail 29 (2013) Habitat specialisation in the Reed Parrotbill Porodoxornis heudei 69 important to understand the relationships between the species and the reed vegetation. Studies have shown that Reed Parrotbills feed on insects, insect eggs and larvae on reed shoots year round, in a tidal marsh in Changjiang Estuary (Xiong etal. 2007, 2010). The species may have special morphological adaptations in its bill, which facilitate breaking reed stems to retrieve insects within (Xiong et al. 2010) but which compromise its ability to use food resources in other vegetation. Similar extreme specialisation is seen amongst the bamboo-specialist insectivores, which feed on insects in, on and around living bamboo (Cockle et al. 2009). Although the Reed Parrotbill is limited to reed vegetation and it feeds on insects in and on reeds (Xiong et al. 2007), it is not yet certain that their prey is also restricted to reed vegetation. It is not known whether the two recognised subspecies differ in habitat use or other life history characteristics. Subspecies beudei occurs on Chongming Island, where the fieldwork was carried out. At sites where polivanovi is found, such as Zhalong National Nature Reserve (site 3 in Figure 2), Longfeng Wetland Nature Reserve (4) and Xingkaihu National Nature Reserve (6), the Reed Parrotbill habitat was described as reed vegetation or reed marsh. Thus, it seems likely that the two subspecies have the same habitat requirements. CONCLUSIONS This review of Reed Parrotbill distribution in China revealed that its range and the number of locations where it occurs are larger than previously thought. This does not indicate an improvement in the conservation status of the Reed Parrotbill, as the study also confirmed the species’s strict habitat specialisation. Given that existence of Common Reeds is a precondition for Reed Parrotbills to survive, more attention must be paid to the conservation of reed- dominated habitat, such as coastal wetlands, lakeside wetland and marshes, and corridors in reed-dominated habitats should be designed and maintained to reduce the effects of habitat fragmentation. Corridors might also be useful to link areas of reedbeds without Reed Parrotbill populations to nearby locations that are already populated and this could be used in the selection and development of protected areas. It would be useful to learn more about how the Reed Parrotbill uses and has adapted to reed vegetation, to help understand the evolutionary history of the species and likely threat mechanisms. The distribution of Reed Parrotbills may be predicted based on its strong relationship to the Common Reed. Large areas of reed vegetation close to or linked by corridors to reed vegetation with Reed Parrotbill populations could be potential habitat. ACKNOWLEDGEMENTS We thank Jia Liu, Lingdong Li, Guolang Zhang for sharing their birding records. We thank the staff at Chongxi Wetland Research Centre for assistance in the field. We especially thank two anonymous reviewers and the editors for their valuable comments. REFERENCES Ai, D., Wang, X. M., Jia, J. L. & Chang, Y. H. (2001 ) Status and conservation of animal species biodiversity in Honghe Natural Reserve. Territory and Natural Resources Study 1: 54-56. (In Chinese with English abstract.) Bai, L. & Bai, Y. K. (1993) Records of Reed Parrotbill in Shandong province. Chinese Journal of Zoology 28: 44. (In Chinese.) Bi rd Life International (2013) Species factsheet: Paradoxornis heudei. Downloaded from hhttp://www.birdlife.org on 5/08/2013. Cheng, T. H. ( 1 987) A synopsis of the avifauna of China. Beijing: Science Press. Cockle, K., Bodrati, A. & Areta, J. i. (2009) Bamboo specialist birds of the Atlantic forest. In E. Hirschfeld ed. Rare birds yearbook. Cambridge UK: BirdLife International. Cofre, H.L., Bohning-Gaese, K.& Marquet, P. A. (2007) Rarity in Chilean forest birds: which ecological and life-history traits matter? Diversity and Distributions 1 3: 203-212. David, A. (1872) Sur une espece nouvelle de Paradoxornis. Comptes Rendus de TAcademie des Sciences. Serie III, Sciences de la Vie, IXXIV: 1 449-1450. Dong, B„ Wu, D., Song, G. X., Xie, Y. M., Pei, E. L. & Wang,T. H. (2010) Research on the habitat-selection of Reed Parrotbill (Paradoxornis heudei) during the winter in Chongming Dongtan, Shanghai. Acta Ecologica Sinica 30: 4351-4358. (In Chinese with English abstract.) Feng, Z. J., Xu, Q. F„ Wang, G. B., Zhao, Y. Y. & Zhou, H. (2006) Bird resource in wetland of Luoma Lake and its conservation .Sichuan Journal of Zoology 25: 564-569. (In Chinese with English abstract.) Futuyma, D.J.& Moreno, G. (1988) The evolution of ecological specialization. Annual Review of Ecology, Evolution, and Systematics 1 9: 207-233. Gan, X. J., Choi, C. Y„ Wang, Y„ Ma, Z. J., Chen, J. K. & Li, B. (2010) Alteration of habitat structure and food resources by invasive smooth cordgrass affects habitat use by wintering saltmarsh birds at Chongming Dongtan of east China. Auk 1 27: 317-327. Gan, X., Cai, Y„ Choi, C., Ma, Z., Chen, J. & Li, B. (2009) Potential impacts of invasive Spartina alterniflora on spring bird communities at Chongming Dongtan, a Chinese wetland of international importance. Estuarine, Coastal and Shelf Science 83: 211-218. Gao, Q. H. (2003) The research of the diversity, population dynamics and breeding ecology of the dominant species of wetland birds in the Plengshui lake. Shijiazhuang: Masters thesis, Hebei Normal University. (In Chinese with English abstract.) Gee, N. G. & Moffett, L. I. (1 91 7) A key to the birds of the Lower Yangtse Valley, with popular descriptions of the species commonly seen. Shanghai: Shanghai Mercury Limited Press. Han, J. G„ Lu, Y. M„ Li, H. K„ Liu, Z. J., Dong, L. W„ Ma, Q. & Wu, Y. H. (2007) A survey of bird species in Hengshui Lake National Nature Reserve. Journal of Fujian Forestry Science and Technology 34: 144-150. (In Chinese with English abstract.) Han, L. P„ Fu, X. Y„ Huo, Y. S„ Li, S. & Liang, C. S. (2011) Research of bird resource inTanghai wetland and bird natural reserve. Journal of Hebei Forestry Science and Technology 5: 20-27. (In Chinese with English abstract.) Hou, J. H„ Wu, M. L„ Li, J. T„ Hu, Y. F. & Zhang, X. Z. ( 1 997a) New records of birds from Hebei Province — Reed Parrotbill and Himalayan Griffon. Chinese Journal of Zoology 32:26. (In Chinese.) Hou, J. H„ Wu, M. L„ Li, J. T„ Hu, Y. F. & Zhang, X. Z. (1997b) A research of avian fauna in Saihanba, Hebei. Chinese Journal of Zoology 32: 20-25. (In Chinese.) Huang, H. M., Zhang, L. Q. & Gao, Z. G. (2005) The vegetation resource at the intertidal zone in Shanghai using remote sensing. Acta Ecologica Sinica 25:2686-2693. Ji, T. (2007) Biodiversity and ecological planning of Hongze Lake National Nature Reserve. Nanjing: Masters thesis, Nanjing Forestry University. (In Chinese with English abstract.) Jin, L. K., Zhang, Y. W. & Liang, Y. (1991 ) New record of birds from Liaoning province — Reed Parrotbill. Journal of Liaoning Forestry Science and Technology 6: 28. (In Chinese.) Julliard, R„ Clavel, J., Devictor, V., Jiguet, F. & Couvet, D. (2006) Spatial segregation of specialists and generalists in bird communities. Ecology Letters 9: 1237-1244. Kong, W. Y. (2006) The relationship between spring avian community and environment during the initial stages of water compensation in Zhalong. Ha'erbin: Masters thesis, Northeast Forestry University. (In Chinese with English abstract.) Krauss, J., Steffan-Dewenter, I. & Tscharntke, T. (2003) Local species immigration, extinction, and turnover of butterflies in relation to habitat area and habitat isolation. Oecologia 137:591-602. 70 LI-HU XIONG &JIAN-JIAN LU Forktail 29 (2013) La Touche, J. D. D. (1906) Field-notes on the birds of Chinkiang, Lower Yangtse Basin - Part I. Ibis 48: 427-450. La Touche, J. D. D. (1925) A handbook of the birds of eastern China (Chihle, Shantung, Kiangsu, Anhwei, Kiangsi, Chekiang, Fohkien, and Kwangtung Provinces). London: Taylor & Francis. Lei, F. M. & Lu, T. C. (2006) China endemic birds. Beijing: Science Press. Li, F„ Lu, C. FI., Yang, H. J. & Yang, S. Z. (1998) Community diversity about breeding birds in reed marsh of Zhalong Natural Reserve. Journal of Northeast Forestry University 26: 68-72. (In Chinese with English abstract.) Lin, J., Li, Y„ Zhang, E. & Wu, X. S. (2008) Reed Parrotbill. China Nature 2: Il¬ ls. (In Chinese.) Liu, Z. G„ Wang, W. & Wang.T. FI. (2009) A preliminary study on introduced woody engineering plants in Chongxi tidal wetlands. Forestry Science and Technology 34: 1-5. (In Chinese with English abstract.) Lu, J. J. (1990) Wetland in China. Shanghai: East China Normal University Press. (In Chinese.) Lynes, FI. (1914) Some notes on the habits and distribution of Paradoxoris heudei David. Ibis 10:177-185. Ma, S. Q. (1988) Ecological research on the population of Reed Parrotbill Paradoxornis heudei. Zoological Research 9: 217-224. (In Chinese with English abstract.) Ma, S. Q. & Sun, Z. Q. (1988) Reed Parrotbills in coast of Shanghai. Chinese Wildlife 4: 46-47. (In Chinese.) Ma, Z., Gan, X., Choi, C., Jing, K., Tang, S., Li, B. & Chen, J. (2007) Wintering bird communities in newly-formed wetland in the Yangtze River estuary. Ecological Research 22: 1 1 5-1 24. MacKinnon, J. & Phillipps, K. (2000) A field guide to the birds of China. Oxford: Oxford University Press. Niu, J. Y. (2007) Study on avian diversity and dynamics change in Plenan Yellow River National Nature Reserve. Shijiazhuang: Masters thesis, Flebei Normal University. (In Chinese with English abstract.) Rank, M. (1989) Eye-browed Parrotbill. Oriental Bird Club Bull. 1 0: 22-27 . Shaw, Z. FI. (1934) Notes on the birds of Cheking .Bulletin of the Fan Memorial Institute of Biology (Zoology) 5: 285-338. Shi, Z. R. & Cui, G. FI. (1989) Ecological observation of Reed Parrotbills. China Nature 2: 1 2-1 3. (In Chinese.) Sol, D„ Timmermans, S. & Lefevre, L. (2002) Behavioural flexibility and invasion success in birds. Animal Behaviour 63: 495-502. Styan, F. W. (1891 ) On the birds of the lower Yangtse Basin -Part I. Ibis 33: 316-359. Su, L. Y„ Ma, J. Z., Xu, J., Jiang, X. X. & Wu, C. S. (1987) Preliminary study on bird fauna of Zhalong nature reserve. Journal of Northeast Forestry University 1 5: 62-73. (In Chinese with English abstract.) Tang, J. (2007) Winter and spring avian communities in southern Flongze Lake wetland. Nanjing: Masters thesis, Nanjing Forestry University. (In Chinese with English abstract.) Wang, F. Q„ Yuang, L. & Liu, L. S. (2002) New records of birds from Tianjin. Animal Science and Veterinary Medicine 19: 12-13. (In Chinese.) Wang, S., Wang, Y. L. & Jian, X. (201 1) A new bird record of Chinese Crowtit Paradoxornis heudei in Anhui Province. Journal of Anhui Science and Technology University 25: 16-18. (In Chinese with English abstract.) Wang, Z. Y. (1990) A trial in feeding and training of Reed Parrotbill. Chinese Journal of Zoology 25: 34-35. (In Chinese.) Wang, Z. Y. & Tian, G. W. (1988) Reed Parrotbill in Lianyungang. Chinese Journal of Zoology 23: 39-40. (In Chinese.) Wang, Z. Y. & Zhou, Y. S. (1988) Habits and breeding of Reed Parrotbills in Lianyungang. Chinese Journal of Zoology 9: 216. (In Chinese.) Xiong, L. FI., Wu, X., Gao, W„ Zhou, J. & Lu, J. J. (2007) Impact of reed cutting on foraging of Reed Parrotbill Paradoxornis heudei. Chinese Journal of Zoology 42: 41-47. (In Chinese with English abstract.) Xiong, L. FI. & Lu, J. J. (2008) Seasonality and flock size affect response of the Reed Parrotbill Paradoxornis heudei to human approach. Acta Zoologica Sinica 54: 622-629. (In Chinese with English abstract.) Xiong, L. FI., & Lu, J. J. (2013) Exploitation of reedbeds by specialist passerines: Reed Parrotbill and Oriental Reed Warbler. Wilson Journal of Ornithology 125: 165-173. Xiong, L. FI., Wu, X. & Lu, J. J. (2010) Bird predation on concealed insects in a reed-dominated estuarine tidal marsh. Wetlands 30: 1 203-121 1. Xu, L., Li, B„ Yuan, X. & Xu, FI. F. (2006) The characteristics of the avian community in Chongming Dongtan spring 2003. Chinese Journal of Zoology 41: 120-126. (In Chinese with English abstract.) Zhai, S. J„ Hu, W. P. & Qian, Y. (2008) Evaluation of the ecological service function of the Hongze Lake wetland nature reserve in Sihong, Jiangsu. Journal of Ecology and Rural Environment 24: 24-28. (In Chinese with English abstract.) Zhang, Y. W. (2004) Features of avian community and their relationships with habitats in Nandagong wetland. Shijiazhuang: Masters thesis, Hebei Normal University. (In Chinese with English abstract.) Zhao, Q., Zhao, Q. L. & Deng, Z. H. (2004) Species diversity change of birds inhabiting during autumn and winter at island Xinglongsha (Qidong, Jiangsu) in the ten year period. Chinese Journal of Zoology 39: 63-68. (In Chinese with English abstract.) Zheng, G. M. (2005) A checklist on the classification and distribution of the birds of China. Beijing: Science Press. Zhou, L. Z. (2010) A new bird record of Chinese Crowtit Paradoxornis heudei heudei in Anhui Province, China. Journal of Anhui University 34: 91-92. (In Chinese with English abstract.) Zhu, S. Y„ Lu, J. Z., Yu, H. L., Shan, K„ Zhang, A. F„ Wang, Y. Z„ Du, X. H„ Dong, X. H., Wang, L. D. & Song, S. W. (2001 ) Research on the distribution and amount of Reed Parrotbill Paradoxornis heudei in Yellow River Delta Nature Reserve, Shandong. Shandong Forestry Science and Technology 5: 34-35. (In Chinese.) Zou, S. C. & Qin, D. R. (1989) Investigation for the birds near Xuzhou city. Journal of Xuzhou Teachers College 7: 46-63. (In Chinese with English abstract.) Li-Hu XIONG, Zhejiang Institute of Hydraulics & Estuary, 50E Fengqi Rd, Hangzhou, Zhejiang Province 310020, China. Email: lihuxiong@126.com Jian-Jian LU, Chongxi Wetland Research Centre, The State Key Laboratory of Estuarine and Coastal Research, East China Normal University, 3663N Zhongshan Rd, Shanghai 200062, China. FORKTAIL 29 (2013): 71-77 Increasing variation in population size and species composition ratio in mixed-species heron colonies in Japan MIYUKI MASHIKO& YUKIHIKOTOQUENAGA Long-term population dynamics of colonial herons and egrets are well documented in Europe and the USA, but not in East and South-East Asia. Here the population dynamics of mixed-species colonies from 2002 to 201 1 in Ibaraki prefecture, east Kanto, Japan, are reported. From censuses based on a combination of aerial and ground surveys, the number of breeding colonies was found to vary from 1 5 to 20. The population dynamics of Great Egret Casmerodius albus and Black-crowned Night Heron Nycticorax nycticorax remained relatively constant, while Grey Heron Ardea cinerea and Intermediate Egret Mesophoyx intermedia increased, but Little Egret Egretta garzetta and Cattle Egret Bubulcus ibis decreased. When data for the six species were combined, the sum of their populations was almost constant, but variation increased in colony size, species composition ratio and the number of years that individual colonies existed. The population of colonies typically ranged from 200 to 2,000 individuals up to 2004, but smaller (under 200 individuals) and larger (over 2,000 individuals) colonies appeared after 2006. Increased variation in the number of consecutive years colonies existed was closely related to increased variation in colony size. Increased variation in species composition ratios in colonies was not a by-product of the increased variation in colony size, and the occurrence of colonies dominated by Grey Heron, Intermediate Egret or Black-crowned Night Heron after 2006 played an important role in the structural changes of mixed-species colonies. INTRODUCTION H erons and egrets (Ardeidae) are commonly found in aquatic habitats worldwide (Kushlan & Hafner 2000). In Europe, long¬ term population trends of such species have been well investigated, and some factors that explain how and why population sizes fluctuate at regional level have been revealed: cold winters (Stafford 1971, Reynolds 1979, Hafner & Fasola 1997, Fasola et at. 2010), rainfall (McKilligan 2001), water level (Griill & Ranner 1998), habitat conditions (Tourenq etal. 2000, 2004), aquaculture (Fleury & Sherry 1995) and human disturbance (Fasola et al. 2010). In East and South-East Asia, long-term records of breeding populations of colonial nesting herons and egrets only exist in Hong Kong and Vietnam (Kushlan & Hafner 2000, Wong & Young 2006). Fack of local information makes it difficult to assess the current status of these birds. In Japan Grey Heron Ardea cinerea. Great Egret Casmerodius albus, Intermediate Egret Mesophoyx intermedia. Little Egret Egretta garzetta , Cattle Egret Bubulcus ibis and Black-crowned N ight Heron Nycticorax nycticorax breed in mixed-species colonies. Nationwide research was carried out in 1980 and 1992 (Research Division of the Wild Bird Society of Japan 1981, Environmental Agency ofjapan 1994), and it was reported that single- and mixed- species colonies were distributed throughout Japan’s lowlands. Although there are many observations of colonies in various areas, long-term local population trends have only been reported by Narusue (1992) and Matsunaga et al. (2000). Narusue (1992) argued that both the populations and the average colony size of these species declined from the 1 940s to 1992 in Saitama prefecture, west Kanto Plain, due to loss of foraging areas and use of agricultural chemicals. Changes in the irrigation of rice fields from shallow earth ditches to deep concrete -walled channels and the decline in aquatic prey caused the decline of Intermediate Egret (Narusue & Uchida 1993, Lane & Fujioka 1998), the commonest egret until the 1960s, but now categorised as ‘near threatened’ (Ministry of the Environment 2002). In contrast, a long-term study of Grey Herons in Hokkaido by Matsunga et al (2000) reported an increase in population and the number of colonies. There are currently no other reliable data to assess population trends of herons and egrets in Japan. The Environmental Agency ofjapan (1994) showed that in Ibaraki prefecture, east Kanto, Japan, both the average colony size and the population of Intermediate Egret were large compared to elsewhere in Japan, and this suggested that data from this area could provide important information for future assessment of populations of colonial breeding herons and egrets in Japan and other parts of Asia. In this study, colony censuses were carried out from 2002 to 2011 in Ibaraki prefecture to investigate trends in these populations using a combination of aerial and ground surveys. The changes are discussed here with reference to the trends in population dynamics of each species, changes in the nesting vegetation and the number of consecutive years that colonies existed. METHODS Study area The study focused on Ibaraki prefecture and parts of Tochigi and Chiba prefectures in Honshu, central Japan (35.783°-36.767°N 139.767°-140.683°E) (Figure 1). The area is in the east Kanto Plain, near Lake Kasumigaura, and includes six major rivers: Kuji, Naka, Sakura, Kokai, Kinu and Tone. The north is mountainous, but the predominant land use in other areas is farming, with large areas along the rivers being used for rice production. There are also lotus fields near Lake Kasumigaura, areas of lowland forest and human habitations. Japan started a national project to consolidate rice production in 1 963; this included extending irrigation ditches, improving service roads, and enlarging fields to facilitate mechanised farming equipment. It was largely complete by 1980 (Himiyama & Kikuchi 2007), but continued in part ol the study area into the last decade, being 78% complete by 2010 in Ibaraki prefecture. The climate of the region is moderate with an annual average air temperature of 14.0 ± 0.1°C and an annual precipitation of 1,388.2 ± 54.3 mm. Despite a small annual decrease in rice cultivation, neither climate nor land use showed obvious changes during the study period (Figure 2). The herons and egrets breed from March to August, but there is considerable variation from species to species (Figure 3). The Grey Heron arrives first in March, Great Egret, Little Egret, and Black-crowned Night Heron arrive in April; these species are residents and wanderers, and some individuals winter in this area. Finally, the migrant species arrive. Intermediate Egret in late April and Cattle Egret by early May (A. Abe in litt. 200 6). Usually Grey 72 MIYUKI MASHIKO& YUKIHIKO TOQUENAGA Forktail 29 (2013) Figure 1. Locations of colonies from 2002 to 201 1 . Grey regions show an altitude greater than 1 00 m where the distribution of egrets is lower. Dots enclosed by a circle are considered to be historically identical colonies. Exceptionally, there are two cases in which the nearest- neighbour distance is shorter than 6.47 km; ( 1 ) l20 — L30: because L was newly established in 2010 and consisted of Grey Herons Ardea cinerea and Great Cormorants Phalacrocorax carbo, we assumed L was different from I. (2) Q40-R41 42: Q was newly established in 2006, whereas Koshida (2007) reported that R has existed since 1984. It is difficult to accept that Q and R are one colony. N Figure 2. Changes in a. mean temperature, b. annual rainfall, and c. areas of six land-use types from 2002 to 201 1 in Ibaraki prefecture, which was the main region of the study area (6,096 km2). 'Paddy field' includes both rice paddies and lotus fields, and 'others' includes parks, golf courses, and uncultivated fields. Data were downloaded 8 June 2013 from http: //www. data.jma.go.jp/obd/ stats/etrn/index.php for climate and http: //www. pref.ibaraki.jp/bukyoku/kikaku/mizuto/ ibarakinotochi/25/ibarakinotochi.htm for land use. Year Heron and Black-crowned Night Heron are nocturnal, but during the breeding season they are also active during the day. Colonies were located in bamboo thickets, trees or a mixture of both. Bamboo thickets were composed of Moso Bamboo Phyllostachys pubescens, Simon Bamboo Pleioblastus simonii or Dwarf Bamboo P. chino. Coniferous tree sites consisted mainly of Japanese Red Pine Pinus densiflora, Japanese Cedar Cryptomeria japonica and Japanese Cypress Cbamaecyparis obtusa-, broadleaftree sites were mainly Japanese Zelkova Zelkova serrata, Japanese Oak Qiiercus serrata and Yoshino Cherry Prunus x yedoensis. Census of colonies Colonies have been recorded in the area over the last 25 years (Koshida 2007) and have high site fidelity (Custer et al. 1980, Frederick et al. 1996); between March and early May, 93 ± 0.02% of the colony sites were found by checking the places where colonies had been located in previous years. When a colony was abandoned, checks were made to determine whether other colonies had formed nearby. Local literature and personal communication were used to locate colonies that had not been found during the authors’ own field work. Site vegetation was recorded by identifying bamboo or tree species holding at least one nest. In small colonies of fewer than about 50 nests, if all the nests were visible from outside or within the site, the nests of each species were counted directly from the ground, and the breeding population estimated by doubling the number of nests counted. In most cases, ground-based counts were impossible due to colony size, the impenetrable nature of dense bamboo thickets, or other vegetation, such as tall trees, that made nests invisible. Hence, counts were made using a combination of aerial and ground surveys, following the method of Fujioka et al. (2001). For aerial surveys, a small ‘Sky Surfer’ radio-controlled paraglider was used (Green Corporation, Japan, Plate la). This equipment is quieter than fixed-wing aircraft or helicopters commonly used for bird colony censuses (Kushlan 1979, Rodgers et al. 2005), and very suitable in this case where more than 80% of the colonies were close to residential areas. Aerial photographs of each colony were taken at an altitude of 30- 50 m just before sunrise (about 04h00) when most birds were in the colony. Photography was started in mid-May after arrival of Cattle Egrets, and was continued until early July when distinguishing between growing nestlings and adults became difficult (Figure 3). Aerial photographs were taken once during that period at each site. All individuals of the four light-coloured species (Great Egret, Intermediate Egret, Little Egret, and Cattle Egret) in the images were counted (Plate lb). For large colonies, several photographs were used to obtain a complete composite image of the colony. Figure 3. Breeding period of each species in the study area showing the timetable of aerial surveying and species composition counts. White, grey, and black shading show arrival and nest building, incubation, and chick-rearing periods, respectively. The parallelogram shapes indicate the variation in individual breeding periods. After breeding is over, some birds continue to roost in the colonies but all disperse by October. Mar. Apr. May Jun. Jul. Aug. Grey Heron Black-crowned Night Heron^- Great Egret, Little Egret s' Intermediate Egr^^ZI Cattle Egret ^^s' Aerial photography Species composition counting GREEN CORPORATION Forktail 29 (2013) Variation in population size and species composition ratio in mixed-species heron colonies in Japan 73 Plate 1. a. Sky Surfer in flight. b. An example of aerial photographs that show one part of a composite photograph. Because it was not feasible to identify the light-coloured species only from the photographs, and because the two dark-coloured species (Grey Heron and Black-crowned Night Heron) were difficult to count in the photographs, species composition ratios — the proportion of each species within a colony — were estimated using data from ground surveys. The number of individuals of all species going in and out of each colony for a period of 30 minutes in the daytime were counted using binoculars. It was very difficult to identify to species level white egrets flying in and out at the same time, so the viewing range of each observer was restricted by setting a common range of observation, approximately a 30° field of vision. In the case of large colonies, surveys were carried out from two or three different directions. Ground surveys were made once or twice in June, the peak chick-rearing period (Figure 3), when all species engage in frequent foraging flights. Thirty minutes is much shorter than the typical duration of one foraging flight and it was assumed that each individual counted, whether arriving or departing, was observed only once during the observation period and therefore the observed proportion of each species reflects the species composition ratio of the colony. The total estimated number of individuals in the colony (T = colony size) was calculated using t = -- where A is the actual number of light-coloured individuals counted from aerial photographs, and x is the sum of the proportion of Grey Herons and Black-crowned Night Herons obtained from the ground survey. The estimated population size of each species in the colony was obtained by multiplying the colony size by the proportion of each species. Data analyses To determine the number of colonies each year, the number of observed colony sites was first counted. But the number of sites itself was not taken to be the number of colonies because a few colony sites were very close to each other despite foraging ranges having radii of about 10 km, and sometimes over 20 km (Nabeya 2011). An earlier study showed that heron colonies are evenly distributed to avoid overlap of foraging sites (Gibbs et al. 1987). Consequently if colony sites were located close together, they were grouped together and counted as a single colony because their foraging areas overlapped substantially. To determine which sites should be counted as a single colony, the half of the mean nearest- neighbour distance (ND) of observed colony sites for each year was used. If more than one colony site was located within the ND, colony censuses were carried out at each site, the data were combined and it was counted as a single colony. Colonial birds have been found to have high site fidelity (Custer et al. 1980, Frederick et al. 1996); every year some colonies returned to the same locations as the previous year, some birds were abandoned and new ones were established. To obtain the number of consecutive years (NCY) each colony existed, the number of years from first establishment at the location was counted. The movement of a colony was also considered and when abandonment and new establishment occurred in neighbouring locations in successive years, the new site was assumed to be a descendant of the abandoned one, e.g. abandonment was sometimes caused by vegetation loss through natural causes or felling and the colony was often re-established nearby. The ND was used to determine a reasonable displacement distance of a colony and it was assumed that each colony had a domain of attraction of half the average distance between the next nearest sites. Thus, colonies consecutively established at the same site or at a different site within a radius of the ND were counted as a single colony. Koshida (2007) was used as data source of the NCY of colonies established before 2002; consequently the NCY ranged from 1 to 36 years rather than being limited to the period of this study. The population of each species was calculated annually using the mean population size per colony rather than summing the population sizes for all colonies with census data. This approach was used because aerial and ground surveys produced only partial data due to practical difficulties — problems in taking aerial photographs and/or delays in detecting colony sites. The simple sum of colony population sizes would have been inappropriate because it is an increasing function of the number of colonies with census data. The percentage of colonies surveyed increased from 78% in 2002-2004 to 94% in 2006-201 1. (Data from 2005 were excluded because aerial and ground surveys were limited to only 5 out of 1 8 colonies.) Thus, the overall total population of the target M. MASHIKO 74 MIYUKI MASHIKO& YUKIHIKOTOQUENAGA Forktail 29(2013) species reflect the mean colony sizes rather than the total number of individual birds in the study area. To evaluate difference in colony size, species composition ratio, and NCY among colonies, the coefficient of variance (CV) for each year was calculated. For species composition ratio, the proportional similarity index (Whittaker 1952) was calculated for each colony every year as Lif,'0~f)-> where p, is the proportion of species i in one colony and p~t is the mean proportion of that species in all colonies surveyed in that year. The index ranges between zero and unity: zero means completely different and unity means completely equal. Then the CV of proportional similarity of the species composition ratio was obtained for each year. Ten years is too short for ordinal time series analyses, so randomisation tests were done to assess temporal trends in the number of colonies, population sizes of each species, sum of the population sizes of the six species, colony sizes and CVs of three variables (colony size, species composition ratio and NCY). In a randomisation test, the linear regression coefficient ((3) of a target variable based on the original data was obtained first. Next the data were shuffled 30,000 times and compared beta with the linear regression coefficients (|3’s) of the shuffled data to obtain one-sided P-values to assess whether the target variable was increasing or decreasing. Sensitivity analyses of the population of each species against the three CVs (colony size, species composition ratio and NCY) were performed. Generalised linear models specifying population sizes of the species as explanatory variables and CVs as dependent variables were constructed, using Gaussian distribution with an identity link function for all model fitting. The most suitable models based on Akaike’s information criterion values were chosen and the coefficients of explanatory variables of the models as sensitivity against dependent variables were considered. If the CVs of colony size and species composition ratio show parallel changes, there is a possibility that the variation in species composition ratios increased as a by-product of the increase of variation in colony sizes. To examine this possibility, a randomisation test was performed to determine whether the variation in species composition ratios was solely caused by a sampling bias according to the variation in colony size. First a hypothetical total number of herons that consisted of the six species was prepared. The species composition ratio of the whole number of herons was arbitrary. Next multiple colonies with equal colony sizes from the total number of herons were sampled. Then proportional similarities of species composition ratios of these hypothetical colonies against the species composition ratio of the whole number of herons were calculated. Proportional similarities for hypothetical colonies of the same number but with different colony sizes were also calculated. Finally, the variance of the proportional similarities between equal and unequal size colonies were compared, and the probability that proportional similarities of unequal size colonies were larger than or equal to those of equal size colonies with 10,000 iterations was obtained. To evaluate changes in nesting vegetation, the Friedman test was used to analyse whether the vegetation of colony sites changed from year to year. Nesting vegetation consisted of one or a mixture of the following three types: bamboo thickets, coniferous trees and broadleaf trees; there were seven types in total. Finally, a randomisation test was performed to determine whether there was a positive correlation between NCY and colony size among colonies by reshuffling the year record so as to randomise the consecutive colony-size dynamics of each colony. All statistical analyses were conducted using R ver. 2.13.0 (R Development Core Team 2011). Data are presented as mean ± SE throughout. The randomisation test on the relationship between the variation of population sizes and that of species proportion ratios was also conducted with R. All R scripts for the above statistical analyses are available from the authors. RESULTS During the 10-year period, there was an average of 19 colony sites in the study area every year (19.10 ± 0.72 colony sites, n = 10); cumulatively 191 colony sites were used over the 10 years. Some colonies were in the same locations for more than one year, and a total ol 62 colony sites were used (1 to 62 in Figure 1). Colony sites were separated by an average of 13 km (mean ND over 10 years 12.95 ± S.39 km, n — 191), so the ND was defined as within 6.47 km. Hence, these 62 colony sites were categorised into 27 colonies (A to a in Figure 1) because colony sites consecutively established at different locations within a 6.47 km radius were considered a single colony. Six of 27 colonies were made up of two or three colony sites in at least one breeding season, and the median distance between them was 1.44 km (range: 0.32-5.12 km, 10 combinations of colony sites in all). Finally, the annual number of colonies increased gradually ((3 = 0.382, P = 0.006) from 15 to 20 (Figure 4a). In the case of Intermediate Egret and Black-crowned Night Heron, the average population per colony was relatively large (about 300 individuals) and these species remained dominant throughout the 10-year period (Figure 4b). Conversely, it was small (about 50 individuals) for Great Egret and Grey Heron, and intermediate (about 100 individuals) for Cattle Egret and Little Egret. The sum of the population of the six species (mean colony size) ranged from 726 to 966 individuals and remained almost constant ((3 = -4.301, P = 0.342). The population trends of each species varied (Figure 4b). Grey Heron and Intermediate Egret increased (Grey Heron: (3 = 9.575, P < 0.001; Intermediate Egret: (3 = 9-519, P = 0.033), whilst Little Egret and Cattle Egret decreased steadily (Little Egret: (3 = -2.069, P = 0.002; Cattle Egret: (3 = -20.672, P < 0.001). The Black- crowned Night Heron population fluctuated over the years but remained almost constant ([3 = 9.3 1 1 , P = 0. 145). The Great Egret population was small but almost constant ((3 = 0.036, P = 0.492). Colonies were very variable in size, and the CV of colony size continuously increased (Figure 4c) over the ten years ((3 = 7.510, P < 0.001). Colonies ranged from 200 to 2,000 individuals until 2004, while smaller (under 200 with minimum 8 individuals) and larger (over 2,000 with maximum 3,280 individuals) colonies appeared after 2006. Between 2008 and 20 1 1, the smaller and larger colonies increased from 33% to 41% of colonies surveyed. In parallel with the increase in the CV of colony size, the CV of proportional similarity of species composition ratios increased (Figure 4c), especially after 2006 ((3 = 7.002, P < 0.001). Until 2004, most colonies consisted of five species (Great Egret, Intermediate Egret, Little Egret, Cattle Egret and Black-crowned Night Heron), and the composition ratio was similar among surveyed colonies (mean proportional similarity = 0.86 ± 0.02, n = 43). Grey Heron bred in only three, six and seven colonies in 2002, 2003 and 2004, respectively. Until 2004, the composition ratios of the Intermediate Egret and the Cattle Egret were higher than those of other species in half of the surveyed colonies in accordance with their large population (Figure 4b), but no species became dominant (over 50% of the composition ratio). After 2006, 37% of all surveyed colonies were dominated by the Grey Heron, Intermediate Egret or Black-crowned Night Heron, and differences in the species composition ratios among colonies increased. The CV of the NCY also increased gradually (Figure 4c) ((3 = 2.453, P < 0.001). Eight colonies persisted between 2002 and 2011; the remainder were abandoned or newly established. Every year 1-4 colonies were abandoned and 0-3 were established. Considering the period prior to this study, 14 out of 27 colonies had existed before 2002 and 4 had persisted for over 25 years. Table 1 shows the results of the sensitivity analyses of population of each target species against three CVs. Increase in the Forktail 29 (2013) Variation in population size and species composition ratio in mixed-species heron colonies in Japan 75 Figure 4. a. Changes in the number of colonies between 2002 and 2011. b. Changes in population of each species per colony of and the sum of the six species. c. Changes in the coefficient of variation (CV) of colony sizes, number of consecutive years (NCY), and species composition ratios. For the changes in population size and CVs of colony sizes and species composition ratios, the year 2005 is not shown because aerial and ground surveys were limited to only 5 out of 18 colonies. Year CV of colony sizes was explained by the increasing Grey Heron population, and the increase in the CV of species composition ratios was also explained by the increasing Grey Heron population, and marginally explained by the increasing Intermediate Egret population. The increase in the CV of NCY was explained by the increasing Grey Heron population and the fluctuating, though statistically constant overall, population trend of the Black- crowned Night Heron. The randomisation test to determine whether the variation in species composition ratios increased as a by-product of the increase in variation of colony sizes did not reveal a significant result: the probability that the proportional similarities between unequal size colonies would be larger than or equal to those of equal size colonies was almost even (0.538). The increase of variation in species composition ratios could not solely be caused by the increase of variation in colony sizes. Changes in vegetation of the colonies were significant over the years (%: = 25.2, df = 6, P < 0.001) (Figure 5). While the vegetation in most colonies included bamboo until 2004, after 2008 more than half the colonies were located in trees. The slope obtained by a linear regression analysis of colony sizes against NCY (43.05 ± 5.43) was significantly larger (P < 0.001) than slopes obtained by the randomisation test where the year record was shuffled for each colony so as to randomise consecutive colony-size dynamics (Figure 6). This randomisation test indicates that there was a positive correlation between colony sizes and the NCY for the colonies. Table 1. Sensitivity of population sizes against CVs. [5s are coefficients of the best fit generalised linear model with Gaussian distribution and identity link function. R2 = (null deviance - residual deviance)/(null deviance). CS: colony size, SCR: species composition ratio, NCY: number of consecutive years. CV of CS P P CV of SCR P P CVofNCY P P Grey Heron 0.351 0.054 0.449 0.009 0.307 0.016 Great Egret Intermediate Egret 0.168 0.077 0.047 0.193 Little Egret -0.133 0.169 0.124 0.101 Cattle Egret -0.119 0.179 Black-crowned Night Heron 0.081 0.116 0.072 0.020 Akaike's information criterion 56.329 65.833 59.157 R2 0.975 0.920 0.902 Figure 5. Changes in colony vegetation. See description of study area for details of species. Bamboo tree 2002 2003 2004 2005 2006 2007 2008 Year Figure 6. Relationship between colony size and longevity. Each dot represents a colony censused in a particular year (n = 141). The regression line was obtained by a linear regression analysis of colony sizes against longevity assuming that each annual colony was established independently. 76 MIYUKI MASHIKO& YUKIHIKOTOQUENAGA Forktail 29 (201 3) DISCUSSION The survey in Ibaraki prefecture from 2002 to 20 1 1 indicated that the number of breeding colonies (average 19) increased slightly and mean colony size was almost constant. These results accord well with the report by the Environmental Agency of Japan (1994): there were 20 colonies in 1992 in Ibaraki prefecture, ranging in size from 15 to 2,990 individuals (CV = 1 12.5), and the population of these species has been relatively constant in the area for at least two decades to 2011. However, variations in size, species composition ratio and NCY among colonies increased significantly. Colony vegetation changed from predominantly bamboo thickets to tall trees. Trends in population dynamics differed: Grey Heron and Intermediate Egret increased. Little Egret and Cattle Egret decreased; and Great Egrets and Black-crowned Night Herons were relatively constant — the population of both the latter species were similar to previous reports (Research Division of the Wild Bird Society of Japan 1981, Environmental Agency of Japan 1994) and unchanged for three decades. Overall, there was no significant change in population of these colonial species in the study area during the decade, but variation in the structure of colonies and population dynamics clearly increased. In contrast to Great Egret and Black-crowned Night Heron, the population of the other four species changed during the period (Figure 4b). Grey Heron showed the greatest population growth, which is in line with earlier reports that its population is growing in other parts of Japan (Narusue 1992, Environmental Agency of Japan 1994, Matsunaga et al. 2000, Sasaki 2001). In Hokkaido, Matsunaga et al. (2000) suggested that recent climatic warming and increase in aquaculture have provided the species with additional food resources. It is not known whether the increase of this species in other more temperate parts of Japan also depends on these factors, but its ability to respond quickly to changes in food availability (Adams & Mitchell 1995) would be expected to boost populations. The other increasing species, Intermediate Egret, was a predominant species in this area even though it has been designated as a ‘near threatened’ species in Japan (Ministry of the Environment 2002). Owing to the lack of current data from other parts of Japan, it is not clear whether the population has been recovering, but the abundant population in this area may be of conservation significance in Japan; monitoring of this species should continue. Little Egret and Cattle Egret both showed a steady decline over the period; the Environmental Agency ofjapan ( 1994) considered them to be predominant and numerous throughout Japan, including Ibaraki prefecture in 1992, and the population of both has decreased during the last two decades. Although mild winter weather contributed to their increase in France (Hafner & Fasola 1997) and rainfall drove the changes in Cattle Egret population in Australia (McKilligan 2001) and Hong Kong (Wong & Young 2006), climatic variables are unrelated to the decrease of these species in the study area because both temperature and rainfall have been almost constant (Figure 2). It seems likely that changes in food resources or foraging habitats may be contributory factors. In northern Japan, Shimada et al. (2005) suggested that Little Egrets might be strongly affected by the increase in population of the introduced Black Bass Micropterus psalmoides, which has caused a decrease of the smaller native fish species they prefer. In the absence of historical and quantitative data in Japan, monitoring studies in other regions are needed to make a complete assessment of population dynamics of these declining species. During the study period, variations in size, species composition ratio and NCY increased (Figure 4c). Since these temporal trends showed parallel changes, there is a possibility that the variation in species composition ratios increased as a by-product of the increased variation in colony sizes; but a randomisation test contradicted this possibility, and it was concluded that the observed increased variation in proportional similarities of species composition ratio could not be solely caused by the increased variation in colony size. Another change that coincided with the study period was the change in nesting vegetation; the majority of colonies changed from bamboo thickets to trees (Figure 5). More colonies were newly established in tall trees even though bamboo thickets persisted in the area. The decrease in the number of colonies in bamboo may be due to the increase in Grey Herons because they prefer to nest near the top of tall trees. However, those results contradict the general knowledge that the target species use a wide range of nest sites, including trees, bushes, reeds and on the ground. No other species shows a particular preference for specific substrates (Kushlan & Hancock 2005). Hence, there is no strong support for the possibility that the vegetation of established colony sites affected the size or species composition ratios of colonies. Increasing variation in NCY may help explain the increased variation in colony size and species composition ratios. These results showed that the variation in the NCY among colonies grew from year to year (Figure 4c), and there was a significant positive correlation between colony size and the NCY that a colony existed (Figure 6). Although food availability, measured as the area of potential foraging habitat around the colony, has often been thought to be the most important Factor affecting colony size (Fasola & Barbieri 1978, Gibbs et al. 1987, Gibbs 1991, Baxter & Fairweather 1998), previous studies in this locality showed that variables related to foraging sites (areas around ponds, rivers, paddy fields and lotus fields) did not have a major impact on colony size; instead the NCY had a significant positive relationship with colony sizes (Fujioka et al. 2001, Tohyama 2005). Increasing variation in a colony’s size is therefore closely related to the colony’s longevity. As for increasing variation in species composition ratio among colonies, variations were due to the occurrence of colonies dominated by Grey Heron, Intermediate Egret, or Black-crowned Night Heron after 2006. In particular, the dominance of Grey Heron was notablet in small, recently established colonies (Figure 6). It is well known that the Grey Heron often breeds in small colonies of only 2-10 nests, while the other five species are more gregarious and usually breed in large mixed-species colonies (Kushlan & Hancock 2005). Thus, the Grey Heron population growth after 2007 might contribute significantly to the increasing variation in the species composition ratio and colony size despite its relatively small overall population (Table 1). Overall, the local population of herons and egrets in eastern Japan seems to have remained constant for at least the last decade, in parallel with the constant climate and land use variables. Nonetheless, population dynamics of constituent species have been changing, and variations in colony sizes and species composition ratios have also increased. Such changes are revealed only by long¬ term and comprehensive colony census. Continuing studies are required not only to reveal the factors affecting the population dynamics of each species at a regional level, but also to establish a better understanding of relationships between each species’s population and the sizes or composition ratios of mixed-species colonies. ACKNOWLEDGEMENTS We thank T. Tohyama, T. Yamaguchi, R. Yamagishi, C. Koshida, H. Takeda, A. Abe, K. Nabeya and L. T. Carrasco for colony censuses, and S. Ikeno, M. Seido and K. Takeda for information on the location of colonies. We especially thank Y. Asano and M. Kobayashi for their mechanical advice and maintenance of the Sky Surfer, and K. Ohashi and M. Fujioka for helpful comments on an early draft. We also thank the Kashima Kyodo Shisetsu and the Kubota Tsukuba Plant for access to their premises to make colony censuses. Forktail 29 (2013) Variation in population size and species composition ratio in mixed-species heron colonies in Japan 77 We are grateful to our anonymous reviewers for useful comments. This study was partly funded by Global Environment Research Fund from the Japanese Ministry of the Environment and by a Grant-In-Aid for Scientific Survey from the Japanese Ministry ofEducation, Science and Culture (No. 13740433 and 19570014). REFERENCES Adams, C. E. & Mitchell, J. (1995) The response of a Grey Heron Ardea cinerea breeding colony to rapid change in prey species. Bird Study 42: 44-49. Baxter, G. S. & Fairweather, P. G. (1998) Does available foraging area, location or colony character control the size of multispecies egret colonies? Wildlife Research 25: 23-32. Custer, T. W„ Osborn, R. G. & Stout, W. F. (1980) Distribution, species abundance, and nesting-site use of Atlantic coast colonies of herons and their allies. Auk 97: 591-600. Environmental Agency of Japan (1994) Distribution and population status of colonies and communal roosts of 22 bird species from 1990 to 1992. Tokyo: Wild Bird Society of Japan & the Environmental Agency of Japan. (In Japanese.) Fasola, M.& Barbieri, F. (1 978) Factors affecting the distribution of heronries in northern Italy. Ibis 1 20: 537-540. Fasola, M„ Rubolini, D., Merli, E„ Boncompagni, E. & Bressan, U. (2010) Long¬ term trends of heron and egret populations in Italy, and the effects of climate, human-induced mortality, and habitat on population dynamics. Population Ecology 52: 59-72. Fleury, B. E. & Sherry, T. W. (1995) Long-term population trends of colonial wading birds in the southern United States: the impact of crayfish aquaculture on Louisiana populations. Auk 1 12: 613-632. Frederick, P. C., Towles, T„ Sawicki, R. J. & Bancroft, T. (1996) Comparison of aerial and ground techniques for discovery and census of wading bird (Ciconiiformes) nesting colonies. Condor 98: 837-841 . Fujioka, M., Yoshida, H. & Toquenaga, Y. (2001 ) Research on the dynamic phase of biodiversity in geographic scale and its preservation. (3). Analysis of the dynamic phase wildlife population in geographic scale. (2). Research on the spatiotemporal dynamic phase of bird gathering places. Pp. 75-88 in Global Environment Research Fund: analysis and conservation of biodiversity on a geographical scale. Tokyo: Ministry of the Environment. (In Japanese.) Gibbs, J. P. (1991) Spatial relationships between nesting colonies and foraging areas of Great Blue Herons. Auk 108: 764-770. Gibbs, J. P., Woodward, S„ Hunter, M. L. & Hutchinson, A. E. (1987) Determinants of Great Blue Heron colony distribution in coastal Maine. Auk 104: 38-47. Grull, A. & Ranner, A. (1 998) Populations of the Great Egret and Purple Heron in relation to ecological factors in the reed belt of the Neusiedler See. Colonial Waterbirds 21 : 328-334. Hafner, H.& Fasola, M. (1997) Long-term monitoring and conservation of herons in France and Italy. Colonial Waterbirds 20: 298-305. Himiyama, Y. & Kikuchi, Y. (2007) Agricultural field improvement projects in Japan since 1 980. Reports of the Taisetsuzan Institute of Science 41 : 9-1 8. (In Japanese with English summary.) Koshida, C. (2007) Colony vicissitudes of herons and egrets around southern Ibaraki Prefecture. Unpublished M.Env.Sc. thesis. University ofTsukuba. (In Japanese.) Kush Ian, J. A. (1979) Effects of helicopter censuses on wading birds colonies. J. Wildlife Management A3: 756-760. Kushlan, J. A. & Hafner, H. (2000) Heron conservation. London: Academic Press. Kushlan, J. A. & Hancock, J. A. (2005) The herons. New York: Oxford University Press. Lane, S. J. & Fujioka, M. (1998) The impact of changes in irrigation practices on the distribution of foraging egrets and herons (Ardeidae) in the rice fields of central Japan. Biological Conservation 83: 221-230. Matsunaga, K., Matsuda, A. & Fukuda, H. (2000) Changing trends in distribution and status of Grey Heron colonies in Hokkaido, Japan, 1960-1999 . Japanese J. of Ornithology 49: 9-16. McKilligan, N. (2001) Population dynamics of the Cattle Egret (Ardea ibis) in south-east Queensland: a 20-year study. Emu 101: 1-5. Ministry of the Environment (2002) Threatened wildlife of Japan - Red Data Book. Second edition, Vol. 2. Aves. Tokyo: Japan Wildlife Research Center. (In Japanese.) Nabeya, K. (2011) Egrets do not give up their foraging site after colony extinction. Unpublished M.Sc. thesis, University ofTsukuba. Narusue, M. (1992) Changes in the distribution and extent of breeding colonies of egrets in Saitama Prefecture. Strix 1 1 : 189-209. (In Japanese with English summary.) Narusue, M. & Uchida, H. (1993) The effect of structural changes of paddy fields on foraging egrets. Strix 1 2: 121-1 30. (In Japanese with English summary.) R Development Core Team (201 1) R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Research Division of the Wild Bird Society of Japan (1981) Research for colonies and roosts of herons and egrets by questionnaire. Tokyo: Wild Bird Society of Japan. (In Japanese.) Reynolds, C. M. (1979) The heronries census: 1 972-1 977 population changes and a review. Bird Study 26: 7-12. Rodgers, J. A. J., Kubilis, P. S. & Nesbitt, S. A. (2005) Accuracy of aerial surveys of waterbird colonies. Waterbirds 28: 230-237. Sasaki, M. (2001 ) Distribution of breeding colonies of herons and egrets and their protection in Kyoto Prefecture. 5fr/x 19: 149-160. (In Japanese with English summary.) Shimada.T., Shindo, K„ Takahashi, K. & Bowman, A. (2005) The effects of the increase of Largemouth Bass on a wetland bird community through change of the fish community. Strix 23: 39-50. (In Japanese with English summary.) Stafford, J. (1971) The heron population of England and Wales, 1928-1970. Bird Study 18:218-221. Tohyama, T. (2005) Analysis of factors affecting size and species composition of multispecies heronries. Unpublished M.Sc. thesis, University of Tsukuba. Tourenq, C., Benhamou, S., Sadoul, N„ Sandoz, A., Mesleard, F„ Martin, J.-L. & Hafner, H. (2004) Spatial relationships between tree-nesting heron colonies and rice fields in the Camargue, France. Auk 121:1 92-202. Tourenq, C., Bennetts, R. E., Sadoul, N., Mesleard, F., Kayser, Y. & Hafner, H. (2000) Long-term population and colony patterns of four species of tree¬ nesting herons in the Camargue, South France. Waterbirds 23: 236-245. Whittaker, R. H. (1952) A study of summer foliage insect communities in the Great Smoky Mountains. Ecological Monographs 22: 1-44. Wong, L. C. & Young, L. (2006) Nest numbers of five ardeids in Hong Kong, South China, 1 989-2004: does weather affect the trend? Waterbirds 29: 61-68. Miyuki MASHIKO and Yukihiko TOQUENAGA, Graduate School of Life and Environmental Sciences, University of Tsukuba, Ibaraki 305-8572, Japan. Email: mashiko@pe.ies.Hfe.tsukuba.ac.jp FORKTAIL 29 (2013): 78-87 Phenotypic evidence for the specific and generic validity of Heteroglaux blewitti P. C. RASMUSSEN & N. J. COLLAR The genus Heteroglaux was established for the Forest Owlet H. blewitti when the species was first described, but owing to certain similarities with Spotted Owlet Athene brama, the use of Heteroglaux fell into disuse in the twentieth century until the species was rediscovered in 1 997, and is still not universal; moreover, perceptions appear to linger that blewitti might even be conspecific with brama owing to a recent claim of interbreeding. In reality blewitti is distinct from brama on external morphology (plumage described elsewhere; narial position related to bill width; bill height; more heavily feathered toes; length of middle and hind claws; wing formula) and osteology, in which blewitti is distinct from all three species of Athene (multiple cranial elements, especially the greatly widened and inflated frontal, and the extremely stout tarsometatarsus). Lateral tail-flicking and direct, non-undulating flight further support generic separation. INTRODUCTION For well over a century the Forest Owlet Heteroglaux \Athene ] blewitti remained as much a taxonomic as a conservation enigma. It was discovered in central India 140 years ago, and at once placed confidently in its own genus Heteroglaux by Flume (1873); At first sight it would certainly be classed as an Athene-, but the head is much smaller [ sic. evidently a lapsus for ‘larger’] than in any of the Athene's I possess, viz., brama, radiata, malabarica, cuculoides, castaneonota. The nostrils are not pierced from the front, backwards at the margin of a swollen cere, but are well inside the margin, and are pierced straight in. The upper surfaces of the toes, too, are not covered with bristles, but thickly feathered. In the decade that followed, only six further specimens were taken (enumerated in Rasmussen & Collar 1998), and then the species disappeared. Perhaps as a consequence of this, its generic placement by Hume was never widely accepted. Although Heteroglaux continued to be used by Hume himself and some contemporaries (e.g. Hume 1879, Murray 1887, Sharpe 1891, 1899), from an early stage the Forest Owlet was also treated as congeneric with the Little Owl Athene noctua and Spotted Owlet A. brama, first in the genus Carine — even as early as Ball (1878), writing in Hume’s own Stray Feathers — and more recently Athene. Although Gurney (1894) retained it in Heteroglaux, he cited Hume’s (1873) view that it looks much like A. brama and added that Blanford considered it to belong to Athene-, and the following year Blanford’s (1895) treatment as such appeared. Dubois (1904) retained it in Heteroglaux, albeit without comment, but within a few decades virtually all works treated blewitti as an Athene (e.g. Baker 1934, Peters 1940, Biswas 1953, Ripley 1961, Abdulali 1972, Marshall & King 1988). Without explanation, Wolters (1975) assigned both blewitti and brama to the subgenus Heteroglaux, with noctua and Burrowing Owl A. ( Speotyto ) cunicularia occupying separate subgenera within Athene-, otherwise Heteroglaux has only ever been used for blewitti. Voous (1989: 191) suggested that blewitti ‘might provide a clue to understanding these relationships [between Athene, Ninox and Glaucidium], though the Forest Owlet may already be too close to the Spotted Owlet for that purpose’. Recent molecular phylogenies of owls (e.g. Wink etal. 2004, 2009) have not included blewitti. Indeed, some considered the similarity of brama and blewitti so great as to render them conspecific. Baker (1923) treated blewitti as well as most subspecies of brama — all except, inexplicably (perhaps as a lapsus), A. brama tarayensis of the north-western areas of the subcontinent — as races of A. noctua. This view, although not elsewhere accepted, may explain the listing in the NHMUK specimen register of the Davidson specimen (NHMUK 1925.12.23.958) that was stolen and remade by Richard Meinertzhagen (Rasmussen & Collar 1999) as ‘ Carine noctua blewetti' [sic], although the other blewitti in the same accession (then NHMUK 1925.12.23.1, now MCZ 236630) was listed as Athene blewitti. The fictitious locality of the stolen specimen, which came to the (now) Natural History Museum, Tring, UK, in the late 1960s in Meinertzhagen’s posthumous bequest (Rasmussen & Collar 1999), may in turn explain why it took until 1997 before the Forest Owlet was seen in the twentieth century (King & Rasmussen 1998). Over much of the intervening period, however, in the absence of clear diagnostic illustrations and texts, the species was speculated or judged to be so close in appearance to A. brama that it would be difficult and perhaps impossible to distinguish it (Ripley 1976, Ali 1978, Ali & Ripley 1981), and consequently the few reports or claims of blewitti that appeared in the interim were shown upon scrutiny to be brama (Rasmussen & Collar 1998). Following the rediscovery of the species there has been a degree of conservation-oriented research focusing on its distribution and ecology (Jathar & Rahmani 2002, 2004, Rahmani & Jathar 2004, Ishtiaq & Rahmani 2005, Kasambe et al. 2005, Mehta et al. 2008, Chavan & Rithe 2009, Yosef et al. 2010). However, one aspect of its resurrection has remained unexplored: the issue of its generic identity. Publications at the time of the rediscovery and in its immediate aftermath mentioned both Athene and Heteroglaux in their titles (King & Rasmussen 1998, Rasmussen & Collar 1998, 1999, Rasmussen & Ishtiaq 1999). The two major monographic treatments of owls that appeared at this time (del Hoyo etal. 1999, Konig et al. 1999), both of which cited the preceding references, elected to retain the species in Athene, although the latter entered a caveat that its tail-flicking habit ‘argues against a close relationship with other Athene owls and suggests closer affinity with pygmy owls [Glaucidium)’ , and consequently proposed ‘placing this species in the subgenus Heteroglaux’ . In the twenty-first century the trend has clearly been towards accepting Heteroglaux as a valid monotypic genus. Collar et al. (2001: 1775) remarked that despite the species’s ‘strong superficial resemblance’ toAthene'its original placement in its own genus appears well justified based on osteological evidence (Rasmussen & Collar in prep.) and on recent behavioural observations including flight pattern and song (Rasmussen & Ishtiaq 1999)’. Thereafter, world lists (Dickinson 2003, Gill & Wright 2006), Indian avifaunas (Rasmussen & Anderton 2005, Forktail 29 (2013) Phenotypic evidence for the specific and generic validity of Heteroglaux blewitti 79 Grimmett etal. 2011), one monograph (Mikkola 2012) and many journal papers and reports (preceding paragraph) have used Heteroglaux. Nevertheless, some sources have retained Athene (e.g. Clements 2007, Konigetrt/. 2008, Yosef etal. 2010), one even with the cryptic entry 'Remarks: Spurious use of the generic name Heteroglaux' (Weick 2006). This is perhaps unsurprising given that a clear case for the acceptance of this genus has never been made, and the osteological evidence referred to above never published. Here we seek to rectify these deficiencies. This need is rendered all the more pressing following a recent report (Pande et al. 2011), albeit rejected (Ishtiaq 201 1, Jathar & Patil 2011), of hybrid Forest Owlets x Spotted Owlets. For this reason, we also consider the extensive structural differences between blewitti and brama beyond the plumage distinctions established in Rasmussen & Collar (1998). However, the exercise further requires the osteological analysis to extend beyond differences between these two species to cover not only all members of Athene but key representatives of other related owl genera (including Surnia, Glaucidium , Xenoglaux , Micrathene , Athene, Aegolius and Ninox: Ford 1967, del Hoyo etal. 1999). METHODS We considered two types of evidence: external morphology (focusing on the differences between blewitti and brama) and osteology (considering the differences between blewitti and Athene, thence to other genera) . Plumage comparisons between blewitti and brama have previously been presented in Rasmussen & Collar ( 1 998), and we therefore here restrict our comparisons of external morphology to mensural characters. We also briefly review data reported elsewhere for acoustics and behaviour. External morphology of blewitti and brama For the external morphological analysis we assembled for examination at the Natural History Museum, UK (NHMUK) all known specimens of blewitti (seven; four males, three females), and used the opportunity to compare them with other owl species, most importantly Athene brama, with which blewitti is ostensibly so closely allied as to have been considered conspecific (as noted above). We measured all specimens of blewitti (data in Rasmussen & Collar 1998) and specimens of brama at NHMUK; American Museum of Natural History, New York (AMNH); Academy of Natural Sciences of Philadelphia (ANSP); Museum of Comparative Zoology, Harvard University (MCZ); National Museum of Natural History, Smithsonian Institution, Washington, DC (USNM); University of Michigan Museum of Zoology, Ann Arbor (UMMZ); and Zoological Survey of India, Calcutta (ZSI). This sample includes numerous representatives of each of the races of brama recognised by Peters (1940). The specimens of blewitti were measured, x-rayed, photographed and videotaped in detail. Comparative measurements were also taken from 84 brama skins (37 males, 35 females, 12 unsexed) at AMNH (n = 19), ZSI (10), NHMUK (26) and USNM (29). Of the brama skins measured, 27 originated near known localities for blewitti, but specimens were included from throughout the range of brama. Measurements (in mm) taken from skin specimens were: culmen from base of skull; culmen from distal edge of cere; minimum width between nares; height of upper mandible at distal edge of cere; length of longest rictal bristle (straightened); tarsus; wing (straightened and flattened); tail (callipers inserted between central rectrices at insertion point); middle claw (digit 3) and hindclaw (digit 1, both claw measurements taken from the distal edge of scutes). To compare wing formulae, shortfalls from the wing-tip of each of the primaries (P1-P10, numbered from the outside) were measured (in mm) for six blewitti (one blewitti, NHMUK 1886.2.1.544, was excluded as its wing-tip is heavily worn) and 23 brama from USNM. The distances from the notches in the inner web to the tip of each of the outer four primaries (P1-P4) were also measured (in mm), as was the distance from the distal end (narrowest point) of the emargination on the outer webs to the feather tip for P2-P4. The notch for P4 was often not obvious in specimens of brama, and in these cases it was not measured. Osteology Measurements of skeletal features were taken directly from x-rays of blewitti and brama specimens, in which multiple views taken at various angles allowed direct comparisons with skeletal elements of brama and with the actual skin specimens x-rayed to ensure that the bones were oriented along the correct axes to avoid size distortion due to foreshortening. Only elements lying close to the film surface were measured to minimise parallax. Measurement options were limited by bone preservation and the fact that they are articulated in skin specimens of blewitti. Measurements taken were: greatest width of skull; lengths of humerus and ulna; length of carpometacarpus from proximal end to distal articular surface; length and minimum width of tibiotarsus, and width oi its condylar end; and length and minimum width of tarsometatarsus. Univariate statistics and principal components analysis (PCA) using correlation matrices were done separately on external, skeletal and wing formula measurements using SYSTAT for Windows (version 5). Variables used in PCA were chosen partially to maximise the number of specimens of blewitti that could be included without estimation of missing data. Because of the small sample of blewitti, sexes were combined. Intergeneric skeletal comparisons To allow osteological comparisons, several skeletal elements (the entire humerus, radius, ulna, tibiotarsus and tarsometatarsus; a femur missing the head; and the skull missing part of the posterior and caudal regions) were removed by J. P. Angle from the left side of a blewitti skin specimen (NHMUK 1886.2.1.546) using the techniques in Olson et al. (1987); casts were retained at USNM, as USNM 261299. These elements were compared directly with USNM skeletons of A. brama (n = 6); Little OwlH. noctua (10); Burrowing Owl A. cunicularia (5); Jungle Owlet Glaucidium cuculoides (7); White-browed Owl Ninox superciliaris (l); Brown Hawk OwlIV. scutulata (3); Boreal Owl Aegolius fiunereus (1); and indirectly with noctua (4) and brama (2) from UMMZ. The UMMZ osteological specimens were examined the week following the USNM comparisons, and were videotaped to allow further study. In addition, important osteological features that were observed in the extracted blewitti skeletal elements were then examined (as possible) in the x-rays of all seven blewitti specimens and the x-rayed brama. Osteological terminology follows Howard (1929) and Baumel & Winner (1993). Measurements taken of the above specimens, along with brama (9), noctua (4), N. scutulata (1) and Philippine Hawk Owl N. philippensis centralis (1) were: skull (including culmen) length; minimum widths of the frontal both anterior and posterior to the supraorbital processes; maximum skull width; height of lateral rim of frontal; width of distal half of lacrimal (maximum medio-lateral width); length of lacrimal (maximum antero-posterior length of caudal edge); maximum width across both palatines in situ ; maximum length and minimum width of ulna; for humerus, femur, tibiotarsus and tarsometatarsus, maximum lengths, minimum widths and distal widths, and for the last two elements maximum proximal widths as well. For tibiotarsus, length was from the proximal articular surface, and proximal width did not include the fibula. 80 P. C. RASMUSSEN & N. J. COLLAR Forktail 29(2013) RESULTS Externa! morphology of blewitti and brama The characters that separate blewitti from brama in the held are summarised in Rasmussen & Collar (1998). Even within a race, brama presents great variability in plumage and in most (but not all) ol the characters distinguishing the two species a few individuals of brama closely approach the condition in blewitti, especially when the latter species is in worn plumage. The nares are situated more widely apart in blewitti than in brama, owing largely to the broader culmen ridge of blewitti (Table 1). Moreover, the nares of blewitti are positioned more obliquely, not facing directly anteriad as in brama. The cere of blewitti is less inflated and the nares are situated well inside the cere, instead of right at the edge of the more swollen cere, as in brama. Compared to brama, blewitti has more heavily feathered toes (Hume 1873, Rasmussen & Collar 1998), except in the extremely worn specimen (NHMUK 1886.2.1.544). In most blewitti the white tarsal feathering continues uninterrupted onto the toes, while in brama the more mottled, dingier tarsal feathering stops more abruptly at the top of the toes, with only sparser bristles on the toes themselves. Although the extent of feathering on the tarsus and toes is often highly variable within an owl species (as it is in both Little and Burrowing Owls), the difference in this feature between the Forest and Spotted Owlets seems quite constant, allowing for the effects of wear. The toes and claws of blewitti appear noticeably heavier (and the latter longer; see below) than those of brama. Although the four traditional external measurements (culmen, wing, tarsus and tail lengths) overlap broadly between blewitti and Table 1. Summary statistics for measurements3 (mm) of Heteroglaux blewitti and Athene brama skin specimens (sexes combined) and results of Principal Components Analysis'3 on these variables. blewitti brama Component loadings Measurements Mean SE Range N Mean SE Range N PCI PC2 PC3 External measures Culmen from skull 21.2 0.8 19.9-22.0 7 20.5 1.1 17.4-23.7 83 0.31 -0.44 0.83 Culmen from cere 14.7 0.6 13.7-15.8 7 13.7 0.7 11.9-15.8 81 - - - W between nares 5.4 0.4 5.0-6.0 7 3.4 0.4 2.6— 4.3 63 0.87 0.07 0.02 H upper mandible at cere 10.7 1.0 9.0-11.7 7 7.6 0.3 7.0-8.3 28 0.91 0.15 0.06 Max. rictal bristle 1 18.9 1.2 17.2-20.2 7 18.3 2.4 13.7-23.5 65 - - - Tarsus 1 28.0 1.7 26.2-31.2 7 29.3 2.0 23.8-33.8 81 -0.40 -0.72 -0.26 Wingl 147.8 3.2 144.0-154.0 7 153.2 5.8 139.0-169.0 80 -0.78 0.16 0.44 Taill 69.0 3.4 62.3-72.0 7 72.2 3.6 65.7-82.0 78 -0.69 0.55 0.17 Middle claw (D3) 1 14.1 0.8 13.1-15.2 7 10.6 0.6 9.5-12.1 62 0.92 0.67 0.03 Hindclaw (D1) 1 13.1 0.5 12.6-13.7 6 9.7 0.6 8.5-11.2 58 - - - Skeletal measurements from x-rays Skull w 36.2 1.49 33.8-37.5 5 33.3 1.25 31.4-34.8 16 Humerus 1 53.9 1.64 51.6-55.4 4 46.2 5.92 41.8-52.9 3 - - Humerus distal w 9.5 0.34 9.0-9.8 4 8.5 0.39 7.7— 8.9 9 - - Ulna 1 60.5 4.02 53.3-63.6 6 55.2 3.24 48.0-63.1 17 Carpometacarpus 1 38.6 1.64 27.2-31.4 7 26.6 1.22 23.6-28.2 21 0.73 0.52 Tibiotarsus 1 54.7 3.82 47.9-59.5 7 51.9 2.97 46.1-56.9 11 - - Tibiotarsus w (min.) 3.5 0.35 3. 1-3.9 7 2.9 0.16 2.7— 3.3 17 0.91 -0.06 External condyle w 8.0 0.68 7.5— 9.3 7 6.8 0.61 5.4-8.1 17 0.88 0.22 Tarsometatarsusl 26.8 1.47 24.7-28.2 7 28.9 1.55 25.7-31.8 20 -0.38 0.88 Tarsometatarsus w (min.) 5.1 0.14 4.8— 5.2 6 3.4 0.27 2.9— 3.9 19 0.89 -0.20 Wing formula measurements Shortfalls of each primary from wing point PI 26.7 2.7 23.0-30.1 6 25.5 3.0 19.6-29.5 23 -0.36 0.58 0.36 P2 6.0 1.6 3.8— 8.3 6 5.9 1.9 2.9— 9.8 23 -0.34 0.80 0.07 P3 0.3 0.6 0.0-1. 5 6 0.3 0.6 0.0-2. 1 23 0.02 0.04 0.82 P4 0.2 0.3 0.0-0.5 6 0.5 0.8 0.0-2.5 23 0.34 -0.27 -0.58 P5 3.4 0.7 2.7-4.6 6 4.6 1.7 1. 0-7.9 23 0.67 -0.51 0.23 P6 12.2 2.1 9.8-15.3 6 14.5 1.8 11.7-18.2 23 0.81 -0.47 0.08 P7 19.3 2.3 16.7-22.2 6 24.3 2.8 20.1-30.7 24 0.87 -0.29 0.15 P8 23.4 2.3 21.0-26.3 6 31.2 3.5 23.3-37.1 22 0.90 -0.13 0.21 P9 27.8 3.2 24.0-33.5 6 36.2 3.4 29.7-42.0 22 0.92 -0.06 0.21 P10 32.3 2.5 29.0-35.6 6 41.3 2.7 34.4-47.0 22 0.92 0.16 0.16 Distance from notch to tip PI 32.3 4.2 28.0-39.7 6 39.1 3.0 31.0-45.8 23 0.82 0.08 -0.29 P2 35.8 2.1 33.0-38.1 6 42.6 2.3 39.0-48.2 23 0.78 0.34 -0.17 P3 33.3 1.9 30.9-36.0 6 39.4 2.2 35.4-43.4 23 0.78 0.42 -0.23 P4 27.0 2.5 25.2-32.0 6 32.3 2.9 28.3-36.8 9 - - - Distance from emargination to tip P2 50.5 1.4 49.0-52.1 6 56.2 2.5 52.6-61.4 21 0.84 0.11 0.19 P3 40.5 1.8 38.0-43.0 6 47.9 2.1 43.9-51.4 21 0.84 0.49 -0.10 P4 31.8 1.6 30.0-34.0 6 39.1 1.6 35.5-41.5 21 0.81 0.52 -0.07 a I = length, w = width, h = height b Eigenvalues and percent variance explained for PCI-3 on external measures: 4.0, 44.3%; 1 .6, 17.9%; 1 .3, 14.3%, respectively; for PCI-2 on skeletal measures: 3.1, 61.6%; 1.1, 23.0%, respectively; for PCI-3 on wing formula measures: 8.7, 54.4%; 2.5, 1 5.6%; 1 .6, 9.8%, respectively. Forktail 29(2013) 81 Phenotypic evidence for the specific and generic validity of Heteroglaux blewitti 2 - rsi 1 - u_ 0 - -1 - -2 Factor 1 Figure 1. Graph of individual component scores on PC 1 and 2 for principal components analysis on measurements of Heteroglaux blewitti (diamonds) and Athene brama (circles). (A) external; (B) wing formula; (C) skeletal from x-rays. Figure 2. Wing formulae of Heteroglaux blewitti (solid lines, A,C,E) and Athene brama (dotted lines, B,D,F). (A,B) shortfalls from wing point of PI-10; distance from tips of individual feathers to (C,D) notches on inner webs of PI -4 and (E,F) emarginations on outer webs of P2-5 (descriptive statistics presented in Table 1). brama, the taxa differ strongly in several other external mensural characters, even though sexes were combined owing to the small sample of blewitti (Table 1). External measurements that do not overlap between the two species are: width between nares, bill height and lengths of middle and hindclaws (Table 1). A PCA of external measurements (Table 1, Figure 1A) shows that by far the greatest proportion of the variance is explained on PC 1 by a contrast between middle claw length, width between nares, and upper mandible height vs tail and wing lengths. Complete separation between the species is attained on this axis (FigurelA). Several differences exist between the wing formulae of blewitti and brama, although most measurements overlap at least minimally between the species, and the sample of blewitti is small (Table 1, Figure IB). In blewitti, P7-P10 each have a smaller shortfall, i.e. the feather tips fall closer to the wing-point, making the inner wing broader than in brama (Figure 2A,B). In blewitti, P7 is never as short as PI, while in brama PI and P7 are approximately equal. The emarginations on the outer webs of P2-P4 and the notches on the inner webs of P 1 -P4 are all closer to the tips of the individual feathers in blewitti (Figure 2C-F); measurements of emargination position did not even overlap between the species (Table 1). Finally, 14 of 23 brama do not show a distinct notch on P4, whilst all six blewitti examined have a definite notch on the inner web of this primary. A PCA of wing formulae showed that by far the greatest percentage of the variance was explained by Factor 1 (Table 1), which was mainly a size axis, on which shortfalls of P3 and P4 were not strongly correlated, and those of PI and P2 were weakly negatively correlated. All blewitti had negative Factor 1 scores, while the scores of all brama fell above -1 on Factor 1, reflecting the smaller shortfalls of the inner primaries and notch and emargination distances of blewitti (Figure IB). 82 P. C. RASMUSSEN & N. J. COLLAR Forktail 29 (2013) Osteology Despite the similarity in plumage of blewitti and brama , there are several major osteological differences (Plate 1 ) between blewitti and all three species normally recognised in Athene (including the highly polytypic A. \Speotyto\ cunicularia). Relative to Athene , the nasal process of the premaxillary of blewitti (Plate 1 A,B) is expanded anteriorly; the culmen ridge (Os nasale) of the premaxillary is more arched; the distal tip of the premaxillary is longer and more caudally directed, so the rostrum of blewitti is heavier and more strongly hooked; the narial openings are larger and more ovoid; the mandibular symphysis is broader; and the entire mandible is somewhat heavier. The frontals of blewitti are much broader both anterior and posterior to the supraorbital process than for any Athene (Plate 1 A,B), so that the skull of blewitti strikingly resembles that of Glaucidium and Ninox superciliaris ; the latter Malagasy species has been wrongly placed in Ninox , and is closer to Athene-. H. F. James and S. L. Olson, pers. comm. 1997; Wink etal. 2004). The posterior portion of the interorbital roof of blewitti is not wider than the anterior portion, unlike Athene. The lateral rim of the frontal anterior to the supraorbital process is greatly inflated in blewitti compared to members of Athene, similar to but even more so than in G. cuculoides, N. superciliaris and most other small owls. The great inflation of this region is visible in x-rays of other blewitti specimens as well. The lacrimals of blewitti are very large relative to those of Athene, but like them (Ford 1967) are short and do not contact the jugal bar; those of the extracted skull of blewitti are detached from the skull, but their position relative to the jugal bar is confirmed by x-rays of all blewitti specimens. The maxillopalatines of blewitti are large, with straight medial edges that nearly contact each other for most of their length, unlike in Athene , where the maxillopalatines are more triangular in shape so that they only contact each other at the apex. The palatines of blewitti are relatively short anteroposteriorly as in Athene but are more expanded posteriorly, in the latter respect being similar to G. cuculoides. The supraorbital processes of blewitti are better developed than in most other Athene specimens we examined. The temporal fossa is much deeper in posterior view in blewitti than in brama. The quadrate of blewitti has no intercapitular groove, the lack of which is apparently otherwise autapomorphic for Athene including Speotyto (Ford 1967); but the articular surface of the external capitulum is longer than in Athene, being similar to that of G. cuculoides. The otic process of the quadrate of blewitti is longer than in other owls examined, and the mandibular articulation is broad. The socket for the quadratojugal of blewitti is long and more strongly twisted externally than in brama, similar to that of G. cuculoides. The humerus of blewitti is slightly longer and heavier than in brama, while the ulna of blewitti , although not longer, is substantially more robust. The leg proportions of blewitti (Table 2, Plate 1 C,D) are unusual in that the hindlimb is much more gracile proximally than distally: the femur and the proximal end of the tibiotarsus are heavier than those of brama, but not markedly so; however, the distal end of the tibiotarsus and the entire tarsometatarsus are greatly enlarged and especially broadened relative to those of brama. In addition, the single measurable femur of blewitti is longer than that of brama-, the tibiotarsus is approximately the same length in both; and the tarsometatarsus of blewitti is shorter than that of brama, while the combined length of these three elements is roughly the same for the two species. The shaft widths of each of the leg elements in brama are very similar to one another, in strong contrast to the situation of blewitti, in which the tarsometatarsus shaft is much wider than that of the femur. In comparison to blewitti , G. cuculoides\\2.s all leg elements more uniformly stout; the femur and proximal tibiotarsus heavier, the tarsometatarsus similar in breadth but considerably longer. The Factor 1 Factor 1 Figure 3. PCAs of skeletal measurements of selected surniine owls. Athene noctua, noc; A. brama, bra; Heteroglaux blewitti, ble; A. cunicularia, cun; Ninox philippensis, phi; N. solomonis, sol; N. scutulata, scu -, Aegolius funereus, fun; and Glaucidium cuculoides, cue. (A) skull; (B) forelimb; (C) hindlimb. H. TAYLOR, ©NATURAL HISTORY MUSEUM, LONDON Forktail 29(2013) Phenotypic evidence for the specific and generic validity of Heteroglaux blewitti 83 Table 2. Limb proportions of Heteroglaux blewitti, Athene brama, Athene noctua, Glaucidium cuculoides, and Ninox superciliaris. Ratios are of mean measurements. Species blewitti (n = 1) brama (n = 6) noctua ( n = 10) cuculoides (n = 7) superciliaris (n = 1) Wing proportions Hum l/distal w 5.76 5.64 5.64 5.37 Hum l/ulna 1 0.91 0.83 0.84 0.82 Wing la 107.8 104.5 106.3 119.0 Leg proportions Femur I/distal w 4.88 5.23 5.12 4.78 Tibiotarsus I/distal w 6.78 8.01 7.94 7.36 7.45 Tarsometatarsus I/distal w 2.95 4.33 4.53 3.32 3.71 Femur l/tib. 1 0.74 0.67 0.68 0.71 Tib. I/tar. 1 1.98 1.71 1.66 1.96 1.91 Femur l/tarsometatarsusl 1.46 1.15 1.13 1.39 Leg lb 115.6 115.7 123.5 125.8 Wing l/leg 1 0.93 0.90 0.86 0.95 ‘Humerus I + ulna I + carpometacarpus I Temur I + tibiotarsus I + tarsometatarsus Plate 1. Comparisons between skeletal elements of Heteroglaux blewitti (NHMUK 1886.2.1.546) and Athene brama (NHMUK S/1 989.25.4). (A,B) Skulls of (A) blewitti and (B) brama in cranial (top), lateral (middle), and caudal (bottom) views; (C,D) left femora, tibiotarsi, and tarsometatarsi of (C) blewitti and (D) brama in posterior view. H. TAYLOR, ©NATURAL HISTORY MUSEUM, LONDON 84 P. C. RASMUSSEN & N. J. COLLAR Forktail 29 (2013) Table 3. Summary statistics of measurements3 (mm) and results of Principal Components Analysis1 selected owls. b on osteological measurements of Variable blewitti Q 5 -C5 =3 ts o c cunicularia cuculoides funereus superciliaris scutulata philippensis (s.l.) novaezeelandiae solomonis Component loadings N 1 10-15 16 10-11 4-11 6-7 1 1-5 2-4 1 2-3 PCI PCII Skull Total T 47.1 + 46.1 ±1.3 48.0±0.9 49.0±3.2 50.6±0.8 47.8±0.5 52.5 51.4±5.5 46.4±0.3 47.8 51.3±0.5 0.73 0.53 44.0-49.0 46.7-49.8 45.7-55.7 49.4-52.4 47.0-48.4 47.5-59.4 46.1-46.8 5.8-51.8 Anterior frontal w 13.0 10.1±0.7 11.7±0.6 11.1±1.1 14.3±1.8 11.0±0.8 13.6 10.6±1.3 9.4±1 .2 9.2 11.0±0.3 0.62 0.27 8.9-11.4 10.6-12.9 9.2-13.1 11.7-17.0 9.5-12.0 12.3-14.4 13.2-15.1 10.8-13.3 Posterior frontal w 13.0 10.1 ±0.9 8.5±0.8 8.5±1.2 15.8±1.2 15.7±0.6 14.6 13.5±1.0 14.4±0.9 11.7 12.0±1.2 0.79 -0.53 8.9-12.1 7.3— 9.8 7.1-11.1 13.7-17.5 14.7-16.5 12.3-14.4 13.2-15.1 10.8-13.3 Maximum w 36.4 34.7±1.0 35.2±0.9 37.1 ±2.4 35.9±0.9 39.5±0.8 37.5 35.9±2.0 34.8±0.6 34.7 38.9±0.4 0.69 0.08 33.2-35.9 33.1-36.8 34.5-41.6 33.7-37.0 38.6-40.6 34.0-39.0 34.4-35.5 38.4-39.1 Lat. rim frontal ht 4.4 2.5±0.4 1.9±0.2 1.9±0.4 3.5±0.3 4.3±0.5 4.6 2.5±0.3 3.0±0.2 2.8 3.3±3.4 0.75 -0.61 1.8-3. 1 1. 5-2.4 1. 5-2.7 3.2-4.1 37-4.8 2.1-2.8 2.8— 3.2 3.3— 3.4 Distal lacrimal w 4.4 2.5±0.5 2.6±0.3 2.7±0.6 2.5±0.3 3.2±0.4 2.7 2.5 2.6±0.3 2.9±0.2 2.1-3.8 1. 8-3.0 1. 8-4.1 2.2-3.0 27-3.8 2.4— 2.8 2.8-3. 1 Lacrimal 1 9.5 7.8±0.5 8.1±0.8 8.9±0.7 9.6±0.9 9.5±0.8 10.1 7.6 6.3±1.1 8.8±1.0 6.9— 8.6 6.5-9.1 8.1-10.3 8.6-10.5 8.7-10.7 5.6-7. 1 8.1-9.5 Palatine w 11.8 10.1 ±0.8 11.5±0.6 11.3±0.7 11.8±0.6 11.9±0.3 12.9 12.7±0.3 10.7±0.2 12.8±0.4 0.72 0.38 8.5-11.1 10.0-12.6 10.1-12.7 10.8-13.0 11.4-12.2 12.4-12.9 10.5-10.8 12.6-13.3 Humerus Total 1 51.3 47.4±1.2 48.5±1.7 55.9±4.0 53.7±1.4 47.6±0.5 60.9±3.0 52.3±1.1 55.3 59.9±0.9 0.96 0.12 45.6-49.9 45.8-51.5 49.7-62.5 50.9-55.3 46.8-48.3 56.6-64.1 51.1-53.4 59.1-60.8 Shaft w 3.4 3.5±0.2 3.4±0.2 3.6±0.2 3.8±0.2 3.2±0.1 4.1 4.4±0.3 3.5±0.6 4.0 4.2±0.1 0.93 -0.20 3. 1-3.8 3.0-3. 7 3.3— 3.9 3.6— 4.2 3.0-3.4 3.9-47 3.5— 3.6 4. 1-4.3 Distal w 8.9 8.4±0.3 8.6±0.5 9.2±0.6 10.0±0.4 8.4±0.1 10.4 11.0±0.8 9.3±0.2 10.1 10.7±0.4 0.94 0.01 7.9— 8.9 7.8— 9.6 8.2-10.2 9.4-10.5 8.2— 8.6 9.8-11.6 9.1-9.5 10.3-11.0 Ulna Total 1 56.5 57.1±1.6 57.8±2.0 70.0±4.9 65.3±1.9 54.4±1.0 68.0 70.7±4.4 60.1 ±2.1 63.4 68.4±1.3 0.92 0.06 55.2-59.4 54.9-61.7 61.5-76.4 62.3-68.5 53.0-56.1 63.6-74.5 58.0-62.3 67.0-69.7 Shaft w 2.6 2.5±0.2 2.5±0.2 2.7±0.2 2.8±0.1 2.3±0.2 3.1 3.0±0.1 2.4±0.2 2.6 2.6±0.1 0.81 -0.33 2.2— 2.9 2. 1-3.0 2.4-3.0 2.6-3.0 2.2-27 2.9-3.1 2.3-27 2.5-27 Femur Total 1 38.1 34.5±0.6 36.9±1.5 39.1 ±2.6 40.2±1.0 36.3±0.7 38.5±0.8 34.6±1.3 37.6 40.8±1.4 0.88 -0.10 33.1-35.4 35.0-39.6 35.8-44.2 38.3-42.4 35.1-37.3 37.3-39.5 33.1-35.6 39.6-42.4 Shaft w 3.5 2.9±0.1 3.1±0.3 3.3±0.2 3.6±0.2 2.7±0.1 3.30.1 2.8±0.1 3.1 3.6±0.2 0.94 0.03 2.7-3.1 2.7— 3.7 2.9— 3.6 33-3.9 2.5— 2.9 3.2— 3.5 27-2.9 3.4— 3.8 Distal w 7.8 6.6±0.2 7.2±0.4 7.2±0.4 8.4±0.4 6.2±0.1 8.0±0.3 6.9±0.2 7.3 8.2±0.3 0.96 0.03 6.4-7.1 6.5-8.0 6.5-8.0 7.8-9.0 6.0-6.4 7.6— 8.4 5.5— 5.8 7.9— 8.5 Tibiotarsus Total 1 51.5 51.3±1.6 54.0±2.7 67.7±3.0 56.7±1.3 48.0±0.5 59.6 60.2±2.2 55.7±1 .8 59.3 61.4±17 0.64 -0.73 48.7-54.3 50.2-59.2 62.9-73.5 55.2-58.8 47.4-49.0 57.3-63.3 53.5-57.8 60.3-63.4 Prox. w 6.4 53+0.4 5.8±0.3 6.1 ±0.6 6.8±0.4 5.2±0.1 6.1 6.5±0.3 5.6±0.2 5.9 6.9±0.3 0.90 0.08 4.7— 5.9 5.3-6 .2 5.4-7.0 6.2-7 .7 5. 1-5.4 62-6.9 55.5-5.8 6.6-7.1 Shaft w 3.4 2.7±0.2 2.9±0.3 3.2±0.5 3.5±0.1 2.5±0.1 3.0 3.2±0.2 2.7±0.1 3.1 3.2±0.4 0.89 0.03 2.4-3. 1 2.6— 3.5 3.0-3.5 33-3.7 2.4— 2.8 3. 1-3.5 2.6— 2.8 2.9— 3.6 Distal w 7.6 6.4±0.2 6.8±0.4 7.1±0.5 7.7±0.3 5.9±0.2 8.0 6.7±0.33 6.1±0.3 3.4 7.5±0.4 0.93 0.01 6.1-6.7 6.2— 7.5 6.5-8.1 73-8.2 57-6.2 6.4-7.1 5.8— 6.6 7.1-8.0 Tarsometatarsus Total 1 26.0 29.9±1.1 32.6±2.7 47.7±3.1 28.9±0.8 22.1±0.3 31.2 28.6±2.5 30.9±1 .7 31.6 31.0±1.0 0.34 -0.90 28.5-31.7 28.9-38.0 41.5-53.1 28.0-30.7 21.8-22.6 25.3-32.3 29.1-33.0 30.1-32.1 Prox. w 7.8 6.7±0.2 7.0±0.6 7.3±0.6 8.2±0.4 6.1 ±0.1 7.5 7.3±0.4 6.3±0.5 7.0 7.8±0.3 0.93 0.09 6.4-7. 1 6.2-8.1 6.5— 8.5 7.7— 8.7 5.9-63 7.0-7.9 5.6— 6.6 7.6-8. 1 Shaft w 5.3 3.2±0.1 3.3±0.3 2.9±0.4 5.0±0.2 3.4±0.2 4.3 3.8±0.2 3.0±0.1 3.4 4.0±0.3 0.65 0.72 2.9— 3.4 2.9-39 2.5— 3.6 4.8— 5.2 3.0-37 3. 5 -4.0 3.0-3. 1 37-4.3 Distal w 8.8 6.9±0.2 7.2±0.4 7.6±0.7 8.7±0.6 6.5±0.1 8.4 7.9±0.3 6.8±0.1 7.3 7.9±0.4 0.92 0.19 6.5-7.1 6.6-8.0 6.9— 8.9 7.2-93 6.2-66 7.6— 8.4 67-6.9 7.6— 8.3 a 1 = length, w = width, d = depth, ht = height, lat = lateral, prox = proximal b eigenvalues and percent variance explained for PCI-2 on skull measures: 3.1, 51.5%; 1.1, 19.2%, respectively; on wing and skull length: 4.2, 70.8%; 1.0, 17.3%, respectively; on leg measures: 7.7, 70.1%; 1.9, 17.4%, respectively. ' also included in PCA of wing measures; component loadings on PC 1, 0.29; on PC 2, 0.93. Forktail 29(2013) Phenotypic evidence for the specific and generic validity of Heterogtaux blewitti 85 tibiotarsus and tarsometatarsus ofiV. superciliaris are much longer and more gracile than for any of the above species. The only osteological feature of the limbs listed by Ford (1967) as diagnostic o i Athene (including Speotyto) is the pointed posterior edge of the outer rim of the middle trochlea of the tarsometatarsus; however, unlike Athene and like other owls examined, this is rounded in the single blewitti specimen. Skull widths, length of wing elements (humerus, ulna and carpometacarpus) and widths of leg elements (tibiotarsus and tarsometatarsus) are all considerably larger in blewitti than in brama (Tables 1 and 3, Figure 1C). The tarsometatarsi of blewitti are no longer (averaging shorter) than in brama, but are considerably more robust (Tables 1 and 2). In a PCA of skeletal measurements (Table 1, Figure 1C), Factor 1 is astrongsize axis on which tarsometatarsus length is negatively correlated; all blewitti scores fall above 0 on Factor 1, while most brama scores fall below 0. Intergeneric skeletal comparisons In PCAs of skull and hindlimb measures (Table 3, Figure 3A-C), scores for the single available blewitti skeleton fall well apart from those of any Athene species. On skull measurements (Table 3), blewitti is closest to Aegolius funereus and Glaucidium cuculoides , both of which have relatively large measurements on the variables included in the analysis (particularly so for width of frontal posterior to supraorbital process and height of lateral rim of frontal), and all have Factor 1 scores well above 0. Athene species, conversely, have small measurements on these variables, and all but a few large cunicularia fall below 0 on Factor 1. Factor 2 principally contrasts skull length with posterior frontal width and lateral rim height, and on this axis blewitti and A. funereus differ strongly from large cunicularia. Wingproportions of Athene species and blewitti are very similar (Table 2, Figure 3B), with scores of all but large cunicularia falling below 0 on Factor 1, a strong size axis on which skull width is uncorrelated, indicating that, compared to other genera sampled, most Athene and blewitti have small wings relative to head size. On a PCA of hindlimb measures (Table 3, Figure 3C), Factor 1 is a general size axis on which tarsometatarsus explains most variance, and on this axis cunicularia is the most distinctive group due to its extremely long legs, while blewitti is well separated from Athene and close to G. cuculoides. Factor 2 is basically a contrast between tibiotarsus and tarsometatarsus length with tarsometatarsus width, and on this axis blewitti has the highest score, again reflecting the stoutness of this element (see Table 2). DISCUSSION The specific validity of blewitti There ought to be no question about the status of blewitti as a full species, but in the light of a recent report of a pair composed of a male brama and a female blewitti producing a supposedly fertile offspring, and indeed of a population of hybrids which ‘may have a much wider distribution that could equal or surpass the very limited one of the Forest Owlet’ (Pande et al. 2011), all of which could be taken to imply the conspecificity of the taxa, we briefly here recapitulate and expand the evidence. First, blewitti differs in plumage and external structure from brama on multiple characters, including the narial and cere characters given in the species description by Hume (1873), and others enumerated in Rasmussen & Collar (1998); our elaboration above of the external structural differences involves much greater width between nares, bill height and claw length, plus a somewhat different wing-shape. It is interesting to note how, albeit with tiny sample sizes, the seemingly small size advantage of blewitti over brama translates into a doubling of body mass (241.0 g,n= 1 [based on ‘8.5 oz’ on label of type specimen and in Hume 1873] vs 110- 1 14 g, n = 2) (Dunning 1993). Second, blewitti possesses a wide range of osteological distinctions from not only brama but all members of Athene, including cunicularia. These involve many cranial characters (some on fused elements) on the nares, premaxillary, frontal, mandible, lacrimal, maxillopalatine and quadrate. Of these, the much broader frontal (on cranial view) with its greatly inflated lateral rim anteriorly is most striking. There are also differences between blewitti and Athene in hindlimb proportions, especially the short, very stout tarsometatarsus of blewitti, and in the conformation of the middle trochlea of the tarsometatarsus. Third, the song of blewitti is very dissimilar to vocalisations of brama, and does not support its treatment as a close relative (Rasmussen & Ishtiaq 1999, Jathar & Rahmani 2002, Rasmussen & Anderton 2005). Fourth, behavioural differences include the direct flight and lateral tail-flicking habits of blewitti (King& Rasmussen 1998, Konig^r al. 1999, Rasmussen & Ishtiaq 1999, Pande et al. 2011, Mikkola 2012). Finally, even if these many differences between the taxa are somehow not considered sufficient justification for the specific distinctness of blewitti from brama, the two forms are broadly sympatric: brama occurs wherever blewitti occurs. While the microhabitat where he collected them probably differed, J. Davidson collected at least 1 1 brama (all now in NHMUK) in the same region where he took his five blewitti (and other unattributed Khandesh and ‘Candesh’ specimens in the Hume and Seebohm collections may have also been collected by Davidson). Although Davidson took numerous egg sets of brama (Davidson MS) he never found a nest of blewitti (Barnes 1888), unsurprisingly since even in Khandesh in the 1880s the Spotted Owlet was clearly much the commoner species. Sympatric occurrence is, of course, a prerequisite of hybridisation, but the evidence presented in Pande et al. (2011) is impossible to interpret owing to shortcomings in figure labelling, description and photograph quality. In our experience no specimens of blewitti or brama can be said to be intermediate in more than a few characters, and we have seen no specimens of either for which there is any doubt as to their specific identity. Based on the data presented above on plumage, other external morphological, mensural, mass and osteological differences, and the lack oi intermediates, it is evident that blewitti is a well-marked, distinct species. Its coexistence with brama in areas where both are resident strongly reinforces this view. Thus, under any widely accepted species concept, the Forest Owlet must be considered specifically distinct from the Spotted Owlet. Until far better evidence is produced we take the view that hybridisation is unproven and, given the wide range of differences between the taxa, very unlikely. Even if occasional hybridisation were proven, wild intrageneric hybrids are known for several avian taxa (see McCarthy 2006), and these are not generally taken as evidence of exceptionally close relationship. Baker’s (1923) notion of the conspecificity of blewitti with noctua has never been taken seriously elsewhere, but has never been explicitly dealt with; hence, we do so here. Although A. noctua is a variable species, it is considerably more different in appearance from blewitti than is brama. This is most evident in: the pattern oi the underparts (streaked in noctua, barred in blewitti ); crown pattern (streaked in noctua, nearly unmarked in blewitti)-, broader, spotted frontal semi-collar; less white on face; less banded wings and tail; smaller bill and claws; and tarsal length and shape (long and much more gracile in noctua, short and stout in blewitti). We can find no features in which blewitti resembles noctua more than it does brama, except for the streaked underparts of juvenile blewitti (Rasmussen & Anderton 2005). In osteology, brama, noctua and, except ior tarsometatarsus length, cunicularia are very similar to each other, far more so than any is to blewitti. 86 P. C. RASMUSSEN & N. J. COLLAR Forktail 29(2013) The generic validity of Heteroglaux Clearly the plumage differences between blewitti and established members of Athene are insufficient alone to justify the maintenance o f Heteroglaux. Moreover, while the song of blewitti is very different from brama in its high pitch, tone and modulation, it does resemble the male song of noctua in overall quality; and even the song of cunicularia is somewhat intermediate between blewitti and brama. As a consequence we abandon the pursuit of generic limits in blewitti through acoustic evidence: such a line of taxonomic inquiry is untried elsewhere for fairly obvious reasons of interpretation (especially as convergence may play a part), and would require the discovery of very strong differences to be considered in any way informative. However, we note that while some vocalisations have been documented for blewitti (Rasmussen & Ishtiaq 1999, Ishtiaq & Rahmani 2005), to our knowledge only one recording is available online (AV 16764; http://avocet.zoology.msu.edu/ recordings/ 16764), so detailed further analysis of vocalisations of blewitti cannot in any case be made until a more complete sample of recordings becomes available. This then leaves morphological and behavioural differences to consider. In terms of external morphology, blewitti would not appear out of place within Athene. However, many osteological differences involving multiple cranial elements (especially the greatly widened and inflated frontal and the large, straight-edged maxillopalatines) and the hindlimb (the extremely stout tarsometatarsus) separate it from other Athene, and indicate that plumage convergence or perhaps even mimicry may have resulted in the relative similarity of external phenotype between blewitti and brama. It is unusual to find many marked qualitative (as opposed to mensural or quantitative) osteological differences within an avian genus, and it is even occasionally difficult to distinguish closely related genera osteologically. Extinct island owls placed in Athene show great variation in size and length of extremities: one {A. cretensis of Crete) was relatively large, with very long tarsometatarsi (Weesie 1982), while another {A. angelis of Corsica) had unusually long femora and robust tarsometatarsi (Mourer-Chauvire et al. 1997), and a third (A. vallgornerensis of Mallorca) was small, with short, robust tarsometatarsi (Guerra etal. 2012). However, judging from figures (Guerra et al. 2012), even A. vallgornerensis had a distinctly less robust tarsometatarsus than does blewitti. Living and fossil island owls show no trend in overall size, but do tend to have somewhat larger feet and claws than closely allied continental species (Louchart 2005), as does the mainland blewitti to a striking degree. Behaviourally, blewitti differs from other members of Athene in its direct, non-undulating flight, and in its lateral tail-flicking. While these differences by themselves may not suggest distinctness at the generic level, they provide significant corroborating evidence to the osteological data. In summary, the numerous (and in some cases major) differences in skull and tarsometatarsus morphology between blewitti and all other species of Athene (including Speotyto ) — involving many cranial elements, especially the frontals, and the extremely stout tarsometatarsus — indicate that (although a rapid evolution cannot be excluded) blewitti seems likely to be distantly related to the others. Because of this distinctness, coupled with its unusual flight and tail-flicking behaviours, we consider the resurrection of the monotypic genus Heteroglaux fully justified, and far from ‘spurious’ (Weick 2006). Even those predisposed to recognise very broad genera, and who may thus prefer to maintain blewitti in Athene , should at least be aware that it is osteologically much the most distinctive of the group and quite possibly evolved from an ancient divergence event. Further study involving more owl taxa, such as phylogenetic analyses based on morphology and / or DNA, is likely to shed more light on the relationships of blewitti , but in the meantime we contend that the generic distinctiveness of blewitti only increases the urgency with which the conservation needs of this Critically Endangered species must be addressed. ACKNOWLEDGEMENTS Special thanks are due staff of the N atural H istory Museum, UK, for allowing extraction of skeletal elements from one blewitti specimen, and toj. P. Angle for removing the blewitti elements. Assistance of various kinds was given by R. P. Prys-Jones, M. P. Walters, !. H. Cooper, H. van Grouwand O. Crimmen, the Natural History Museum, UK; M. LeCroy, P. Sweet, G. Barrowclough and M. N. Feinberg, American Museum of Natural History; the late R. A. Paynter and A. Pirie, Museum of Comparative Zoology, Harvard University; S. L. Olson, G. R. Graves, R. L. Zusi, H. F. James, J. P. Angle, F. Grady, C. M. Milensky and J. T. Marshall, Jr., National Museum of Natural History, Smithsonian Institution; the late R. W. Storer andj. Hinshaw, University of Michigan Museum of Zoology; and M. Holmes, National Museum of Ireland. Funding for travel to museums was provided by the Research Opportunities Fund of the National Museum of Natural History, and by British Airways, through M. Sitnick and T. Shille, Office of Biodiversity Programs, Smithsonian Institution. We are most grateful to referees J. H. Cooper and H. James for comments that improved the manuscript. REFERENCES Abdulali, H. (1972) A catalogue of the birds in the collection of the Bombay Natural History Society — 1 1 . J. Bombay Nat. Hist. Soc. 69: 1 02-129. Ali, S. (1978) President's letter. Hornbill No. 8 (July-September): 3-4. Ali, S. & Ripley, S. D. (1981 ) Handbook of the birds of India and Pakistan, 3. Second edition. Delhi: Oxford University Press. Baker, E. C. S. (1923) A hand-list of genera and species of birds of the Indian Empire. Bombay: Bombay Natural History Society. Baker, E. C. S. (1934) The nidification of birds of the Indian Empire. Vol. Ill - Ploceidae — Asionidae. London: Taylor and Francis. Ball, V. (1878) From the Ganges to the Godaveri. On the distribution of birds, so far as it is present known, throughout the hilly region, which extends from the Rajmehal Hills to the Godaveri Valley. Stray Feathers 7: 191-235. Barnes, H. E. (1888) Nesting in western India. J. Bombay Nat. Hist. Soc. 3: 205-224. Baumel, J. J. & Witmer, L. M. (1993) Osteologia. Pp.45-1 32 in J. J. Baumel, ed. Handbook of avian anatomy: Nomina anatomica avium. Second edition. Cambridge, Mass: Publ. Nuttall Ornithol. Cl. No 23. Biswas, B. (1953) A check-list of genera of Indian birds. Rec. Indian Museum 50: 1-62. Blanford, W.T. (1895) The fauna of British India, including Ceylon and Burma. Birds — Vol. 3. London: Taylor and Francis. Chavan, R. A. & Rithe, K. D. (2009) Occurrence and breeding record of the Forest Owlet Heteroglaux blewitti from Yawal Wildlife Sanctuary, Maharashtra, India. J. Bombay Nat. Hist. Soc. 1 06: 207-208. Clements, J. F. (2007) The Clements checklist of birds of the world. Sixth edition. Ithaca, NY: Cornell University Press (Comstock Publishing Associates). Collar, N. J., Andreev, A. V., Chan, S., Crosby, M. J., Subramanya, S. & Tobias, J. A. (2001 ) Threatened birds of Asia: the BirdLife International Red Data Book. Cambridge U.K.: BirdLife International. Davidson, J. (MS) Notes on Khandeish birds. Unpublished manuscript in NHMUK Archives, Tring. Dickinson, E. C., ed. (2003) The Howard & Moore complete checklist of the birds of the world. Third edition. London: Christopher Helm. Dubois, A. (1904) Synopsis avium. Nouveau manuel d'ornithologie. Deuxieme Partie. Bruxelles: H. Lamertin. Dunning, Jr., J. B. (1993) CRC handbook of avian body masses. Boca Raton: CRC Press. Ford, N. (1967) A systematic study of the owls based on comparative osteology. Ph.D. dissert., Univ. Michigan, Ann Arbor. Forktail 29(2013) Phenotypic evidence for the specific and generic validity of Heteroglaux blewitti 87 Gill, F. & Wright, M. (2006) Birds of the world: recommended English names. Princeton: Princeton University Press. Grimmett, R., Inskipp, C.& lnskipp,T. (201 1) Birds of the Indian subcontinent. London: Christopher Flelm. Guerra, C., Bover, P. & Alcover, J. A. (2012) A new species of extinct little owl from the Pleistocene of Mallorca (Balearic Islands). J. Orn. 1 52: 347- 354. Gurney, J. H. (1894) Catalogue of the birds of prey (Accipitres and Striges), with the number of specimens in Norwich Museum. London: R. FI. Porter. Floward, H. (1929) The avifauna of Emeryville Shellmound. Univ. Calif. Publ. Zool. 32:301-394. del Floyo, J., Elliott, A. & Sargatal, J., eds. (1999) Handbook of the birds of the world, 5. Barn-owls to hummingbirds. Barcelona: Lynx Edicions. Hume, A. 0.(1873) Novelties. Stray Feathers 1:464-483 [Heteroglaux p.467; Heteroglaux blewitti pp.468-469]. Hume, A. O. (1879) A rough tentative list of the birds of India. Stray Feathers 8: 73-150. Ishtiaq, F. (201 1) Response to'Discovery of possible hybrid of the Critically Endangered Forest Owlet Athene blewitti and Spotted Owlet Athene brama from northern Maharashtra' by Pande et al. J. Threatened Taxa 3(5): 1798. Ishtiaq, F. & Rahmani, A. R. (2005) The Forest Owlet Heteroglaux blewitti: vocalization, breeding biology and conservation. Ibis 147: 197-205. Jathar, G. & Patil, D. (201 1 ) A review of 'Discovery of possible hybrid of the Critically Endangered Forest Owlet Athene blewitti and Spotted Owlet Athene brama from northern Maharashtra'. J. Threatened Taxa 3(5): 1800-1803. Jathar, G. A. & Rahmani, A. R. (2002) Ecological studies on the Forest Spotted Owlet Athene (Heteroglaux) blewitti. Annual report 2001-2002 (Bombay Natural History Society), unpublished. Jathar, G. A. & Rahmani, A. R. (2004) Status of the critically endangered Forest Owlet Heteroglaux blewitti (Hume) in Maharashtra. Survey report (Bombay Natural History Society), unpublished. Kasambe, R., Wadatkar, J., Bhusum, N. S. & Kasdekar, F. (2005) Forest Owlets Heteroglaux blewitti in Melghat Tiger Reserve, Dist. Amravati, Maharashtra. News!. Birdwatchers 45(3): 38-40. King, B. F. & Rasmussen, P. C. (1998) The rediscovery of the Forest Owlet Athene ( Heteroglaux ) blewitti. Forktail 14: 51-53. Konig, C., Weick, F. & Becking, J.-H. (1999) Owls: a guide to the owls of the world. Robertsbridge, East Sussex, U.K.: Pica Press. Konig, C„ Weick, F. & Becking, J.-H. [name omitted on title page] (2008) Owls of the world. Second edition. London: Christopher Helm. Louchart, A. (2005) Integrating the fossil record in the study of insular body size evolution: example of owls (Aves, Strigiformes). In J. A. Alcover & P. Bover, eds. Proc. International Symposium'InsularVertebrate Evolution: the Paleontological Approach'. Monografies de la Societat d'Histdria Natural de les Balears 12: 155-174. Marshall, J.T. & King, B. (1988) Genus Otus. In D. Amadon & J. Bull, eds. Hawks and owls of the world: an annotated list of species. Proc. Western Foundation Vertebrate Zoology 3(4): 296-357. McCarthy, E. M. (2006) Handbook of avian hybrids of the world. Oxford: Oxford University Press. Mehta, P., Kulkarni, J. & Patil, D. (2008) A survey of the Critically Endangered Forest Owlet Heteroglaux blewitti in central India. BirdingASIA 10: 77- 87. Mikkola, H. (2012) Owls of the world: a photographic guide. London: Christopher Helm. Mourer-Chauvire, C., Salotti, M„ Pereira, E., Quinif, Y., Courtois, J.-Y., Dubois, J. N. & La Milza, J. C. (1997) Athene angelis n. sp. (Aves, Strigiformes), nouvelle espece endemique insulaire eteinte du Pleistocene moyen et superieur de Corse (France). C. R. Acad. Sci. Paris 324 (Ser. Ila): 677-684. Murray, J. A. (1887) A rough distribution table of the birds of British India and its dependencies. The Indian Annals and Magazine of Natural Science 1(1). Bombay: Education Society's Press. Olson, S. L., Angle, J. P., Grady, F. V. & James, H. F. (1987) A technique for salvaging anatomical material from study skins of rare or extinct birds. Auk 104: 510-512. Pande, S. A., Pawashe, A. P., Kasambe, R. & Yosef, R. (201 1 ) Discovery of a possible hybrid of the Critically Endangered Forest Owlet Athene blewitti and Spotted Owlet Athene brama (Aves: Strigiformes) from northern Maharashtra, India. J. Threatened Taxa 3(4): 1 727-1 730. Peters, J. L. (1940) Check-list of birds of the world, 4. Cambridge, Massachusetts: Harvard Univ. Press. Rahmani, A. R.& Jathar, G. A. (2004) Ecological studies of the Forest Spotted Owlet Athene ( Heteroglaux ) blewitti. Final report (Bombay Natural History Society), unpublished. Rasmussen, P. C. & Anderton, J. C. (2005) Birds of South Asia: the Ripley guide. Washington, D.C. and Barcelona: Smithsonian Institution and Lynx Edicions. Rasmussen, P. C. & Collar, N. J. (1 998) Identification, distribution and status of the Forest Owlet Athene (Heteroglaux) blewitti. Forktail 14: 41-49. Rasmussen, P. C. & Collar, N. J. (1999) Major specimen fraud in the Forest Owlet Heteroglaux (Athene auct.) blewitti. Ibis 141: 1 1-21. Rasmussen, P. C. & Ishtiaq, F. (1 999) Vocalizations and behaviour of the Forest Owlet Athene (Heteroglaux) blewitti. Forktail 1 5: 61-65. Ripley, S. D. (1 961 ) A synopsis of the birds of India and Pakistan together with those of Nepal, Sikkim, Bhutan and Ceylon. Bombay: Bombay Natural History Society. Ripley, S. D. (1976) Reconsideration of Athene blewitti (Hume). J. Bombay Nat. Hist. Soc. 73: 1-4. Sharpe, R. B. (1891) Scientific results of the Second Yarkand Mission; based upon the collections and notes of the late Ferdinand Stoliczka, Ph.D. Aves. London:Taylor & Francis. Sharpe, R. B. (1899) A hand-list of the genera and species of birds [Nomenclator Avium turn Fossilium turn Viventium.]\/o\. 1. London: Trustees. Voous, K. H. (1989) Owls of the Northern Hemisphere. Cambridge, Massachusetts: MIT Press. Weesie, P. D. M. (1982) A Pleistocene endemic island form within the genus Athene: Athene cretensis n.sp. (Aves, Strigiformes) from Crete. Proc. Koninkl. Nederl. Akad. Wetensch., ser. B, 85(3): 323-336. Weick, F. (2006) Owls (Strigiformes): annotated and illustrated checklist. Berlin: Springer-Verlag. Wink, M., Sauer-Gurth, H. & Fuchs, M. (2004) Phylogenetic relationships in owls based on nucleotide sequences of mitochondrial and nuclear marker genes. Pp.483-495 in R. D. Chancellor & B.-U. Meyburg, eds. Raptors worldwide. Berlin: World Working Group on Birds of Prey and Owls. Wink, M„ El-Sayed, A. -A., Sauer-Gurth, H. & Gonzalez, J. (2009) Molecular phylogeny of owls (Strigiformes) inferred from DNA sequences of the mitochondrial cytochrome b and the nuclear RAG-1 gene. Ardea 97: 581-591. Wolters, H. E. (1975) Die Vogelarten der Erde. Hamburg: Paul Parey. Yosef, R„ Pande, S. A., Pawashe, A. P., Kasambe, R. & Mitchell, L. (2010) Interspecific interactions of the critically endangered Forest Owlet ( Athene blewitti). Acta Ethologica 1 3: 63-67. P. C. RASMUSSEN, Department of Zoology and Michigan State University Museum, Michigan State University, East Lansing, Ml 48824, USA; and Bird Group, Department of Life Sciences, Natural History Museum, Akeman St, Tring, Herts HP23 6AP, UK. Email: rasmus39@gmail.com N. J. COLLAR, BirdLife International, Wellbrook Court, Girton Road, Cambridge CB3 ON A, UK; and Bird Group, Department of Life Sciences, Natural History Museum, Akeman St, Tring, Herts HP23 6AP, UK. Email: nigel.collar@birdlife.org FORKTAIL 29 (201 3): 88-93 The Moluccan Woodcock Scolopax rochussenii on Obi Island, North Moluccas, Indonesia: a 'lost' species is less endangered than expected H. EDEN W. COTTEE-JONES, JOHN C. MITTERMEIER & DAVID W. REDDING The Moluccan Woodcock Scolopax rochussenii is an enigmatic forest wader endemic to the North Moluccas, Indonesia. Until recently, the species was known from fewer than ten confirmed records and it is currently considered to be endangered under the criteria of the International Union for the Conservation of Nature (IUCN). In July-August 201 2, field surveys were conducted at 20 sites and semi-structured interviews held in seven villages to assess the status of the Moluccan Woodcock on Obi Island, Maluku Utara province. Field surveys resulted in 51 records of minimum 13 individuals, and the findings suggest that this species is widespread on Obi, occurring from 15-1,150+ m. Contrary to the existing assumption that the Moluccan Woodcock is a montane species, the data indicate that it is primarily a lowland species, and that population densities decline with altitude. The species tolerates minor habitat disturbance, such as selective logging and small-scale agriculture, and does not appear to be hunted or frequently trapped by local people. A Maxent species distribution model indicates that Moluccan Woodcock distribution correlates strongly with the presence of rivers and streams and predicts 9,530 woodcock territories on Obi. The primary threats to the species are severe habitat disturbance from mining and logging, and better environmental mining regulations need to be enforced to safeguard habitat on Obi. It is recommended that the Moluccan Woodcock be reassessed as vulnerable following IUCN criteria, and that surveys following the same protocol should be conducted on Bacan to clarify the status of the species on that island. INTRODUCTION The woodcock (genus: Scolopax ) consist of eight extant species; two migratory species with large ranges in North America (American WoodcockA minor ) and Eurasia (Eurasian Woodcock A rusticola ), and six inhabiting islands in tropical East Asia and New Guinea (Olson 1979, Piersma 1996). Of these, the Ryukyu Woodcock S. mira is confined to a handful of small islands in the Ryukyu archipelago, Japan, and another four species (Bukidnon Woodcock S. bukidnonensis. New Guinea Woodcock S. rosenbergii , Javan Woodcock S. saturata , Sulawesi Woodcock S. celebensis) are restricted to montane forest on large islands in the Philippines, Indonesia and New Guinea (Piersma 1996, Kennedy et al. 2001). The final species, the Moluccan Woodcock S. rochussenii , is known from two small islands in the North Moluccas, Indonesia, and is the largest and least known member of the genus (Hayman et al. 1991, Coates & Bishop 1997). The Moluccan Woodcock was first collected by Heinrich Bernstein, who obtained a single male specimen from Obi in 1862, but did not live to see the species named (Jansen 2008). Bernstein died of illness in New Guinea in 1 865 and ‘ Scolopax rochussenii was not described until 1 866 (Schlegel) when his specimen arrived back at the Museum of Natural History in Leiden, the Netherlands. Over the next 150 years, only seven additional individuals were recorded, six from Obi and a single individual from Bacan in 1902, and following two birds collected in 1982, the species disappeared for nearly 30 years. Ornithologists visiting Obi in 1 989, 1 992 and 2010 (Lambert 1992, Linsley 1994, Bashari 2011) failed to record the bird and a targeted search for the species on Bacan in 2010 also produced no records (Lagerveld 2010). In this same year, however, the species was ‘rediscovered’ at two localities on Obi by M. Thibault et al. (2013) and its vocalisations were recorded for the first time. Given this paucity of records, virtually nothing is known of the distribution, breeding behaviour or feeding habits of the species. Most information regarding its ecology has been based on assumed similarities to other Scolopax species, in particular those on neighbouring islands in Indonesia and New Guinea. Despite the collection of at least one Moluccan Woodcock in lowland habitats (BirdLife International 2001), the species has been assumed to be a montane species restricted to high elevation forest, a distribution that would seemingly explain why it has been recorded so infrequently (White & Bruce 1986, Coates & Bishop 1 997, BirdLife International 2013). As a result of the few known records and the relatively small area of montane habitat on Obi and Bacan, the Moluccan W oodcock is currently considered Endangered (BirdLife International 2013). From 5 July to 27 August 2012 the first field study of the Moluccan Woodcock on Obi Island was conducted and the species was observed on 51 occasions. Here the distribution, display behaviour and population size of this enigmatic species are reported and the impact of these findings on its conservation status are discussed. METHODS Field surveys Field surveys were conducted at 20 sites around Obi. Localities included all major habitat types on the island, and covered an elevational range from sea level to 1,550 m (Mittermeier et al. 2013). At each site, dawn and dusk survey points (n=60) were coupled with daily field observations (total 630 hours). Surveys were designed to maximise the possibility of encountering a Moluccan Woodcock and, when possible, dawn and dusk surveys were carried out along ridgelines or in open areas (such as river beds or forest clearings) where observers could scan for displaying birds. Morning observations began about 10 minutes before first light and continued until about 10 minutes after sunrise (05h30- 06hl0); evening observations were from about 20 minutes before sunset to about 10 minutes after dark ( 1 8h40— 1 9h 15). At each location, the number of individuals, detection method, and observation times were recorded along with habitat information including elevation, level of disturbance and the presence of nearby streams or swamps. Displaying woodcock were almost always detected by their call, and therefore the field of view at survey points could not be controlled. At three locations where it was possible to track a Moluccan Woodcock over the course of its entire display, the GPS points at the territory boundaries were Forktail 29(2013) The Moluccan Woodcock Scolopax rochussenii on Obi Island, North Moluccas, Indonesia 89 marked in order to estimate the size of the territory (see Discussion). Sites with swamps, streams with a width of greater than 3 m, or swamp forest located within 1 00 m of the point count were classified as wetlands. Level of disturbance was qualitatively assessed with primary forest defined as ‘undisturbed’, areas with small agricultural clearings and light logging defined as ‘minor disturbance’ and areas with mining, extensive cultivation or extensive and recent logging (within the last 5 years) defined as ‘major disturbance’. A generalised linear model with a log-link function and a Poisson error structure was used to identify any significant relationships between these variables and the number of Moluccan Woodcock recorded at the survey sites. Distribution and population assessment Field data were used to construct a model predicting the distribution of the Moluccan Woodcock on Obi and Bacan. The data for the model was downloaded as raster layers: 1 9 climatic variables relating to temperature and rainfall (Worldclim, 30 arc second resolution, WGS84 projection; Hijmans et al. 2005), altitudinal data (Worldclim, 30 arc second resolution, WGS84 projection; Hijmans etal. 2005), land cover data (Globcover, 300 m x 300 m resolution, WGS84 projection; Globcover 2009) , world geopolitical boundaries (Digital Chart of the World, 1 km2 resolution, WGS84 projection; Danko 1992), and hydrological information (Hydro IK, 1 km2 resolution, WGS84 projection; Verdin etal. 201 1). All input layers were resampled to 300 m x 300 mcell size using linear interpolation (resample, Rpackage raster; Hijmans & van Etten 2012) to maintain the resolution of the finest scale data (Globcover), and cropped with a bounding box of latitude 0-2°S and longitude 127-129°E (crop, R package raster; Hijmans & van Etten 2012). Rather than using raw values for two of the Hydro IK data layers, the distance of each grid cell to either a river (flow accumulation) or wet area (compound topographic index) was calculated (distance, R package raster; Hijmans & van Etten 2012). Species distribution models were estimated using Maxent (Phillips et al. 2004) with presence-absence points taken directly from the field data. All the 19 Worldclim, Hydro lk and land cover data layers were entered as predictor variables; atotal of21 variables. A total of 100 Maxent runs were done, each time using a random subset of the data as either training (4/5) or testing sets (1/5), and then a mean probability surface was calculated across those 100 runs. The ability of each of the 100 training datasets to predict the locations of the corresponding test datasets was measured using the ‘area under operating curve’ approach. This gives a value between 1, where the locations of testing sets are perfectly predicted by the niche model that was created using the training set, and 0 where the probability of occurrence of the niche surface is random with respect to the testing set. Given that the approximate area of each territory estimated in the field was close to 1 0 ha, the probability of occurrence grid output from Maxent was used to create a rough estimate of population size. The assumption was made that every thousand grid cells covering land was a potential range site and that the probability of occurrence taken from Maxent was the chance that this site was occupied by a single individual. Based on these assumptions, therefore, the summed probability from the model equated to an estimate of minimum population on the island. Although ranges will not, in reality, be shaped or arranged in such a uniform manner, given that occupancy will in many cases be greater than a single individual, this approach remains a conservative estimate of population size and means that errors resulting from the assumptions will be unlikely to cause an overestimation of the number of woodcock on the island. Interviews Field surveys were supplemented by 46 semi-structured interviews in seven villages around Obi. Interviews were conducted in Bahasa Indonesia with the help of three students from the University of Indonesia (Christine Endang Purba, Eka Hesdianti and Nova Maulidina Ashuri). Interviewees were selected opportunistically or following recommendations from the local village head. Interviews commenced with several background questions including religious affiliation, age, hometown, activities pursued in the forest, estimated frequency of visits to the forest and the amount of time spent there. Next, interviewees were shown a series of pictures of Indonesian birds and asked whether the species occurred on Obi and if so what the local name was and where it could be found, and whether they hunted or caught it. Pictures were colour photocopies taken from plates in Coates & Bishop (1997) and featured the Moluccan Woodcock, Drummer Rail Habroptila wallacii , White Cockatoo Cacatua alba, Dusky Megapode Megapodius freycinet, Maleo Macrocephalon maleo and Common Sandpiper Actitis bypoleucos , in no particular order. Respondents’ familiarity with the Moluccan Woodcock illustration was analysed for significance using a chi-squared test, and their ability to correctly identify the Moluccan Woodcock compared to the other illustrated species was analysed using Cochran’s Qtest. Using a binary logistic regression model, several variables were tested against the ability to correctly identify the Moluccan Woodcock, namely: whether someone had lived their entire life on Obi, how often they visited the forest, how well they identified other species in the interview, whether they hunted or trapped birds, their religion, and their village of origin. RESULTS Display behaviour During display a single bird would fly quickly with shallow wing beats at a height of about 10 m above the canopy, vocalising at regular intervals (Plate 1). Vocalisations consisted of an explosive even trill, lasting 0. 1 -0.6 seconds in duration and given at intervals of 1.9-3. 2 seconds. When two Moluccan Woodcock encountered each other during display flights (believed to be territorial conflicts), interactions included short parallel flights and a descending, twittering call, without overlapping into the adjacent bird’s display area (Macaulay Library of Natural Sounds LNS 182223). Display areas followed the course of a river or stream or circled above areas of swampy habitat. In the highlands, displaying birds were observed flying up to the headwaters of a stream before looping back to follow the course of a valley. Displays covered a large area, and flight patterns were either generally circular (in more open environments) or linear (following narrow valleys), but did not appear to follow a consistent pattern. In open areas, displaying birds were recorded passing above an observer on average every 3.8 minutes (n=15 display flights), and disappearing out of sight in the intervening period. In other conditions, birds would double back and fly over more frequently. Displaying Moluccan Woodcock were not disturbed by people, and would occasionally fly directly overhead with heads tilted to look down at observers. While there was a slight variation in timings, morning display flights would typically run from 05h25-06h02 (mean length 21 minutes), and evening display flights would usually take place from 18h47-19h06 (mean length 13 minutes). Ambient recordings of complete dawn and dusk display flights recorded by JCM are available online from the Cornell Lab of Ornithology Macaulay Library of Natural Sounds (http://macaulaylibrary.org) . Distribution and population The Moluccan Woodcock was recorded on 51 occasions at 1 1 sites (Figure 1). They were recorded almost exclusively during their dawn and dusk display flights, during which they were both consistent and conspicuous. At seven sites where several surv ey days JOHN C, MITTERMEIER JOHN C. MITTERMEIER 90 H. EDEN W. COTTEE-JONES, JOHN C. MITTERMEIER & DAVID W. REDDING Forktail 29 (2013) Plate 1 . Moluccan Woodcock photographed during display flights over the Cabang River, south of Kampon Buton, Obi, 26 August 201 2. were spent near a displaying bird, it was recorded every day at both dawn and dusk regardless of weather conditions. Non-displaying birds, on the other hand, were extremely inconspicuous and difficult to locate. During 630 hours of field surveys, only one non¬ displaying individual was observed. This bird was flushed byJCM from an area of mossy boulders and pools along the edge of a stream, in primary montane forest at 930 m on 3 August (Plate 2). Figure 1. A species distribution model for the Moluccan Woodcock on Obi Island. Darker colours indicate areas of more suitable habitat, and circles identify field survey sites where woodcock were present (closed circles) and not recorded (open circles). The Moluccan Woodcock was recorded at sites between 15 and 1,150 m elevation in a range of habitats including primary lowland forest, selectively logged secondary forest, swamp forest, secondary forest with small agricultural clearings and montane forest. A generalised linear regression model found no significant difference in Moluccan Woodcock densities in areas with minor disturbance relative to undisturbed habitats (Table 1); indeed the birds were frequently present in selectively logged forest and areas with small-scale agriculture and agroforestry. Moluccan Woodcock were not, however, recorded in areas with major disturbance (e.g. extensive agriculture and mining). While the model found no significant relationship with altitude, raw point count data illustrate that the Moluccan Woodcock was more frequently recorded in the lowlands than the highlands (Figure 2). The strongest correlation identified by the generalised linear model was with the presence of streams and rivers (Table 1). The Maxent models created using the training data showed a good ability to predict the testing data, with a mean receiver operating characteristic (ROC) value of 0.833 with a standard deviation of 0.11 over 100 replicates. The influence of waterways was also reflected in this distribution model (Figure 1) where ‘distance to major river’ explained more than 80% of the variation using both the Hydro IK data set and the Digital Chart of the World. The remaining variation in the model was explained by land cover type (~20%), and in a few models, by mean diurnal temperature range (~5%). The variable ‘Distance to rivers’ was consistently the highest loading variable (with an average of 95%), with habitat type (2%) and daily temperature range (2.5%) also being consistently important. Based on data from individual displaying birds at three sites, the mean territory size for a Moluccan Woodcock was estimated to be 10.67 ha (SE = 2.3, range=7.6-13). By applying this territory size to the modelled distribution, a mean estimate of 9,530 JOHN C. MITTERMEIER JOHN C. MITTERMEIER JOHN C. MITTERMEIER Forktail 29(2013) The Moluccan Woodcock Scolopax rochussenii on Obi Island, North Moluccas, Indonesia 91 Plate 2. Two habitats where Moluccan Woodcock were found on Obi: a) a fast-flowing mountain stream in primary forest at 930 m and b) logged forest bordering the Cabang River at 35 m, July 201 2. Figure 2. The mean number of Moluccan Woodcock recorded at survey sites in lowland and highland elevations. 0.9 - Table 1. Generalised linear regression model results for Moluccan Woodcock habitat associations. Parameter fi(SE) p-value Intercept 0.71 (0.79) 0.37 Altitude -0.001 (0.0) 0.17 Water absence -28.9 H - Light disturbance 0.03 (0.8) 0.97 Severe disturbance -2.22(1.26) 0.08 Deviance=0.83, Pearson Chi-square|H|=10 .17, p=0.73 100 89 Dusky Megapode Common Sandpiper Drummer Rail White Cockatoo (standard deviation=282, n= 1 00) woodcock territories on Obi was obtained. Interviews Interview results indicate that the majority of Obi residents are unfamiliar with the Moluccan Woodcock. With the exception of Maleo, which does not occur on Obi, all species in the survey were identified significantly more often than the woodcock (Figure 3). This even included Drummer Rail, a notoriously secretive species that until this study was unknown on Obi (Mittermeier et al. 2013). Almost 83% of respondents did not identify the Moluccan Woodcock, while nearly 11% identified it as a coastal shorebird found on the beach or in open areas along waterways (X2(1)=34.78, p<0.005). The lack of respondents accurately identifying the species constrained the ability of the model to select predictors, and none of the explanatory variables was significant. In total, only three people stated that they were familiar with the woodcock and that it was found in the forest on Obi. Of these, one reported that he had caught a bird in a snare and then released it because it did not look good to eat, and a second said that he saw it frequently flying at dawn and dusk near his rice fields, but that he believed it ate fruit in the treetops and that he had never seen it on the ground. The third, a parrot trapper and a very astute observer, noted that he occasionally saw the species eating worms in muddy areas along the margins of rivers and that he called the woodcock wapichu (transcribed phonetically). No interviewees reported hunting or eating this species. Figure 3. The percentage of interviewees (n=46) who correctly identified Indonesian birds using picture prompts taken from plates in Coates & Bishop (1997). Moluccan Woodcock vs Dusky Megapode: Cochran's Qn =38.00*, vs Common Sandpiper: Cochran's Q(] =29.00*, vs Drummer Rail: Cochran's Qm=16.00*, vs White Cockatoo: Cochran's Q . =32.00*, vs Maleo: Cochran's Q(1 =0.20. * p<0.005. 7 7 Maleo Moluccan Woodcock JOHN C. MITTERMEIER 92 H. EDEN W. COTTEE-JONES, JOHN C. MITTERMEIER & DAVID W. REDDING Forktail 29 (2013) DISCUSSION Ecology and display The results indicate that the Moluccan Woodcock occurs throughout Obi. It is found at higher densities in the lowlands and favours areas near water particularly in the vicinity of streams and rivers. This close association with waterways reflects observations of the Sulawesi Woodcock, which has been reported to forage along forest stream banks (Mole & Wangko 2006). The elevational distribution of the Moluccan Woodcock, with the species notably more common in lowland habitats, contrasts with that of related Scolopitx species on the larger islands of Sulawesi, Java, New Guinea and the Philippines. As the Moluccan Woodcock associates strongly with waterways, it is possible that this distribution reflects the lower density of large rivers at higher elevations. While this lowland distribution overturns existing assumptions, it is not overly surprising; many birds in the Indo-Pacific that are restricted to the highlands on large islands are found in lowland habitats on oceanic islands (Mayr & Diamond 2001). The display flights of different woodcock species show substantial plasticity. The American Woodcock performs a unique display involving a terrestrial ‘peenting’ call followed by a vertical display flight (Duke 1966), several male Eurasian Woodcock perform ‘roding’ display flights over the same area of woodland competing for females polygamously (Hirons 1980, Hoodless et al. 2009), and the Ryukyu Woodcock does not seem to perform any display flight at all (BirdLife International 2001). In the absence of any evidence to suggest otherwise, this study indicates that the Moluccan Woodcock performs territorial display flights, and suggests that territories abut without overlapping, similar to the Bukidnon Woodcock (Kennedy et al. 2001). Population The lack of ecological data on the Moluccan Woodcock and the degree of variation in the breeding behaviour of other woodcock species makes it difficult to establish how many individual woodcock might be present in a single territory. At the most conservative, an estimate of one individual per territory predicts a total population of 9,530 individuals. A more realistic estimate, though still very conservative, would be two individuals per territory for a minimum population of 19,059 individuals. However, the relationship between the number of displaying individuals and the total population is unknown even in well- studied woodcock species (Hoodless et al. 2009), and so this population estimate for the Moluccan Woodcock should be regarded as preliminary. Obtaining a more accurate population estimate for the species will require a better understanding of the relationship between the number of displaying birds and the total number of individuals in a given area, the degree of variation in territory size, and clarifying the status of the bird in several areas not surveyed. Neither the eastern side of Obi nor the outlying island of Bisa were visited, as part of eastern Obi had been surveyed by F. Lambert in 1992, and current reports indicated that the rest had been converted into coconut groves. Due to the lack of rivers in these two areas, the distribution model indicates that they are unsuitable for woodcock. Whether this is accurate should be confirmed by future field surveys. The status of the Moluccan Woodcock on Bacan is also important. Bacan (1,900 km2) is smaller than Obi (2,500 km2) and could theoretically hold a similar population of the woodcock. However, the only known record for Bacan is a female collected from an unspecified location in 1902 (BirdLife International 2001). Bacan has been surveyed more often than Obi (White & Bruce 1986), Alfred Russel Wallace spent six months collecting there in 1858-1859 (Wallace 1869) and it is rather surprising that it has not been recorded there again for more than a century. That said the distribution model does identify significant areas of suitable habitat on Bacan, particularly along large rivers in the central part of the island. Surveys on Bacan, and also Halmahera, using the same methodology should be undertaken to ascertain whether the Moluccan Woodcock occurs on other islands in the North Moluccas. If Moluccan W oodcock are not present on these islands, it would be appropriate to revert to the former common name of Obi Woodcock. Local knowledge People on Obi often had a detailed knowledge of the local avifauna, particularly of species that were trapped, such as parrots, and terrestrial birds caught in snares, such as rails and megapodes. For example, several forest workers near Kampon Buton and Wayaloar readily identified 5-6 species of rail. In this context, the limited knowledge of Moluccan Woodcock is somewhat surprising. This may partly be due to the survey method (Diamond & Bishop 1 999). The illustration of Moluccan Woodcock in Coates & Bishop (1997) displays the species in daylight on the ground, a context in which it is apparently rarely seen. In the wild, the vast majority of observations are of a silhouetted bird, flying in poor light and giving its distinctive call. It is recommended that any future interviews to determine local knowledge of Moluccan Woodcock use photographs and sound recordings rather than the standard illustrations. An important implication of this result, however, is strong support for the fact that people do not regularly hunt or eat the species and therefore have limited opportunities to see it in the hand. This is critical to the conservation of the species. In other parts of the world, woodcock are frequently hunted and considered a good source of meat. Given the conspicuous nature of the Moluccan Woodcock’s display flights, a change in attitudes coupled with an increase in hunting on Obi could lead to a rapid decline in the bird’s population. Conservation The population and distribution estimates for the Moluccan Woodcock indicate that the number of individuals and area of occupancy for this species are both above the established thresholds for listing as Endangered (IUCN 20 1 2) . With an area of occupancy of more than 2,000 km2 on Obi and the number of territories estimated to be 9,530, this species qualifies to be reassessed as Vulnerable under the IUCN criteria. Although the species can tolerate minor habitat disturbance, the current spread of mining on the island could represent a significant threat. The ultrabasic rock formations are rich in nickel ore, and nickel mining currently takes place on a large scale in the Kawasi area of western Obi. Moluccan Woodcock was not found in this degraded mining landscape, and the expansion of mining on Obi poses a major threat to the Moluccan Woodcock and other species. Industrial nickel mining was due to expand into the foothills north of Tapaya village, but this has been suspended, possibly due to a recent government quota on the export of unprocessed material; however, it seems likely that it may soon commence. More sophisticated approaches to the regulation of the impact of mining on biodiversity and the restoration of mined areas need to be a priority for conservation on Obi. A protected area has been proposed in the mountainous centre of Obi, but these results demonstrate that a highland protected area is unlikely to contain a high density of Moluccan Woodcock. Conservation efforts need to consider both montane and lowland habitats (see Mittermeier et al. 2013). Fortunately, Moluccan Woodcock appear to tolerate a relatively high degree of habitat degradation including selective logging, agroforestry, and small- scale agriculture, and these habitats should also be considered in devising a conservation plan for the species. In addition the display behaviour of the Moluccan Woodcock makes it potentially Forktail 29 (2013) The Moluccan Woodcock Scolopax rochussenii on Obi Island, North Moluccas, Indonesia 93 vulnerable to hunting. While no evidence of hunting was observed, high rates of immigration to the island, linked to the expansion of mining, could change this. ACKNOWLEDGEMENTS Our expedition to Obi was generously supported by a National Geographic Society/Waitt Grant, a Ron & Mary Neal LSU Graduate Fellowship, a Thesiger Award from the Old Etonian Association, a small fieldwork grant from the Royal Geographical Society (with IBG), a Graham Hamilton travel grant from St Edmund Hall, the Oxford University Expeditions Council and A.J. Tours & Travel. Our grateful thanks go to Christine Endang Purba, Eka Hesdianti and Nova Maulidina Ashuri for their hard work in the field. Thanks go to our supervisors and referees, in particular: Frederick Sheldon, Robb Brumfield, Robert J. Whittaker, Pauljepson, Rich Grenyer, Shonil Bhagwat, Maan Barua, Stuart Butchart,Jatna Supriatnaand Kate Jones. Frank Lambert, Hanom Bashari, Marc Thibault, Diah Asri, David Bishop, Dewi Prawiradilaga, Mohammad Irham, Richard Noske and Jared Diamond provided advice on the conduct of fieldwork on Obi. Finally, our thanks go to the people of the island for their hospitality, specifically BambangSetiawan, Pak Uspa, Pak La Gode, La Ham, Adam, Ikhsan and Nisha and her family. H. Eden W. Cottee-Jones and John C. Mittermeier contributed equally to this work. REFERENCES Bashari, H. (2011) Rediscovery of Carunculated Fruit Dove Ptilinopus granulifrons on Obi, North Moluccas. Birdi rig ASIA 16: 48-50. B i rd Life International (2001) Threatened birds of Asia: The BirdLife International Red Data Book. Cambridge UK: BirdLife International. BirdLife International (2013) Species factsheet: Scolopax rochussenii. Downloaded from www.birdlife.org on 10 February 2013. Coates, B. J. & Bishop, K. D. (1997) A guide to the birds ofWallacea. Alderley: Dove Publications. Danko, D. M. (1992) The digital chart of the world project. Photogramm. Eng. Rem. S. 58:1 1 25-1 1 28. Diamond, J. M. & Bishop, K. D. (1999) Ethno-ornithology of the Ketengban people, Indonesian New Guinea. Pp.1 7-46 in D. Medin & S. Atran, eds. Folkbiology. Cambridge Mass: MIT Press. Duke, G. E. (1966) Reliability of censuses of singing male Woodcocks. J. Wildl. Manage. 30: 697-707. Globcover (2009) GlobCover Land Cover v2 2008 database. Downloaded from http://ionia1.esrin.esa.int/index.asp on 20 February 2013. Flayman, P., Marchant, J. & Prater, T. (1 991 ) Shorebirds: identification guide to the waders of the world. London: Christopher Flelm. Fiijmans R. J. & van Etten, J. (201 2) Raster: Geographic analysis and modeling with raster data. R package version 2.0-12. Downloaded from http:// CRAN.R-project.org/package=raster on 20 February 201 3. Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G. & Jarvis, A. (2005) Very high resolution interpolated climate surfaces for global land areas, /nr. J. Climatol. 25: 1965-1978. H irons, G. (1980) The significance of roding by Woodcock Scolopax rusticola: an alternative explanation based on observations of marked birds. Ibis 122:350-354. Floodless, A. N., Lang, D„ Aebischer, N. J., Fuller, R. J. & Ewald, J. A. (2009) Densities and breeding estimated of Eurasian Woodcock Scolopax rusticola in Britain in 2003. Bird Study 56: 15-25. IUCN (2012) IUCN Red List categories and criteria: version 3.1. Gland, Switzerland & Cambridge UK: IUCN. Jansen, J. (2008) Fleinrich Bernstein. BirdingASIA 10: 103-107. Kennedy, R. S., Fisher, T. H„ Harrap, S. C. B„ Diesmos, A. C. & Manamtam, A. S. (2001) A new species of woodcock (Aves: Scolopacidae) from the Philippines and a re-evaluation of other Asian/Papuan woodcock. Forktail 17: 1-12. Lagerveld, S. (2010) Op zoek naar de Molukse Houtsnip. (Looking for the Moluccan Woodcock) Downloaded from http://dutchbirding.nl/ news.php?ntype=1 7&id=71 2&lang=en on 10 November 2011. (In Dutch). Lambert, F. R. (1994) Notes on the avifauna of Bacan, Kasiruta, and Obi, Northern Moluccas. Kukila 7: 1-9. Linsley, M. D. (1995) Some bird records from Obi, Maluku. Kukila 7: 1 42— 151. Mayr, E. & Diamond, J. M. (2001 ) The birds of northern Melanesia: speciation, dispersal and biogeography. Oxford: Oxford University Press. Mittermeier, J. C„ Cottee-Jones, FI. E. W„ Purba, E. C., Ashuri, N. M., Flesdianti, E. & Supriatna, J. (201 3) A survey of the avifauna of Obi Island, North Moluccas, Indonesia. Forktail 29: 128-137. Mole, J. & Wangko, M. F. (2006) FHabitat of the Sulawesi Woodcock Scolopax celebensis in Lore Lindu National Park. Kukila 13: 64-66. Olson, S. L. (1979) Fossil woodcocks: an extinct species from Puerto Rico and an invalid species from Malta (Aves: Scolopacidae: Scolopax). Proc. Biol. Soc. Wash. 89: 265-274. Phillips, S. J., Dudik, M. & Schapire, R. E. (2004) A maximum entropy approach to species distribution modeling. Pp: 655-662 in Proceedings of the 21 st International Conference on Machine Learning. New York: ACM Press. Piersma, T. (1996) Family Scolopacidae (sandpipers, snipes & phalaropes). Pp.444-533 in J. del Floyo, A. Elliott & J. Sargatal, eds. Handbook of the birds of the world, 3. Barcelona: Lynx Edicions. Schlegel, FI. (1866) Observations zoologiques. Ned. Tijdschr. Dierk. 3: 257- 258. Thibault, M., Defos du Rau, P., Pineau, O. & Pangimangen, W. (2013) New and interesting records for the Obi archipelago (north Maluku, Indonesia), including field observations and a first description of the vocalisation of Moluccan \NoodcockScolopaxrochussenii. Bull. Brit. Orn. Club 133:83-1 15. Verdin, K. L. (2011) ISLSCP II HYDROlk Elevation-derived Products. Downloaded from http://daac.ornl.gov/ on 20 February 201 3. Wallace, A. R. (1 869) The Malay Archipelago. Singapore: Graham Brash. White, C. M. N. & Bruce, M. D. (1986) The birds ofWallacea. Tring: British Ornithologists' Union (Checklist 7). H. Eden W. COTTEE-JONES, Conservation Biogeography and Macroecology Programme, School of Geography and the Environment, Oxford University Centre for the Environment, South Parks Road, Oxford OX1 3QY, UK. Correspondence address: H.E.W. Cottee-Jones, St Edmund Hall, Queen's Lane, Oxford, OX1 4AR, UK. Email: henry.cottee-jones@seh.ox.ac.uk John C. MITTERMEIER, Museum of Natural Science, 1 1 9 Foster Hall, Louisiana State University, Baton Rouge, LA, 70803, USA. Email: john.mittermeier@gmail.com David W. REDDING, Department of Genetics, Evolution and Environment, University College London, Gower Street, London WC1E 6BT, UK. Email: dwredding@gmail.com FORKTAIL 29 (2013): 94-99 Taxonomic status of Blackthroat Calliope obscura and Firethroat C. pectardens PER ALSTROM, GANG SONG, RUIYING ZHANG, XUEBIN GAO, PAUL I. HOLT, URBAN OLSSON & FUMIN LEI The Chinese endemic breeders Blackthroat Calliope obscura and Firethroat C. pectardens are two of the world's rarest and least known 'chats' (Muscicapidae). They have been considered colour morphs of the same species (Firethroat, by priority), although they are nowadays usually treated as separate species. The taxonomic status of these two taxa is here investigated based on analyses of mitochondrial and nuclear DNA, vocalisations and reassessed distributions. Phylogenetic analysis confirms that they are sisters. Their genetic divergence (cytochrome b 6.4%, GTR+r+l corrected) is comparable to several other species pairs of 'chats'. Discriminant function analysis of songs correctly classified 88% of the recordings. The breeding ranges appear to be mainly parapatric. Based on congruent differences in morphology, songs and molecular markers, it is concluded that Blackthroat and Firethroat are appropriately treated as separate species. INTRODUCTION The Blackthroat Calliope obscura and Firethroat C. pectardens are two of the world’s rarest and least known ‘chats’. Males are uniformly blue-grey above, with blackish tail with white sides basally. As the names imply, male Blackthroat has a black throat, breast and side of the head, whereas male Firethroat has a ‘shining’ orange throat and breast with black sides, and black sides of the head. Male Firethroat also has a small white patch on the side of the neck. Females are, as usual in chats, much more cryptically coloured (Meyer de Schauensee 1984, MacKinnon & Phillipps 2000, Collar 2005, Rasmussen & Anderton 2012, Song et al. in press). Both breed in the mountainous regions of central China, Firethroat also in south-east Tibet and perhaps Arunachal Pradesh (Meyer de Schauensee 1984, Cheng 1987, MacKinnon & Phillipps 2000, Collar 2005, Rasmussen & Anderton 2012, BirdLife International 2013a, b). The non-breeding ranges are poorly known, but there are records of Blackthroat from north-west Thailand and of Firethroat from north-east India, northern Myanmar and Bangladesh (Cheng 1987, Collar 2005, Rasmussen & Anderton 2012, BirdLife International 2013a, b). Both species were described in the late nineteenth century, but there have been rather few records since then, especially of the Blackthroat, whose breeding grounds were only rediscovered in 2011 (Song et al. in press). The Blackthroat is classified as Vulnerable and Firethroat as Near Threatened (BirdLife International 2013 a,b). Blackthroat and Firethroat are usually placed in either Luscinia (Sibley & Monroe 1990, Dickinson 2003, Collar 2005, Gill & Donsker2013) or Erithacus (Ripley 1964). However, Luscinia sensu Dickinson (2003) was recently shown to be non-monophyletic and proposed to be divided into the genera Luscinia ( sensu stricto ), Larvivora , Tarsiger and Calliope , with Firethroat in the Calliope clade (Sangster et al. 2010). Blackthroat was not included in that study, but it was placed in Calliope due to its assumed close relationship with Firethroat (Sangster etal. 2010). The species status of Blackthroat has been questioned, and it has been considered a colour morph of Firethroat (Goodwin & Vaurie 1956, Cheng 1958, Vaurie 1959, Etchecopar & Hire 1983). This view was rebutted by Ripley (1958) based on morphological differences, and later authors have treated it as a distinct species (e.g. Ripley 1964, Sibley & Monroe 1 990, Dickinson 2003, Collar 2005, Gill & Donsker 2013). However, the relationship between the Blackthroat and Firethroat has not yet been properly studied. As both are very rare (BirdLife International 2013a,b, Song et al. in press), it would be helpful for conservation purposes if their taxonomic status could be clarified. Here the taxonomic status of the Blackthroat and Firethroat is discussed, based on analyses of mitochondrial and nuclear DNA and songs, and the distributions of the two species are reviewed. It is concluded that Blackthroat and Firethroat are best treated as separate species. MATERIAL AND METHODS Sequencing and phylogenetic analyses Total genomic DNA was extracted from an adult male Blackthroat collected in Foping, Shaanxi province (33.693°N 107.849°E), in June 201 1 (Institute of Zoology, Chinese Academy of Sciences, Beijing No. IOZ 62531) using the QIAamp DNA Mini Kit (Qiagen) following the manufacturer’s protocol. Partial sequences of the mitochondrial cytochrome b and flanking tRNA-Thr (hereafter cyt b) were obtained through PCR amplification. The primer pair L14851 and H 16058 (Groth 1998) was used for cyt b, with annealing temperature 46-48°C. We also amplified two nuclear markers, myoglobin intron 2 (Myo) and ornithine decarboxylase (ODC). Primer pair myo2 and myo3F (Kimball et al. 2009) were used for Myo, and primer pair OD6 and OD8R for ODC (Allen & Omland 2003), with annealing temperatures 55°C and 59°C, respectively. PCR products were purified using QIAquick PCR purification Kit (Qiagen). Sequencing was carried out using an ABI 3730 automatic sequencer following the ABI PRISM BigDye Terminator Cycle Sequencing protocol. Both strands were sequenced using the same primers as for PCR. All sequences have been submitted to GenBank (Table SI). For the phylogenetic analyses, sequences of Firethroat and 10 other chats, all of which have been placed in the genus Luscinia (Dickinson 2003), and two more distant outgroup species (choice based on Sangster et al. 2010), were obtained from GenBank (Table Si). Sequences were aligned and checked manually in MEGA4 (Tamura et al. 2007). The phylogeny was estimated by Bayesian inference using MrBayes 3.2 ( H uelsenbeck & Ronquist 200 1,2010; Ronquist et al. 2011). All loci were analysed separately, as well as concatenated. In the multi-locus analyses, the data were partitioned by locus, using rate multipliers to allow different rates for the different partitions (Ronquist & Huelsenbeck 2003, Nylander et al. 2004). Appropriate substitution models were determined based on the AIC (Akaike information criterion: Akaike 1973) as calculated in MrModeltest2 (Nylander 2004). For cy tb, the general time-reversible (GTR) model (Lanave et al. 1984, Tavare 1986, Rodriguez et al. 1990), assuming rate variation across sites according to a discrete gamma distribution with four rate categories (T; Yang 1994) and an estimated proportion of invariant sites (I; Gu etal. 1995), was selected. For Myo, the HKY model (Hasegawa Forktail 29 (2013) Taxonomic status of Blackthroat Calliope obscura and Firethroat C. pectardens 95 et al. 1985) + T and for ODC the GTR + T model were selected. Two simultaneous runs of four incrementally heated Metropolis- coupled MCMC (Markov Chain Monte Carlo) chains were run for 5 million generations and sampled every 1,000 generations in MrBayes. Convergence to the stationary distribution of the single chains was inspected using a minimum threshold for the effective sample size. Joint likelihood and other parameter values were inspected in Tracer 1.5.0 (Rambaut & Drummond 2009) and indicated large effective sample sizes (>1,000). Good mixing of the MCMC and search reproducibility were established by running each analysis at least twice, and topological convergence was examined by eye and by the average standard deviation of split frequencies (<0.0l). The first 25% of the generations was discarded as ‘burn-in’, well after stationarity of chain likelihood values had been established, and the posterior probabilities were calculated from the remaining samples. Pairwise sequence divergences among all 12 chats were calculated in PAUP* (Swofford 2002) for all individual loci, following the recommendations of Fregin et al. (2012), i.e. by comparing homologous parts of the genes (same parts, same lengths), deleting all positions with any uncertain base pairs from the matrix, and using the best-fit model (same as the model used in phylogenetic analyses; choice of model determined in MrModeltest2 [Nylander 2004] with the two distant outgroup species, Oriental Magpie Robin Copsychus malabaricus and Spotted Flycatcher Muscicapa striata , excluded) . The shape parameter alpha and estimated proportion of invariable sites were obtained from a Bayesian Inference, since PAUP* cannot estimate these parameters under the distance criterion. Vocalisations Songs of a total of 18 male Blackthroats (17 from Foping, one from Changqing) were recorded in June 201 1 and May 2012 using a Marantz PMD661 solid state recorder and a Telinga Pro Twin Science microphone (five individuals, by PA), a Marantz PMD661 solid state recorder and Sennheiser MKFI416 (five individuals, by XG), and a Sound Devices 722 hard drive recorder with a Telinga Pro 7 parabolic microphone (eight individuals, by PIH). These recordings were compared by ear and by inspection of sonograms to 1 1 recordings of songs of Firethroat, all from Sichuan, China (seven from Wolong, one from Longcangguo, two from Jiuding Shan and one from Hailuogou — six by PA, three by PIH and two from www.xeno-canto.org). Sonograms were produced in Raven Pro 1.4 (Cornell Laboratory of Ornithology 2012). The following measurements were taken on entire song strophes of 14 Blackthroats and eight Firethroats: A time (s) = duration; A frequency (Hz) = frequency range; minimum frequency (Hz); maximum frequency (Hz); and number of elements. Discriminant function analysis (DFA) of the song variables was carried out in SPSS Statistics version 20 (IBM Corp.) ; mean values for each male were used as input in these analyses, as there were few individuals but several different strophes recorded per individual. Three of PA’s recordings of Blackthroat are available on-line at www.xeno- canto.org (XC9 1 803, XC 1 804, XC 143220) as well as www.slu.se/ per-alstrom-research; two of PA’s recordings of Firethroat are available on Xeno-canto (XC 143218, XC 1432 19); and all of PIH’s recordings have been deposited at the National Sound Archive, London. Distributions Records of Blackthroat were taken from Song et al. (in press). To revise the distribution of Firethroat we reviewed the literature, as well as the BirdLife International species database (http:// www.birdlife.org/), China Bird Report (http://birdtalker.net/ birdtalker/report/index.asp?lan = 0), a database of the birds of China (http://www.cnbird.org.cn/first.htm), Chinese bird gallery (http://www.wwfchina.org/birdgallery), Oriental Bird Images (http://orientalbirdimages.org), the Internet Bird Collection (http://ibc.lynxeds.com) and museum collections that we thought might hold specimens of Blackthroat or Firethroat (museums in the USA searched through ORNIS: http://ornis2.ornisnet.org). RESULTS Molecular analyses We obtained 1,076 bp of cyt£>, 664 bp (674 bp aligned) of Myo and 705 bp (737 bp aligned) of ODC from Blackthroat. The tree based on the concatenated sequences (Figure 1) showed Blackthroat and Firethroat to be sisters with strong support. These were, in turn, inferred to be sisters to Siberian Rubythroat Calliope calliope and White-tailed Rubythroat C. pectoralis, with high support. The sister relationship between Blackthroat and Firethroat was strongly supported in single-locus analyses of all three markers (not shown). Figure 1. Blackthroat Calliope obscura and Firethroat C. pectardens are sister species, as shown in this phylogenetic tree of all Calliope species sensu Sangster et al. (2010) (grey shade) and a selection of outgroup species. The tree is based on concatenated mitochondrial cytochrome b and nuclear myoglobin intron 2 and ODC introns 6-7 sequences (see Sangster etal. [2010] and Zuccon & Ericson [2010] for a broader taxon sampling within Muscicapidae). Values at nodes represent Bayesian posterior probabilities; * indicates posterior probability 1 .00. - 0.99 * — Calliope pectoralis - Calliope pectardens * - Calliope obscura 0.96 * * 0.0080 - Tarsiger hyperythrus - * - Tarsiger rufilatus - Luscinia luscinia — * - Luscinia megarhynchos - Larvivora akahige * - Larvivora komadori - Larvivora brunnea - * - Larvivora cyane - Copsychus malabaricus Muscicapa striata The genetic divergences between Blackthroat and Firethroat (cytb 6.4%, Myo 0.32%, ODC 0.35%) were considerably lower in all three loci than in the majority of pairwise comparisons between the chats in the present dataset (Figure 2). However, they were comparable to the divergences between the well-accepted species pairs Larvivora cyane/ L. brunnea, L. akahige / L. komadori, Luscinia luscinia! L. megarhynchos and Tarsiger rufilatus/T. hyperythrus (Figure 2). Vocalisations The song of Blackthroat consists of rather short, rapidly delivered, varied strophes that include both whistles and harsh notes, and masterful mimicry of other species (Song et al. in press; Figure 3). The song of Firethroat (Figure 3) is very similar, and due to the complexity of the songs (including mimicry), fairly large repertoire sizes of individual males (Blackthroat 5-22 strophes recorded per male, mean 10.1 ±4.2; Firethroat 4-30, mean 13.6±7.9) and pronounced variation among males in both species, as well as limited sample sizes, we were unable to differentiate safely between 96 PERALSTROM etal. Forktail 29 (2013) Figure 2. The genetic distances between Blackthroat Calliope obscura and Firethroat C. pectardens are comparable to the divergences between other chat sister species (highlighted by blue lines). The symbols represent genetic distances for all pairwise comparisons among the 12 chats in Figure 1; distances corrected (cytb GTR + T + I, Myo HKY + T, ODC GTR + D. As expected, the mitochondrial cytb is overall more divergent than the two nuclear introns Myo and ODC. Figure 3. Songs of Blackthroat (top) and Firethroat are hard to distinguish owing to the fairly large repertoire sizes of individual males and extensive variation among males of both species (differences hence exaggerated in sonograms; a sonogram of another male Blackthroat appears in Song et at. in press). Flowever, statistical analyses of songs reveal differences (see Table 1 ). Blackthroat recorded at Foping, Shaanxi province, China (June 201 1) and Firethroat at Wolong, Sichuan province, China (May 1990), both by Per Alstrom (recordings available at www.xeno-canto.org, XC1 43220 and XC143219, respectively). Pauses between strophes have been shortened (indicated by dots). Blackthroat kHz 8-i 4 4 4 I I* 4 4 10 16 20 Firethroat * "tom *4 * i m%l * 4. \ ** s A* "hit K A. !> hihiitiiii). i MMHMMli'n 18 \ A* 20 them by ear or by sonograms. However, DFA correctly classified 88% of the recordings (Wilks’s Lambda = 0.379, Chi-square = 20.364, P = 0.000), and identified the top frequency and number of elements as the variables most important in the discrimination (Table 1). Table 1. Standardised canonical discriminantfunction coefficients for song variables in Blackthroat and Firethroat. Function 1 Mean low frequency 0.201 Mean top frequency 0.844 Mean delta time -0.972 Mean number of elements 0.855 Eigenvalue 1.637 Variance explained 100% Distributions The breeding areas of Blackthroat and Firethroat appear to be mainly non-overlapping (Figure 4, Table S2). Blackthroat has been recorded at presumed or proven breeding localities in southern Shaanxi (Qinling Mountains), southern Gansu and northern Sichuan, whereas records of Firethroat at presumed breeding localities are from central Sichuan, north-western Yunnan, south¬ east Xizang (Tibet) and Arunachal Pradesh (a single record from latter area). One old and one recent record of Firethroat were made in Shaanxi, at localities where Blackthroat has been found to breed (Figure 4, Table S2). Records from wintering areas are completely segregated, with Blackthroat only observed in Thailand and Firethroat reported from north-east India, Bangladesh and northern Myanmar (Figure 4, Table S2). Observations of birds assumed to be migrating (at places unsuitable for breeding, during August to October and March to May) were made of Blackthroat in Thailand Forktail 29(2013) Taxonomic status of Blackthroat Calliope obscura and Firethroat C. pectardens 97 Gansu Shaanxi Sichuan Xizang Bhutan Yunnan Myanmar Breeding season Non-breeding season Thailand Breeding and non-breeding season Breeding season Non-breeding season Figure 4. Blackthroat and Firethroat are mainly parapatric, with only limited evidence of sympatric breeding. Breeding season is defined as May-August, but records during this period of birds obviously on migration have been placed in the 'non-breeding season' category; for further details, see Table S2. Paintings by Hilary Burn from Collar (2005), reproduced with permission from the publishers. and Yunnan, China, and of Firethroat in north-east India and Sichuan, China (Figure 4, Table S2). DISCUSSION AND CONCLUSION The close relationship between Blackthroat and Firethroat, which has been assumed by all previous workers (e.g. Ripley 1964, Dickinson 2003, Collar 2005, Sangster et al. 2010), is confirmed by the molecular data and further supported by the similarity in song between these two taxa. Their mainly parapatric distributions (see below) might call into question their status as separate species, and support earlier suggestions that they are conspecific (Goodwin & Vaurie 1956, Cheng 1958, Vaurie 1959, Etchecopar & Hue 1983). However, the cyt b divergence is actually greater than in the two sympatrically breeding species pairs Luscinia luscinia/ L. megarhynchos and Tarsiger hyperythrus! T. rufilatus , and only marginally lower than the allopatric Larvivora cyanelL. brunnea and parapatric L. akahige/L. komadori. The cyt b divergence also agrees well with pairwise comparisons between 69 parapatric, non¬ hybridising species (mean 6.17% Kimura 2-parameter [K2P] corrected distances: Aliabadian et al. 2009) (however, as different correction methods and different datasets have been used in these two studies, the genetic distances are not directly comparable; see Fregin et al. 2012). Although the present genetic analyses are based on only one individual per taxon, the cyt b divergence between Blackthroat and Firethroat is far greater than normal intraspecific variation in birds (mean 0.74% K2P divergence in 656 species investigated by Aliabadian et al. 2009). The plumage differences between Blackthroat and Firethroat are of a similar magnitude to the differences between L. cyane and L. brunnea and between L. akahige and L. komadori , and much more pronounced than the difference between L. luscinia and L. megarhynchos. The songs of Blackthroat and Firethroat may seem surprisingly similar for different species (see Alstrom & Ranft 2003). However, 88% of the analysed recordings were correctly identified by the DFA, and it should also be noted that especially L. cyane and L. brunnea have closely similar songs (Rasmussen & Anderton, 2012, pers. obs.; recordings on www.xeno-canto.org). Based on present knowledge, the breeding distributions of Blackthroat and Firethroat appear to be mainly parapatric. There are two undocumented records of Blackthroat from potential breeding sites in Sichuan province, a female in June 1931 and a male in May 1991 (Song etal. in press). The latter is from Wolong, which is a stronghold for Firethroat, indicating potential sympatry. However, as both are undocumented, and the earliest record is the first report ever of a female Blackthroat, they should be considered uncertain. Moreover, a ‘May’ record could represent a bird on migration (Song et al. in press). According to Goodwin & Vaurie (1956), on 12 July 1905 a Firethroat and four male Blackthroats were collected at the same locality and by the same collectors on Taibai Shan, Shaanxi province. On 8 and 24 May 2013, a singing male Firethroat was observed in Changqing, Shaanxi province, in close proximity to singing Blackthroats (P. Morris and T. Townshend in lift). These reports indicate sympatric breeding of the two taxa, although both Firethroat records could have concerned individuals that had strayed north of their usual breeding 98 PER ALSTROM etal. Forktail 29 (2013) range, and nothing is known of the interactions between Blackthroat and Firethroat in these places. Although hybridisation is a possibility, it seems unlikely in view of the large genetic distances between them. In conclusion, the congruence between morphological, vocal, genetic and distributional data show that Blackthroat and Firethroat represent independently evolving lineages, and it is reasonable to treat them as separate species under both the ‘phylogenetic’ (Cracraft 1983, 1989) and ‘biological’ (Mayr 1942, 1963) species concepts. More research is needed on their distributions and possible geographical overlap, as well as on their numerical status and potential threats. ACKNOWLEDGEMENTS We are indebted to Xiangzu Ma, Naxun Zhao, Ruiqian Sun and Yongwen Zhang in Changqin National Nature Reserve Management Bureau andGaodi Dang in Foping National Nature Reserve Management Bureau for their kind help and support facilitating our fieldwork. We also thank Qing Quan for his help with Figure 4. Field survey was supported by the Major International (Regional) Joint Research Project (No. 31010103901) to FL and PA and Shaanxi Normal University (No. 2012CXB008) to FL. We are grateful to the Jornvall Foundation and the Chinese Academy of Sciences Visiting Professorship for Senior International Scientists (No. 2011T2S04) for support to PA, and NSFC Young Scientists Fund (31 101630) to GS. Mark Adams, Nick Bray, Peter Clement, Lei Dong, Sid Francis, Peng Li, Yang Liu, Vladimir Loskot, Pete Morris, Pamela Rasmussen, Barry Reed, Nate Rice, George Sangster, Paul Thompson, Terry Townshend and Lei Zhu provided other assistance or information; Josep del Hoyo and Andrew Elliott gave permission to use two paintings from the Handbook of the birds of the world in Figure 4; and Nigel Collar and Fred Sheldon provided helpful comments on the manuscript as referees. Per Alstrom and Gang Song contributed equally to this work. REFERENCES Akaike, H. (1973) Information theory as an extension of the maximum likelihood principle. Pp. 267-281 in B. N. Petrov & F. Csaki, eds. Second International Symposium on Information Theory. Budapest: Akademiai Kiado. Aliabadian, M., Kaboli, M., Nijman, V. & Vences, M. (2009) Molecular identification of birds: performance of distance-based DNA barcoding in three genes to delimit parapatric species. PLoS ONE 4(1): e41 19. doi: 1 0.1 371 /journal. pone.00041 1 9. Allen, E. S. & Omland, K. E. (2003) Novel intron phylogeny supports plumage convergence in orioles (Icterus). Auk 120: 961-969. Alstrom, P. & Ranft, R. (2003) The use of sounds in avian systematics, and the importance of bird sound archives. Bull. Brit. Orn. Club Supplement 123A: 114-135. Bi rd Life International (2013a) Species factsheet: Luscinia pectardens. Downloaded from http://www.birdlife.org on 10/04/2013. B i rd Life International (2013b) Species factsheet: Luscinia obscura. Downloaded from http://www.birdlife.org on 10/04/2013. Cheng, T. (1958) A checklist of the distribution of the birds of China. Beijing: Science Press. (In Chinese.) Cheng, T.-h. (1987) A synopsis of the avifauna of China. Beijing & Flam burg & Berlin: Science Press & Paul Parey. Collar, N. J. (2005) FamilyTurdidae (thrushes). Pp.5 14-807 in J. del Floyo, A. Elliott & D. A. Christie, eds. Handbook of the birds of the world, 10. Barcelona: Lynx Edicions. Cornell Laboratory of Ornithology (2012) Raven Pro 1.4. Ithaca: Cornell Laboratory of Ornithology. Cracraft, J. (1983) Species concepts and speciation analysis. Pp.1 59-187 in R. F. Johnston, ed. Current ornithology. New York: Plenum Press. Cracraft, J. (1989) Speciation and its ontology: the empirical consequences of alternative species concepts for understanding patterns and processes of differentiation. Pp. 28-59 in D. Otte & J. A. Endler, eds. Speciation and its consequences. Sunderland, Mass.: Sinauer. Dickinson, E. C., ed. (2003) The Howard and Moore complete checklist of the birds of the world. Third edition. London: Christopher Plelm. Etchecopar, R. D. & Hue, F. (1983) Les oiseaux de Chine, de Mongolie et de Coree. Passereaux. Paris: Boubee. Fregin, S., Haase, M„ Olsson, U. & Alstrom, P. (201 2) Pitfalls in comparisons of genetic distances: a case study of the avian family Acrocephalidae. Mol. Phylogenet. Evol. 62: 319-328. Gill, F. & Donsker, D., eds. (2013) IOC World Bird List (v. 3.3). Available at http://www.worldbirdnames.org [Accessed 10 April, 2013]. Goodwin, D. & Vaurie, C. (1956) Are Luscinia pectardens (David and Oustalet) and Luscinia obscura (Berezowsky and Bianchi) colour phases of a single species? Bull. Brit. Orn. Club 76: 141-143. Groth, J. G. (1998) Molecular phylogenetics of finches and sparrows: consequences of character state removal in cytochrome b sequences. Mol. Phylogenet. Evol. 10: 377-390. Gu, X., Fu, Y.-X. & Li, W.-H. (1995) Maximum likelihood estimation of the heterogeneity of substitution rate among nucleotide sites. Mol. Biol. Evol. 12:546-557. Hasegawa, M., Kishino, H. & Yano.T. (1985) Dating of the human ape splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol. 22: 160-1 74. Huelsenbeck, J. P. & Ronquist, F. (2001) MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17: 754-755. Huelsenbeck, J.P. & Ronquist, F. 2010. MRBAYES version 3.2. Available from: URL http://mrbayes.scs.fsu.edu/download.php. Kimball, R.T., Braun, E. L., Barker, F. K., Bowie, R.C. K., Braun, M J., Chojnowski, J. L., Hackett, S. J., Han, K. L„ Harshman, J., Heimer-Torres, V., Holznagel, W., Huddleston, C. J., Marks, B. D., Miglia, K. J., Moore, W. S., Reddy, S., Sheldon, F. H„ Smith, J.V., Witt, C. C. & Yuri,T. (2009) A well-tested set of primers to amplify regions spread across the avian genome. Mol. Phylogen. Evol. 50: 654-660. Lanave, C., Preparata, C., Saccone, C. & Serio, G. (1984) A new method for calculating evolutionary substitution rates. J. Mol. Evol. 20: 86-93. MacKinnon, J. & Phillipps, K. (2000) A fieldguideto the birds of China. Oxford: Oxford University Press. Mayr, E. (1942) Systematics and the origin of species. New York: Columbia Univ. Press. Mayr, E. (1963) Animal species and evolution. Cambridge, Mass.: Harvard University Press. Meyer de Schauensee, R. (1984) The birds of China. Washington, D.C.: Smithsonian Institution Press. Nylander, J.A.A. (2004) MRMODELTEST2 version 2.2. Available at http:// www.abc.se/~nylander. Nylander, J.A.A., Ronquist, F., Huelsenbeck, J.P. & Nieves-Aldrey, J.L. (2004) Bayesian phylogenetic analysis of combined data. Syst. Biol. 53: 47-67. Rambaut, A. & Drummond, AJ. (2009) Tracer, version 1 .5. Available at http:/ /beast. bio. ed.ac.uk. Rasmussen, P. C. & Anderton, J. C. (2012) Birds of South Asia: the Ripley guide. Second edition. Barcelona: Lynx Edicions. Ripley, S. D. (1958) A note on the Firethroat and the Blackthroated Robin. Postilla 37. Ripley, S. D. (1964) Subfamily Turdinae, thrushes. Pp.1 3-227 in E. Mayr & R. A. Paynter, eds. Check-list of birds of the world, 10. Cambridge, Mass: Museum of Comparative Zoology. Rodriguez, J., Oliver, L„ Marin, A. & Medina, R. (1990) The general stochastic model of nucleotide substitution./. Theoret. Biol. 142: 485-501. Ronquist, F. & Huelsenbeck, J. P. (2003) MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 1 9: 1 572-1 574. Ronquist, F„ Teslenko, M., van der Mark, P„ Ayres, D. L., Darling, A., Hohna, S„ Larget, B„ Liu, L„ Suchard M. A. & Huelsenbeck, J. P. (201 1) MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Sysr. Biol. 61 : 539-542. Forktail 29(2013) Taxonomic status of Blackthroat Calliope obscura and Firethroat C. pectardens 99 Sangster, G., Alstrom, P., Forsmark, E. & Olsson, U. (2010) Multilocus phylogenetic analysis of Old World chats and flycatchers reveals extensive paraphyly at family, subfamily and genus level (Aves: Muscicapidae). Mol. Phylogen. Evol. 57: 380-392. Sibley, C. G. & Monroe, B. L.(1990) Distribution and taxonomy of birds of the world. New Haven: Yale University Press. Song, G„ Alstrom, P„ Zhang, Y., Gao, X., Gong, H„ Holt, P.I., Qing, Q„ Yin, Z. & Lei, F. (in press) Rediscovery of an enigmatic Chinese passerine, the Blackthroat Calliope obscura : plumage, vocalizations, distribution, habitat choice, nesting and conservation. J. Orn. Swofford, D. L. (2002) PAUP*: phylogenetic analysis using parsimony (*and other methods). Sunderland, Mass: Sinauer Associates. Tamura, K„ Dudley, J., Nei, M. & Kumar, S. (2007) MEGA4: Molecular evolutionary genetics analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24: 1596-1599. Tavare, S. (1986) Some probabilistic and statistical problems on the analysis of DNA sequences. Lee. Math. Life Sci. 1 7: 57-86. Ul Haque, E., Dymond, N., Round, P. & Thompson, P. (2012) Ringing and recent ornithological exploration in Bangladesh. BirdingASIA 17: 78- 79. Vaurie, C. (1959) The birds of the Palearctic fauna: order Passeriformes. London: H.F. & G. Witherby. Yang, Z. (1994) Maximum likelihood phylogenetic estimation from DNA sequences with variable rates over sites: approximate methods. J. Mol. Evol. 39: 306-314. Zuccon, D. & Ericson P. G. P. (2010) A multi-gene phylogeny disentangles the chat-flycatcher complex (Aves: Muscicapidae). Zool. Scripta 39: 21 3- 224. SUPPLEMENTARY ONLINE MATERIAL Available on Oriental Bird Club website, links at http:// www.orientalbirdclub.org/ publications/ forktail29 Table SI. Sequences used in the present study Table S2. Records of Firethroat Per ALSTROM, Key Laboratory of Zoological Systematics and Evolution, Institute of Zoology, Chinese Academy of Sciences, I Beichen West Road, Chaoyang District, Beijing 100101, China; and Swedish Species Information Centre, Swedish University of Agricultural Sciences, Box 7007, SE-750 07 Uppsala, Sweden. Email: per.alstrom@slu.se (corresponding author) Gang SONG and Ruiying ZHANG, Key Laboratory of Zoological Systematics and Evolution, Institute of Zoology, Chinese Academy of Sciences, 1 Beichen West Road, Chaoyang District, Beijing 100101, China. Email: songgang@ioz.ac.cn, zhangry@ioz.ac.cn Fumin LEI, Key Laboratory of Zoological Systematics and Evolution, Institute of Zoology, Chinese Academy of Sciences, 1 Beichen West Road, Chaoyang District, Beijing 100101, China; and School of Life Sciences, Shaanxi Normal University, Xi'an 710062, China. Email: leifm@ioz.ac.cn (corresponding author) Xuebin GAO, Shaanxi Institute of Zoology, No. 88, Xingging Road , Xi'an 710032, China. Email: gaoxb63@163.com Paul I. HOLT, Bracken Dean, Pendleton, Clitheroe, Lancashire, BB7 1PT, England. Email: piholt@hotmail.com Urban OLSSON, Systematics and Biodiversity, Department of Biology and Environmental Sciences, University of Gothenburg, Box 463, SE-405 30 Goteborg, Sweden. Email: urban.olsson@bioenv.gu.se FORKTAIL 29 (2013): 100-106 Bonelli's Eagle Aquila fasciata renschi in the Lesser Sundas, Wallacea: distribution, taxonomic status, likely origins and conservation status COLIN R.TRAINOR, STEPHEN J. S. DEBUS, JERRY OLSEN, JANETTE A. NORMAN & LES CHRISTIDIS Records of Bonelli's Eagl e Aquila fasciata renschi on 18 islands in the Lesser Sundas, from Lombok to the Tanimbar islands, in Indonesia and Timor-Leste are reviewed, and its taxonomic status examined. The species is resident on many islands, known to breed on several of the larger islands and is most abundant on Flores and Timor. It appears to be rare on Lombok and Sumba. There is minimal genetic differentiation between local subspecies A. f. renschi and nominate A. f. fasciata, suggesting only recent geographic isolation and that it should not be afforded species status. The species may have been introduced to the Lesser Sundas by traders or colonists in the past. The species's biology and ecology are poorly known in Wallacea. It occurs in a wide range of sites from sea level to about 2,000 m, in wooded habitats with a preference for tropical forest. The sparse data on diet show that it feeds on introduced wild junglefowl Gallus sp., but presumably it also feeds on large-bodied frugivorous pigeons, other birds and small mammals. The two reports of nesting were in May and June-July. The frequency of records on Flores and Timor suggests that these populations are currently secure, but may be threatened by hunting, capture and deforestation. A predator of village chickens, it is likely to be targeted by local communities. Conservation priorities include distribution and population density surveys and awareness projects throughout its range. INTRODUCTION It has been proposed that the isolated Lesser Sundas population of Bonelli’s Eagle Aquila fasciata renschi should be regarded as a distinct species (Thiollay 1994, Ferguson-Lees & Christie 2001), partly based on biogeography. However, this has not prompted specific field or taxonomic studies of this taxon. According to recent evaluations the Wallacean subspecies is found on the Lesser Sunda islands of Sumbawa, Komodo, Flores, Besar, Timor, Wetar and Luang, and Yamdena in the Tanimbar islands (White & Bruce 1 986, Coates & Bishop 1997, Ferguson-Lees & Christie 2001). The nominate form fasciata has a wide but fragmented distribution; it is resident, with little or no evidence of migration, in North Africa, the Iberian peninsula, the Mediterranean, parts of the Arabian Peninsula, the Middle East, Afghanistan, Pakistan, India and disjunctly to north Indochina and southern China (Thiollay 1994, Hernandez-Matias etal. 201 1). The Lesser Sundas are about 3,000 km from the nearest Asian population in Vietnam (Ferguson-Lees & Christie 200 1 ). In Vietnam, Laos, Thailand and Myanmar there are very few records and it appears to be sedentary (Duckworth et al. in press). Although the species has a relatively distinctive appearance there have been identification problems in the Lesser Sundas. The type specimen was originally identified by Rensch (1931) as a Changeable Hawk Eagle Nisaeetus cirrhatus — the Lesser Sunda population of this species is now known as the Flores Hawk Eagle N. floris (Gjershaug et al. 2004), and a Wetar island record of ‘Changeable Hawk Eagle’ (Hartert 1904) was later identified as Bonelli’s Eagle (Mees 2006). Bonelli’s Eagle is now in the genus Aquila rather than Hieraaetus, based on DNA data ( W ink & Sauer- Giirth 2004, Helbig et al. 2005, Lerner & Mindell 2005) demonstrating that Aquila and Hieraaetus as conventionally circumscribed were paraphyletic. Recent observations confirming the species’s presence on Lombok and Sumba, and on several other islands and islets in the Lesser Sundas, are reviewed, and substantial populations on at least Flores and Timor are documented. Records were collected from published and unpublished trip reports, bird tour reports and grey literature, and by canvassing individuals and the Orientalbirding e-group. The taxonomic status of renschi is assessed using DNA evidence. Some of the text of this paper was originally submitted as a book chapter in 2007 (still unpublished), which was split into two chapters by the editor (Christidis etal. in press, Debus etal. in press) and is expected to be published in late 2013. This opportunity is taken to substantially update those manuscripts, particularly in relation to distribution, ecology and taxonomy. Figure 1 . Map of the Lesser Sundas region, showing the islands mentioned in the text. Forktail 29(2013) Bonelli's Eagle Aquila fasciata renschi in the Lesser Sundas, Wallacea 101 Study region The Lesser Sundas are located in the extreme south-east of Asia, between continental Java and Bali and the Australo-Papuan region (Figure 1). Lying west to east in a 1,700 km arc, they comprise about 20 large oceanic islands and several hundred small islands and islets. The inner Banda arc includes young volcanic islands (Monk et al. 1997) front Lombok to Damar in Maluku province and the outer Banda arc, dominated by sedimentary rocks such as limestone, extends from Sawu (Sabu), Roti, Timor, Moa and Babar to the Tanimbar islands. Suntba is considered to be a continental fragment. Most islands are only a few kilometres apart and would have been contiguous during the last glacial period 9,000 to 18,000 years ago, thus aiding avian dispersal (Voris 2000), but Suntba, Wetar, the Tanimbar islands and Damar are separated from their nearest neighbours by tens of kilometres. The natural vegetation of the islands is closed-canopy tropical forest (tropical dry to evergreen) and various savannahs, including Eucalyptus, but on many islands, agriculture has repeatedly changed this to regrowth forest and savannah woodland (Monk etal. 1 997). The main islands have been cleared to varying extents, many now being essentially deforested (Table 1). On large islands, forest fragments are often restricted to steep mountain slopes and peaks, but on some isolated islands in the Banda Sea (e.g. Wetar, Romang, Babar and Damar) forest cover is extensive. Table 1. Area, climate (in relative terms), estimated remaining forest cover and relative biological survey effort, on Indonesian islands inhabited by Bonelli's Eagle (source: CRT unpubl. data, who has visited all islands mentioned, but has only sailed past Luang Island). Island Area (km2) Climate Forest cover (%) Survey effort Lombok 4,625 wet (dry on coast) 10 Moderate Sumbawa 14,386 dry 15 Moderate Komodo 330 dry 40 Moderate Flores 14,154 wet (dry on coast) 15 High Besar 64 dry 75 High Adonara 509 dry 5 Moderate Pantar 728 dry 15 Low Alor 2,864 dry 3 Moderate Sumba 10,711 dry 12 High Timor 28,418 dry 10 High Wetar 2,684 dry 97 Moderate Atauro 147 dry 10 High Romang 184 wet 75 Low Luang 5 wet 25 Moderate Sermata 103 wet 35 Low Damar 198 wet 75 Moderate Yamdena 3,333 (+ satellites cl, 500) wet 75 High RESULTS Summary of Bonelli's Eagle records in the Lesser Sundas Islands are listed west to east. Lombok: adult pairs observed in Gn Rinjani National Park, and between Sembalum and Sapit, in June-July 2003 (J. M. Thiollay in litt. 2007); the first records for the island. None was reported in a review by Myers & Bishop (2005) or during surveys for the Critically Endangered Flores Hawk Eagle (Raptor Conservation Society 2011)." Sumbawa: a single historical specimen is the renschi holotype (Stresemann 1932), a male (Mees 2006). Two birds were recorded by Johnstone et al. (1996), who considered it rare. J. M. Thiollay {in litt. 2007) observed it at four locations (Alas, Reloka, Ropang valley and Tambora peninsula), where it seemed common in June- July 2003. One bird was seen at Teluk Saleh in 2009 (V. Dinets in litt. 2013). Moyo: a pair recorded in December 1999 (Trainor et al. 2006); first record for this almost land-locked island. Sumba: observed at Lake Pambotanjaraand Lewaon 14 July 1991 (Dreyer 1993), apparently the first records for the island. One bird was observed at Lewapaku in October 1998 (Trainor et al. 2006). It was not observed in four weeks by Olsen & Trost (2007), suggesting that it is uncommon or rare. None have been recorded by recent bird tour visits to the island. Sumba and Flores are visible from each other and separated by 45 km of sea, which should present little barrier to an eagle’s flight. Sumba appears to hold suitable habitat: a dry limestone island cut by canyons, with abundant parrot, pigeon and junglefowl prey (JO, pers. obs.). Komodo: recorded in semi-deciduous forest and savannah (Butchart etal. 1994). At least two observed in June-July 2003 (J. M. Thiollay in litt. 2007). Flores: two fairly recent specimens, collected in 1971, identified as male and female (Mees 2006). Coates & Bishop (1997) considered it locally moderately common. Fourteen records mostly of pairs, mainly in hilly and mountainous terrain with cultivation, coconut plantations, scrub and secondary forest; also on a cultivated and scrubby plain (V erhoeye & King 1990). Recorded in moist and semi- deciduous forest, thorn scrub and montane forest, from sea level to 2,000 m (Butchart etal. 1994). Two observed in coastal scrub, gallery forest and grassland (Gibbs 1990). Verhoeye & Holmes (1999) reviewed about 20 records from cultivated and wooded hills; there are five additional records for forest and cultivated land (Pilgrim et al. 2000), and eight sightings (including a pair twice) from six localities in September-November 1998 (R. Drijvers unpubl. data). Recorded at 12 locations by Trainor & Lesmana (2000). Observed at eight locations in June-July 2003 and found to be widespread on the island; one nest found (J. M. Thiollay in litt. 2007). Observed throughout the island, from sea level to at least 1,600 m in forest, rice fields and valleys; common in central Flores; a captive juvenile (Plate 1) observed on 14 October 2004 originated from the slopes of Gn Iya, near Ende (M. Schellekens in litt. 2007). One bird was photographed at Riung in April 2012 (O. Hidayat in litt. 2013). The species is recorded regularly by tour groups and individual birdwatchers at several sites. Besar: considered locally moderately common (Coates & Bishop 1997). Recorded in semi-deciduous, deciduous and coastal forest (Butchart etal. 1994). Adonara: two over dry agricultural land and closed forest on mountain slopes, December 2000 (Trainor 2002); first record for the island. Pantar: the only record was an adult observed by P. Verbelen at Gn Wasbila on 3 September 2009 (Trainor et al. 2012). Alor: observed in Eucalyptus savannah in May 2002 (Trainor 2005a); first record for the island. Some records in 2002, 2009 and 102 COLIN R.TRAINOR et al. Forktail 29 (2013) 2010 by CRT and P. Verbelen may have been confused with Flores Hawk Eagle, but pairs and singles were seen and confirmed with photographs at several sites up to about 1,100 m (Trainor et al. 2012). Timor: single historical specimen, a male (Mees 200 6). One sighting, over mountain forest and peaks at 2,000 m, East Timor, in 1972 (White & Bruce 1986). Three records in East Timor in 1974 (H. Thompson, J. McKean & I. Mason unpubl. data). Since the 1 990s it has been regularly observed in West Timor, particularly at Bipolo and Camplong (Verbelen 1996, Mauro 1999, Van Biers 2004, N. Kemp in litt. 2007). Noted at four forested localities in West Timor by Noske & Saleh ( 1 996). CRT had 36 sightings, 25 of single birds, 7 of two birds and 4 of three birds, from eight districts in Timor-Leste over four years 2003-2006; from sea level to 1,200 m in habitats ranging from coastal flats to village cultivation, freshwater lakes/ swamps, secondary forest and primary forest (dry deciduous, semi-evergreen, evergreen and montane). Atauro: pair over montane forest in December 2003, and a captive juvenile was said to have originated from a nest on Atauro (T rainor & Soares 2004); first records for the island. Wetar: single historical specimen, originally sexed as male but probably a female (Mees 2006). Considered locally moderately common (Coates & Bishop 1997), although this assessment was based on a half day observation in west Wetar. In 2008, a 44-day survey recorded it from 8 of 12 sites up to about 500 m: the island is one of the least disturbed in insular South-East Asia (Trainor et al. 2009); there is at least one additional record of a bird photographed at sea level in September 2010 (CRT unpubl. data). Romang: during the first ornithological visit to the island since 1 902, a single bird and a displaying pair were observed over tropical forest at two sites at about 300 m, during two weeks in October 2010 (Trainor & Verbelen in press). Luang: two historical specimens, both males (Mees 2006). Sermata: one adult bird photographed at forest edge in November 2010 during the first ornithological exploration of the island since 1906 (Trainor & Verbelen in press). Damar: two sightings over forest and forest edge, August 2006 (Trainor 2007); first records for the island. Tanimbar islands: first observed on Yamdena, Tanimbar, by F. Rozendaal between August and November 1985 (F. G. Rozendaal unpubl. data). Also observed on Yamdena in August 1994 by Verbelen (1996); and in October 1998 by Mauro (1999). A pair was observed over tall subcoastal primary and secondary semi-evergreen forest and woodland in January 1996 (Coates & Bishop 1997, Bishop & Brickie 1999). Most recent tour group records are from the Lorulun area, 20 km north of Saumlaki, but there are records closer to Saumlaki. Taxonomic status In using levels of DNA differentiation to assess taxonomic assignments of species and subspecies, Norman et al. (1998) and Christidis & Norman (2010) advocated the requirement to include a relative hierarchical perspective of DNA divergences in the genus of interest. The relevant DNA data dealing with the taxonomy of th e Aquila fasciata species-complex are summarised below. The taxonomy of Bonelli’s Eagle and the African Hawk Eagle Aquila spilogaster has been a contentious issue. Long regarded as a single species (Brown & Amadon 1968), the recent tendency is to treat the two as separate species (Thiollay 1994, Ferguson-Lees & Christie 2001). The consistent morphological, plumage and behavioural differences have been cited as evidence for species-level separation. Lerner & Mindell (2005) give a molecular perspective through their examination of mitochondrial DNA differentiation in a range of birds of prey, including Bonelli’s Eagle and African Hawk Eagle. Between the two species there were 16 and 18 base- pair differences in cytochrome b and ND2, respectively. These figures were larger than those recorded between other species-pairs identified in the study: Wedge-tailed Eagle Aquila audax and Gurney’s Eagle A. gurneyi-, and Little Eagle Hieraaetus morphnoides and Booted Eagle H. pennatus. Although not conclusive, the DNA data support separate species treatment for Aquila fasciata and A. spilogaster. Apart from the three individuals of A. fasciata examined by Lerner & Mindell (2005), cytochrome b data are available for a further six individuals. Haring etal. (2007) lodged a 264-base-pair fragment on GenBank (accession numbers EF459628-EF459631) from two individuals of A.f fasciata (one from Italy and one with no locality information), and two individuals of A.f. renschi from Flores. Helbig etal. (2005) examined a 1,143-base-pair fragment from an individual of A.f. fasciata from Israel, and Bunce et al. (2005) examined 1,017 base pairs in another A.f. fasciata individual (no locality data). In addition, JAN & LC sequenced 409 base pairs of an individual renschi from Timor, feathers of which were collected by JO. A 217- base-pair fragment was common to all five studies, and this was compared across the three individuals of A.f. renschi and seven individuals of A. f. fasciata examined. There were only three variable sites, and this variation was limited to a unique base change in each of three individuals of A.f. fasciata. By excluding the four samples from Haring et al. (2007), it was possible to compare a 267 base-pair fragment across the remaining six individuals, but this did not reveal any additional variation. There were no differences recorded between A.f. renschi and the common A.f. fasciata haplotype. The negligible cytochrome b variation recorded was, therefore, limited to comparisons within A. f. fasciata. This lack of any molecular differentiation between the subspecies A.f. fasciata and A.f. renschi is consistent with a very recent separation. Variation in a 253-base-pair fragment of the mitochondrial control region was assessed in 72 individuals of A. f. fasciata from Spain, Portugal and Morocco by Cadahia etal. (2007). They found four mitochondrial types each differing from the other by a single base-pair change. Moreover, there did not appear to be any geographic structure to the genetic variation observed across the populations surveyed. One explanation offered for the low levels of genetic variation was a loss of genetic variation caused by population reduction during the Pleistocene glaciations, and more recently through human activities such as habitat clearance and hunting. Control-region sequence data for A. fasciata have also been lodged on GenBank (Accession Numbers EF459585-459588) by Haring et al. (2007). Unfortunately, the 237-base-pair fragment does not correspond to the region examined by Cadahiaer^/. (2007). The control-region data of Haring etal. (2007) was obtained from the same specimens that cytochrome b data were obtained (see above): two individuals of A.f. fasciata (one from Italy and one with no locality information) and two individuals of A.f. renschi from Flores. The A. f. fasciata individual from Italy differed by 5-6 changes from the other three individuals. The remaining three individuals differed from each other by 1-2 changes. Although the level of variation is low, there is nevertheless more variation recorded within A.f. fasciata than between A.f. fasciata and A. f. renschi. Both the cytochrome b and control-region DNA datasets showed similar patterns: low levels of DNA variation across the range of A.fasciata\ no diagnostic DNA marker distinguishing^/. f. fasciata from A.f. renschi ; and more variation within A.f fasciata than between the two subspecies. Although this pattern of variation MARKSCHELLEKENS Forktail 29(2013) Bonelli's Eagle Aquila fasciata renschi in the Lesser Sundas, Wallacea 103 Plate 1. Captive juvenile Bonelli's Eagle A. f. renschi on Flores, Indonesia, 14 October 2004. The captive bird was kept tethered to a wooden plank, by a short rope tied to one ankle; the bird had a small wound on the underside of its carpal joint, and an overgrown bill. could be indicative of a slow rate of mitochondrial evolution in Aquila, the comprehensive cytochrome b and ND2 datasets of Lerner & Mindell (2005) do not provide any such indication. The widely disjunct distributions of A.f fasciata and A.f renschi also make it unlikely that a lack of differentiation is caused by past bottlenecks. It is difficult to envisage similar genetic bottlenecks occurring in such widely separated populations. Juvenile morphology Colour photographs (Plates 1 & 2) and other unpublished images of the same birds show that juveniles of A.f renschi are similar in plumage to juveniles of the nominate subspecies, with few evident differences (V. Hernandez in litt. 2007). The photographic evidence shows colour variation within the range of that of juvenile Eurasian Bonelli’s Eagles. However, Wallacean birds are more lightly built than Eurasian birds. Juveniles of each subspecies would be indistinguishable, and only separable by measurement (V. Hernandez in litt. 2007). Biology There is little information on the feeding ecology or breeding biology of Bonelli’s Eagle in Wallacea. A bird was observed feeding on a Green Junglefowl Gallus varius carcass at Gn Ranaka, Flores, in August 2007 (Myers 2007), an adult bird was photographed holding a village chicken Gallus sp. near Gn Ranaka in 20 1 1 (Plate 3), and in September 2011 an adult Bonelli’s Eagle at Pagal, Flores, delivered a chicken Gallus sp. or arallid to a juvenile (Robson 2011). In Ruteng, west Flores, Bonelli’s Eagle were twice (in separate years) observed flying low over the town, and were suspected to be searching for village chickens (J. Eaton in litt. 2013). Other likely prey within their range includes Pink-headed Imperial Pigeon Plate 2. Captive juvenile Bonelli's Eagle in Timor-Leste, 10 April 2003. Plate 3. Adult Bonelli's Eagle in flight near Mt Ranaka, Flores, holding a village chicken, 30 August 2011. MARC THIBAULT COLIN R. TRAINOR 104 COLIN R.TRAINOR etal. Forktail 29(2013) Ducula rosacea , Green Imperial Pigeon D. aenea , Timor Black Pigeon Turacoena modesta and other forest pigeons, cuscus Phalanger sp., rats (Muridae) and medium-sized fruit-bats (Pteropodidae) that roost in caves, forests and savannah palms. A nest was found on 12July2003on Gn Ranaka, Flores, at 1,420 m: an adult was sitting on the nest in a tree and the mate flew in with prey (J . M. Thiollay in lift. 2013). A second nest was found at Lermatangon Yamdena, Tanimbar islands, on 22 May 2008 (Yong & Lee 2008): it was about 20-25 m up in a forest tree, approximately 1.5 m in diameter and consisted of sticks and vines. It was unclear whether the pair were sitting on eggs or had young, but they were actively managing the nest. DISCUSSION This review confirms that Bonelli’s Eagle is more widespread in the Lesser Sundas than previously believed; recent new records from nine islands in addition to the nine where it was previously recorded have extended its range to a land area of about 87,400 km2. The distinctiveness of the isolated Lesser Sunda population has been subject to ongoing speculation, due to its smaller size and plumage differences compared with fasciata ; the tail being more strikingly barred and belly, thighs and crissum more boldly marked (Ferguson-Lees & Christie 2001). However, the negligible levels of genetic differentiation between A.f. renscbi and A.f. fasciata do not support the contention of Thiollay (1994) and Ferguson-Lees & Christie (2001) that renschi should be accorded full species status, although it is harder to argue against subspecific recognition for both forms. The patterns of DNA variation are more consistent with the relatively recent arrival of renschi in the Lesser Sundas and it may have been introduced from Eurasia (see below). Accordingly, the smaller size of renschi and the plumage differences between it and fasciata would also have evolved relatively rapidly. Within the archipelago Flores and Timor appear to be strongholds, with many records from human-modified landscapes, but this may be partly a result of greater observer effort on these islands compared with elsewhere. The lack of earlier records from Sumba and Sumbawa, and recent records from other islands, may partly reflect bias of historical collectors and recent increase in survey effort (M. Bruce in litt. 2007). For example, there was a reluctance to collect cumbersome large specimens including raptors because ofthe relatively high shipping costs (Hartert 1904). Sumba is relatively well surveyed, and tour parties now visit annually, so the paucity of records suggests that it is either a rare resident, or that birds are occasional visitors from nearby islands. Knowledge of the avifauna of Roti is improving (Verheijen 1975, Trainor 2005b, Collaerts et al. 2011, P. Verbelen in litt. 2010), but there are no records of Bonelli’s Eagle from this largely deforested island. Lack of records from other largely deforested islands (Sawu [Sabu], Semau and Kisar) suggests that a minimum level of forest cover is necessary to sustain populations of the species. This could be associated with the scarcity of large prey species, such as frugivorous pigeons (Newbold et al. 2013), in agricultural land or savannah woodland. There are insufficient data to comment on population trends, but Bonelli’s Eagle appears to be holding its own at present, although the extensive and rapid deforestation in Indonesia (Brooks et al. 1999) may adversely affect it. The species’s Indonesian (at the time including East Timor) conservation status was assessed as ‘no immediate danger’ (van Balen 1994), and there seems no reason to amend this at present. The species currently occurs in cultivated lands and secondary forest as well as natural habitats. Hunting either for food, or to reduce the perceived impact on village livestock, might also affect populations of Bonelli’s Eagle. Hunting is part of life for many villagers in the Lesser Sundas. On some islands, hunters are often armed with powerful air-rifles (comparable with a .22 rifle) and children have powerful slingshots. They shoot a wide range of wildlife, including raptors, and climb trees to collect nestlings for lood. As a predator of village chickens, Bonelli’s Eagle is likely to be targeted to reduce the perceived impact on economically important village livestock. Raptor nestlings are commonly taken captive, and suffer casual, habitual cruelty in captivity. A captive juvenile Bonelli’s Eagle was photographed tethered by a metal ring on one ankle attached to a short rope and, not surprisingly, the bird had bumblefoot (a bacterial infection and inflammatory reaction) in the shackled foot as well as cere damage and abraded carpals (Plate 1). Another captive juvenile/subadult bird owned by a foreign defence worker in Timor-Leste was housed in a chicken wire cage at a United Nations military compound for months. It had serious cere damage (CRT unpubl. data). CRT has also seen at least two other captive Bonelli’s Eagles in Timor-Leste, although there may be many more out of sight. Conservation priorities include further field surveys on the large islands of Sumba, Sumbawa, Wetar and Yamdena and on islands where there are no records (e.g. Babar, Moa, Roti, Solor and Lembata). The single largest tropical forest in the Lesser Sundas is in west Sumbawa (about 2,000 km2) (Jepson et al. 2001), whilst Wetar retains more than 97% forest cover; both deserve specific surveys for Bonelli’s Eagle. The breeding biology of this subspecies is essentially unknown, so it would be useful to monitor population levels and breeding success at selected sites on various islands. There is also a need for an environmental education campaign to discourage persecution ol eagles in general, improve the lot of captive birds, and encourage local people to ‘own’ and value such iconic species (Burnham et al. 1994, Salvador 1994). In Timor- Leste, for example, the campaigns should also target foreign nationals (military and embassy staff). Effective conservation of Bonelli’s Eagle and other raptors in Wallacea is likely to deliver broader biodiversity benefits (Sergio et al. 2006). Origins ofthe Waflacean population The isolated populations of Bonelli’s Eagle, and Short-toed Snake- Eagle Circaetus gallicus , in Wallacea stand out as zoogeographically anomalous — usually explained as relicts from past climatic and sea- level changes (Voris 2000). But it may be important to consider the human history of these islands; the first Dutch ships arrived in Indonesia (East Indies) in 1596 and determined exploitation started around 1830. The first Europeans to visit Timor were Portuguese, perhaps as early as 1512, and the Dutch occupied Kupang in present-day West Timor in the mid-seventeenth century, beginning a long conflict for control of the sandalwood trade. The Dutch controlled most of the Lesser Sundas from the 1600s, but the Portuguese held Flores (especially east Flores with forts on nearby Adonara and Solor) and East Timor, including the Ambeno (Oecusse) enclave, for long periods (Fox 2003). Before assuming that the eagle occurs naturally on these islands, it is important (but difficult) to rule out the possibility that Bonelli’s Eagles were transported from Europe or South Asia (notably India where the Portuguese also had colonies) by Dutch or Portuguese traders or settlers. Europeans may have introduced eagles to their Indonesian colonies as mascots, pets or falconry birds, perhaps from Iberia, North Africa or South Asia. Other birds, junglefowl Gallus spp., and Red Avadavat Amandava amandava are thought to have been introduced to the Lesser Sundas several centuries ago. The Red Avadavat is represented in the Lesser Sundas by the subspecies flavidiventris , which occurs naturally in South Yunnan (China), Thailand and Myanmar (White & Bruce 1986). There was also much movement of various animals, by Asian and Melanesian peoples, between Asia, Wallacea and New Guinea (Heinsohn 2003). Alternative hypotheses that need to be Forktail 29(2013) Bonelli's Eagle Aquila fasciata renschi in the Lesser Sundas, Wallacea 104 investigated include vagrant Bonelli’s Eagles from Asia settling in the Lesser Sundas. ACKNOWLEDGEMENTS We thank Vladimir Dinets, James Eaton, Raf Drijvers, Hank Hendriks, Oki Hidayat, Rob Hutchinson, Nev Kemp, Frank Rozendaal, Mark Schellekens, Greg Smith, Brian Sykes, Marc Thibault, Jean-Marc Thiollay and Philippe Verbelen for their unpublished sightings and data, Rui Pires for assistingJO in the field, and Victor J. Hernandez and James Eaton for evaluating images of eagles. Thanks go to Will Duckworth and Phil Round for access to an unpublished manuscript on the South-East Asian population of Bonelli’s Eagle, and more recent observations. We also thank Mark Schellekens and Marc Thibault for permission to use their images, and Murray Bruce for commenting on a draft. SD thanks Sofia Debus for facilitating the initial drafting of this paper. JO thanks Luis Palma and Marcos Moleon for helpful discussions and for showing him Bonelli’s Eagles in Iberia. REFERENCES van Balen, S. (1994) The status and conservation of birds of prey in the Sundaic and Wallacean regions of Indonesia. Pp. 245-254 in B.-U. Meyburg & R. D. Chancellor, eds. Raptor conservation today. Mountfield, UK: World Working Group on Birds of Prey and Owls/Pica Press. Bishop, D. & Brickie, N. W. (1999) An annotated checklist of the birds of the Tanimbar Islands. Kukila 10: 115-1 50. Brooks, T. M., Pimm, S. L., Kapos, V. & Ravilious, C. (1999) Threats from deforestation to montane and lowland birds and mammals in insular South-East Asia./. Anim. Ecol. 68: 1 061-1 078. Brown, L. H. & Amadon, D. (1968) Eagles, hawks and falcons of the world. London: Country Life Books. Bunce, M., Szulkin, M., Lerner, H. R. L., Barnes, l„ Shapiro, B., Cooper, A. & Holdaway, R. N. (2005) The evolutionary history of New Zealand's extinct giant eagle revealed by ancient DNA. PLoS Biology 3: (e9) 0044-0046. Burnham, W., Jenny, P. & Whitacre, D. (1994) The Maya Project: use of birds of prey as tools for conservation and ecological monitoring of biological diversity. Pp. 257-264 in B.-U. Meyburg & R. D. Chancellor, eds. Raptor conservation today. Mountfield, UK: World Working Group on Birds of Prey and Owls/Pica Press. Butchart, S. H. M., Brooks, T. M., Davies, C. W. N., Dharmaputra, G., Dutson, G. C. L., Lowen, J. C. & Sahu, H. (1994) Preliminary report of the Cambridge Flores / Sumbawa Conservation Project 1 993. Cambridge University. Cadahia, L., Negro, J. J. & Urios, V. (2007) Low mitochondrial DNA diversity in the endangered Bonelli's Eagle ( Hieraaetus fasciatus) from SW Europe (Iberia) and NW Africa. J. Orn. 148: 99-104. Christidis, L. & Norman J. A. (2010) Evolution of the Australasian songbird fauna. Emu 1 10: 21-31. Christidis, L., Norman, J. A. & Olsen. J. (in press) An assessment of molecular data and the resolution of the taxonomic status of the Lesser Sunda race of Bonelli's Eagle, Aquila fasciata renschi. In V. J. Plernandez, ed. The Bonelli's Eagle: ecology, behaviour and conservation/ El aguila perdicera: ecologia, comportamiento y conservacion. Madrid: Tundra Ediciones. Coates, B. J. & Bishop, K. D. (1997) A field guide to the birds of Wallacea. Alderley: Dove Publications. Collaerts, E., Collaerts, P. & Theuwis, T. (2011) Birding trip report to the Lesser Sundas. Privately published. Available at http:// www.travellingbirder.com/. Debus, S. J. S., Trainor, C. R. & Olsen, J. (in press) Distribution, status and conservation priorities of Bonelli's Eagle Aquila fasciata renschi in Wallacea. In V. J. Hernandez, ed. The Bonelli's Eagle: ecology, behaviour and conservation/ El aguila perdicera: ecologia, comportamiento y conservacion. Madrid: Tundra Ediciones. Dreyer, N. P. (1993) Trip Report: Sumba & Timor (Indonesia), July 13-24, 1991. Unpublished report, available at http \l lwww.camacdonald.com). Duckworth, J. W., Round, P. D„ Russell, D. G. D„ Kasorndorkua, C. & Robson, C. R. (in press) Bonelli's Eagle Aquila fasciata in South-East Asia. In V. J. Hernandez, ed. The Bonelli's Eagle: ecology, behaviour and conservation/ El aguila perdicera: ecologia, comportamiento y conservacion. Madrid. Tundra Ediciones. Ferguson-Lees, J.& Christie, D. A. (2001) Raptors of the world . London: Helm. Fox, J. (2003) Tracing the path, recounting the past: historical perspectives on Ti mor. Out of the ashes: destruction and reconstruction of East Timor. Canberra: ANU Press. Gibbs, D. (1990) Wallacea: a site guide for birdwatchers. Privately published. Gjershaug, J. O., Kwaloy, K., Rov, N., Prawiradilaga, D. M., Suparman, U. & Rahman, Z. (2004) The taxonomic status of the Flores Hawk-eagle Spizaetus floris. Forktail 20:51-61. Haring, E., Kvaloy, K„ Gjershaug, J.-O., Rov, N. & Gamauf, A. (2007) Convergent evolution and paraphyly of the hawk-eagles of the genus Spizaetus: phylogenetic analyses based on mitochondrial markers. Unpublished sequences submitted to GenBank (28-Feb-2007). Hartert, E. (1904) On the birds of the south-west islands of Wetter, Roma, Kisser, Letti and Moa. Novit. Zool. 11:1 74-221. Heinsohn, T. (2003) Animal translocation: long-term human influences on the vertebrate zoogeography of Australasia (natural dispersal versus ethnophoresy). Ausf. Zool. 32: 351-376. Helbig, A. J., Kocum, A., Seibold, I. & Braun, M. J. (2005) A multi-gene phylogeny of aquiline eagles (Aves: Acciptriformes) reveals extensive paraphyly at the genus level. Mol. Phylogenet. Evol. 35: 147-164. Hernandez-Matias, A., Real, J., Pradel, R. Ravayrol, A. & Vincent-Martin, N. (201 1) Effects of age, territoriality and breeding on survival of Bonelli's Eagle Aquila fasciata. Ibis 1 53: 846-857. Jepson, P., Agista, D., Trainor, C., Lesmana, D., Setiawan, I. & Sujatnika (2001) Identification and boundary proposals for a new wildlife sanctuary covering the Tatar Sepang drainage in Southwest Sumbawa: An area with international significance for the conservation of global biodiversity. BirdLife International-lndonesia Program Technical Memorandum No. 22. Johnstone, R. E„ Jepson, P., Butchart, S. H. M., Lowen, J. C. & Prawiradilaga, D. (1996) The birds of Sumbawa, Moyo and Sangeang Islands, Nusa Tenggara, Indonesia. Rec. West. Austral. Mus. 18: 157-178. Lerner, H. R. L. & Mindell, D. P. (2005) Phylogeny of eagles, Old World vultures, and other Accipitridae based on nuclear and mitochondrial DNA Mol. Phylogenet. Evol. 37: 327-346. Mauro, I. (1999) Preliminary report on birds recorded from Wallacea: Sulawesi, Moluccas & Lesser Sundas. Unpublished. Mees, G. F. (2006) The avifauna of Flores (Lesser Sunda Islands). Zool. Meded. (Leiden) 80(3): 1-261. Monk, K. A., de Fretes, Y. & Lilley, G. (1997) The ecology of Nusa Tenggara and Maluku. Singapore: Periplus Editions. Myers, S. D. (2007) Indonesia: Lesser Sunda islands scouting report. Ventbird. Available at: http://www.ventbird.com. Myers, S. D. & Bishop, K. D. (2005) A review of historic and recent bird records from Lombok, Indonesia. Forktail 21 : 147-160. Newbold, T„ Scharlemann, J. P. W, Butchart, S. H. M., Sekercioglu, C. H„ Alkemade, R., Booth, H. & Purves, D. W. (2013) Ecological traits affect the response of tropical forest bird species to land-use intensity. Phil. Trans. R. Soc. Lond. B Biol. Sci. 280: 2012-2131. Norman, J. A., Christidis, L., Westerman, M. & Hill, F. A. (1998) Molecular data confirms the species status of the Christmas Island Hawk-Owl Ninox natalis. Emu 98: 1 97-208. Noske, R. A. & Saleh, N. (1996) The conservation status of forest birds in West Timor. Pp.65-74 in D. Kitchener & A. Suyanto, eds. Proceedings of the first international conference on Eastern Indonesian-Australian vertebrate fauna, Manado, Indonesia, November 22-26, 1994. Olsen, J. &Trost, S. (2007) Diurnal raptors on the island of Sumba, Indonesia, in June/July and December/January 2001-2002. Austral. Field Orn. 24: 158-166. 106 COLIN R. TRAINOR et al. Forktail 29 (2013) Pilgrim, J. D., Leadley, J. D. & Saifuddin (2000) Bird surveys and conservation status of four forests on Flores. Cambridge UK: CSB Conservation Publications. Raptor Conservation Society (2011) Study on distribution, population, habitats and ecological aspect of Flores Plawk Eagle ( Nisaeetus floris) on Rinjani National Park and other protected areas on Lombok. Final report. Cipanas, Java: Raptor Conservation Society. Rensch, B. (1931) Die Vogelwelt von Lombok, Sumbawa und Flores. Mitt. Zool. Mus. Berlin 17:451-637. Robson, C. R. (201 1 ) The Lesser Sundas 4-22 September 2011. Tour report. Birdquest. Available at http://www.birdquest-tours.com. Salvador, D. J. I. (1994) Socio-economic incentives for conservation of rainforest habitat of the Philippine Eagle Pithecophaga jefferyi. Pp.277- 282 in B.-U. Meyburg & R. D. Chancellor, eds. Raptor conservation today. Mountfield, UK: World Working Group on Birds of Prey and Owls/Pica Press. Sergio, F., Newton, I. Marchesi, L. & Pedrini, P. (2006) Ecologically justified charisma: preservation of top predators delivers biodiversity conservation. J. Appl. Ecol. 43: 1 049-1 055. Thiollay, J. M. (1994) Family Accipitridae (hawks and eagles). Pp. 52-205 in J. del Hoyo, A. Elliott & J. Sargatal, eds. Handbook of the birds of the world, 2. Barcelona: Lynx Edicions. Trainor, C. R. (2002) The birds of Adonara, Lesser Sundas, Indonesia. Forktail 18: 93-100. Trainor, C. R. (2005a) Species richness, habitat use and conservation of birds of Alor Island, Lesser Sundas, Indonesia. Emu 105: 127-135. Trainor, C. R. (2005b) Birds of Tapuafu Peninsula, Roti Island, Lesser Sundas, Indonesia. Forktail 21: 121-131. Trainor, C. R. (2007) Birds of Damar Island, Banda Sea, Indonesia. Bull. Brit. Orn.Club 127 (3): 8-28. Trainor, C. & Lesmana, D. (2000) Exploding volcanoes, unique birds, gigantic rats and elegant ikat: identifying sites of international conservation significance on Flores, East Nusa Tenggara. Bogor, Indonesia: PKA/ BirdLife International/WWF Report no. 1 1 . Trainor, C. R. & Soares, T. (2004) Birds of Atauro Island, Timor-Leste (East Timor) Forktail 20: 41-48. Trainor, C. R. & Verbelen, P. (in press) New distributional records from forgotten Banda Sea islands:the birds of Babar, Romang, Sermata, Leti and Kisar, Maluku, Indonesia. Bull. Brit. Orn. Club. Trainor, C. R., Benstead, P. J., Martin, K., Lesmana, D., Agista, D„ Benstead, M. C„ Drijvers, R. & Setiawan, I. (2006) New bird records for Nusa Tenggara Islands: Sumbawa, Moyo, Sumba, Flores, Pulau Besar and Timor. Kukila 1 3: 6-22. Trainor, C. R„ Imanuddin, Aldy, F. & Walker, J. S. (2009) The status and conservation of the Endangered Wetar Ground-dove (Gallicolumba hoedtii) and other wildlife on Wetar Island, Indonesia, 2008. Final Technical Report. UK: Columbidae Conservation. Trainor, C. R., Verbelen, P. & Johnstone, R. E. (2012) The avifauna of Alor and Pantar, Lesser Sundas, Indonesia. Forktail 28: 77-92. Van Biers, M. (2004) The Lesser Sundas Southeast Indonesia 7-24 July 2004. Tour Report. BirdQuest. Available at http://www.birdquest-tours.com. Verbelen, F. (1996) Birding in Sumba and Timor, Lesser Sundas. Indonesia. Unpublished. Verheijen, J. A. J. (1975) Some data on the avifauna of Roti, Lesser Sunda Islands, Indonesia. Zool. Meded. (Leiden) 50: 1-21 . Verhoeye, J. & Flolmes, D. (1999) The birds of the islands of Flores - a review. Kukila 10:3-59. Verhoeye, J. & King, B. (1990) Notes on three raptor species new to Flores. Kukila 5: 61 -68. Voris, FI. K. (2000) Maps of Pleistocene sea levels in Southeast Asia: shorelines, river systems and time durations J. Biogeogr. 27: 1 1 53-1 1 67. White, C. M. N. & Bruce, M. D. (1986) The birds ofWallacea (Sulawesi, the Moluccas & Lesser Sunda Islands Indonesia): an annotated check-list. London: British Ornithologists Union (Check-list No 7). Wink, M. & Sauer-Gurth, FI. (2004) Phylogenetic relationships in diurnal raptors based on nucleotide sequences of mitochondrial and nuclear gene markers. Pp.483-498 in B.-U. Meyburg & R. D. & Chancellor, eds, Raptor worldwide. Budapest: World Working Group on Birds of Prey / MME. Yong, D. L., & Lee, T. T. (2008) Nesting of Bonelli's Eagle Hieraaetus fasciatus in Yamdena, Tanimbar, Indonesia. BirdingASIA 10: 93. Colin R. TRAINOR, Research Institute for the Environment and Livelihoods, Charles Darwin University, Northern Territory 0909, Australia. Email: halmahera@hotmail.com Stephen J. S. DEBUS, Division of Zoology, University of New England, Armidale, NSW 2351, Australia. Email: sdebus@une.edu.au Jerry OLSEN, Institute for Applied Ecology, University of Canberra, ACT 2601, Australia. Email: jerry.olsen@canberra.edu.au Janette A. NORMAN, Department of Genetics, University of Melbourne, Parkville, Victoria, 3001, Australia. Email: janorman@unimelb.edu.au Les CHRISTIDIS, National Marine Science Centre, Southern Cross University, Coffs Harbour, NSW 2450, Australia. Email: les.christidis@scu.edu.au FORKTAIL 29 (2013): 107-119 The species of white-nest swiftlets (Apodidae, Collocaliini) of Malaysia and the origins of house-farm birds: morphometric and genetic evidence EARL OF CRANBROOK, GOH WEI LIM, LIM CHAN KOON & MUSTAFA ABDUL RAHMAN The taxonomy of South-East Asian swiftlets (Apodidae, Collocaliini) has proved challenging because of their limited variation in size and plumage colouration. Of particular interest are 'white-nest' swiftlets, whose nests, built almost entirely of hardened secretions from paired sublingual salivary glands, are valued in the edible birds'-nest trade. The natural breeding sites of white-nest swiftlets are caves or grottoes but, for over a century, there has been a progressive increase in numbers occupying man-made structures. Through most of South-East Asia there is now a developed industry, utilising sophisticated practices to attract and retain white-nest swiftlets in purpose-made buildings, known as'house-farms' — a novel form of domestication. A review of the systematics of wild populations based on museum skins collected in late nineteenth and early twentieth centuries, before the expansion of house-farms, concludes that there are two largely allopatric species of white-nest swiftlet in Malaysia, identified as Grey-rumped Swiftlet Aerodramus inexpectatus, with subspecies A. /. germani and A. /. perplexus, and Thunberg's or Brown-rumped Swiftlet Aerodramus fuciphagus, with subspecies A. f. fuciphagus and A. f. vestitus. During 2003 to 2010, house-farm swiftlets in southern Thailand, east and west coasts of Peninsular Malaysia, Sarawak, Java and southern East Kalimantan, Indonesia, were photographed to show variability in plumage of the rump. House-farm birds of Sarawak resembled neither of the wild species occurring naturally in the state. Tissue samples from embryos in eggs were collected for genetic studies from house-farms in Medan, Sumatra, west and east coasts of Peninsular Malaysia, and Sibu, Sarawak. Results of phylogenetic analyses, AMOVA and pairwise Fsrcomparison based on the partial cytochrome-fa sequence are presented. Of the 11 haplotypes identified, two are restricted to a wild population of Brown-rumped Swiftlets A. f. vestitus of Middle Baram, Sarawak, thereby shown to be genetically distinct from house-farm birds. One haplotype is common among all house-farm birds, two are unique to Medan, three and one to Kuantan and Endau-Rompin, respectively. The birds from Sarawak share haplotypes with all other house-farm populations in Peninsular Malaysia and Medan, Sumatra. The evidence for two clades within house-farm samples indicates that Peninsular Malaysian birds combine genetic components from north (A. inexpectatus germani) and south (A. f. fuciphagus). Sarawak house-farm birds are similar to east coast Peninsular Malaysian populations in plumage characters and genes, and apparently arrived by spontaneous immigration from Peninsular Malaysia. If hybrids have arisen among Malaysian house-farm white-nest swiftlets, they are excluded from regulation by the International Code of Zoological Nomenclature. INTRODUCTION Swiftlets are small swifts Apodididae, subfamily Apodinae, tribe Collocaliini (Chantler 1999), inhabiting the Indo-Pacific region and reaching greatest diversity in South-East Asia. A shared character of most swifts, including swiftlets, is the production of a dense secretion from a pair of sublingual salivary glands that serves as structural or binding material to form the nest (Chantler 1999). Termed ‘nest-cement’, this salivary secretion is the edible component, and is sufficiently copious in the nests of some swiftlets to make them commercially valuable. Edible birds’-nests have been esteemed in Chinese society since at least the late sixteenth century, and there is a long history of harvesting from natural wild colonies (Medway 1963, Lim & Cranbrook 2002). Most sought-after and expensive are ‘white’ nests, composed wholly of the edible salivary material with, at most, the incorporation of a few small feathers from the body plumage of the adult birds, probably adhering accidentally. Sequencing of genetic material (mitochondrial DNA; mtDNA) derived from commercial edible birds’-nests has distinguished authentic nests of Indonesian white-nest swiftlets from counterfeit products derived from nests ofEfouse Swift Apus affinis = nipalensis (Lin et al. 2009). However, this study did not attempt to discriminate between the nests of different swiftlet species. One, two or three species of white-nest swiftlet? Lack of distinctive external characters has caused persistent difficulty in defining species limits among swiftlets. For many years all were included in a single genus Collocalia. A series of papers by Stresemann ( 1914, 1925, 1926) culminated in a revision of species in the Malaysian subregion (Stresemann 1931). In this paper, the author acknowledged the loan of swiftlet skins from the Raffles Museum, Singapore, supplemented by specimens in the museums at Tring, Leiden and Berlin. Basing his taxonomy chiefly on wing length, tail length and furcation, and tarsal feathering, Stresemann ( 1931) combined a group of dull blackish-brown swiftlets in a single widespread Indo-Malayan species for which the prior name was Collocalia francica (Gmelin, 1789), the Mascarene Swiftlet. He noted that the type of nest was variable within this species, as defined, and listed subspecies building white nests: germani , inexpectata, javensis, vestita and micans. Of these, three occurred in localities now within Malaysia and Singapore. First, Germain’s Swiftlet Collocalia francica germani Oustalet, 1876, type locality Pulau Condore (=Con Son island), Vietnam (Plate 1A), was seen by Stresemann (1931) in the form of skins collected in 1913 by H. C. Robinson on Koh Pennan (= Koh Phangan), east coast of peninsular Thailand (Plate IB). He characterised these birds as having tarsus invariably unfeathered, and rump much paler than the back, ‘whitish grey with blackish shafts’; wing 113-121 mm, tail 5-53 mm, furcation 5-7 mm. Thus defined, C. f. germani extended through southern (peninsular) Thailand and Peninsular Malaysia ‘nearly as far as Johore’. At this point, Stresemann considered that C. f germani intergraded with a subspecies having rump ‘as a rule of the same colour as the back’. However, in the transition zone, ‘individual variation is great in some localities, specimens with dark rumps being found together with light-rumped ones’ (Stresemann 1931: 87). The dark-rumped subspecies was identified as C. f vestita (Lesson, 1843), type locality Sumatra, and the variable population in the transition zone zs germani > < vestita. This nomenclature indicated a north-south cline among white-nest swiftlets in Peninsular Malaysia, from a subspecies that was pale grey-rumped with dark 108 EARL OF CRANBROOK etal. Forktail 29 (201 3) shaft-streaks to a uniformly dark-rumped subspecies, across a transition zone in the south where individuals of both patterns were mixed. Although shown below to be erroneous, this interpretation by a respected ornithologist proved influential on subsequent opinion. Stresemann (1931) also applied the name vestita to dark- rumped specimens from Borneo, of which he saw six in the Berlin Museum from Tamaluang cave. East Kalimantan, and ten in the Raffles Museum from eastern North Borneo (now Sabah). He found no valid name for the dark-rumped white-nest swiftlets of Java, which he described as a new subspecies C. francica javensis , type locality Ceribon (Stresemann 1931: 89-90), distinguished by rump ‘a little paler than the back but by no means as light as in germani , wing 109-1 16 mm, tail 49-53 mm, furcation 4—7 mm (n=6). He also noted that a series of eight swiftlets collected by Chasen in Singapore had ‘mostly a very great similarity with the Javanese C. f. javensis' , wing 113-118 mm, tail 47-52 mm, furcation 4-7 mm (Plate 2D). The first modification of Stresemann’s (1931) scenario followed a survey of the birds’-nest caves of Sabah by Chasen (1931). New specimens, not seen by Stresemann, showed that grey- rumped swiftlets occupied small caves and grottoes on the Mantanani Islands (Plate 1 D), off the west coast, and Berhala Island in Sandakan harbour, on the east coast (Plate IE), while the white- nest swiftlets in caves at Gomantong, ‘only a few miles away and within sight of Berhala’, were dark-rumped (Plate 2F). On the grounds that, despite the close proximity of Berhala and Gomantong, the grey-rumped and dark-rumped white-nest swiftlets remained distinct, Chasen (1935) treated the two populations as separate species. The grey-rumped swiftlets from Sabah islands he considered to be to be ‘absolutely inseparable from tru e. germani' (Chasen 1935), and followed Stresemann (1931) in listing these under the trinomial C. francica germani. He also recognised that the distinct dark shaft-streaks of the dull brownish grey rump of C. francica perplexa Riley, 1927 of Maratua Islands, East Kalimantan, Indonesia, confirmed affinity with germani and therefore included this as a subspecies among the grey-rumped swiftlets. For the dark-rumped birds, he raised the name vestita to species rank, with the English name Brown-rumped Swiftlet. He also noted that Brown-rumped Swiftlets occurred at other inland caves in Sabah, at Baturong, Madai, Tapadong and, once again not far from the coast, near Lahad Datu. In Sarawak, white-nest swiftlets of the two kinds were recorded by Banks (1935), again separated by habitat but nonetheless treated as a single species. Grey-rumped Swiftlets (as C. francica germani) occurred ‘in several suitable places around the coast, such as the two Pulo Satang and Pulo Lakei, nesting in the soft sandstone crevices’. At inland localities in Sarawak, Banks (1935) recorded dark-rumped white-nest swiftlets (as C. francica vestita ) in limestone caves of the Middle Baram. The only other locality for vestita in Sarawak known to Banks (1935) was a small colony in a sandstone cave in Ulu Suai, yielding ‘a couple of katties’ of nests (i.e., around 140 nests). In Peninsular Malaysia ail nesting records ofwhite-nest swiftlets were from coastal or island locations. No occupied inland caves were known (and none has since been discovered). On the west coast Chasen ( 1 935, 1 939) and his successor at the Raffles Museum, Gibson-Hill (1948, 1949), agreed that white-nest swiftlets from peninsular Thailand and islands of northern Peninsular Malaysia were identical with topotypes of Germain’s Swiftlet (Plate 1A, IB), displaying a pale grey rump, almost white, with distinct, broad dark longitudinal stripes that involve both shafts and vanes of the rump feathers. The west coast range of these ‘Northern Grey-rumped Swiftlets’ (C. francica germani) included Penang and Selangor. On the evidence of Allen (1948), Gibson-Hill (1949) provisionally added the Sembilan Islands, Perak. White-nest swiftlets of the south of Peninsular Malaysia, including east coast islands and rocky stacks of the Pahang-Johor archipelago (specifically, Tioman, Tinggi and Tokong Gantong), were characterised by Chasen (1939) as having the rump darker than Northern Grey-rumped Swiftlets. Judging that this character justified separation at subspecies level, Chasen (1939: 123) called these birds ‘Southern Grey-rumped Swiftlets’, and ‘found it convenient to use for them the name proposed by Dr H. C. Oberholser, amechana' . At the same time, echoing Stresemann ( 1 93 1 ), he reiterated the mixed appearance of the swiftlets in south Peninsular Malaysia: ‘There is a considerable amount of variation in the colour of the rump: in some birds it is almost as pale as in the northern subspecies, C. f germani, but in other specimens it is much darker and only slightly paler than the back’. In a later survey of the east coast islands Gibson-Hill (1948) found white-nest swiftlet colonies from Pulau Nyireh in the Tenggol group, Terengganu, through the Tioman archipelago, Pahang, to the Pulau Tinggi group and Pulau Batu Gajah, Johor. Following Chasen, he too identified these as C. francica amechana (Gibson-Hill 1949). To be consistent with his discoveries in Borneo, Chasen (1935, 1939) recognised dark-rumped birds sympatric with Southern Grey-rumped Swiftlets in the south of Peninsular Malaysia as a second species, Brown-rumped Swiftlet Collocalia v. vestita , conspecific with those of interior caves of Borneo to which he applied the trinomial C. vestita maratua Riley, 1 927. However, he was unwilling to overturn the views of Stresemann on the north- south cline in Peninsular Malaysia. Commenting on his decision to recognise the species C. vestita , Chasen (1935) wrote: ‘but otherwise, in our arrangement of this very difficult genus, we follow the latest reviewer, Dr E. Stresemann in Bull. Raffles Mus. 6. 1931, p. 83.’ Gibson-Hill (1949: 110) took a narrower view, identifying Brown-rumped Swiftlet ‘only from Tioman [island], where it is breeding in the neighbourhood ofjuara Bay, and the adjacent coast of Johore’. Opinion subsequently remained unsettled on species limits and nomenclature of the white-nest swiftlets of territories now comprising Malaysia. In Borneo, Smythies (1957) recognised two species, noting that among the grey-rumped group Hume’s Swiftlet Collocalia inexpectata Hume, 1873, type locality Andaman Islands, had priority and therefore naming the birds of Sarawak and Sabah C. inexpectata germani, restricting C. i. perplexa to the type locality, Maratua Island. For the Brown-rumped Swiftlets, Smythies (1957) restricted Collocalia vestita vestita to the Natuna Islands, Indonesia, and C. v. maratua to Maratua Island, applying C. vestita mearnsi Oberholser, 1912 to birds of mainland Borneo. Later, Smythies (i960) retained this treatment of Brown-rumped Swiftlets, but placed the Grey-rumped Swiftlets in Collocalia francica , and subsequently in C. fuciphaga (Smythies 1968). Meanwhile, Medway (1966a) showed that the type of nest is a reliable taxonomic indicator among swiftlets, and that an unmistakable illustration of a white edible nest accompanied the description of Hirundo Fuciphaga Thunberg, 1812, overlooked by Stresemann (1914). This is therefore the oldest available systematic name forwhite-nest swiftlets ofjava, reducing Stresemann’s javensis to synonymy. Nuclear and mitochondrial DNA sequencing has subsequently confirmed that Mascarene Swiftlet {now Aero dramas francicus ) is a distinct species, confined to Mauritius and Reunion, Indian Ocean (Johnson & Clayton 1999). Medway (1966a) accepted the existence of a north-south cline through Peninsular Malaysia to Singapore, linking Germain’s or Northern Grey- rumped Swiftlets with the dark-rumped swiftlets ofjava, but differed from previous opinion by proposing that sympatry of grey- rumped and brown-rumped taxa in north and north-west Borneo could be explained if the two forms were the ends of a Rassenkreis or ‘circle of overlap’ (Mayr 1942), thereby justifying their inclusion in a single ‘ring’ species. Forktail 29 (2013) White-nest swiftlets (Apodidae, Collocaliini) of Malaysia and the origins of house-farm birds 109 1 A IB 1C ID IE IF Plate 1 . Grey-rumped Swiftlets Aerodramus inexpectatus from caves. (1A) Topotype A i. germani from Pulau Condore, Vietnam. 1882, USNM. (IB) Koh Phangan, Thailand. 1912, AMNH. (1C) Satang Kechil, Sarawak. 1932, RMBR. (1 D) Manttanani, Sabah. 1 931, RMBR. (IE) Berhala, Sabah. 1931, RMBR. (IF) A. /'. perplexus from Maratua. 1 927, RMBR. 2A 2B 2C 2D 2E Plate 2. Thunberg's Swiftlet A. f. fuciphagus and Brown-rumped Swiftlets A. f. vestitus from caves. (2A) Thunberg's Swiftlet from inland cave at Jampea, Java. 1960, NHMUK. (2B) Thunberg's Swiftlet from coastal cave at Karangbolong, Java. 1960, NHMUK. (2C) Topotype of Brown-rumped Swiftlet from Sumatra. USNM. (2D) Thunberg's Swiftlet from Singapore. 1931, RMBR. (2E) Brown- rumped Swiftlet from Baram, Sarawak. 1957, NHMUK. (2F) Brown-rumped Swiftlet from Gomantong, Sabah. 1958, NHMUK. no EARL OF CRANBROOK et al. Forktail 29 (2013) 3A 3B 3C 3D 3E 3F Plate 3. Sympatric specimens of Grey-rumped Swiftlet and Thunberg's Swiftlet collected around 3°N in Peninsular Malaysia. (3A) A. inexpectatus germani from Malacca. 1953, RMBR. (3B) A. inexpectatus from Selangor. 1879, NHMUK. (3C) A. fuciphagus from Selangor. 1887, NHMUK. (3D) A. inexpectatus from Tioman. 1907, RMBR. (3E) A. fuciphagus from Tioman. 1907, RMBR. (3F) A. amechanus topotype from Anamba Is., Indonesia. 1899, USNM. 4G 4H 41 4J 4K 4L Plate 4. Variations in rump shade in house-farm birds. (4A) Bukit Imbiah, Singapore. (4B) Sajira, Java. (4C) Pak Phanang, Thailand. (4D) Miri, Sarawak (4E) Kuching, Sarawak. (4F) Penang. (4G) Penang. (4H) Kota Bharu. (41) Pusing, Perak. (4J) Johor Bahru. (4K) Johor Bahru. (4L) East Kalimantan. Forktail 29 (201 3) White-nest swiftlets (Apodidae, Collocaliini) of Malaysia and the origins of house-farm birds 111 Brooke (1970, 1972) divided the swiftlets into three genera, recognising the Giant Swiftlet (now Waterfall Swift) as monotypic Hydrochous gigas and, among the remainder, restricting Collocalia to the small swiftlets with glossy plumage and separating as Aerodramus the group of middle-sized drab blackish-brown species, to which white-nest swiftlets belong. Until the discovery that the Pygmy Swiftlet Collocalia troglodytes utters an echolocating call (Price et al. 2004), it was thought that the capacity to orientate in darkness by echolocation was a further defining character of Aerodramus. Monroe & Sibley (1993), Inskipp et al. (1996) and, following these checklists, regional field guides by Lim & Gardner (1997) and Robson (2002) continued to combine all species except the Waterfall Swift in the genus Collocalia. However, molecular studies have confirmed genetic boundaries between Hydrochous, Aerodramus and Collocalia (Lee etal. 1996, Thomassen et al. 2003, 2005), and these genera were recognised by Chantler (1999), Smythies (1999), Wells (1999), Strange (2001), Mann (2008) and Phillipps & Phillipps (2009). Salomonsen (1983: 65) suggested that there could be three white-nest species: Collocalia fuciphaga (with vestita, dammermani, micanszndinexpectata as subspecies), C.germani (with amechana) and possibly C. perplexa with amelis of the Philippines. Monroe & Sibley (1993) recognised two species: Collocalia fuciphaga (including inexpectata and vestita ) and C. germani. In recent publications, Robson (2002) and Phillipps & Phillipps (2009) followed, listing two species: Grey-rumped ( germani ) and Brown- rumped ( vestita grouped with fuciphaga), whereas others including Chantler (1999), Smythies (1999), Wells (1999) Lim & Cranbrook (2002) and Jeyarajasingam (2012) have treated all white-nest swiftlets as a single species under the prior name Aerodramus fuciphagus. Wells (1999: 459) criticised the arbitrary nature of species boundaries within dines of changing rump colouration, and called for more research where different-looking populations meet. Origins of house-farming and house-farm white-nest swiftlets The propensity of swiftlets to select hollows, rock-shelters or caves as nest sites is reflected throughout their range by many instances of occupation of similar man-made structures, such as culverts, multi-storey car-parks, houses, barns or other buildings. White- nest swiftlet ‘farming’ began with the spontaneous occupation of buildings by birds and the responses of people. The earliest instances arose in Java, with the first reputedly in 1880 at Sedayu, East Java (Lim & Cranbrook 2002). In western Java, in I960 Medway (1961) was told that the birds nesting in outbuildings around three sides of a courtyard of a large country house, near Jakarta, had been present for about 60 years. Elsewhere in Java by that time there were already many buildings, domestic or industrial, in which colonising swiftlets had been encouraged by a variety of modifications to thrive and increase. From such beginnings, enterprises steadily developed. The buildings involved, whether modified from existing structures or purpose-built, have become known in English as ‘house-farms’, and the management of the swiftlet colonies within them as ‘house-farming’ (e.g. Nugroho & Whendatro 1994). The swiftlets occupying house-farms are normally allowed free egress to forage for food and water (Marzuki 1994). An important advance in Java was the discovery that eggs of house-farm swiftlets could successfully be transferred to nests of Linchi Swiftlet Collocalia linchi, which will hatch and rear the fostered chicks. The procedure was widely promoted and became standard practice (Nugroho et al. 1994). In Peninsular Malaysia, an early house-farm colony in Penang was studied by Langham (1980). Although wildlife protection legislation covered all swiftlets, thereby rendering illegal any operation involving the handling of the birds or interference with their nests, clandestine house-farm developments continued. Trailing the process in Indonesia, the great expansion of swiftlet house-farming in Peninsular Malaysia was a phenomenon of the last decade of the twentieth century. The town of Sitiawan, Perak, became the foremost mainland centre, with more than 50 shophouses undergoing conversion by the end of 1 999 (Ng 2000a). Simultaneously, public health and nuisance concerns were being raised (Ng 2000b) . It was claimed that the repeal of Malaysia’s Rent Control Act with effect from 1 January 2000 incentivised the process (Tan 2000). At present, in 20 1 3, few towns are without modified or specially constructed premises and, with government encouragement, others have been erected in rural areas. On the internet, many sites provide video clips of the birds and bird-houses, and several offer consultancies on management and manuals in English, Bahasa Malaysia and Chinese. Active associations of bird-house owners and nest traders have been established in most Malaysian States. A report on the industry by Merican (2007) provided guidance through current complexities and, following an initiative of the Federal Veterinary Department (Fadzilah A’ini 2007), in 2009 the Malaysian Department of Standards published provisional guidelines on good husbandry practice (MS2273:2009P). In the history of the relationship between humans and animals, house¬ farming of swiftlets has become a prominent and novel form of domestication. Where a systematic name is required, it has been customary to identify house-farm birds as Aerodramus fuciphagus or Collocalia fuciphaga. The multiplication of house-farms has not been restricted to Malaysia. Through much of tropical South-East Asia there have been entrepreneurial developments in the adaptation of existing structures and the construction of new, purpose-designed buildings, coupled with practices to attract and hold new colonists, especially the use of recorded vocalisations. Many urban house- farms now exist in Vietnam, notably in Khanh Hoa and Tien Giang provinces and Ho Chi Minh City (Phach & Voisin 2007), and between 2003 and 2009 activity developed in Cambodia (Poole 2010). The increase in numbers and expanding geographical range of house-farm white-nest swiftlets raise questions on the origins of these birds and their relations with natural wild populations. In Vietnam, Phach & Voisin (2007) found that urban house-farms swiftlets were not the native Germain’s Swiftlets of island caves (Phach etal. 2002), but resembled the house-farm birds of Sumatra and Malaysia. They concluded that immigration and colonisation of buildings in towns occurred spontaneously during the 1970s. Occupying separate nesting habitats, with different breeding seasonality and dissimilar diets, the two forms behave as separate species. Yet in southern Thailand Aowphol et al. (2008), finding very low genetic diversity of mtDNA among swiftlets of ten house- farms along the coasts of the Gulf of Thailand and the Andaman Sea, concluded that this was a single panmictic population, and attributed the observed genetic homogeneity to regular mixing by natal dispersal between wild population in natural sites on coastal islands and house-farm birds on the adjoining mainland. It is an aim of the present paper to discover which, if either, of these contrasting scenarios prevails in Malaysia. Since the skies are now crowded with house-farm swiftlets, evidence to determine the identity of potential wild ancestors must rely on collections made before the practice was so prevalent, i.e. before the mid-twentieth century. Thanks to good curation, many specimens on which taxonomic judgments can be based still exist in museums in USA, Europe and South-East Asia. A review of historic museum specimens, notably from the overlap zone in southern Peninsular Malaysia, leads to clarification of the original geographic boundaries of wild species and subspecies. A photographic survey of house-farm swiftlets of Malaysia has 112 EARL OF CRANBROOK et at. Forktail 29 (2013) illustrated plumage variation within and between colonies that can be compared with museum skins. The extent to which this variation is matched by genetic diversity was investigated by sequencing mtDNA cytochrome-^ (cyt -b). From the combined data, it becomes possible to form a view of the relations of house-farm white-nest swiftlets of Malaysia with putative source species. Other than countries, provinces or states, localities mentioned are listed in a gazetteer (Appendix 1). METHODS Morphometric studies Skins of swiftlets collected in the nineteenth and twentieth centuries, before the expansion of house-farming, were examined in the following museums: American Museum of Natural History, New York (AMNH), United States National Museum, Washington (USNM), National Museum ofNatural History, Paris (MNHN), Naturalis, Leiden (RMNH), Sarawak Museum, Kuching (SM), Raffles Museum of Biodiversity Research, University of Singapore (RMBR), and the Natural History Museum, Tring (NHMUK), where loans from the Academy of Natural Sciences, Philadelphia (ANSP) were also seen. Particularly crucial have been skins in RMBR which include those originally seen byStresemann ( 1931), Chasen (1935, 1939) and Gibson-Hill (1949). These are now very fragile, and liable to shed feathers at the lightest touch. It has been possible to take photographs but not to risk the handling necessary to check wing or tail measurements. Between 2003 and 2010, with the agreement of owners or managers, juvenile house-farm birds were photographed on the nest at Pak Phanang, Thailand, and Miri and Kuching, Sarawak. To ensure that they were fully fledged, other birds were caught in flight inside, emerging from or returning to house-farms located in Peninsular Malaysia in the states of Penang, Perak, Kelantan, Terengganu and Johor, and in Sarawak at Miri, Bintulu and Sarikei; also in Indonesia at Sajira, Banten, Java and southern East Kalimantan. The number of swiftlets caught at each house-farm varied from one to four. Birds were held singly in cloth bags for short periods. Standard procedure was then to measure wing-length and tail, closed, note the state of moult in the primary tract and rectrices, photograph the dorsal and ventral aspects, and the feet, and then to release the bird. A dead bird from a new house-farm in Sulawesi was also measured and photographed. In addition, swiftlets in natural colonies occupying the former underground military works at Bukit Imbiah, Sentosa Island, Singapore, were caught and handled by these procedures. Genetic studies Eggs or nestlings ofwhite-nest swiftlets were collected from: house- farms at Medan, Sumatra, Indonesia (nine individuals); the west coast of Peninsular Malaysia at Sitiawan, Perak, and Selangor (12 individuals); the east coast of Peninsular Malaysia at Kuantan ( 1 1 individuals) and Rompin (five individuals), Pahang, and Endau, Johor (six individuals); and Sibu, Sarawak (four individuals). Six samples were also taken from wild white-nest swiftlets occupying Salai cave, Middle Baram, Sarawak. The collected specimens were kept in 70% ethanol at room temperature at the sampling site and later at -20°C in the laboratory. Total genomic DNA was extracted from tissue using Promega Wizard' Genomic DNA Purification Kit following manufacturer’s instructions. The partial cyt b sequence was amplified using the primers Cyt523 (forward) and Thr (Reverse) (Thomassen et al. 2003). The polymerase chain reaction (PCR) mixture contains a final concentration of 0.5 |iM of each primer, 1 x reaction buffer, 2.5 jiM MgCL, 0.2 |iM of each dNTP, and 2.5 unit of Taq polymerase and ~60 ng of DNA template. The reaction was run using a Perkin Elmer GeneAmp 9600 Thermocycler with the programme set at 94°C for 3 minutes; 29 cycles of 94°C for 35 seconds, 55°C for 45 seconds and 72°C for 1 minute; 72°C for 5 minutes; hold at 4°C. The PCR products were purified using the Promega PCR Clean-Up System following the manufacturer’s instructions. Direct sequencing was commercially done by First Base Laboratories Sdn. Bhd. (Malaysia) (Goh 2007). The DNA sequences were trimmed to readable bases on both ends of the strands. In most cases the scoring of the bases started by the light-strand complementing the light-strand towards the centre. All sequences obtained were deposited with GenBank (JF269226-JF269236). The cyt b haplotypes were defined by ARLEQUIN 3.1 (Excoffier et al. 2005) and DNaSP (Rozas et al. 2003). Haplotype sequences were aligned using the ClustalX vl.81 (Thompson et al. 1997). The neighbour-joining (NJ) and most parsimonious (MP) trees were reconstructed using 1,000 bootstrap replicates in Molecular Evolutionary Genetic Analysis (MEGA) 4 (Tamura et al. 2007) and Phylogenetic Analysis Using Parsimony (PAUP) v4.0b (Swofford 2002), respectively, based on the cyt b haplotype matrix. The cyt b sequence of two white-nest swiftlet individuals, named as Aerodramus fucipbagus germani (DHC04; Price etal. 2004) and Aerodramus fucipbagus vestitus (DHC40; Price et al. 2004), were retrieved from GenBank (accession numbers AY294429 and AY294428, respectively) and incorporated into the phylogenetic analyses. Black-nest Swiftlet Aerodramus maximus lowi (Thomassen et al. 2003; Genbank accession number AY1 35623) was included as the outgroup in the phylogenetic trees. The genetic structure of the white-nest swiftlets was estimated using the analysis of molecular variance (AMOVA; Excoffier et al. 1992) and the pairwise comparison Fsr. Both analyses were performed using 10,000 permutations in the ARLEQUIN software. RESULTS Plumage characters and species limits Historic collections confirm the presence of Grey-rumped Swiftlets on the Mantanani Islands (Plate ID) and Berhala (Plate IE) and Brown-rumped Swiftlets in Gomantong caves (Plate 2F), Sabah (NHMUK, RMBR, USNM). Further observations have found only Grey-rumped Swiftlets on other islands of north-west and north Borneo. Sabah records have confirmed Mantanani Islands (Sheldon et al. 1983), and Francis (1987) added Batu Mandi, off Kudat, Balambangan Island, and Gaya (Bodgaya) and Si Amil, Sempurna bay. Francis (1987) also noted that birds from the Mantanani Islands had a slightly paler back and whiter rump than those of Berhala, assigning the former to the subspecies german i and the latter, by implication, to perplexus (Plate IF). No specimens are available of grey-rumped swiftlets from Gaya or Si Amil, but on geographical grounds these are also likely to be attributable to perplexus. RMBR holds two skins taken in 1932 by Banks on Pulau Satang Kecil, Sarawak (Plate 1C), confirming his record of germani from this group of islands (Banks 1935). A specimen was obtained on Satang Kecil in 1957 (NHMUK); it is poorly skinned but nonetheless shows a distinct whitish rump. Tom Harrisson, quoted by Smythies (1957: 653), reported that ‘about fifty pairs [have nested] most years since 1947 on Satang Besar and Kechil (two sea caves)’. Repeated searches around both islands from 1998 to 2008 by Lim & Cranbrook (pers. obs.) have failed to find occupied sea caves. Pulau Lakei, a site also mentioned by Banks (1935), and the nearby islet Batu Sarang, were investigated by Lim & Cranbrook (pers. obs.), but only Black-nest Swiftlets were found. These Sarawak colonies of grey-rumped white-nest swiftlets may now be extinct. Forktail 29 (2013) White-nest swiftlets (Apodidae, Collocaliini) of Malaysia and the origins of house-farm birds 113 Banks’s (1935) record of Brown-rumped Swiftlets in limestone caves of the Middle Baram is confirmed by specimens (Plate 2E). Lim (in Lim & Cranbrook 2002) has provided many photographs of this population. In November 1957, Cranbrook visited the sandstone cave in Ulu Suai noted by Banks (1935), and confirmed the presence of white nests. Two skins collected (NHMUK) are indistinguishable from Middle Baram Brown-rumped Swiftlets. In the altered landscape of modern Sarawak, the site has not since been rediscovered. Skins in RMBR collected in 1953 at Melaka (Malacca), although faded and foxed, show the characteristic pale rump with dark shaft-streaks (Plate 3A), thereby extending the historic range of Germain’s or Northern Grey-rumped Swiftlet southwards of previous records on the west coast of Peninsular Malaysia. In April 2009, on a brief visit to the Sembilan Islands, Cranbrook saw no swiftlets around Pulau Rembia, the site of Allen’s (1948) observations. However, on the rocky islet known as Batu Putih, underneath the tumble of huge, angular granite boulders, there were separate groups of seven black nests and 1 1 white nests. There were no eggs, and (around midday) no swiftlets present in the vicinity, leaving the identity of the white-nest builders unverified. Further south and east, a specimen from Horsburgh Light considered a stray by Gibson-Hill (1949) is still in RMBR. This lighthouse (now commonly known at Pedra Branca) was visited on 28 August 2012 when about 40 nests, with young, were present in the building. All were Black-nest Swiftlets and there was no evidence of any other resident swiftlet species (Geoffrey Davison pers. comm.). Skins in NHMUK are from Selangor around latitude 3°N, near the coast at Kelang and at interior locations. Of six collected (presumably shot in flight) in the vicinity of Kelang by W. Davison in 1879, mostly part of the Hume collection (Collar & Prys-Jones 2012), three (reg. nos 1887.8.1.297, 298 and 299), although faded and foxed with age, show pale rumps with distinct, dark longitudinal shaft-streaks, identifying them as Grey-rumped Swiftlets (Plate 3B). In three others, (1887.8.1.272, 300 and 301), the rump is uniformly coloured with the back, or slightly paler, with only the feather shafts dark, and no dark colour extending to the vanes (Plate 3C). Two other skins from interior Selangor also have dark rumps: 1887.8.1 .296 collected April 1 879 in Ulu Langat and 1908.12.15 collected in March 1907 (by H. C. Robinson) on Mengkuang Lebar at 4,300 ft (1,310 m) elevation. On the east coast islands, three birds were collected in Juara bay, Tioman Island, Pahang, in September 1907 (RMBR), of which two have the characteristic streaked rump of Grey-rumped Swiftlets (Plate 3D) but one is dark-rumped (Plate 3E), likely to be the specimen identified by Gibson-Hill (1949) as Brown-rumped Swiftlet. Despite the assertion of breeding by Gibson-Hill, there is no indication on the labels that any of these birds was taken at the nest. Medway (1966b) was told that white-nest swiftlets nested on Tioman in sea-caves, but failed to find any, and Lee (1977) repeated this assertion, again without location. There is also in RMBR a dark-rumped bird collected by Robinson in 1915 on Tokong Gantong, Johor, presumably the specimen noted by Chasen (1939). In Java, wild white-nest swiftlets collected in caves at coastal and inland sites in I960 (Plate 2A & 2B) have rump feathers the same colour as the back or slightly paler, without prominent dark shaft-streaks, matching the description of C. francica javensis Stresemann, 1931, now recognised as a junior synonym of Thunberg’s Swiftlet Aerodramus fuciphagus fuciphagus. Although old and faded, the dark-rumped swiftlets of southern mainland and islands of Peninsular Malaysia, identified by Chasen and Gibson-Hill as vestita (RMBR), are similar. As noted by Stresemann (1931), skins collected by Chasen in Singapore, in 1930-1931 (RMBR) are indistinguishable from Javan Thunberg’s Swiftlets. Photographs of white-nest swiftlets occupying former military underground emplacements at Bukit Imbiah, Sentosa Island, Singapore (Kang et al. 1991, Kang & Lee 1993: 18) and measurements and photographs of living adults mist-netted at this site in 2005 (Plate 4A) show that, by plumage character, these white-nest swiftlets of a natural colony are also identifiable as Thunberg’s Swiftlet. Re-examination of historic collections has therefore confirmed that, as in the Borneo States, there are two original wild white-nest swiftlet species in Peninsular Malaysia, grey-rumped and dark- rumped, evidently sharing the same diurnal habitat in a zone around 3°N on the mainland and east coast islands. The former are confirmed as nesting on the Pahang-Johor islands of Peninsular Malaysia, but not at Horsburgh Light (Pedra Branca). The latter nest on Singapore, but there is no confirmation that they do also on the most southerly Johor rocky stacks. The white-nest swiftlets of house-farms Java was the site of multiple early instances of spontaneous occupation of buildings by white-nest swiftlets of the native population of Thunberg’s Swiftlets. House-farm swiftlets of western Java, such as those handled in 2005 at Sajira, Banten (Plate 4B), are similar in size, plumage characters and tarsal feathering to wild Thunberg’s Swiftlets from caves at interior sites, such as Jampea (Plate 2A), or on the south coast at Karangbolong (Plate 2B). By the transportation and cross-species fostering of eggs in the nests of Linchi Swiftlets, the distribution of house-farm swiftlets has been enlarged to many new areas within the island of Java. Eggs from Java have also been traded, to an unrecorded and unknown extent, to localities beyond the natural range of the subspecies A. f. fuciphagus. In Kalimantan successful fostering of eggs from Java by White-bellied Swiftlets Collocalia esculenta cyanoptila is known as far north on the west coast as Singkawang, West Kalimantan (Charles Leh pers. comm. 2006) and on the east coast at Bayangkara, East Kalimantan (Lim & Cranbrook 2002: 149). In Singapore, Chasen observed prospecting swiftlets in the 1930s: ‘In January of two years I have found large numbers seeking the shady shelter of large stone-walled rooms, or vaults in buildings, in the late afternoon for roosting purposes: they were then easily caught with a large butterfly net.’ In a footnote he added: ‘Later. There is now a breeding colony of these birds in a much-frequented large building in Singapore’ (Chasen 1939: 119). These remarks are supported by skins in RMBR, collected on Singapore Island at various dates in January 1931, with a note on one label: Taken in a large building’. The dark rumps of these skins, concolorous with or slightly paler than the back, identify them as Thunberg’s Swiftlets. Later, Gibson-Hill (1948, 1949) reported swiftlets occupying an office building on Robinson Road, Singapore. The fate of this colony is not known but it is clear that in Singapore, by this time, there had been more than one spontaneous occupation of buildings by Thunberg’s Swiftlets. In north-west Peninsular Malaysia the pioneer birds occupying buildings were grey-rumped swiftlets. Gibson-Hill (1949: 110) reported Northern Grey-rumped Swiftlets (as C. francica germani) nesting in a godown in Penang, first noticed in 1 947, and ‘Southern Grey-rumped Swiftlets’ in the Federal Survey Office, Kuala Lumpur, alongwith grey-rumped swiftlets of uncertain subspecies in a building in Teluk Anson, Perak. In the 1960s, white-nest swiftlets (identity not determined) occupied government buildings in (then) Mountbatten Road, Kuala Lumpur, ultimately being excluded by the advent of air-conditioning and hence the glazing of all apertures (Cranbrook pers. obs. 1968). In the 1970s a small colony, defiantly persistent in the face of repeated nest removal, occupied the porcb of Kuala Lumpur Town Hall (Medway & Wells 1976); no specimens were collected. On the east coast ofPeninsular Malaysia, by 1 974 swiftlets were nesting in six sea-front shophouses 114 EARL OF CRANBROOK et al. Forktail 29 (2013) in Kuala Terengganu (Cranbrook pers. obs.). Specimens were not collected at that time, but the presumed origin of these birds would be ‘Southern’ Grey-rumped Swiftlets of the Redang or Tenggol groups of islands (Gibson-Hill 1949, Wells 1999). Swiftlet house-farming is a private and confidential enterprise, and in Peninsular Malaysia there is no authoritative data source for innovation or development in husbandry. There is, however, no evidence that the progressive increase in house-farm colonies in Malaysia has involved egg-transfer and fostering to a significant extent. One case of cross-species fostering in the nests of White- bellied Swiftlets reported to us was carried out at the town of Bentong, Pahang, around 2000-2002. An established population persisted in 2012 in the building used. In addition, other colonies have established themselves in this town, probably involving birds fledged from this source. There are no colonies of wild white-nest swiftlets in interior Peninsular Malaysia and, so far, no confirmed instance of swiftlets of the house-farm type establishing breeding colonies in natural sites. For instance, in the environs of Ipoh, Perak, there are numerous house-farms and abundant limestone caves that so far remain unoccupied (Cranbrook pers. obs., Tou Jing Yi in litt. 2011). The expanding population of house-farm swiftlets into new areas in Peninsular Malaysia therefore reflects an upsurge in recruits from pre-existing house-farms, reinforced by the imprinting of buildings as potential nest sites and the attraction of acoustic stimulus in the form of recorded swiftlet calls, now universally employed. No doubt, the increasing architectural sophistication of house-farm design also plays a part. But, essentially, Malaysian- fledged house-farm white-nest swiftlets seek familiar constructions to occupy, and do not look for natural sites. This behavioural trait can lead to ecological separation within common activity space, as has occurred in Vietnam (Phach & Voisin 2007). As among house-farm birds in Vietnam (Phach & Voisin 2007), throughout their range from southern Thailand, at Pak Phanang (‘Birds nest city’), through Peninsular Malaysia, and in Sarawak, at Miri and Kuching, nestling house-farm swiftlets in their first plumage have pale grey rumps (Plate 4C, 4D & 4E). Among adult house-farm swiftlets of Malaysia, our accumulated photo-record shows variability in rump colouration between and within colonies. At Penang, three from the same farm-house showed minor variation in rump shade, in all cases with moderately defined shaft- streaks (Plate 4F & 4G). At Kota Bharu, Kelantan, all three birds caught showed similar pale, brownish rumps with lightly defined shaft-streaks (Plate 4H). At Kuala Terengganu, poor pictures of four birds are sufficient to confirm similar rump patterns, varying slightly in lightness of shade. On the west coast, at Pusing, Perak, the general tone was darker, with two of four birds showing rump the same shade as the back but one paler, with dark shaft-streaks (Plate 41). In southern Peninsular Malaysia, five birds from house- farms in the neighbourhood ofKotaTinggi and Johor Bahru, Johor, all had rumps more or less mottled with darker feather centres; one was distinctive, with a uniformly pale band and narrow dark shaft-streaks (Plate 4J & 4K). In Sarawak, although there is anecdotal report of successful hand-rearing in Kuching of nestlings from an outside source (reputedly from Pontianak, West Kalimantan), house-farm owners have testified that there have been no transfers of eggs from Java or elsewhere. The dramatic spread of house-farm swiftlets into this state initially occurred in coastal locations, starting in the north¬ east. The first house in Miri was occupied in the mid-1990s. In Bintulu the first colonists noted were a pioneer group of 18 nests in the eaves of the MAS building in 1997 (Lim and Cranbrook pers. obs.), and by 2000 Mukah was colonised. These three towns now support many large colonies. The spread to south-west Sarawak was later: in 2000, an informant went every weekend all the way along the coast from Kuching westward to Sematan, testing with sound replay, and found no evidence of swiftlets (Tsai Mui Leong in litt. 2010). By 201 1, this coastline contained at least live house-farms with substantial colonies. Adult house-farm swiltlets in Miri, Bintulu, Sarikei and Kuching do not resemble either of the wild species of Sarawak, i.e. Germain’s or Grey-rumped on the islands, or Brown-rumped of interior caves. The house-farm swiftlets of Sarawak appear to be generally uniform in appearance, in rump colouration resembling most closely those of east coast Peninsular localities such as Terengganu and southern Johor. The similarity in appearance and size points to a common origin, leading to the conclusion that pioneer birds crossed the South China Sea from west to east, i.e. from Peninsular Malaysia to northern Sarawak. In the Kalimantan provinces of Indonesia, outside Malaysian borders, specialised house-farms have been constructed at many localities, urban and rural, not infrequently on a trial basis. Swiftlets from a house-farm on the coast of southern East Kalimantan, near Balikpapan, resemble the house-farm swiftlets of Sarawak (Plate 4L). A carcass from Sulawesi, brought from a new house-farm by Anton Hoo, was similar in size and appearance, representing a further trans-marine range extension by swiftlets of house-farm type. Genetic studies Cyt-b haplotypes and data matrix Eleven haplotypes are defined among the 55 sequences obtained (Table 1). Elaplotype 5 (H05) is the most common, shared by 31 individuals from all house-farm populations, but not by wild Brown-rumped Swiftlets^. f. vestitus of Middle Baram, Sarawak. Elaplotypes H04 and E107 are unique to the Medan house-farm population; H02, H03 and H 1 1 unique to that of Kuantan; H07 to Endau-Rompin; and H09 and El 10 to the wild swiftlets of Middle Baram. The Sibu birds share haplotypes with all other house-farm populations. The aligned DNA matrix is 558 bp in length, with 20 variable sites and no alignment gap. Among the variation sites, 10 sites are parsimony-informative (Table 1). Phylogenetic analyses based on the cyt-b haplotypes As the NJ tree shows no major topological difference from the MP tree, the NJ bootstrap values were mapped on the MP tree (Figure 1 ). Both NJ and MP trees recover two moderately supported major clades, Clade 1 and Clade 2, among the ingroups. Together, both clades include all haplotypes of house-farm birds, but none of the wild swiftlets of Middle Baram. Haplotypes H09 and H10 are exclusive to these swiltlets of Middle Baram. The specimen DHC04, which was identified as A. fuciphagus germani in Price et al. (2004), is included in Clade 2, while the specimen DHC40, which was identified as A. j. vestitus in Price et al. (2004), is unresolved among the ingroups (Figure 1). AMOVA and pairwise F ST comparison As there are two major clades of house-farm swiftlets recovered in the phylogenetic analyses (Figure 1), pairwise FST comparison and AMOVA were used to test the genetic structure suggested by the clustering pattern in the phylogenetic trees. Individuals represented by the haplotypes in Clade II were combined to define a population, while the remaining individuals define the other six populations according to their sampling sites (which are combined into six area groups), i.e. (1) Middle Baram, Sarawak, (2) Medan, North Sumatra, (3) combined west coast locations in Perak and Selangor of Peninsular Malaysia, (4) Kuantan, the central east coast of Peninsular Malaysia, (5) Endau-Rompin, the southern east coast of Peninsular Malaysia, and (6) Sibu, Sarawak. Pairwise comparison analysis shows that FST values are significant between the Middle Baram population and all other populations, and between the Clade 2 population and all other Forktail 29 (2013) White-nest swiftlets (Apodidae, Collocaliini) of Malaysia and the origins of house-farm birds 115 Table 1 . S ummary of the parsimony-informative sites and the distribution of the cyt b haplotypes in white-nest swiftlet. Site numbers of the parsimony-informative characters are shown vertically; dots indicate identity with DHC04 sequence and letters designate base substitutions. Parsimony-informative characters 111 1 2 2 3 3 4 3 0 1 3 7 4 9 5 8 7 Sampling areas Haplotype 3 6 1 2 4 6 7 1 1 7 West Coast Kuantan Endau-Rompin Sibu Middle Baram Sumatra DHC04 G T G A T c C G G C - - - - - H11 - 1 - - HOI A G C T T A T 2 2 3 1 - H02 A G C T T A T - 1 - H03 C A G C A A T - 1 - - - H04 C A G C A A T - , - - 1 H05 C A G C A A T 10 7 5 3 - 7 H07 C A G C A A T - 1 2 - - H08 C A G C A A T - - 1 - 1 H06 C A G c A A T - - - - - H10 A G c A A T - - - 5 - H09 A A G c A A T - - - 1 - DHC40 A A c A A T _ _ _ _ _ 68/51 56/54 \ 61/63 - H03 (Kuantan) - H04 (Sumatra) H05 (Kuantan, Endau-Rompin, West Coast, Sibu) - H07 (Endau-Rompin) - H08 (Sumatra) - H06 (Kuantan, Endau-Rompin) 99/99 f~ DHC04 (Balambangan Island) CLADE 1 House-farm swiftlets of unknown origin C 94/8 H11 (Kuantan) HOI (Kuantan, Endau-Rompin, West Coast, Sibu) H02 (Kuantan) CLADE 2 House-farm swiftlets showing affinity with germani - H10 (middle Baram) H09 (middle Baram) L- DHC40 (Gomantong) — — - Aerodramus maximus lorn Wild populations of Aerodramus fuciphagus vestitus Possible misidentification Figure 1. The phylogram of the most parsimonious (MP) tree based on cyt b haplotype sequence rooted by A. maximus lowi. Refer to Table 1 for the haplotype distribution. Figures next to the nodes indicate the NJ bootstrap values / MP bootstrap values. DNA sequences obtained from Genbank are shown as highlighted individuals. Table 2. Matrix of pairwise Revalues among six populations of the white-nest swiftlets based on cyt b sequence. Figures with asterisk indicate the values which are significant at p = 0.05. Cladell Endau- Rompin Kuantan Sibu Sumatra West Coast Endau-Rompin 0.00010* Kuantan 0.00000* 0.71201 Sibu 0.00356* 0.82398 0.99990 Sumatra 0.00000* 0.16929 0.46481 0.99990 West Coast 0.00000* 0.06871 0.08910 0.99990 0.20364 Middle Baram 0.00000* 0.00020* 0.00040* 0.00980* 0.00010* 0.00010* Table 3. Hierarchical AMOVA of the white-nest swiftlet populations. Fixation indices, i.e. the total variance (FSJ), the among population within group variance (F ) and among group variance (Fa), are shown for the various structures tested. Figures with asterisk indicate the values which are significant at p = 0.05. The maximum FCJ is highlighted in bold. Structure Groups Fsc Fa 1 (Clade 2), (Endau-Rompin, Kuantan, Sibu, Sumatra, West Coast), (Middle Baram) 0.77595* -0.02580 0.78158* 2 (Clade 2), (Endau-Rompin, Kuantan, West Coast), (Sibu), (Sumatra), (Middle Baram) 0.69453* 0.00268 0.69371 3 (Clade 2), (Endau-Rompin, Kuantan), (West Coast, Sumatra), (Sibu), (Middle Baram) 0.67924* -0.04615 0.69338* 4 (Clade 2), (Endau-Rompin, Kuantan, West Coast, Sumatra, Sibu), (Middle Baram) 0.74161* 0.60571* 0.34468 116 EARL OF CRANBROOK et al. Forktail 29 (2013) populations (Table 2). Among the various groupings tested in AM OVA, Structure 1 has the highest statistically significant FCT value (Table 3), suggesting that it is the most plausible genetic structure among the white-nest swiftlets based on the cyt b sequence. DISCUSSION Stresemann (1931) considered the variable population of white- nest swiftlets of the south of Peninsular Malaysia to be transitional members of a north-south cline, germani >< vestita. From experience in the field and with skins before them, Chasen, Gibson- Hill and Banks recognised two species of white-nest swiftlet in this area, as well as in the Borneo territories, Grey-rumped and Brown- rumped. Re-examination of historic museum specimens has confirmed that the two species overlapped in diurnal activity range in the south of Peninsular Malaysia. Rather than a clinal transition, a zone around 3°N therefore represents an area of contact where the two species shared a common feeding zone. Sympatric breeding ranges are not proven. The single dark-rumped bird shot on Tioman many years ago may have nested on that island as asserted by Gibson-Hill (1949) but, given the mobility and extensive daily foraging ranges of all swiftlets, it could equally have originated from Singapore or elsewhere within the range of Thunberg’s Swiftlet. Medway’s (1966a) suggestion that the situation in Borneo could be explained in terms of a Rassenkreis is redundant. Moreover, the classic example of a supposed ring species, the Great Tit Parus major , has been invalidated by morphological, acoustic and molecular data (cyt -b sequences) by Packert et al. (2005), thereby strengthening doubts about the place of this mechanism in speciation (Mayr 2002: 183). Available molecular evidence reinforces this conclusion. With samples from Sabah, Grey-rumped Swiftlets of Balambangan Island (as A. f. germani) and Brown-rumped {A. f vestitus ) from Gomantong caves, Lee et al. (1996) showed separation equivalent to the genetic distance between morphological species (with an anomalous result suggesting possible misidentification). Thomassen (2005: 161, Fig. 1) amplified the results of Price et al. (2004), again showing as great or greater genetic distance between the two as between many clades recognised on behavioural and morphological grounds as distinct species. The prior specific name for the dark- or brown-rumped swiftlets is Aerodramus fuciphagus. The observations of Stresemann (1931) are supported by historic specimens and recent photographs, confirming that Singapore white-nest swiftlets are indistinguishable from those ofjava, and are therefore A. fuciphagus fuciphagus. The dark-rumped swiftlets in historic collections from the south of Peninsular Malaysia, in NHMUK and RMBR, are also identifiable as A. f. fuciphagus. The type of Collocalia vestita maratua Riley, 1927 has been shown to be a Mossy-nest Swiftlet Aerodramus salanganus (Medway 1966a). This name is therefore not available for a Borneo subspecies of white-nest swiftlets, as proposed by Chasen (1935). Measurements and plumage characters do not distinguish the Brown-rumped Swiftlets of Borneo from those of interior Sumatra, type locality of Salangana vestita Lesson. Adthough nominate fuciphagus appears to intervene between these two separate populations, many authors, including latterly Smythies (1999) and Mann (2008), have used the nameH. fuciphagus vestitus for Borneo Brown-rumped Swiftlets. Further clarification, particularly genetic evidence, is needed to define the relationship of Bornean Brown-rumped Swiftlets with Thunberg’s Swiftlets ofjava and topotypical vestitus of Sumatra. In Peninsular Malaysia, both Chasen (1935, 1939) and Gibson- Hill ( 1 949) observed a darker and more variable rump-band among grey-rumped swiftlets of the east coast islands. As a subspecific name, Chasen (1935, 1939) chose Collocalia fuciphaga amechana, described by Oberholser (1912: 13) on the basis of two skins collected on Pulau Jemaja, Anamba Islands, Indonesia, by Dr W. L. Abbott in 1899. Oberholser compared these birds with white- nest swiftlets ofjava (known by him as typical Collocalia fuciphaga ), noting in particular that they were darker on the upperparts, with a metallic greenish sheen. This green sheen is clearly evident in a third skin, also from Pulau Jemaja (therefore a topotype), kindly loaned by ANSP (Plate 3F). Although Oberholser described the rump as ‘decidedly paler’ than the back, there is no demarcated pale rump-band with dark shaft-streaks. As Oberholser remarked, amechana is characterised by its unusual glossy colouration and, until details of its biology are known including the type of nest built, it is best regarded as an endemic of the Anamba Islands. If separable, the ‘Southern Grey-rumped Swiftlet’ of Peninsular Malaysia lacks a systematic trinomial. Among the grey-rumped swiftlets, while the diagnostic dark shaft-streaks remain distinctive, there is a peripheral cline from the palest, most contrastingpattern of the rump of germani ofVietnam and peninsular Thailand to a darker background shade of grey of the rump-band. In northern Borneo this is evident from the Mantanani group, Sabah, eastwards to perplexus in the Maratua Islands, Indonesia, and in Peninsular Malaysia from west and north to the southern islands of the Pahang-Johor archipelago. An extreme westerly outlier, with the rump marked by the distinctive blackish shaft-streaks on a dark grey background colour, is Hume’s (Edible Nest) Swiftlet Aerodramus inexpectatus of the Andaman and Nicobar Islands. As Smythies (1957) recognised, inexpectatus has priority as species name of the grey-rumped swiftlets. Malaysian representatives are therefore Germain’s or Northern Grey-rumped Swiftlet Aerodramus inexpectatus germani and, on the eastern islands of Sabah, Riley’s Swiftlet Aerodramus inexpectatus perplexus. Historical sources show that, in the region, wild white-nest swiftlets spontaneously colonised urban buildings at multiple sites. In Singapore, colonies of Thunberg’s Swiftlet were established in the 1930s. In Peninsular Malaysia, by 1949 grey-rumped swiftlets Aerodramus inexpectatus already occupied buildings in Penang, Telok Anson and Kuala Lumpur, and at Kuala Terengganu before 1974. There is no evidence that similar events occurred in the Borneo states and, in plumage characters, the house-farm swiftlets appearing in Sarawak during the 1 990s resemble neither of the wild species of Borneo. Although receiving only moderate statistical support, the genetic comparisons using mitochondrial cyt b sequence emphasise the distinctiveness of Brown-rumped Swiftlets from the Middle Baram caves, Sarawak (Figure 1). The uniqueness of this wild population is reflected in the pairwise distance matrix (Table 2) and the observation that the Middle Baram population shares no haplotypes with house-farm populations. Molecular analysis therefore matches plumage comparisons, and serves to stress that the lineage of house-farm swiftlets of Sarawak is distinct from the inland wild population of Bornean Brown-rumped Swiftlets. It is, however, of note that these results show a more distant relationship between the Middle Baram Brown-rumped Swiftlets and the Genbank specimen DHC40 from Gomantong, Sabah (identified as A. f. vestitus by Price et al. 2004). This apparent anomaly is possibly due to limitations of sampling design and molecular methods, but could also indicate misidentilication ol the specimen DHC40. It is not easy to distinguish in the hand between Brown- rumped and Mossy-nest Swiftlet A. salanganus , both of which occur at Gomantong, and the possibility of erroneous identification of the specimen from which the Genbank sequence derived has been raised elsewhere (Lee et al. 1996). Among the sample ol 49 house-farm individuals, phylogenetic and population genetic structure analyses show substantial gene- flow, but also suggest the existence of two clades. These clades, 1 Forktail 29 (2013) White-nest swiftlets (Apodidae, Collocaliini) of Malaysia and the origins of house-farm birds 117 and 2 (Figure 1), represent the grouping of house-farm swiftlets in the most plausible genetic structure (Table 3). Clade 1 includes house-farm swiftlets from the entire geographical range sampled, broadly between 2-4°N and 99-1 14+°E, coveringNorth Sumatra, across Peninsular Malaysia and Sarawak, but excludes haplotypes of all wild birds, represented by Brown-rumped Swiftlets of Middle Baram, Sarawak, and the two GenBank sequences from Sabah. This result is evidence that the wild swiftlet population of the Borneo states was not implicated in the ancestry of this clade. Clade 2 is significantly different from all separate populations sampled (Table 2). This clade includes nine house-farm swiftlets from the west and east coasts of Peninsular Malaysia and Sibu, Sarawak, i.e. approximately 2-4°N 100-1 14°E, along with specimen DHC04, collected on Balambangan Island, Sabah, 7.267°N 1 16.917°E, and reported to be Germain’s Swiftlet (as A. f.germani ) by Price etal. (2004). One individual from Kuantan (haplotype H 1 1 ) shows a strong genetic relationship with DHC04, while the other eight from both coasts of Peninsular Malaysia and Sibu (haplotypes H01 and H02) show a moderately close relationship with DHC04 (Figure 1). The inference is that Germain’s Swiftlet was implicated in the ancestry of Clade 2. The existence of two clades is likely to reflect diversity of origins among the house-farm swiftlets. As well as Java, where houses were first occupied more than a century ago and many innovative management processes originated, the range of Thunberg’s Swiftlets included Singapore, where buildings were occupied in the 1930s, and (at least in diurnal activity) southern Peninsular Malaysia to about 3°N as well. It is therefore expected that Thunberg’s Swiftlets contributed to the genetic diversity of modern Malaysian house-farm populations, possibly augmented by the transportation of Javan genetic material as eggs or fostered young. At the same time, or a little later, on the west coast of Peninsular Malaysia the first records of white-nest swiftlets occupying buildings, in Penang, and at inland localities in Perak and at Kuala Lumpur, were attributed to Grey-rumped Swiftlets of two subspecies by Gibson-Hill (1949). Peninsular Malaysia, therefore, appears to have become a mixing ground where house- farm lineages from two species have met. Such a mixed ancestry is reflected in observed variation in plumage, notably in rump colouration (Plate 4), and is supported by the recognition of two genetic lineages. In the Kalimantan provinces of Borneo, it is known that genes of Thunberg’s Swiftlets were introduced in house-farms by the transfer of eggs for fostering in the nests of the local White-bellied Swiftlet at more than one location. Nonetheless, Sarawak house- farm swiftlets resemble those of Peninsular Malaysia, and genetic studies confirm that this is the case. It appears that Sarawak birds arrived by immigration from west to east across the South China Sea, not later than 1990. After the immigration event (or events) to the north-east of Sarawak, the population of house-farm genotypes expanded south-westwards along the coast. It is no longer possible to test the extent to which the progressive increase in the population of swiftlets drew solely on locally bred recruits or was augmented by supplementary immigration. Long-distance movements across seas are not unexpected among swiftlets. The global distribution of Aerodramus species, embracing many remote islands from the western Indian Ocean to the Pacific (Chantler 1999), illustrates the natural mobility of this group of birds. The inclusion of Medan house-farm swiftlets in Clade 2 confirms genetic exchange across the Straits of Malacca. Phach & Voisin (2007) concluded that the colonisation of urban buildings in Vietnam by house-farm swiftlets was unassisted, representing a displacement of some 1,000 km, possibly including a sea-crossing. Further expansion in continental South- East Asia is shown by the appearance of house-farm birds in Cambodia (Poole 2010), in one direction, and eastwards to Sulawesi, Indonesia, again involving a sea crossing if not assisted by human intervention. In Sarawak, there has been one observation of one pair of swiftlets of the house-farm type being found nesting in caves, in Batu Lebik at Bukit Sarang, Tatau. However, the pair did not return the following season. In Peninsular Malaysia, there is so far no confirmed record of white-nest swiftlets of the house-farm type occupying caves. That this has not occurred in more than half a century suggests decisive imprinting of many successive house-farm generations, to seek only buildings as nesting sites. FUTURE PROSPECTS This study has shown the potential of the mtDNA cyt-b gene as a marker in assessing genetic relationships among swiftlets, including comparisons between wild and house-farm populations. Lirmer conclusions on the ancestry of Malaysian house-farm swiftlets could be achieved by sampling wild colonies of Grey-rumped Swiftlets of the east coast islands of Peninsular Malaysia and Sabah islands. As openness develops in the industry, it is to be hoped that there will be greater appreciation of the value of research and forthcoming sponsorship. As it was, our studies were self-funded, and therefore under-resourced. Results generated were limited, partly due to the small number of molecular markers and the lack of comprehensive sampling. Further sampling of adult birds is needed to test the relations between plumage character and genetics. Investigation is needed to determine the number of independent entries from wild sources in different parts of Malaysia, and to discover the extent to which these have generated genetically distinct lineages of house-farm birds. Future studies should incorporate longer DNA sequences and more DNA regions so that the bootstrap support values can be improved. Understanding the genetics of house-farm swiftlets could assist stakeholders in other ways. In the scenario of this newest domestication, with the backing of sound husbandry and good science, rational planning will be beneficial to ensure the perpetuation and sustainable management of this important avian resource. It may become possible to identify and propagate genotypes that show advantageous characters — for instance, those that are particularly productive, make nests of exceptional size or quality, or display strong fidelity to their home site. With disease inevitably threatening any birds kept in large numbers in close quarters, lineages offering genetic resistance may be identifiable. With enhanced understanding of the genome, it may even prove feasible to engineer deliberate crosses and thereby introduce other desirable characters. An aspiration of this study was to decide the correct systematic name for house-farm swiftlets of Malaysia. A firm decision is prevented by evidence that the original pioneers were drawn from at least three wild sources of two species: Northern Grey-rumped Swiftlets Aerodramus inexpectatus germani in Penang and Southern Grey-rumped Swiftlets A. inexpectatus subsp. in Kuala T erengganu, and Thunberg’s Swiftlet A. fuciphagus fuciphagus in Singapore, as also in Java. Further genetic evidence is needed, in particular from wild colonies of these three taxa. Future research may then provide a clearer understanding of the genetic relations between wild progenitors and, possibly, between local stocks of house-farm birds. Nuclear DNA markers will also be informative in determining whether house-farm swiftlets are products of hybridisation. II hybrids have been generated, they are excluded from regulation under the International Code of Zoological Nomenclature (ICZN 1999) Art. 1.3.3. Nonetheless, as a fertile, stable domesticate, a distinctive new form could be identified by an informal varietal name. We leave the choice of this name to the discretion of stakeholders. 118 EARL OF CRANBROOK et al. Forktail 29 (2013) ACKNOWLEDGEMENTS Our thanks are expressed to the respective directors and curators of the museum collections for facilities provided. House-farm owners are often anxious to prevent disturbance of their colonies, and uneasy at admitting strangers. It is therefore a rare privilege to be allowed access to premises occupied by white-nest swiftlets, and exceptional to be permitted to catch and handle living birds. Thanks to the kindness of owners, too many to be named individually, we have been able to measure and photograph house- farm white-nest swiftlets in most states ol Malaysia, and in parts of Indonesia. Special thanks to Datuk Seri Lim Chong Keat and Dr Tan Boon Siong, who have been of great assistance in providing introductions to swiftlet-house owners in Malaysia, and to Dr Boedi Mranata and Anton Hoo who kindly allowed birds to be netted in their swiftlet houses in Indonesia. Anton also provided a specimen from his house in Sulawesi. Samples for DNA analysis by Goh Wei Lim were provided byjohannes Siegfried (Medan), Kebing Wan (Baram), Dr Charles Leh (Sibu) and members of the Malaysia Birds’ Nest Merchants Association, namely Tan Y. T. and Lim Thiam Siew (Setiawan, Perak), Yap Kuang (Selangor), Sim Ee Poh and Lee Kong Heng (Kuantan, Endau-Rompin). Dr D. R. Wells and Dr G. W. Davison kindly read and commented on a draft of this paper. Dr Svetlana Nikolaeva kindly advised on the nomenclatural status of hybrids. We are also grateful to two anonymous reviewers, whose comments on a first submission have greatly improved the revised version. REFERENCES Allen, E. F. (1948) Nidification and other field notes on some Malayan birds. Malayan Nature J. 3: 82-86. Aowphol, A., Voris, H. K., Feldheim, K.A., Harnyuttanakorn, P. &Thirakhupt (2008) Genetic homogeneity among colonies of the White-nest Swiftlet (Aerodramus fuciphagus) in Thailand. Zoological Science 25: 372-380. Banks, E. (1935) Notes on birds in Sarawak, with a list of native names. Sarawak Mus.J. 4(14): 267-325. Brooke, R. K. (1970) Taxonomic and evolutionary notes on the subfamilies, tribes, families and subfamilies of the swifts (Aves: Apodidae). Durban Mus. Novlt. 9 (2): 1 3-24. Brooke, R. K. (1972) Generic limits in Old World Apodidae and Flirundinidae. Bull. Brit. Orn. Club 92: 53-57. Chantler, P. (1999) Family Apodidae (swifts). Pp. 388-457 in J. del Hoyo, A. Elliott & J. Sargatal, eds. Handbook of the birds of the world, 5. Barcelona: Lynx Edicions. Chasen, F. N. (1 931 ) Report on the birds' nest caves and Industry of British North Borneo, with special reference to the Gomanton caves. Jesselton: Government Printer. Chasen, F. N. (1 935) A handlist of Malaysian birds. Bull .Raffles Mus. 1 1 : i-xx, 1-389. Chasen, F. N. (1939) Birds of the Malay Peninsula, 4. London: H. F. & G. Witherby. Collar, N. J. & Prys-Jones, R. P. (2012) Allan Octavian Flume. BirdingASIA 17: 17-43. Excoffier, L, Laval, G & Schneider, S (2005) Arlequin ver. 3.0: An integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online 1 : 47-50. Fadzilah A'ini, A. K. (2007) Good animal husbandry practice for edible-nest swiftlets Aerodramus species: ranching and its premises. Department of Veterinary Services, Ministry of Agriculture, Malaysia. Francis, C. M. (1987) The management of edible bird's nest caves in Sabah. Sandakan: Wildlife Section, Sabah Forest Department. Gibson-Hill, C. A. (1 948) The Malayan swiftlets. Malayan Nature Jl. 3: 1 90-200. Gibson-Hill, C. A. (1949) An annotated checklist of the birds of Malaya. Bull. Raffles Mus. 20: 1-299. Goh Wei Lim (2007) Phylogeography and differentiation of white-nest swiftlet (Aerodramus fuciphagus) in Malaysia. M Sc thesis, Faculty of Science and Technology Resources, University Malaysia Sarawak, Sarawak, Malaysia. ICZN (International Commission on Zoological Nomenclature) (1999) International Code of Zoological Nomenclature 4th edn. London, International Trust for Zoological Nomenclature. Inskipp, T., Lindsey, N. & Duckworth, W. (1996) An annotated checklist of the birds of the Oriental region. Sandy, UK: Oriental Bird Club. Jeyarajasingam, A. (201 2) A field guide to the birds of Peninsular Malaysia and Singapore. Second edition. Oxford: Oxford University Press. Johnson, K. P. & Clayton, D. FI. (1999) Swiftlets on islands: genetics and phylogeny of the Seychelles and Mascarene swiftlets. Phelsuma 7: 9-1 3. Kang Nee & Lee Peng Guam (1993) Bird's nest soup: panacea or ill? Nature Watch 1: 15-18. Kang Nee, Flails, C. J. & Sigurdsson, J. B. (1 991 ) Nest construction and egg- laying in edible-nest swiftlets Aerodramus spp. and implications for nest harvesting. Ibis 1 33: 1 70-177. Langham, N. (1 980) Breeding biology of the edible-nest swiftlet Aerodramus fuciphagus. Ibis 122: 447-460. Lee, D.W. (1977) Animals of PulauTioman. Pp.20-27 in D.W. Lee, B. C. Stone, M. Ratnasabapathy &T.T. Khoo, eds. The natural history of Pulau Tioman. Merlin Samudra Tioman. Lee, P. L„ Clayton, D. FI., Griffiths, R. & Page, R. D. (1996) Does behaviour reflect phylogeny in swiftlets (Aves: Apodidae)? A test using cytochrome b mitochondria DNA sequences. Proceedings of the National Academy of Sciences, USA 93(14): 7091-7096. Lim, Chan Koon & Cranbrook, Earl of (2002) Swiftlets of Borneo: builders of edible nests. Kota Kinabalu: Natural History Publications (Borneo). Lim, K S & Gardner, D. 1 997 . An illustrated field guide to the birds of Singapore. Singapore: SunTree Publishing. Lin Jie-Ru, Zhou Hua, Lai Xiao-Ping, Hou You, Xian Xiao-Min, Chen Jian- Nan, Wang Pei-Xun, Zhou Lian& Dong Yan (2009) Genetic identification of edible birds' nest based on mitochondrial DNA sequences. Food Research International 42: 1 053-1 061 . Mann, C. F. (2008) The birds of Borneo: an annotated checklist. BOU Checklist No. 23. Peterborough: BOU. Marzuki, H.A.F. (1994) Prinsip-prinsip budidaya pemeliharaan burung walet. Surabaya, Biro Pusat Rehabilitasi Sarang Burung. Mayr, E. (1942) Systematics and the origin of species. New York: Columbia University Press. Mayr, E. (2002). What evolution is. London: Weidenfeld & Nicolson. Medway, Lord (1961 ) Birds' nest businessmen. The Sarawak Gazette February 28, 1961: 20-21. Medway, Lord (1 963)The antiquity of trade in edible birds'-nest. Federation Museums Journal 8 (N.S.): 36-47. Medway, Lord (1966a) Field characters as a guide to the specific relations of swiftlets. Proc. Linn. Soc. London 1 77: 151-172. Medway, Lord (1 966b) The fauna of PulauTioman: 4. The birds. Bull. Raffles Mus. 34: 39-52. Medway, Lord & Wells, D. R. (1976) The birds of the Malay Peninsula, 5. London: H. F. & G. Witherby. Merican, H. S. (2007) The 2007 Malaysian Swiftlet farming industry report summary and synopsis. Agricultural and Agro-based Businesses Sub- Committee, SMI Association of Penang. Monroe, B. L., Jr., & Sibley, C. G. (1993) A world checklist of birds. Newhaven: Yale University Press. Ng, S. H. (2000a) Birds nest business at'swallow hotels'. The Star January 19, 2000. Ng, S. H. (2000b) Group to check out 'swift hotels'. The Star January 21 , 2000. Nugroho, E. & Whendatro, I. W. (1994) The farming of Edible-nest Swiftlets Aerodramus fuciphagus in Indonesia. Semarang: Indonesian Swiftlet Lovers Association. Nugroho, E., Whendatro, I. & Madyana, I. M. (1994) Merubah rumah seriti menjadi rumah walet. Semarang: Eka Offset. (In Indonesian). Phillipps , Q. & Phillipps, K. (2009) Phillipps' field guide to the birds of Borneo. Oxford: Beaufoy Books. Oberholser, H. C. (1912) A revision of the forms of the edible-nest swiftlet, Collocalia fuciphaga (Thunberg). Proceedings of the U. S. National Museum 42 (1881): 11-20. Forktail 29 (2013) White-nest swiftlets (Apodidae, Collocaliini) of Malaysia and the origins of house-farm birds 119 Packert, M., Martens, J., Eck, S.„ Nazarenko, A. A., Valchuk, O. P., Petri, B. & Veith, M. (2005) The great Tit ( Parus majorha misclassified ring species. Biol. J. Linn. Soc. 86: 1 53-1 74. Phach Nguyen Quang & Voisin, J-P (2007) On an ecological form of the white- nest swiftlet Aerodramusfuciphagus (Aves, Apodidae) breeding in houses in Vietnam. Revue Ecologique (Terre and Vie) 62: 49-57. Phach Nguyen Quang, Yen Vo Quang & Voisin, J-P (2002) The white-nest swiftlet and the black-nest swiftlet: a monograph. Paris: Boubee. Poole, C. (2010) Swiftlet farming comes to Cambodia. BirdingASIA 1 3: 62-63. Price, J. J., Johnson, K. P.& Clayton, D. H. (2004) The evolution of echolocation in swiftlets. J. A vian Biol. 35: 135-143. Robson, C. (2002) A field guide to the birds of South-East Asia. London: New Holland. Rozas, J., Sanchez-DelBarrio, J. C., Messeguer, X. & Rozas, R. (2003) DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 1 9: 2496-2497. Salomonsen, P (1983) Revision of the Melanesian swiftlets (Apodes, Aves) and their conspecific froms in the Indo-Australian and Polynsian region. Noona Dan Papers no, 141. Kongelige Danske Videnskabernes Selskab Biologiske Skrifter 23 (5): 1-112. Schneider, S., Roessli, D., & Excoffier, L. (2000 )Ariequin version 2.000: a software for population genetics data analysis. Genetic and Biometry Laboratory, University of Geneva, Switzerland. Sheldon, F. H., King, B. F., Yong, D. & Francis, C.M. (1983) The birds of the Mantanani Islands. Sabah Society J. 7: 1 65-1 75. Smythies, B. E. (1 957) An annotated checklist of the birds of Borneo. Sarawak Museum JournaU: 523-818 Smythies, B.E. (1960) The birds of Borneo. Edinburgh & London: Oliver & Boyd. Smythies, B.E. (1968) The birds of Borneo. Second edition. Edinburgh & London: Oliver & Boyd. Smythies, B.E. (1999) The birds of Borneo. Fourth edition. Revised by G. W. H. Davison. Kota Kinabalu: Natural History Publications (Borneo). Strange, M. (2001) A photographic guide to the birds of Indonesia. London: Christopher Helm. Stresema n n, E. ( 1 9 1 4) Was ist Collocalia fuciphaga (Thun berg)? Verhandlungen Ornithologischen Gesellschaft Bayern 12: 1-14. Stresemann, E. (1 925) Bruchstucke einer Revision der Salanganen ( Collocalia ). Mitteilungen Zooiogischen Museum Berlin 1 2: 1 79-190. Stresemann, E. (1 926) Bruchstucke einer Revision der Salanganen (Collocalia) II. Mitteilungen Zooiogischen Museum Berlin 1 2: 349-354. Stresemann, E. (1931) Notes on the systematics and distribution of some swiftlets ( Collocalia ) of Malaysia and adjacent subregions. Bull. Raffles Mus. 6: 83-101. Swofford, D. L. (2002) PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sunderland, Massachusetts.: Sinauer Associates. Tamura, K., Dudley, J., Nei, M. & Kumar, S. (2007) MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Molecular Biology and Evolution 24: 1 596-1 599. Tan, J. (2000) 'Swift' way to earn money at home. The Star January 21, 2000. Thomassen, H. A. (2005) Swift as sound: design and evolution of the echolocation system in swiftlets (Apodidae: Collocaliini). Thesis, Leiden University. Enschede, Netherlands: Print Partners Ipskamp B.V., Thomassen, H. A., Wiersema, A.T., de Bakker, M. A. G„ de Knijff, P., Hetebrij, E„ & Povel, G. D. E. (2003) A new phylogeny of swiftlets (Aves: Apodidae) based on cytochrome-b DNA. Mol. Phylogenet. Evol. 29: 86-93. Thomassen, H. A., Tex, R.-J., Bakker, M. A. G.& Povel, G. D. E. (2005) Phylogenetic relationships amongst swifts and swiftlets: a multi locus approach. Mol. Phylogen. Evol. 37: 264-277. Thompson, J. D., Gibson, T. J„ Plewnial, F., Jeanmougin, F. & Higgins, D. G. (1997) The Clustal X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 25:4876-4882. Wells, D. R. (1999) Birds of the Thai-Malay peninsula, 1 . London: Academic Press. EARL OF CRANBROOK, Great Glemhom House, Saxmundham IP 17 1LP, UK. Email: lordcranbrook@greatglemhamfarms.co.uk (Corresponding author) GOH Wei Lim, Institute of Biological Sciences, Faculty of Science, University of Malaya, 50603 Kuala Lumpur, Malaysia. Email: weilim_goh@yahoo.com LIM Chan Koon, 1 55 Lorong 4 A, OffJalan Stampin Timur, 93350 Kuching, Sarawak, Malaysia. Email: limchankoon@yahoo.com Mustafa Abdul RAHMAN, Department of Zoology, Faculty of Resource Science and Technology, Universiti Malaysia Sarawak (UNIMAS), 94300 Kota Samarahan, Sarawak, Malaysia. Email: rmustafa@rimc.unimas. my Appendix 1 . Gazetteer: the coordinates of localities mentioned in the text. Balambangan L, Sabah 7.267°N 1 16.917°E Kuala Lumpur (Mountbatten Road) 3.133°N 101.683°E Redang 1., Terengganu 5.717°N 103.800°E Batu Mandi, Kudat, Sabah 6.917°N 116.950°E Kuala Terengganu (sea front) 5.317°N 103.150°E Robinson Road, Singapore 1.267°N 103.833°E Batu Putih, Sembilan 1. 4.000°N 100.500°E Kuantan, Pahang 3.81 7°N 103.317°E Rompin, Pahang 2.800°N 103.483°E Baturong caves, Sabah 4.700°N 118.017°E Kuching, Sarawak 1.550°N 110.350°E Sajira, Banten, Java 6.483°S 106.367°E Bayangkara, East Kalimantan 2.850°N 117.283°E Lahad Datu, Sabah 5.11 7°N 1 1 8.300°E Sarikei, Sarawak 2.1 17°N 111.517°E Belitung 1., Indonesia 2.900°S 107.933°E Lakeil., Sarawak 1.750°N 110.483°E Satang Besarl., Sarawak 1.783°N 1 10.150°E Bentong, Pahang 3.517°N 1 01 .900°E Malacca (Melaka) 2.250°N 102.233°E Satang Kecil 1., Sarawak 1.750°N 110.150°E Berhala L, Sandakan, Sabah 5.867°N 118.133°E Mantanani L, Sabah 6.700°N 116.333°E Sematan, Sarawak 1.800°N 109.767°E Bintulu, Sarawak 3.167°N 113.033°E Maratua 1., East Kalimantan 2.233°N 1 18.567°E Sembilan 1., Perak 4.000°N 1 00.533°E Bukit Imbiah, Sentosa 1., Singapore 1.250°N 103.800°E Medan, Sumatra 3.583°N 98.667°E SiAmill., Sabah 4.283°N 118.850°E Bukit Sarang (Batu Lebik), Sarawak 2.650°N 113.033°E Mengkuang Lebar, Genting Highlands 3.433°N 101.783°E Sibu, Sarawak 2.300°N 111.317°E Endau-Rompin, Johor 2.667°N 103.600°E Middle Baram, Sarawak 3.650°N 1 1 4.41 7°E Singkawang, West Kalimantan 0.900°N 108.983°E Gaya (Bodgaya) L, Sabah 4.617°N 1 18.733°E Miri, Sarawak 4.250°N 1 1 3.950°E Sitiawan, Perak 4.200°N 100.700°E Gomantong caves, Sabah 5.533°N 118.067°E Pak Phanang, Nakhon Si Thammarat 8.350°N 100.200°E Suai, Sarawak 3.783°N 1 13.617°E Horsburgh Light (Pedra Branca) 1.333°N 104.400°E Pontianak, West Kalimantan 0.033°S 109.317°E Tapadong caves, Sabah 5.083°N 108.133°E Ipoh, Perak 4.600°N 101.100°E Pulau Batu Gajah, Johor 2.483°N 103.850°E Tamaluang cave, East Kalimantan 0.100°S 115.700°E Jakarta, Indonesia 6.283°S 106.833°E Pulau Jemaja, Anamba Is., Indonesia 2.917°N 105.750°E Teluk Anson (Teluk Intan), Perak 4.000°N 101.033°E Johor Bahru, Johor 1 .550°N 103.800°E Pulau Nyireh, Terengganu 4.867°N 103.067°E Tenggol 1., Terengganu 4.783°N 103.950°E Koh Phangan, Surat Thani 9.750°N 1 00.01 7°E Pulau Rembia, Sembilan 1., Perak 4.000°N 100.533°E Tioman 1., Pahang 2.783°N 104.167°E Kota Bharu, Kelantan 7.417°N 102.250°E Pulau Tinggi, Johor 2.300°N 104.1 17°E Kota Tinggi, Johor 1 .71 7°N 103.900°E Pusing, Perak 4.467°N 101.000°E FORKTAIL 29 (2013): 120-123 Nesting period and breeding success of the Little Egret Egretta garzetta in Pattani province, Thailand SOMSAK BUATIP, WANCHAMAI KARNTANUT & CORNELIS 5WENNEN Nesting of Little Egret Egretta garzetta was studied between October 2008 and September 2009 in a colony near Pattani, southern Thailand, where the species is a recent colonist. Nesting was bimodal over a 12-month observation period. The first nesting period started in December in the middle of the rainy season (November-December).The second period started in March during the dry season (February-April). In the second period, nesting began in an area not occupied during the first period but gradually expanded into areas used in the first period. Egg and chick losses were high; the mean number of chicks that reached two weeks of age was 1.0 ± 1 .2 In = 467 nests), based on nests that had contained at least one egg. Considerable heterogeneity of clutch size and nest success was apparent between different locations within the colony. The main predator appeared to be the Malayan Water Monitor Varanus salvator. INTRODUCTION The breeding range of the Little Egret Egretta garzetta extends from western Europe (northern limit about 53°N) and North Africa across Asia south of the Himalayas to east Asia including Korea and Japan (northern limits about 40°N), with some isolated areas in southern Africa, the Philippines and north and east Australia (Hancock et al. 1978, 'Wongetal. 2000). Thus, the breeding range covers temperate, subtropical and tropical climate zones. The Little Egret is a colonial nesting species, constructing nests in trees, low shrubs and reedbeds. Several nesting studies have been conducted throughout the species’s breeding range, e.g. in France (Hafner et al. 2008), Greece (Kazantzidis et al. 1997), Israel (Ashkenazi & Yom-Tov 1997), India (Hilaluddin et al. 2003), China (Ruan etal. 2003, Wei et al. 2003, Wong 2003), and South Korea (Kim et al. 2006). Nesting success of the Little Egret in central Thailand was studied in the Wat Tan-en Non-Hunting Area (Keithmaleesatti et al. 2007) while the seasonality of breeding had previously been studied in the Thale Noi Non-Hunting Area, southern Thailand (Kaewdee 1999). Prior to this last-cited study, Little Egret was known only as as a winter visitor in the southern provinces of Thailand. However, in the second half of the 1 990s it expanded its breeding range 160 km to the south of Thale Noi and started nesting near Pattani (Figure 1). Here, information on the breeding of the Little Egret in this relatively new colony in southern Thailand is presented. The objectives were to obtain descriptive metrics for breeding success; understand nesting synchrony; and finally document if breeding success parameters varied spatially within the focal colony. MATERIALS AND METHODS Study area The study was conducted at the Pattani waterbird colony, which is located next to the local Central Prison (6.867°N 101.250°E) near Pattani Bay, Gulf of Thailand (Figure 1). Pattani is a mixed colony which includes Little Egrets, Cattle Egrets Bubulcus ibis and Little Cormorants Phalacrocorax niger. About 4,000 Little Egret nests are located in this colony. Within the fence enclosing the prison is a small, brackish wetland measuring about 180 x 240 m (approx. 4.3 ha) with a maximum water depth of 0.8 m in the rainy season. The wetland contains short stature White Mangroves Avicennia marina , some Red Mangroves Rhizophora mucronata and open spaces. The area is surrounded on three sides by a wall with barbed wire on top and on the fourth by the high wall around the prison buildings. The whole area is flat and largely covered by Holocene sand and clay deposits mainly of marine origin. The area has a tropical monsoon climate with the south-west monsoon from mid-May to mid-October and the north-east monsoon from mid-October to mid-February. The driest months are February to April, followed by moderate rain in May to September, while most precipitation occurs from October to December. Figure 1. Location of the study area near Pattani, Thailand. Figure 2. Spatial and temporal expansion by nesting Little Egrets of the Pattani colony during the two nesting periods in the 2008-2009 nesting season. Week 1 I! Week 2 HI Week 3 Forktail 29 (2013) Nesting period and breeding success of the Little Egret Egretto garzetta in Pattani province, Thailand 121 The nesting area was not homogeneous, with variation in both the density of the woodland and the tree species present. Prior to the nesting season, it was measured and divided into three sub- areas A, B and C of 1.44 ha (180 x 80 m). The outer two sub-areas were further subdivided into two parts of 0.72 ha each (90 x 80 m; see Figure 2). The middle part was not subdivided because it had large open spaces without trees. A total of five sections were therefore recognised, Al, A2, B, Cl and C2. This stratification was necessary to assess if breeding success parameters varied between strata. Data collection The colony was studied from October 2008 to September 2009. A fixed survey route that criss-crossed all sub-areas was delineated. A sample of nests were surveyed in each of the sub-areas in each nesting period. Sample sizes were determined prior to the study based on rough estimates of nest density. The sample sizes were as follows: Nestingperiod 1: 50, 100 and 100 nests for B, Cl and C2, respectively (there were no nests in the other sections during the first round of nesting); Nestingperiod 2: 100, 50, 30, 50 and 50 nests for Al, A2, B, Cl and C2, respectively. New nests without eggs were marked in each section until the predetermined sample size was reached. The colony was surveyed once every three days in the morning during the nesting season. Occupied nests along the route were marked with a numbered plastic tag placed below the nest to allow determination of nest outcomes. Surveys were temporarily stopped during brief, light rain showers. However, during heavy continuous rain, surveys were rescheduled for the next day. Surveys to check all marked nests took about four hours to complete. Nest content was checked using a mirror attached to a 2 m pole. The number of eggs and nestlings were recorded for each marked nest. Nestlings were aged each survey and placed in three age classes: hatchlings (1-4 days), young nestlings (5-9 days) and old nestlings (10-14 days) respectively. This classification by age is arbitrary. Surveys of nests containing nestlings older than 14 days were discontinued because these nestlings could move out of nests preventing individual identification. A nest was deemed to be successful if it contained at least one egg. The total number of Little Egret nests in the colony was estimated towards the end of each of the two nesting periods by delimiting the proportion of each sub-area where nesting had occurred. For this purpose two possible states, ‘nesting’ or ‘not- nesting’, were assumed. The mean density of nests within areas identified as ‘nesting’ was estimated for each sub-area randomly using two 10 x 10 m survey plots within such areas (80 x 90 m) since the focus was not on individual trees. The total number of nests for each species was calculated by multiplying the mean density within each sub-area by the corresponding area that contained suitable nesting trees and subsequently summed over all sub-areas. Estimates are therefore very crude. Data analysis Homogeneity of variance was tested using Levene’s test (SAS 2009). Generally, it was found that variances of measurements of nesting success (number of eggs and hatchlings) for each period and sub-area were homogeneous. Wilk-Shapiro tests were used to test for normality and it was found that measurements of success often deviated from normality. Therefore, non-parametric one-way analyses ofvariance (ANOVA) were also used (Kruskal- Wallis test) to test for differences among areas and among areas by nesting period for nest success (Tables 1 and 2). The results of these non- parametric tests were the same as the (parametric) analyses of variance results presented in this paper. This is not surprising, given the fact that the (parametric) analysis of variance is robust with respect to the assumption of the underlying populations’ normality (Zar 1984). A considerable body of literature (see Zar 1984) has concluded that the validity of the ANOVA is affected only slightly by even considerable deviations from normality, especially with increasing sample sizes. Thus, given the fact that variances of populations are (generally) homogeneous, that parametric analysis of variance are robust (especially to even considerable deviations from normality), supported by the fact that results of non- parametric analyses of variances provided similar results, it is believed that both the one-way and two-way ANOVA results presented in this paper are accurate. RESULTS Little Egrets started using the study area as a night roost early in October 2008 and abandoned it in late July 2009. The first nests were built by 1 December 2008 in the middle of the rainy season. The first eggs were found on 5 December 2008 and the first chicks were seen on 30 December 2008. During this period, more pairs initiated nest building, laid eggs and hatched chicks. The last eggs were reported on 28 J anuary 2009. Thus, the laying period extended over about 54 days, which includes replacement clutches after early egg loss. About 3.5 months after the start of the first nesting period (Nesting period 1), a second nesting period (Nesting period 2) began around 12 March 2009 in the dry season. New nests were constructed in sub-areas not used during the first period, but soon thereafter also expanded into sub-areas that had been used in the previous nesting period (Figure 2). During both nesting periods colony growth occurred mainly during the first three weeks. The first eggs of Nesting period 2 were found on 15 March 2009 and the last eggs were laid around 12 May 2009, resulting in a laying period of 58 days, similar to Nestingperiod 1. Clutch sizes ranged from 1 to 6 eggs, with an average clutch size of 2.8 ± 0.9 eggs. Clutch size was significantly different across nesting periods (two-way AN OVAs; Ps<0.0001) but not by sub- area, while chick rearing (all three stages) was significantly influenced by sub-area but not nestingperiod (Table 1). For Nestingperiod 1, clutch size did not differ among sub-areas (Table 2). Although number of young hatched, and 7 and 14 day old nestlings did not differ between sub-areas Cl and C2, nest success in these sub-areas was significantly higher than in sub-area B (Table 2). Nestingperiod 2 showed somewhat different results for nest success among sub-areas (Table 2; Figure 3). Generally, the highest nest success was found for sub-areas Cl and C2, followed by Al, A2, and B (Table 2, Figure 3). Specifically, for all nesting stages, sub-areas A2 and B showed significantly lower nest success than sub-areas Cl and C2. Clutch sizes were significantly lower in Nesting period 2 compared to Nestingperiod 1 for all sub-areas (B: FL59 = 22.27, P<0.0001; Cl F 1-I34 = 22.27 P<0.0001; C2 Flil3’3= 22.27, P<0.0001). No differences were found in nest success between Nesting periods 1 and 2 for either number of young hatched, or the number of 7- and 1 4-day-old young, respectively, for sub-areas B, Cl andC2 (Ps>0.10). The low nest success in sub-area B appears not to be the result of an initial small clutch size, but could be due Table 1 . Two way ANOVAs to determine effects of nesting period and sub-area on nesting stages of the Little Egret in the Pattani colony. Nesting stage Nesting period Fv P F,7 Sub-area P Clutch size 128.48 <0.0001 3.24 0.01 Hatching 1.69 0.19 11.17 <0.0001 Nestlings 7-days 0.15 0.70 10.29 <0.001 Nestlings 14-days 0.14 0.71 8.44 <0.0001 122 SOMSAK BUATIP, WANCHAMAI KARNTANUT & CORNELI5 SWENNEN Forktail 29 (2013) Table 2. Mean (± SD) clutch size, young hatched and nestlings 7 and 14 days old, based on successful nests, during the two successive nesting periods of the Little Egret in the Pattani colony during the 2008-2009 nesting season. Nesting period Sub- area Nt rfil Successful Clutch size (x ± SD) Young hatched (x ± SD) Nestlings (7 days) (x ± SD) Nestlings (14 days) (x ± SD) 1 B 33 3.3 ± 0.9a21 0.8 ± 1.4b 0.6± 1.3b 0.5 ± 1.0b Cl 87 3.3 ± 0.8a 1.7 ± 1.6a 1.5 ± 1.6a 1.1 ± 1.3a C2 84 3.3 ± 0.8a 2.0 ± 1.6a 1.6 ± 1.6a 1.2 ± 1.3a All 204 3.3 ±0.8 1.7 ± 1.6 1.4 ±1.6 1.0 ± 1.3 2 Al 92 2.5 ± 0.8a, b 1.5 ± 1.1a,b 1.3 ± 1.1a, b 0.9 ± 1.0a, b A2 46 2.2 ± 0.7b 1.0 ± 1.2b,c 1.0 ± 1.1b, c 0.7 ± 1.0b,c B 27 2.3 ± 0.7a, b 0.7± 1.1c 0.4 ± 1.0c 0.3 ± 0.7c Cl 48 2.7 ± 0.7a 2.0 ± 1.2a 1.9 ± 1.2a 1.5 ± 1.2a C2 50 2.7 ± 0.7a 1.2 ± 0.2a 1.8 ± 1.2a 1.3 ± 1.2a, b All 263 2.5 ±0.7 1.5 ± 1.3 1.3 ±1.2 1.0 ± 1.1 1) Successful nests contained at least 1 egg. 2) Clutch sizes, young hatched and nestlings 7 and 14 days old with the same letter do not differ significantly within a nesting period among areas surveyed (Bonferroni multiple range test, P = 0.05). to higher nest predation (Figure 3). In both periods, sub-areas Cl and C2 showed substantially higher success rates for young to 14 days averaging 45% survival for Nesting period 1 and 62% for Nesting period 2 (Figure 3). Sub-areas A1 and A2 showed survival rates intermediate between sub-areas B and Cl, and C2 (Figure 3). Little Egrets built nests more frequently in White Mangroves than in Red Mangroves in both nesting periods. Little Egrets nests had long thick twigs in the base and long thin twigs in the upper layer and were built at the lowest levels in the trees. Cattle Egrets constructed their nests of tiny twigs and in the middle layer of the foliage. Little Cormorants used thick short twigs in the base layer and twigs with leaves on top and placed their nests highest in the trees. Various predators noted in the colony were suspected of predating eggs and nestlings, including Fishing Cat Felis viverrina, Brahminy Kite Haliastur Indus, Large-billed Crow Corvus macrorhynchos, Malayan Water Monitor Varanus salvator and Siamese Cobra Naja kaouthia. Actual predation was not observed, but fresh nail scrapes on the bark of trees where nests were destroyed strongly suggesting that a large Malayan Water Monitor had climbed the tree and predated the nests. The effect of destruction of nests and the differences between sub-areas in different stages of the breeding process is summarised in Figure 3. Predation of nests in the sub-areas showed similar patterns between Nesting periods 1 and 2 (Figure 3). The highest predation rates were found in sub-area B: only 12.0% and 13.3% survival rates of the selected nests that produced 14-day-old young for Nesting periods 1 and 2, respectively. DISCUSSION In temperate climatic regions, the reproductive season of Little Egrets starts in spring when increased temperature and day length induce nesting. The species has typically one brood per year, but re-nesting may occur after clutch loss (Bauer & Glutz von Blotzheim 1966). In tropical areas daylight and temperature do not fluctuate much over the year and the nesting period is related largely to the rainy season that varies both temporally and regionally (del Hoyo etal. 1992). The nesting seasons summarised in del Hoyo et al. (1992) also indicate a unimodal and not bimodal nesting pattern. No studies were found that describe the bimodal pattern observed in Thailand. Nesting periods 1 and 2 were similar in length (54 and 58 days, respectively). The interval between occupation of the colony and the start of egg-laying in both nesting periods was about 100 days. Hancock et al. (1978) and del Hoyo et al. (1992) estimated the interval between egg-laying and independence of the nestlings to be least 68 days (incubation about 23 days, hatching to independence/fledging of young about 45 days). In Nesting period 2, nesting started in the sub-areas not occupied in Nesting period 1 and gradually expanded to all sub-areas. Egrets avoided nesting in sub-areas B, Cl and C2 where some fledglings of relatively late broods of Nesting period 1 were present (pers. obs.). Ali and Ripley (1987) summarised information from ‘Egret Farms’ in Sind (India), in which captive egrets were maintained to harvest valuable egret plumes. Captive, well-fed Little Egrets produced up to four or even five clutches between March and September when one-week old chicks were removed for hand rearing. This suggests that Little Egrets are neither genetically nor physiologically predisposed to one brood per year. Thus, food availability seems to be the driver for the start and continuation of nesting. There is little reason to presume that the Pattani colony was used by two different populations of Little Egrets. The fact that there were many more nests in Nesting period 2 than in Nesting period 1 may suggest that older and therefore more experienced birds nested in Nesting period 1. More experienced Little Egrets are likely to initiate the nesting cycle early and be able to nest twice, while less experienced birds start later and nest only Figure 3. Little Egret nest success in the two nesting periods at the Pattani colony during the 2008-2009 nesting season. Points on x-axis, 1 : nests surveyed; 2: nests with at least one egg; 3; nests with hatchlings; 4: nests with 7-day-old chicks; 5: nests with 1 4-day old chicks. Forktail 29 (2013) Nesting period and breeding success of the Little Egret Egretta garzetta in Pattani province, Thailand 123 once per year (Nesting period 2). The slightly larger eggs and clutch sizes during the first breeding period might support these assumptions, which could be verified by longer-term studies using colour-banded birds. It is unclear why there is not a gradual transition between these two nesting populations, resulting in a single long breeding season. However, there seems a benefit of synchronous nesting that is triggered by an environmental cue that signals the start of a nesting season. Several authors have suggested a relationship between the onset of the rainy season or water conditions in tropical areas (Hancock et al. 1978, Ali & Ripley 1987, del Hoyo et al. 1992). In this area, precipitation patterns in November and December cause local flooding that may improve feeding conditions and induce birds to start nesting. The heavy rain showers did not seem to affect nesting in Nesting period 1. The adults protected their eggs well during incubation and the rainy season had ended by the time the young hatched. Weather conditions were dry during Nesting period 2 and it was not clear what prompted egrets to begin a second round of nesting. Clutch sizes observed (average 2.8 ± 0.9 eggs) were similar to clutch sizes reported from other tropical areas summarised by Hancock et al. (1978). Nest, egg and chick losses were high at Pattani, resulting in a low number of nestlings surviving beyond 14 days. Most predatory attacks resulted in complete loss of a clutch or chicks and often partial destruction of the nest structure. The result was that a relatively low number of pairs reared chicks to 14 days (1.0 ± 1.2young). Hilaluddin et al. (2003) reported a slightly higher success of 1.74 nestlings up to 15 days from India. The highest successes were reported from China, 3.86 young by Ruan etal. (2003) and 3.96 young by Zhang etal. (2000), but the authors did not report whether nest loss was incorporated in these numbers. Partitioning the colony in sub-areas and carefully designing a survey route through the entire colony proved to be useful to determine spatial and temporal differences in nesting and nest success in the colony, even when initial clutch sizes did not differ by sub-area. This was an unexpected outcome, but suggests that studies of colonies should take into account that differences in nest initiation and establishment of pairs in a colony can have a strong spatio-temporal component which should be addressed in study design. Spatial heterogeneity within nesting sites of E. garzetta has not been reported, but has been seen in Cattle Egret Bubulcus ibis (Petry & Fonseca 2005). The cause of the differences in nesting success in different parts of the colony may be due to varying predation rates, but this requires confirmation using improved nest observation methods. In both nesting periods the losses in section B were highest while C 1 and C2 had the best results. The difference may be due to the less dense vegetation in B where there was also a large area of water allowing easier access by the Malayan Water Monitor and raptors. ACKNOWLEDGEMENTS We thank the director and staff of Pattani Central Prison, Department of Correction, for permission to conduct our study on their property, and the staff of the Pattani meteorological station and the Thai Meteorological Department for providing the weather data. Thanks go to the late Nihasem Waesalae for his kind help in constructing the mirror device to check nests, and for transport during the feeding range observations. We are grateful to Suradej Bahem and Muhammaasan Wande who successively helped to collect data in the colony and to Anthonie M. Holthuijzen for statistical analyses and suggesting improvements to the manuscript. REFERENCES Ali, S.& Ripley, S. D. (1987) Compact handbook of the birds of India and Pakistan. Second edition. Delhi: Oxford University Press. Ashkenazi, S.& Yom-Tov,Y. (1997) The breeding biology of the black-crowned night-heron (Nycticorax nycticorax) and little egret (Egretta garzetta) at the Huleh Nature Reserve. Israel J. Zool. 242(4): 623-641. Bauer, K. M. & Glutz von Blotzheim, U. N. (1966) Handbuch der Vogel Mitteleuropas Bd 1 . Frankfurt am Main: Akademische Verlagsgesellschaft. del Floyo, J., Elliott, A. & Sargatal, J., eds (1 992) Handbook of the birds of the world, 1 . Barcelona: Lynx Edicions. Hafner, H„ Dugan, P. J., Kersten, M„ Pineau, O. & Wallace, J. P. (2008) Flock feeding and food intake in Little Egret Egretta garzetta and their effects on food provisioning and reproductive success. Ibis 135: 25-32. Flancock, J., Elliott, H. & Gillmor, R. (1978) The herons of the world. London: London Editions. Hilaluddin, Shah, J. N. & Shawl, T. R. (2003) Nest site selection and breeding success by Cattle Egret and Little Egret in Amroha, Uttar Pradesh, India. Waterbirds 26: 444-448. Kaewdee, W. (1 999) Population study of waterbirds and the assessment of the suitability of Khuan Khi Sian,Thale Noi Non-hunting area as a Ramsar site. Master's Thesis, Inter-department of Environmental Science, Graduated School, Chulalongkorn University. Kazantzidis, S., Goutner, V., Pyrovetsi, M. & Sinis, A. (1997) Comparative nest site selection and breeding success in 2 sympatric Ardeids, black crowned night-heron (Nycticorax nycticorax) and little egret (Egretta garzetta) in the Axions Delta, Macedonia, Greece. Col. Waterbirds 20(3): 505-517. Keithmaleesatti, S„ Thirakhupt, K., Pradatsudarasar, A., Varanusupakul, P„ Kitana, N. & Robson, M. (2007) Concentration of organochlorine in egg yolk and reproductive success of Egretta garzetta (Linnaeus, 1758) at Wat Tan-en non-hunting area, Phra Nakhorn Si Ayuthaya Province, Thailand. Ecotoxicology and Environmental Safety 68: 79-83. Kim J. S., KooT. H., Oh FI. S. & Mori.T. (2006) Clutch size, reproductive success, and growth rate of the Little Egrets Egretta garzetta. J. Faculty of Agriculture, Kyushu University 51: 135-1 38. Petry, M.V. & Fonseca, V.S.D.S. (2005) Breeding success of the colonist species Bubulcus ibis (Linneus, 1758) and four native species. Acta Zoologica (Stockholm) 86:217-221. Ruan, L„ Zhang, Y., Dong,Y.&Mauro, F. (2003) Egretta garzetta as bioindicator of environmental pollution in Tai Lake region .ChineseJ. Applied Ecology 14(2): 263-268. (In Chinese.) SAS Institute Inc. (2009). SAS/STAT ® 9.2 User’s Guide, 2nd Edition. Cary, NC: SAS Institute Inc. Wei G. A., Chen X. L., Flu FI. J. & Chen J. R. (2003) Observation on some activities of reproduction in Little Egrets (Egretta garzetta) at Jiyu Island in Xiamen. Zook Res. 245: 343-347. Wong, L. C„ Corlett, R. T., Young, L„ & Lee, J. S. Y. (2000) Comparative feeding ecology of Little Egret on intertidal mudflats in Flong Kong, South China. Waterbirds 23 (2): 214-225. Wong, L. C. (2003) Egretry counts in Flong Kong, with particular reference to the Mai Poand Inner Deep Bay Ramsar site. The Hong Kong Bird Watching Society, Summer 2002 Report 1 : 1-16. Zar, J. FI. (1984) Biostatistical analysis. Englewood Cliffs, NJ: Prentice Flail. Zhang Y. M„ Li Y. D., Wang FI. & Fasola, M. (2000) Breeding Biology of Night Fleron (Nycticorax nycticorax) and Little Egret (Egretta garzetta) inTaihu Lake of Wuxi, China. Zoo!. Res. 21: 275-278. Somsak BUATIP, Wanchamai KARNTANUT and Cornells SWENNEN, Faculty of Science and Technology, Prince of Songkla University, Pattani 9400, Thailand. Email: kwancha@bunga.pn.psu.ac.th FORKTAIL 29 (2013): 124-127 The status of Brown-chested Jungle Flycatcher Rhinomyias brunneatus in Vietnam SIMON P. MAHOOD, SEBASTIEN DELONGLEE, FLORIAN KLINGEL, FALK WICKER & RICHARD CRAIK The number of records of some migratory species is so low that there are insufficient data to infer status, even in countries within their normal distribution. Brown-chested Jungle Flycatcher Rhinomyias brunneatus, a globally threatened bird, is one such species. We gathered data on the occurrence of this species and 1 3 other migrant flycatchers in the city of Hanoi, Vietnam, throughout autumn 2010. These data include the second to tenth records of Brown-chested Jungle Flycatcher in Vietnam, and it was the fifth commonest flycatcher recorded in Hanoi during autumn 2010. Records of the species spanned the period 2 September-4 October, thus suggesting that it is a relatively early migrant with a narrow migration period. We also comment on the incidence and patterns of occurrence of other flycatcher species in Hanoi. INTRODUCTION Compared with countries in temperate regions, the status of migrant birds in tropical countries is relatively poorly known. For most species, broad patterns of occurrence have been elucidated, and increasingly there are sufficient data to analyse seasonal, geographical and even trend data within certain areas or countries, such as Hong Kong and Thailand (Carey eta/. 2001, Round 2010). In Vietnam, broad patterns of occurrence are known for most migrants, but are based on relatively few data and remain incomplete for some species. One poorly known species in Vietnam is Brown-chested Jungle Flycatcher Rhinomyias brunneatus , which is unique in its genus in being a long-distance migrant (Taylor & Clement 2006). It is considered uncommon and localised within its breeding grounds in south-east China, and this is likely to have contributed to its listing as Vulnerable (BirdLife International 2012a). In common with other members of the genus Rhinomyias , it is a sluggish, unobtrusive forest interior species usually detected by voice (SPM pers. obs.); these traits render it liable to be under-detected. The species spends the non-breeding season in southern Peninsular Malaysia and Singapore (Wells 2007), and within this small range are found primarily in mature lowland moist evergreen forest; they show strong site-fidelity (Wells 2007). Small numbers are recorded annually on passage in Thailand (P. Round in lift. 2012). Assuming that it takes a direct migratory route, much of the global population estimated at 2,500-9,999 (BirdLife International 2012a) would be expected to pass through or over Vietnam. Robson (2011) listed one vagrant record of the species for Vietnam, an individual collected on the campus of the Agricultural University, Hanoi, on 26 April 1981. The bird was initially identified as a Red-eyed Bulbul Pycnonotus brunneus , a species endemic to the Sundaic lowlands of Peninsular Thailand, Malaysia and Indonesia (Stusak & Vo Quy 1986). However, knowing this identification to be untenable, C. Robson examined the specimen and reidentified it as the first, and until 2010, the only record of Brown-chested Jungle Flycatcher for Vietnam (C. Robson in litt. 2011, Robson 2011). The present paper re-evaluates the status of Brown-chested Jungle Flycatcher in Vietnam using data collected in Hanoi during 2010. Data are sufficient to document its status in East Tonkin (north-east Vietnam). Occurrence data for all other migrant flycatchers of the genera Muscicapa, Ficedula, Eumyias, Cyanoptila and Cyornis (genus limits following BirdLife International 2012b) in Hanoi are also presented for the first time, for the purpose of comparison with Brown-chested Jungle Flycatcher. METHODS Data collection Data on the occurrence ol migrant flycatcher species during autumn passage were collected between 27 August and 14 November 2010 in the only two accessible large green spaces in Hanoi, namely the Botanical Gardens (21.040°N 105.830°E) and Thonh Nhat Park, commonly called Lenin Park, (21.015°N 105.846°E). At both sites there are no resident populations of any flycatcher species (all authors pers. obs.), thus all flycatchers recorded can be considered migrants. Data were collected by most of the Hanoi-based birdwatchers (SPM, FW, FK, SD) and occasionally by the Ho Chi Minh City based RC. Observations were collated on the Vietnam Bird News blog (http:// vietnambirdnews.blogspot.co.uk). At least one of the two parks was visited on most days. On each visit the observer (very rarely observers) searched actively for flycatchers and recorded all individuals seen to species level. On the rare occasion that one of the parks was visited twice in one day (either twice by the same person or on separate occasions by different people) the highest single observer tally ol each flycatcher species is used here. There is thought to be no (or negligible) exchange of birds between the two sites, based on observations ol individually identifiable birds. Using the same method it is thought that all or almost all flycatchers remained for only one day. All birds seen were identified to species with reference to Robson (201 1) with the exception ol Blue-and-white Flycatcher Cyanoptila cyanomelana / Zappey’s Flycatcher C. cumatilis. Leader & Carey (2012) demonstrated that Zappey’s Flycatcher is a species distinct from Blue-and-white Flycatcher. The latter is now considered to include only the nominate and C. c. intermedia (Leader & Carey 2012). Since not all males were photographed in 2010, and because identification criteria for females are not yet fully worked out, in this study we assign these birds to Cyanoptila. A more thorough review of the status ol Blue-and-white and Zappey’s Flycatchers in Vietnam is ongoing (Mahood et al. in prep.). Visits to the parks by observers were temporally standardised — almost all visits took place during a one hour period between 07h45 and 08h45 (pre-work, but after the parks have been vacated by people partaking in mass organised exercise sessions), or, occasionally, between 12h00 and 13h00. Habitat in both parks is heterogeneous, but search efforts were spatially standardised because all observers focused on the best areas for flycatchers in the parks. In Lenin Park this was a scrubby area behind a permanently locked toilet block near the south entrance (people unable to access the toilet make use of the area behind it, thus attracting an abundance of flies) whilst in the Botanical Gardens Forktail 29(2013) The status of Brown-chested Jungle Flycatcher Rhinomyias brunneatus in Vietnam 125 this was a quiet scrubby area where a blocked drain overflowed and flooded shallow depressions in the grass, creating pools in which mosquitoes bred. In both of these areas the habitat was relatively open in structure, and consequently we believe that detection probabilities between species and observers were close to equal. Data analysis The study was divided into eight 10-day periods. To allow for variation in survey effort (the parks were not visited every day), data were corrected for number of visits, with each park treated separately. Within each period the number of records of each species in each park was divided by the number of visits to the park during that period, and then multiplied by ten ( the number of days in the period). Corrected data from the two parks were combined to give an incidence of abundance for each species within each 10- day period. For each species, the incidence of occurrence within the 10-day periods was summed to give an incidence of occurrence over the whole study. RESULTS Figure 2. The incidence of occurrence per ten-day period of Brown¬ chested Jungle Flycatcher in Flanoi during autumn 2010, corrected for observer effort. ■27 06-15 16-25 26 11-15 16-25 26 05-14 Aug-D5 S*p Ssp 5ep-Q5 Oct Oct Oct-04 Nov Sep Oct Nov Culicicapa ceylonensis , were also recorded during the study period, the first two as passage migrants and the last as a winter visitor, but were not systematically counted. The Botanical Gardens were visited on 40 days (mean 0.5 visits per day) and Lenin Park on 37 days (mean 0.46 visits per day). Thirteen Muscicapa, Ficedula , Eumyias , Cyanoptila and Cyornis flycatcher species were recorded, consisting of six long-range migrants (originating in Siberian Russia), four medium-range migrants (originating in central or southern China) and three altitudinal migrants (originating from as close as the mountains of northern Vietnam about 50 km to the north and west) (Figure 1) (species limits following BirdLife International (2012b), except where discussed below). Figure 1. The incidence of occurrence of flycatcher species in Flanoi during autumn 2010, corrected for observer effort. Key: a. Brown Flycatcher Muscicapa dauurica; b. Yellow-rumped Flycatacher Ficedula zanthopygia; c. Taiga Flycatcher F. albicilla ; d. Dark-sided Flycatcher M. sibirica; e. Brown-chested Jungle Flycatcher Rhinomyias brunneatus; f. Cyanoptila (see text); g. 'Chinese Blue Flycatcher' Cyornis rubeculoides glaucicomans; h. Flainan Blue Flycatcher C. hainanus; i. Snowy-browed Flycatcher F. hyperythra ;j. Mugimaki Flycatcher F. mugimaki;k. Verditer Flycatcher Eumyias thalassinus; I. Brown-breasted Flycatcher M. muttui; m. 'Green-backed Flycatcher' F. narcissina elisae; n. Ferruginous Flycatcher M. ferruginea. Brown-chested Jungle Flycatcher was the fifth commonest migrant flycatcher in Hanoi during autumn 2010 (Figure l). It is a relatively early passage migrant (Figures 2 & 3). Nine individuals were recorded — in the Botanical Gardens on 2, 9, 14, 23 and 28 September.and in Lenin Park on 21, 23 and 26 September and 4 October. Three additional species often grouped with flycatchers, namely Black-naped Monarch Hypothymis azurea, Asian Paradise- flycatcher Terpsiphone paradisi and Grey-headed Canary- flycatcher DISCUSSION At least during 20 10, Brown-chested Jungle Flycatcher was a fairly common autumn passage migrant in East Tonkin, Vietnam. Data corrected for effort indicate that this species was the most abundant short- or medium-range migrant flycatcher recorded during our study. It is difficult to account for the absence of records in earlier years. It seems unlikely that the recent upsurge in records reflects a genuine increase in abundance of the species on passage in Vietnam. Owing to its superficial similarity to Asian Brown Flycatcher Muscicapa dauurica it is plausible that birdwatchers overlooked the species in the past. However, given the number and quality of birdwatchers resident in or visiting Vietnam over the last 20 years this is unlikely. The almost complete absence of previous records can best be accounted for by a combination of migration strategy and birdwatcher behaviour. Most birdwatching aimed at observing passage migrants in Vietnam has taken place in coastal sites, where Brown-chested Jungle Flycatcher has not been recorded. It is possible that it avoids the coast during migration. The number of Brown-chested Jungle Flycatchers recorded during our study is remarkable, considering that during the last 1 0 years the number of birds recorded in Thailand (where there are considerably more birdwatchers and photographers and a well established network of reporting and disseminating information) is typically less than five annually, and there are still occasionally years when none is recorded (P. Round in litt. 2012). Our data might represent a tiny sample of the number of Brown-chested Jungle Flycatchers that pass through Vietnam every year. The results indicate that the species passes through Hanoi during a relatively short window centred on September. Indeed, over half of the records were made during a one-week period spanning 21- 28 September. However, it is possible that the timing of migration varies between years. Evidence that the occurrence of Brown¬ chested Jungle Flycatchers in Hanoi in 2010 was not a one-off phenomenon was provided in 20 1 2 when two or three individuals were recorded between 13 and 16 September (Le Manh Hung and J. C. Eames in litt. 2012). During the study Asian Brown Flycatcher was the most abundant migrant flycatcher and had a protracted migration period in keeping with a bird with a large source population and wide geographic range (although it was not recorded in August and there was an obvious peak in records in late September); it was followed by Yellow-rumped Flycatcher Ficedula zanthopygia , Taiga 126 SIMON P. MAHOODefa/. Forktail 29 (2013) Figure 3. Incidence of occurrence per ten-day period for all species of flycatcher recorded in Hanoi during autumn 2010, corrected for observer effort. Ten-day periods are the same as in Figure 2, species key as in Figure 1. 1.60 J b Flycatcher F. albicilla and Dark-sided Flycatcher Muscicapa sibirica. All four species have relatively large source populations (Taylor & Clement 2006) and are long-range migrants, although Yellow- rumped Flycatcher breeds as far south as north-east China (Brazil 2009). Timing of migration differs between these species: Yellow- rumped Flycatcher was only recorded during the first half of the study period and Taiga Flycatcher was not recorded before the beginning of October, whilst Dark-sided Flycatcher showed a protracted migration period with a peak in records that corresponded to that of Asian Brown Flycatcher. This peak might represent either a genuine similarity in migration timing or favourable conditions for grounding migrants in Hanoi. The remaining nine flycatcher species were recorded less often. Except for Mugimaki Flycatcher Ficedula mugimaki and potentially Cyanoptila, all of these species are exclusively short- or medium- range migrants. The small number of records of most of these scarcer species allows only tentative conclusions regarding the timing of their migration through Hanoi. Mugimaki Flycatcher records were spread out throughout the study period. In contrast, all of the Cyanoptila records were in mid- to late-October. The single record of 'Green-backed Flycatcher’ Ficedula narcissina elisae was also relatively late ( 1 November 2010). The latter has a similar breeding and wintering distribution to Zappey’s Flycatcher. Subsequently ‘Green-backed Flycatcher’ has been recorded in Hanoi in November 2012 (J. C. Eameshz litt. 2012) and the species was recorded twice in Cambodia on 19 and 20 November 2012 (R. Martin verbally 2012, SPM pers. obs.). These data indicate that this species migrates later than the other northerly breeding species Forktail 29(2013) The status of Brown-chested Jungle Flycatcher Rhinomyias brunneatus in Vietnam 127 in the study. This correlation of migration timing perhaps provides some support for the theory that most of the Cyanoptila records constituted Zappey’s Flycatcher rather than the more north¬ easterly breeding Blue-and-white Flycatcher sensu stricto C. c. cyanomelana and C. c. intermedia. Brown-breasted Flycatcher Muscicapa muttui has an atypical migration strategy for a China/ north Vietnam breeding species in that it overwinters in the Indian subcontinent (Rasmussen & Anderton 2005). Data indicate that it is a very early migrant in Hanoi, and this is reinforced by records made in subsequent years (SD pers. obs.). Our records of Snowy- browed Flycatcher Ficedula hyperythra are noteworthy because they are the first records of the species in the lowlands ofVietnam. They probably represent altitudinal migrants from the hills close to Hanoi. The period of passage for Brown-chested Jungle Flycatcher in Hanoi is earlier than the bulk of the flycatcher species. It fits within the known pattern of occurrence of the species in Thailand, where birds are typically recorded during late September and early October. It is much earlier than other central Chinese breeding flycatchers except ‘Chinese Blue Flycatcher’ Cyornis rubeculoides glaucicomans. The closest known breeding population of Brown¬ chested Jungle Flycatcher to Vietnam is in adjacent Guangxi province, China (BirdLife International 2001). However, it is conceivable that the species breeds in the country close to the international border with China, but owing to a paucity of ornithological survey effort, particularly in extreme north-east Vietnam, this cannot be confirmed. Brown-chested Jungle Flycatcher is currently unrecorded in Vietnam outside Hanoi and it has not been found in Laos or Cambodia. The pattern of occurrence of the species in Hanoi and Thailand indicates that it probably occurs as an autumn passage migrant in central Vietnam and perhaps southern Laos and Cambodia. Birdwatchers resident in or visiting those areas should be vigilant to the possibility of encountering the species in September and October. Brown-chested Jungle Flycatcher is recorded annually on spring passage in Thailand, typically during April and early May. The first record for Vietnam remains the only spring passage record for the country. The date of this record is similar to those in Thailand. The lack of subsequent spring records probably represents the limited observer effort at that time of year. With the benefit of hindsight we should have started the study at the beginning of August, because the passage of flycatchers was already underway when the study began. This should probably not detract from conclusions regarding Brown-chested Jungle Flycatcher, because although passage was fairly high during the first 10-day period, the first record made during that period was quite late and the peak passage period was also the last period in which the species was recorded. There was very little observer attention given to the parks prior to the study, and it is consequently possible that some individuals were missed. The peak passage period lor Yellow-rumped and Brown-breasted Flycatchers in 2010 was probably either during the first 10 days of the study period or preceded the study. However, any conclusions regarding the timing of migration of Brown-chested Jungle Flycatcher and other species refer strictly to these sites in 2010 only, and should be tested in subsequent years. After the study was complete, observers continued to visit the parks often until February and recorded only one or two overwintering Taiga Flycatchers and Asian Brown Flycatchers. Although our study focused on a globally threatened species, data on abundance and distribution of most migrant bird species in Indochina remain sparse. This study demonstrates that useful data on bird species can be obtained even in the most unlikely places. It further indicates that in familiar and unexpected locations interesting species can be found. ACKNOWLEDGEMENTS We are grateful to Craig Robson for providing information on the first record of Brown-chested Jungle Flycatcher in Vietnam. We thank Phil Round for information on the status of the species in Thailand, Andy Symes at BirdLife International for use of the library and John Pilgrim for useful discussions. Le Manh Hung and Jonathan C. Eames contributed information on records of the species in 2012. An anonymous reviewer provided invaluable comments on a draft of the manuscript. REFERENCES BirdLife International (2001) Threatened birds of Asia: the BirdLife International Red Data Book. Cambridge UK: BirdLife International. BirdLife International (2012a) Species factsheet: Rhinomyias brunneatus. Downloaded from http://www.birdlife.org on 24/01/201 2. BirdLife International (2012b) The BirdLife checklist of the birds of the world, with conservation status and taxonomic sources. Version 5. http:// www.birdlife.org/datazone/info/taxonomy Brazil, M. (2009) Birds of East Asia. London: Christopher Flelm. Carey, G. J., Chalmers, M. L., Diskin, D. A., Kennerley, P. Fb, Leader, P. J., Leven, M. R„ Lewthwaite, R.W., Melville, D. S„ Turnbull, M. & Young, L. (2001) The avifauna of Hong Kong. Hong Kong: Hong Kong Birdwatching Society. Leader, P. J. & Carey, G. J. (201 2) Zappey's Flycatcher Cyanoptila cumatilis, a forgotten Chinese breeding endemic. Forktail 28: 121-128. Round P. D. (2010) An analysis of records of three passage migrants in Thailand: Tiger Shrike Lanius tigrinus, Yellow-rumped Flycatcher Ficedula zanthopygia and Mugimaki Flycatcher F. mugimaki. Forktail 26: 24-31. Rasmussen, P. C. & Anderton, J. C. (2005) The birds of South Asia: the Ripley guide. Washington DC & Barcelona: Smithsonian Institution & Lynx Edicions. Robson, C. (201 1) A field guide to the birds of South-East Asia. London: New Holland. Stusak, J. M.& Vo Quy (1986) The birds of the Hanoi area. Prague: University of Agriculture. Taylor, B. & Clement, P. (2006) Family Muscicapidae (Old World flycatchers). Pp.422-427 in J. del Hoyo, A. Elliott & D. A. Christie, eds. Handbook of the birds of the world, 1 1 . Barcelona: Lynx Edicions. Wells, D. R. (2007) The birds of the Thai-Malay peninsula, 2. London: Christopher Helm. Simon P. MAHOOD, Wildlife Conservation Society Cambodia Programme, House 21, Street 21, SangkatTonle Bassac, Phnom Penh, Cambodia. Email: smahood@wcs.org Sebastien DELONGLEE, 50 rue de la Pommerais, 35136 Saint- Jacques-de-la-Lande, France. Email: sebastiendlngl40@gmail.com Florian KUNGEL, Hofstettenstrasse 9, 9012 St. Gallen, Switzerland. Email: florianklingel@gmail.com Falk WICKER, 9/3 Wardens Walk, Coburg, 3058 Victoria, Australia. Email: falk.wicker@gmail.com Richard CRAIK, 3rd Floor, 7 1 -75 Hai Ba Trung Street, District 1, Ho Chi Minh City, Vietnam. Email: richard@vietnambirding.com FORKTAIL 29 (2013): 128-137 A survey of the avifauna of Obi island. North Moluccas, Indonesia JOHN C. MITTERMEIER, H. EDEN W. COTTEE-JONES, ENDANG CHRISTINE PURBA, NOVA MAULIDINA ASHURI, EKA HESDIANTI & JATNA SUPRIATNA The avifauna of eastern Wallacea remains little studied despite high diversity and endemism and basic knowledge of the ecology, taxonomy and distribution of species is lacking. Results of a two-month survey on Obi, North Moluccas, Indonesia, in July and August 201 2 are presented here. General observations, point counts, mist-netting and interviews with villagers were carried out in five areas. A total of 109 species including 89 resident landbirds were recorded, of which 14 were new records for the island. Surveys up to 1,550 m extended the known altitudinal range of several species and resulted in the discovery of three montane species not previously recorded on Obi: Red-breasted Pygmy Parrot Micropsitta bruijnii, Mountain White-eye Zosterops montanus and Mountain Tailorbird Orthotomus (Phyllergates) cucullatus. Other notable records were five species of rail, including a surprising range extension of Drummer Rail Habroptila wallacii and observations of the poorly known Moluccan Woodcock Sco/opoxrocht/ssen/7. The biogeographical and conservation implications of findings and elevational turnover in bird communities on Obi are discussed. In line with recent surveys in other parts of Wallacea, this survey highlights the need for continued ornithological fieldwork in eastern Indonesia. INTRODUCTION Located at the boundary of two major biogeographical regions, Wallacea is exceptional for its diversity of species, conservation significance and opportunity to study evolution and speciation. Wallace’s Line, marking the western boundary of Wallacea, is ‘the most prominent and well-studied biogeographic division in the world’ (Schulte et al. 2003) and the region is well known for inspiring Alfred Russel Wallace to develop his ideas on evolution by natural selection. Despite this, Wallacean birds remain poorly known: even basic distributional and life history information is lacking, and continued fieldwork is important for biological research and conservation in the region. The mountainous island of Obi is the seventh largest in the North Moluccas, just over 2,500 km2 in area and with a maximum elevation of 1,61 1 m. Even by Wallacean standards it has received little attention from ornithologists. It was not visited by Wallace although he spent considerable time on adjacent islands (Wallace 1869). H. A. Bernstein made the first ornithological collections in the early 1860s and there were visits by F. H. H. Guillemard in 1883, W. Doherty in 1897, a ‘Mr. Lucas of Brussels’ in 1898, J. Waterstradt in 1902, W. Goodtellow in 1907, A. M. R. Wegner in 1953, and R. Tatu and Y. Momou in 1982-1983 (White & Bruce 1986, P. M. Taylor pers. comm.). Recent records from Obi include observations by M. D. Linsley in 1989 (Linsley 1995), F. R. Lambert in 1992 (Lambert 1994), H. Bashari in 2010 (Bashari 2011) and M. Thibault and others in 2010 (Thibault et al. 2013). Obi was visited by a joint expedition from the Louisiana State University Museum of Natural Science, the University of Oxford, and the University of Indonesia between 5 July and 27 August 2012; here new and interesting observations resulting from this fieldwork are reported and discussed. STUDY AREA AND METHODS Obi is similar to other islands in the Moluccas in being primarily covered by humid evergreen forest with narrow areas of coastal mangrove, scattered swamp-forests and an area of montane forest in the interior. Historically, Obi was inhabited intermittently with human settlement restricted to a few sites along the coast. Guillemard (1885) noted the island was uninhabited in 1883, but stated that ‘it is said that years ago there were many people living on the island, but pirates caused its desertion,’ and Stibbe (1919) commented that ‘permanent settlements [on Obi] are only found at Lawui river mouth (north coast) and at Akeklamo (south-west coast)’. Recently, however, human activity has translormed Obi — coconut plantations cover many lowland areas, with clove and nutmeg groves on the lower hills; logging has been extensive — no primary lowland forest was Found during the visit, and a logging company manager doubted that any such Forest remains on the island. In the highlands, selective logging was evident up to 1,100 m. Nickel ore deposits in the ultrabasic soils of west and south Obi have brought in large-scale mining operations which remove all native vegetation and topsoil, and have already caused serious degradation around Kawasi in the north-west. In the Moluccas, south-east trade winds prevail in July and August and during the visit, the weather in north Obi — Jikotamo, Cabang Kiri River — was clear and sunny with predictable heavy downpours in the early aFternoon. On the west coast — Danau Sagu, Kawasi — it was hot and dry with little to no rainFall, and in the south — Tanjung Rijang, montane areas north oF Fluk — heavy rain was Frequent, oFten beginning before dawn and continuing all day. Data collected in the southern highlands indicated rain or drizzle every day From 1 July to 4 August, with a total July rainfall oF 792 mm. Lambert (1994) had surveyed east Obi in 1992, and the 2012 expedition focused on the west oF the island where five areas were studied (Jikotamo vicinity, Cabang Kiri River, Tanjung Rijang, Danau Sagu plus Kawasi, and the montane area north of Fluk) and 2-4 sites within each area were surveyed, 13 in all; in addition, more generally, we recorded birds in coastal and marine areas. Coordinates, dates, site details and descriptive notes are given in Table 1; see also Figure 1. Common Forest plant species in the Telagabakti Persada logging concession at T anjung Rijang included Cananum balsamiferum (Burseraceae) and dipterocarps Shorea spp. .Anisoptera thurifera , Hopea spp. and Vatica rassak, similar to those in the lowlands of Seram (Marsden 1998). Opportunistic observations were made by JCM and EC-J at all sites (total 630 hours), usually beginning just before dawn and continuing until alter dark, with a break in the middle of the day. Sound recordings were made and are archived at the Macaulay Library, Cornell University. Point counts were carried out near the Cabang-Sumbali river confluence, at Plasma NutFah, in the montane forest north of Fluk, and in the nickel mining area near Danau Sagu; unFortunately, however, although they contributed Forktail 29(2013) A survey of the avifauna of Obi island, North Moluccas, Indonesia 129 Figure 1. Map of Obi island. North Moluccas, Indonesia, showing locations of sites surveyed. Specific sites (see Table 1) are: (1) Jikotamo town, (2) Jikotamo-Sembiki Road, (3) Kampung Buton plantations, (4) Cabang-Sumbali confluence, (5) Cabang Kuning, (6) base camp Rijang, (7) Plasma Nutfah, (8) Danau Sagu lakeshore, (9) Kawasi town, (10) GPS mining camp, (11) ridge camp, (12) old logging road, (13) summit area. Table 1 . Individual sites visited between 5 July and 27 August 2012, showing dates of fieldwork and brief descriptions of habitat at each location (see Figure 1). Specific site Coordinates Survey dates Description; habitat and altitude (m) Jikotamo vicinity 1. Jikotamo town 2. Jikotamo-Sembiki Road 3. Kampung Buton plantations 1.344°S127.655°E 1.357°S127.670°E 1.346°S127.644°E 6-8 July; 13, 20, 26-27 Aug 6—8 July; 26 Aug 6-9 July; 13 Aug Large town and a principal port; coastline and village gardens (sea level). Road from Jikotamo to Sembiki; coastal mangroves, coconut and clove plantations, secondary forest (sea level-35 m). Agricultural land south of one of the largest towns; coconut plantations and open fields (sea level). Cabang Kiri River 4. Cabang-Sumbali confluence 5. Cabang Kuning 1.398°S127.649°E 1.378°S 1 27.659°E 10-13 July; 23-25 Aug 14-19,21-23 Aug Cabang and Sumbali River confluence; lowland forest heavily logged in late 1990s, open gravel river beds, clove and cacao orchards (35-50 m). Rice fields; wet rice paddies and swampy areas surrounded by flooded forest, clove and cacao plantations (30-45 m). Tanjung Rijang 6. Base camp Rijang 7. Plasma Nutfah 1.703°S127.488°E 1.663°S127.536°E 14-16, 18-22 July 17— 18 July Logging camp by the Rijang River mouth; extensive, selectively logged lowland forest, coconut plantations (sea level-100 m). Lowland primary forest fragment (300 ha) along steep-sided gorge; designated as a seedbank. (1 10-300 m). Danau Sagu and Kawasi 8. Lakeshore 9. Kawasi town 1.512°S127.447°E 1.547°S 1 27.41 3°E 8-11 Aug 1 1—12 Aug Largest freshwater lake; reed fringed open water, narrow strip of swamp forest, extensive dry savannah of the Kawasi nickel mine (160-350 m). Coastal town; coconut groves, freshwater swamp, extensive nickel mining area (sea level). Montane area north of Fluk 10. GPS mining camp 11. Ridge camp 12. Old logging road 13. Summit area 1.651°S127.714°E 1 .604°S 127.721°E 1 .585°S 1 27.703°E 1.541°S127.668°E 23-24 July; 5-6 Aug 26-28 July 28 July-4 Aug 2-3 Aug Small exploratory mining camp; logged forest on sandy soil (370-550 m). Steep forest ridge: montane primary forest (850-950 m). Abandoned highland logging road; montane primary forest, secondary growth along overgrown road (1,050-1,150 m). Ridgelines near the highest elevations in Obi's interior: montane primary forest (1 ,200-1 ,550 m). Coastal and marine areas Obi coastline - 5,13,23 July; 7, 13,20, 28 Aug Coastal boat trips between Fluk and Sembiki; beaches, inshore marine areas (sea level). to the overall survey effort, insufficient data were obtained to do statistically significant analyses owing to the difficult terrain and weather conditions. Mist-netting was carried out in forest habitats at Plasma Nutfah and Tanjung Rijang (17 net hours), in the montane forest north ofFluk (416 net hours) and near Danau Sagu (271 net hours). A total of 46 local people from seven villages, particularly parrot-trappers, were interviewed. They were shown colour plates from Coates & Bishop (1997), asked to indicate species they were familiar with and occasionally also asked about species of specific interest. The interviews formed a part of more extensive, structured surveys focused on assessing local knowledge of the Moluccan Woodcock Scolopax rochusseni (Cottee-Jones et al. 2013) and gathering information about parrot-keeping and trapping on Obi (Cottee-Jones et al. in prep.). RESULTS A total of 109 bird species was recorded including 14 new for the island (Appendix), nine of them resident landbirds: Red-breasted Pygmy Parrot Micropsitta bruijnii , Red-necked Crake Rallina tricolor. Bare-eyed Rail Gy?nnocrex plumbeiventris, White-browed Crake Porzana cinerea, Drummer Rail Habroptila ivallacii. Purple Swamphen Porphyrio porphyrio. Little Black Cormorant JOHN C. MITTERMEIER 130 JOHN C. MITTERMEIER etai Forktail 29 (2013) Phalacrocorax sulcirostris , Mountain White-eye Zoster ops montanus and Mountain Tailorbird Orthotomus ( Phyllergates ) cucullatus. The others were a seabird. Great Frigatebird Fregata minor , and four migratory species: Common Greenshank Tringa nebularia , Wood Sandpiper T. glareola , Australian Hobby Falco longipennis and Intermediate Egret Mesophoyx intermedia. The species observed in each area varied from 42 to 78 with the four lowland areas showinghigher richness (mean 65.5 species) than the one highland area (42 species). In the lowlands, diversity was increased by migratory species: 6-8 migrants were recorded in each area but none was found in the highlands. Nonetheless resident landbird diversity was greater in the lowlands (mean 57.75) than the highlands, with the highest overall species diversity in logged forest with subsistence orchards on the Cabang Kiri River (74) and in selectively logged forest at Tanjung Rijang (69). In lowland areas the number of species recorded correlated directly with survey effort, with the greatest number of species being found where most time was spent. Selected species accounts Moluccan Cuckoo Cacomantis heinrichi The taxonomy of Cacomantis cuckoos in the Moluccas is poorly understood. Recently, however, Thibault et al. (2013) concluded that, based on voice and plumage, the ‘Moluccan Cuckoo’ consists of the taxa aeruginosas (Obi, Buru, Ambon, Seram; usually considered a subspecies of Brush Cuckoo C. variolosus) and heinrichi (Halmahera and Bacan). We tentatively accept this judgement here. In 20 1 2 Cacomantis cuckoos were widespread on Obi, and there appeared to be two, possibly three, distinct types (Plate 1) that differed according to habitat preference, vocalisations, extent of rusty underpart colouration and presence of a yellow eye-ring. These observations suggest that at least two Cacomantis taxa occur on Obi — ‘Moluccan Cuckoo’ and one, possibly two (potentially resident and migratory) subspecies of Brush Cuckoo. In contrast to Halmahera, where ‘Moluccan Cuckoo’ has been recorded or suspected only infrequently in montane forest (White & Bruce 1986, Tebb et al 2008), on Obi it was widespread from sea level to above 1,150m, relatively common and tolerant of moderate habitat disturbance. Further investigation of the taxonomy of this group is clearly necessary. Chattering Lory Lorius garrulus Vulnerable. A distinctive yellow-backed subspecies flavopalliatus is found only on Obi and Bacan. On Obi it is a popular village cage-bird and at least 40 individuals captured from the wild were seen in KampungButon andjikotamo. The species is believed to be declining owing to trapping and habitat loss (BirdLife International 201 3c), but was locally common where trapping was limited, specifically in the inaccessible montane forest north of Fluk and at Tanjung Rijang, where the Telagabakti Persada logging company enforces a trapping ban. It was seen from sea level to 1,100 m (and probably occurred higher), and around Tanjung Rijang was frequently found in selectively logged and primary forest fragments (Table 1). In contrast, it was not seen near Jikotamo or on the Cabang Kiri River. In this area, parrot trappers from Kampung Buton travelled inland beyond the study area to catch Chattering Lories in the mountains. On 13 July, for example, a trapper travelled upriver from the Cabang-Sumbali confluence at 06h00 and returned at 12h00 with three lories. Although trappers here claimed that the species had always been restricted to inland areas, it was common in similar habitat near the coast at Tanjung Rijang, suggesting that the species has been extirpated near Jikotamo, Kampung Buton and Laiwui. Red-breasted Pygmy Parrot Micropsitta bruijnii In Wallacea, this inconspicuous montane species was previously known only from Seram (subspecies pileata ) and Buru (subspecies buruensis). On 3 August a group of six Red-breasted Pygmy Parrots was seen in montane primary forest between 1,350 and 1,550 m in central Obi. One was seen very well, perched in the open for 2-3 minutes. It had emerald-green upperparts and flanks, black spots on the wing-coverts, dull red underparts turning to orange on the undertail-coverts, a diffuse blue-green collar, creamy- Plate 1. Two Cacomantis cuckoos found on Obi: (A) Moluccan Cuckoo Cacomantis aeruginosus and (B) Brush Cuckoo C. variolosus. Although similar in appearance these species differed in habitat use (former in lowland swamp-forest, dense secondary forest and montane forest, latter in open agricultural areas and coconut plantations) and vocalisations. Sonagrams (using Raven Lite 1.0) show: (C) a 50-second segment of vocalisations following playback of the Moluccan Cuckoo and (D) a 50-second segment of vocalisations following playback of Brush Cuckoo (in both cases, calls are from the individual bird shown above). JOHN C. MITTERMEIER Forktail 29 (2013) A survey of the avifauna of Obi island, North Moluccas, Indonesia 131 white throat and cheeks and a whitish-brown cap offset by a grey- brown band extending from the bill through the eye and to the side of the neck. The colour of the collar, crown, cheeks and eye¬ line of this bird differed from descriptions and images of both pileata and buruensis (Coates & Bishop 1997, Arndt & Persulessy 2010) and bruijnii from New Guinea (Juniper & Parr 1998) so this population may be an undescribed taxon meriting conservation concern. Swiftlets Collocalia sp. Apart from the omnipresent Glossy Swiftlet Collocalia esculenta , swiftlets in the north Moluccas are a field identification challenge because of the difficulty in distinguishing between Uniform Swiftlet C. vanikorensis and the dark-rumped infuscata subspecies of Moluccan Swiftlet C. infuscata. Lambert (1994) tentatively identified both Uniform Swiftlet and the white-rumped ceramensis subspecies of Moluccan Swiftlet, sometimes considered a distinct species, on Obi. Dark-plumaged swiftlets were common and frequently seen flying high above the canopy. There appeared to be three morphotypes: (1) blackish upperparts and grey underparts with a distinct white rump-band; (2) brown upperparts and underparts with no obvious rump-band; (3) blackish upperparts and blackish- brown underparts with a hint of a brown rump-band. The white- rumped birds were the least common of the dark swiftlets and usually seen in flocks of 10-30 either alone or in multi-species flocks. Dark- rumped swiftlets occurred both in small flocks and, more frequently, in flocks of several hundred birds. The first two plumage types accord with Lambert’s identification of Moluccan Swiftlet ceramensis and Uniform Swiftlet on Obi. The third type may simply be variation within Uniform Swiftlet or potentially the dark-rumped infuscata subspecies of Moluccan Swiftlet. White-throated Pigeon Columba vitiensis This species is known from all the nearby island groups (White & Bruce 1986), but was not recorded on Obi until Thibault et al. (2013) found it at 1,000 m in 2010. A single individual was seen flying along the forest edge at CabangKuning (20 m). This suggests that it is rare on Obi across a wide altitudinal range. Local parrot- trappers near Jikotamo were familiar with the species and reported that it feeds on the ground in the forest. Scarlet-breasted Fruit Dove Ptilinopus bernsteini Subspecies micrus (Jany 1935) diagnosed principally by its smaller size (White & Bruce 1986), is endemic to Obi. Lambert (1994) recorded it between 1 80 and 600 m. It was uncommon in lowland plantations in the CabangKiri River area (35-50 m) and relatively common in montane forest north of Fluk (800-1,550 m) where two males were mist-netted and measured on 27 July and 2 August. Wing measurements (140 mm) were slightly above the known size range for micrus wing (128-139 mm) (White & Bruce 1986). Caruncuiated Fruit Dove Ptilinopus granulifrons Endemic. Classified as Vulnerable (BirdLife International 2013d), this species was not seen from 1992 until 201 1 (Bashari 2011). It was found to be widespread but inconspicuous in secondary forest near Cabang-Sumbali and in selectively logged forest near Tanjung Rijang. One was also seen in montane forest at 1,100 m. It was usually found feeding in fruiting trees in groups of 2-10 individuals; often the only indication of its presence was a distinctive wing¬ flapping as birds moved between branches. Although only one bird was found in montane forest, given its inconspicuous behaviour it is possibly found at higher altitudes than previously thought. These findings concur with the known distribution and habitat of the closely related Grey-headed Fruit Dove P. hyogaster on Halmahera (Gibbs etal. 2001). Although Caruncuiated Fruit Dove is not listed for the island of Bisa (Gibbs et al. 2001), local people reported its presence there, and this and other outlying islands are worth further investigation. Cinnamon-bellied Imperial Pigeon Ducula basilica The distinctive subspecies obiensis (Ffartert 1898) is endemic to Obi and may warrant recognition as a full species. It was fairly common in disturbed and selectively logged habitats throughout the lowlands, where it appears to tolerate moderate habitat disturbance, and was common in montane forest above 800 m. Pied Imperial Pigeon Ducula bicolor The taxonomy of this species in the Moluccas remains poorly understood owing to confusion between the status of Ducula bicolor and ‘D. melanura (White & Bruce 1986). Coates & Bishop (1997) listed melanura for Obi but Gibbs et al. (2001) concluded that melanura is either a morph of bicolor or the result of genetic introgression between bicolor zn&spilorrhoa of New Guinea. Birds observed on Obi showed some traits of melanura, including a greenish-horn bill and extensive black to the outer rectrices, but lacked black markings on the undertail-coverts. This mix of traits appears to support the conclusion of Gibbs et al. (2001). Red-necked Crake Rallina tricolor In the Moluccas this species is previously known from Ambon and Tayandu (White & Bruce 1986, Taylor 1998). It was common in wet, closed-canopy forest around Cabang Kuning (14-22 August), where substantial rainfall had left pools of standing water in many areas of the forest and uncommon along the Jikotamo to Sambiki road (26 August). Birds vocalised frequently, particularly at dusk, and responded strongly to playback. Elsewhere, Red-necked Crakes apparently migrate from New Guinea to the Cape York Peninsula in the wet season (Taylor 1998), and records from Ambon in june- July were also considered to be migrants from New Guinea (White & Bruce 1986). Given the lack of previous records, it may also be a seasonal visitor to Obi. Bare-eyed Rail Gymnocrex plumbeiventris In the Moluccas, this species occurs on FFalmahera, Bacan and Morotai but has not previously been recorded on Obi. It is also found on Misool and New Guinea (White & Bruce 1986, Taylor 1998). It was uncommon in swamp-forest surrounding the Cabang Kuning rice-fields, where two were recorded on 18-19 and 22 August. It occurred in the same habitat as Red-necked Crake. Vocalisations included steady gulping noises while foraging, and a loud barking call followed by a bizarre, trumpeting woooo-ivooot in response to playback (recordings at: macaulaylibrary.org). White-browed Crake Porzana cinerea This widespread species has been recorded in the Moluccas from Kai, Ambon, Seram, Bacan and Halmahera (White & Bruce 1986, Coates & Bishop 1997). It was heard and seen on 9 and 1 1 August in a dense reedbed on the south-west edge of Danau Sagu, where Purple Swamphen (see below) was also observed. One was flushed from the wet rice-fields at Cabang Kuning on 1 6 August. Although its presence on Obi is not surprising given its distribution, these are apparently the first records for the island. Moluccan Bush-hen Amaurornis moluccana First recorded on Obi by Lambert (1994), one was also heard in 2010 by Thibault etal. (2013). At lowland sites it was relatively common, although inconspicuous — at least four pairs were found along a 2-km stretch of river near Tanjung Rijang, and a minimum of three pairs was present in swamp-forest bordering the Cabang Kuning rice-fields. It was also found in secondary growth along an old logging road in montane forest at 1,150 m. 132 JOHN C. MITTERMEIER etal. Forktail 29 (2013) Drummer Rail Habroptila wallacii This species was previously believed to be endemic to Halmahera (White & Bruce 1986, Taylor 1998). However two birds were observed and recorded in dense swamp-forest near CabangKuning on 17-18 August. Local people were familiar with the species and reported frequently catching it in snares set for scrubfowl. Six interviewees reported it in the Jikotamo-Kampung Buton area, one in Tanjung Rijang and one in Wayloar. This suggests it is relatively widespread in the lowlands. It is considered a delicacy on Halmahera (Taylor 1998, P. M. Taylor pers. comm.) and three local people confirmed eating them although others stated that they released rails from snares because of their strange appearance. Two hunters described collecting eggs from a nest on a palm stump about 0.5 m tall, a description that agrees with the record of a nest from Halmahera (Bashari & van Balen 2011). Given that the Drummer Rail is believed to be flightless (de Haan 1950, Taylor 1998), its presence on Obi is intriguing and warrants further study. Purple Swamphen Porphyrio porphyria White & Bruce (1986) described this species as ‘very local’ in Wallacea — it occurs on most of the large islands (e.g. Sulawesi, Buru, Seram, Halmahera), but is absent from smaller islands such as Bacan, Morotai, Misool and the Sulas (White & Bruce 1986, Taylor 1998). Four Purple Swamphens were seen in dense reedbeds on the south-west edge of Danau Sagu, the only large freshwater lake, on 9 and 1 1 August. Parrot-trappers and local people near Jikotamo and Kampung Buton were not familiar with the species despite its large size and distinctive appearance and it may be very localised or restricted only to the marshes around Danau Sagu. Moluccan Woodcock Scolopax rochussenii Endangered (BirdLife International 2013e). An enigmatic species, known only from Obi (fewer than 10 records) and one specimen collected on Bacan in 1902. It went unrecorded from 1982 until 2010 when Thibault etal. (2013) observed it near Soligi and east of Jikotamo. The species was found to be uncommon but widespread and conspicuous when displaying at dawn and dusk. Birds were observed displaying over swamp-forest and along rivers and stream valleys from 15-1,150 m. The distribution and conservation status of this species are described by Cottee-Jones et al. (2013). Migratory waders A variety of shorebirds migrate through east W allacea but in general the region supports relatively low numbers of migrants and does not appear to be a major wintering area (White 1975, Coates & Bishop 1997). Information on timing and distribution of migrants is sparse. A flock of three Wood Sandpipers Tringa glareola was near Kawasi on 1 1 August and a single on the rice-fields at Cabang Kuning from 18-22 August, with a single Common Greenshank Tringa nebularia. Both species are common passage and wintering migrants in the Moluccas but not previously recorded on Obi (White & Bruce 1986, Coates & Bishop 1997). Common Sandpiper Actitis hypoleucos was seen on 1 1 August at Danau Sagu and again on 13 August at Kampung Buton. Red-necked Phalaropes Pbalaropus lobatus were seen on 27 August, when over 40 were just off the coast near Jikotamo and in the strait between Obi and Bacan, but were not observed on six boat trips through the same area between 6 July and 20 August. Both are common wintering species in the Moluccas, with several previous records from Obi. Australian Hobby Falco longipennis N on-breeding individuals of the nominate subspecies occur in small numbers in the Moluccas and have been recorded on Ternate, Ambon and Seram (White & Bruce 1986). A single bird was seen and photographed at dusk near Kawasi on 1 1 August — the first record for Obi. Little Black Cormorant Phalacrocorax sulcirostris Widespread in the Moluccas with reports from Bacan, Halmahera, Buru and Seram amongst others (White & Bruce 1986). A single flying over a river near the Cabang-Sumbali confluence on 1 1 July is the first record for Obi. In contrast, Little Pied Cormorant P. melanoleucos was relatively common in freshwater habitats around Kawasi and Danau Sagu. Intermediate Egret Mesophoyx intermedia In Wallacea, a widespread but generally uncommon non-breeding visitor from Australia, subspecies plumifera, and the Palearctic subspecies intermedia (White & Bruce 1986). In the Moluccas it has been recorded widely including Bacan, Buru and Seram but until now not Obi. One was foraging in a wet grassy field near the Cabang Kuning rice-fields on 16 August. The black tip to the bill and the entirely black legs suggested the subspecies intermedia. Great and Lesser Frigatebird Fregata minor and F. ariel Both Great and Lesser Frigatebird occur throughout Wallacea (White & Bruce 1986, Coates & Bishop 1997). Lesser Frigatebird appears to be more frequent around Obi and was reported by Linsley (1995). It was seen three times during coastal boat journeys and flying over the shore at both Tanjung Rijang and Kawasi. On 21 July, JCM observed two Great Frigatebirds soaring with six Lesser Frigatebirds over the coast at Tanjung Rijang. Great Frigatebird has been reported from Halmahera, Buru, Ambon and Seram in the North Moluccas but not previously from Obi. Slaty Monarch Myiagra galeata This Moluccan endemic is common on Obi (Coates & Bishop 1997) although information on its nesting behaviour is limited. A pair was nesting in an isolated tree near Jikotamo on 7 July. The tree, roughly 15 m high, was in a cattle pasture about 10 m from the forest edge. The nest was located on a fork in a branch about 12 m above the ground in the subcanopy. It consisted of a small woven cup approximately 5 cm in diameter with sides built up about 5 cm high and was constructed of neatly woven plant fibres mixed with bark and lichen. The male and female took turns in the nest and appeared to be incubating. Mountain White-eye Zosterops montanus This species is found in montane habitats above 1,000 m on Seram, Bacan, Ternate (subspecies obstinatus ) and Buru (subspecies montanus ), but has not previously been recorded on Obi. On 29 July, a dense flock of about 25 individuals was feeding with about 10 Cream-throated White-eyes Z. atriceps. along an old road-cut in montane forest at 1,120 m. The Mountain White-eyes were obviously smaller and more compact and had olive upperparts, head, flanks and undertail-coverts, with a bright yellow throat and underparts, conspicuous broken white eye-rings, dark irises, black legs and bill. On 1 August a flock of about 50 Mountain White-eyes was feeding in the same trees and on 2 August two flocks were seen at 1,100 m and a third flock at 930 m. Cream-throated White-eye Zosterops atriceps First discovered on Obi in 1992 by Lambert (1994) who reported it uncommon between 220 and 700 nr and described birds as closely resembling nominate atriceps from Bacan, it was subsequently seen in 2010 by Thibault etal. (2013). It was common and conspicuous from 500-1,100 m in montane forest north of Fluk, but was not recorded below 500 m. Elsewhere it mainly occurs in lowlands up to 700 m (White & Bruce 1986, Coates & Bishop 1997). Mountain Tailorbird Orthotomus (Phyllergates) cucullatus In Wallacea previously recorded in montane forest on Bacan, (subspecies batjanensis ), Buru and Seram (subspecies dumasi), and A survey of the avifauna of Obi island, North Moluccas, Indonesia 133 Forktail 29 (2013) Sulawesi, four subspecies. White & Bruce (1986) found little difference between these taxa and questioned their validity. Mountain Tailorbird was found in montane forest north of Fluk between 900 and 1,200 m on 27 July-4 August. It was common in patches of dense vegetation, often near old landslides or treefalls. Island Leaf Warbler Phylloscopus poliocephalus The subspecies ivaterstradti found on Obi and Bacan was originally described as a distinct species by Hartert ( 1 903). It was one of the commonest species in forest from 500-1,550 m, where it was a frequent and vocal member of mixed-species flocks. Species not recorded Several species which were not observed deserve mention because they were either (a) familiar to local people but not seen during the fieldwork, or (b) reported or predicted in earlier accounts but were neither seen during fieldwork nor familiar to interviewees. The technique of showing interviewees plates in a field guide is known to have drawbacks (Diamond & Bishop 1999), but in some circumstances the results merit reporting. The people on Obi were particularly knowledgeable about large and conspicuous birds, parrots, terrestrial birds caught in snares, and nectarivores that visited the flowers of clove trees. Species for which there were at least two independent local reports were: a large black eagle, presumably either Gurney’s Eagle Aquila gurneyi or Black Eagle Ictinaetus malayensis , Buff-banded Rail Gallirallus philippensis , Barred Rail G. torquatus , Common Koel Eudynamys scolopacea and Sulawesi Myzomela Myzomela chloroptera. The myzomela was first collected on Obi in 1982 by R. Tatu and Y. Momou (White & Bruce 1986) and was observed in the highlands by Thibault et al. (2013). Clove harvesters near Kampung Buton reported that cui merah (a red sunbird) occasionally visited their trees. Notable species not seen and unfamiliar to local people even after specific questioning, included: White Cockatoo Cacatna alba , Moluccan Scrubfowl Eulipoa wallacii , Goliath Coucal Centropus goliath. Ivory-breasted Pitta Pitta maxima , Red-backed Buttonquail Turnix maculosus , White-breasted Woodswallow Artamus leucorynchus and Long-billed Crow Corvus validus. All these have been listed for Obi in earlier accounts (White & Bruce 1986, Coates & Bishop 1997) and are distinctive in appearance. Failure to record them could be due to several factors, including inaccurate historical records, vagrancy or local extinction and their status warrants further investigation. In the case of White Cockatoo, there has been confusion as to whether the species was once native (White & Bruce 1986) and has been extirpated or only ever occurred as an escaped population (Lambert 1994). Most parrot-trappers were familiar with it, but identified it as being from Bacan. Villagers in Air Mangga Indah in north Obi, however, reported several white cockatoos living in the nearby hills. These observations support Lambert’s suggestion that the species is not native but may occur as small populations of escaped birds near human habitations. DISCUSSION The discovery of 14 new species for Obi is comparable with recent findings on other Wallacean islands (Trainor 2002, Rheindt et al. 2010, Trainor et al. 2012) and re-emphasises the need for on-going fieldwork in the region. On Obi a good example of data deficiency is provided by the Rallidae. Prior to the 2012 fieldwork, the family was known on the island only from records of Moluccan Bush-hen (Lambert 1994, Thibault et al. 2013). In 2012 the bush-hen and an additional five species were recorded, whilst local hunters reported up to four further species. Clearly Obi is not depauperate in rallids and, despite its relatively small size, hosts rail diversity comparable to larger Moluccan islands such as Buru (eight species), Seram (seven), and Halmahera (seven) (Coates & Bishop 1997). From a distributional standpoint, 1 0 of these records are of species that occur on islands both north (e.g. Elalmahera, Bacan) and south (e.g. Seram, Buru) of Obi, so their presence is not surprising. Nine of the new records are resident breeding species and are of particular significance in understanding the island’s ecology and biogeographical relationships. From a distributional standpoint, five of them occur on islands both north (Halmahera, Bacan) and south (Seram, Buru) of Obi, and consequently their presence is unsurprising. As Coates & Bishop (1997) pointed out, however, ‘one of the more striking features of Moluccan birds is the seemingly haphazard occurrence of certain families and species’; therefore, confirming whether apparent range disjunctions are real or a sampling artefact is essential. Obi has been grouped with Halmahera, Bacan and Morotai in a North Moluccan biogeographical unit (White & Bruce 1986, Carstensen & Olesen 2009) and it is therefore tempting to conclude that Obi’s avifauna is a subset of that of Halmahera. The remaining four additions to the Obi list are two species, Drummer Rail and Bare-eyed Rail, which appear to support the connection with Halmahera, but two that are apparently absent from Halmahera with their closest known populations to the south (Red-breasted Pygmy Parrot) and to the east (Red-necked Crake). Elevational turnover The difference between montane and lowland bird communities in Melanesia (Mayr & Diamond 1976) and Wallacea (Poulsen & Lambert 2000) has been of long-standing scientific interest, and in both regions the altitudinal range of a species may vary from island to island (Mayr & Diamond 1976, Arndt & Persulessy 2010). The altitudinal range of each species observed on Obi is reported in the Appendix. Many species recorded up to 1 , 1 50 m are probably found higher, as survey time above this elevation was limited. Seven species were found to be common in the highlands but were not seen in the lowlands, and 25 species seen frequently in lowland forests were not seen in the highlands. The species restricted to the upland forests include three seen down to 300-500 m and four only found in the higher forests above 850 m (see Appendix). This may partly be due to sampling deficiencies, but it is clear that avian communities on Obi change substantially with increasing elevation — in contrast to Halmahera where the lack of a distinct montane bird community is attributed to the limited and fragmented nature of forest cover at higher altitudes (Poulsen & Lambert 2000). Conservation implications Obi is treated as Important Bird Area ID 202 (BirdLife International 2013a) within the Northern Maluku Endemic Bird Area (BirdLife International 2013b). A 45,000 ha nature reserve lying between 500 and 1,61 1 m in the central highlands has been proposed (MacKinnon & Artha 1981). It is unclear whether this reserve has been officially accepted and where exactly it is located. Most people questioned on Obi seemed unaware of it and logging appears to be on-going or to have taken place in most parts of Obi. Four aspects of our results have particular conservation significance. ( 1 ) As elsewhere in the Moluccas, highland and lowland bird communities differ substantially and conservation programmes must take account of habitats at all elevations across the island. Some endemic taxa are rare in or absent from the lowlands e.g. Scarlet¬ breasted Fruit Dove, Pale Cicadabird Coracina ceramensis hoogerwerfi and Island Flycatcher Eumyias panayensis obiensts, whereas others are rare or absent from the highlands e.g. Common Paradise-kingfisher Tanysiptera galatea obiensis and Slender-billed Cicadabird C. tenuirostris obiense. The highlands hold the last 134 JOHN C. MITTERMEIER etoi Forktail 29 (2013) significant area of intact primary forest on the island, but the small size of this forest and logging operations so far as high as 1,100 m place it under threat. (2) Many forest bird species on Obi seem to be resilient to moderate logging and habitat change. In particular, endemic taxa and species of conservation concern including Chattering Lory, Carunculated Fruit Dove and Moluccan Woodcock were observed on many occasions in selectively logged forest. This suggests that these degraded forests should be incorporated into conservation strategies on the island. Conservation measures may benefit from considering existing local cultural values. Survival of forest and reedbeds around Danau Sagu, for example, is primarily due to superstitions surrounding the lake. (3) The environmental impacts of nickel mining on Obi should be carefully evaluated before new areas, particularly in the southern highlands, are exploited. Unregulated nickel mining practices lead to a near-total transformation of the landscape and dramatic changes in the bird community. None of the three species of conservation concern or the endemic taxa with the exception of Drab Whistler Pachycephala griseonota johni and Northern Golden Bulbul Alopboixus afflnis lucasi , was observed in the nickel mining area near Kawasi. (4) Parrot trapping on Obi continues on a significant scale for local and off-island sale and may threaten the survival of some species, in particular Chattering Lory. In 2012, no evidence of monitoring or regulation of the trade was seen. Lambert (1993) conducted an extensive study of the parrot trade in the North Moluccas and outlined clear catch and export quotas for both Chattering Lory and Violet-necked Lory Eos squamata. More recently, Crosby (2003) recommended that ‘zero quotas’ should remain in place for Chattering Lory until a reliable system of management is developed. It seems possible that habitat destruction and trapping have extirpated this species from parts of the island. Future directions It seems likely that more species will be added to Obi’s bird list. Surveys in different seasons and with more focus on the eastern part of the island are recommended, and these should extend to the little- known satellite islands of Bisa, Obilatu, Tapat and Gomumu. From a taxonomic and biogeographic perspective, a top priority should be the targeted collecting of scientific specimens for morphological and molecular analyses — several species were encountered that may prove to be undescribed taxa or require clarification of their taxonomic status. A solid understanding of the taxonomy and biogeographical history of species on Obi will help set conservation priorities. From a conservation perspective, priorities include clarifying the status of the island’s protected area, investigating the value of different types of human-modified habitats for birds, addressing the restoration of habitats following nickel mining, and monitoring and controlling the island’s parrot trade. ACKNOWLEDGEMENTS Our research on Obi was generously supported by a National Geographic Society /Waitt Grant, a Ron & Mary Neal LSU Graduate Fellowship, a Thesiger Award from The Old Etonian Association, a Small Fieldwork Grant from the Royal Geographical Society (with IBG), a Graham Hamilton travel grant from St Edmund Hall, the Oxford University Expeditions Council and A. J. Tours & Travel. We thank our supervisors and referees for their generous help, in particular Frederick Sheldon, Robb Brumfield, Robert J. Whittaker, Paul Jepson, Rich Grenyer, Shonil Bhagwat and Stuart Butchart. For advice on ornithology and logistics in Indonesia we thank Frank Lambert, Hanom Bashari, Marc Thibault, Bruce Beehler, Allison Styring, Diah Asri, David Bishop, Nick Brickie, Dewi Prawiradilaga, Mohammad Irham and Richard Noske. Matthew Medler at the Macaulay Library of Natural Sounds, Cornell University, generously provided sound-recording equipment for JCM. Our work on Obi was only possible with the hospitality of local people, specifically BambangSetiawan, Pak Uspa, the heads of Kawasi, Gambaru, and Kampung Buton villages, Pak La Gode, La Ham and La Raidi of Kampung Buton, Adam and Ikhsan of Wayloar, and Nisha and her family in Jikotamo. John C. Mittermeier and H. Eden W. Cottee-Jones contributed equally to this work. REFERENCES Arndt, T. & Persulessy, Y. E. (2010) Red-breasted Pygmy Parrot Micropsitta bruijnii on Seram and Buru. BirdingASIA 1 3: 49-54. Bashari, H. (2011) Rediscovery of Carunculated Fruit Dove Ptilinopus granulifrons on Obi, North Moluccas. BirdingASIA 1 6: 48-50. Bashari, FI. & van Balen, S. (2011) First breeding record of the Drummer Rail Habroptila wallacii. BirdingASIA 1 5: 20-22. BirdLife International (2013a) Important Bird Areas factsheet: Pulai Obi. Downloaded from http://www.birdlife.org on 24/07/2013. BirdLife International (201 3b) Endemic Bird Area factsheet: Northern Maluku. Downloaded from http://www.birdlife.org on 24/07/2013. BirdLife International (2013c) Species factsheet: Lorius garrulus. Downloaded from http://www.birdlife.org on 03/09/2013. BirdLife International (2013d) Species factsheet: Ptilinopus granulifrons. Downloaded from http://www.birdlife.org on 03/09/2013. BirdLife International (201 3e) Species factsheet: Scolopax rochussenii. Downloaded from http://www.birdlife.org on 03/09/2013. Carstensen, D. W. & Olesen, J. M. (2009) Wallacea and its nectarivorous birds: nestedness and modules. J. Biogeogr. 36: 1 540-1550. Coates, B. J. & Bishop, K. D. (1997) A guide to the birds of Wallacea. Alderley: Dove Publications. Cottee-Jones, H. E. W„ Mittermeier, J. C. & Redding, D. W. (201 3) The Moluccan Woodcock Scolopax rochussenii on Obi Island, North Moluccas, Indonesia: a 'lost' species is less endangered than expected. Forktail 29: 88-93. Crosby, M. J. (2003) Saving Asia's threatened birds: a guide for government and civil society. Cambridge UK: BirdLife International. Diamond, J. M. & Bishop, K. D. (1999) Ethno-ornithology of the Ketengban people, Indonesian New Guinea. Pp.17-46 in D. Medin & S. Atran, eds. Folkbiology. Cambridge Mass: MIT Press. Gibbs, D., Barnes, E. & Cox, J. (2001 ) Pigeons and doves. London: Christopher Flelm. Guillemard, F. H. FI. (1885) Report on a collection of birds obtained during the voyage of the yacht 'Marchesa' Part V. The Molucca Islands .Proc.Zool. Soc.Lond. 1885:561-576. de Plaan, G. A. L. (1950) Notes on the Invisible Flightless Rail of Halmahera (Habroptila wallacii Gray). Amsterdam Naturalist 1 : 57-60. Hartert, E. (1898) On three new pigeons from the island of Obi Major, Moluccas. Bull. Brit. Orn. Club 7: 35. Hartert, E. (1903) The birds of the Obi Group, Central Moluccas. Novit. Zool. 10: 1-17. Jany, E. (1955) Neue Vogel-Formen von den Nord-Molukken. J. Orn. 96: 1 02- 106. Juniper, T. & Parr, M. (1998) Parrots: a guide to the parrots of the world. Robertsbridge: Pica Press. Lambert, F. R. (1993) Trade, status and management of three parrots in the North Moluccas, Indonesia: White Cockatoo Cacatua alba, Chattering Lory Lorius garrulus and Violet-eared Lory Eos squamata. Bird Conserv. Internatn. 3: 145-168. Lambert, F. R. (1994) Notes on the avifauna of Bacan, Kasiruta, and Obi, Northern Moluccas. Kukila 7: 1-9. Linsley, M. D. (1995) Some bird records from Obi, Maluku. Kukila 7: 142-151. MacKinnon, J. & Artha, M. B. (1 981 ) A national conservation plan for Indonesia: Maluku & Irian Jaya. 7. Bogor: UNDP/FAO National Parks Development Project. Marsden, S. J. (1998) Changes in bird abundance following selective logging on Seram, Indonesia. Conserv. Biol. 12: 605-61 1. Mayr, E. & Diamond, J. M. (1976) Birds on islands in the sky: origin of the montane avifauna of northern Melanesia. Proc. Natn. Acad. Sci. USA 73: 1765-1769. Forktail 29(2013) 135 A survey of the avifauna of Obi island, North Moluccas, Indonesia Payne, R. B. (2005) The cuckoos. Oxford: Oxford University Press. Poulsen, M. K. & Lambert, F. R. (2000) Altitudinal distribution and habitat preferences of forest birds on Plalmahera and Buru, Indonesia: Implications for conservation of Moluccan avifaunas. Ibis 142: 566-586. Rheindt, F. E., Verbelen, P, Putra, D. D„ Rahman, A. & Indrawan, M. (2010) New biogeographic records in the avifauna of Peleng Island (Sulawesi, Indonesia), with taxonomic notes on some endemic taxa. Bull. Brit. Orn. Club 130:181-207. Schulte, J. A., Melville, J.& Larson, A. (2003) Molecular phylogenetic evidence for ancient divergence of lizard taxa on either side of Wallace's Line. Proc. Roy. Soc. B—Biol. Sci. 270: 597-603. Stibbe, D. G. (ed.) (1919) Encyclopaedic van Nederlandsch-lndie [Encyclopedia of the Dutch East Indies], Leiden: Martinus Nijhoff. Stresemann, E. (1931) Zur Ornithologie von Flalmahera und Batjan. Ornithologische Monatsberichte 39:1 67-171. Taylor, B. (1 998) Rails: a guide to the rails, crakes, gallinules and coots of the world. New Flaven: Yale University Press. Tebb, G., Morris, P. & Los, P (2008) New and interesting bird records from Sulawesi and Flalmahera, Indonesia. BirdingASIA 10: 67-76. Trainor, C. R. (2002)The birds of Adonara, Lesser Sundas, Indonesia. Forktail 18:93-100. Trainor, C. R., Verbelen, P. & Johnston, R. E. (201 2) The avifauna of Alor and Pantar, Lesser Sundas, Indonesia. Forktail 28: 77-92. Thibault, M., Defos du Rau, P., Pineau, O. & Pangimangen, W. (2013) New and interesting records for the Obi archipelago (north Maluku, Indonesia), including field observations and first description of the vocalisation of the Moluccan Woodcock Scolopax rochussenii. Bull. Brit. Orn. Club 133:83-115 Wallace, A. R. (1869) The Malay Archipelago. Singapore: Graham Brash. White, C. M. N. (1975) Migration of Palearctic waders in Wallacea. Emu 75: 37-39. White, C. M. N. & Bruce, M. D. (1986) The birds of Wallacea (Sulawesi, the Moluccas and Lesser Sunda Islands, Indonesia). London: British Ornithologists' Union. John C. MITTERMEIER, Museum of Natural Science, 1 1 9 Foster Hall, Louisiana State University, Baton Rouge, LA, 70803, USA. Email: john.mittermeier@gmail.com (correspondence) H. Eden I/I/. COTTEE-JONES, Biodiversity Research Group, School of Geography and the Environment, Oxford University Centre for the Environment, South Parks Road, Oxford OX I 3QY, UK. Email: henry.cottee-jones@seh.ox.ac.uk Endang Christine PUR BA, Biology Department, Pascasarjana Program, Faculty of Mathematics and Natural Science, Universitas Indonesia, Gedung E lantai 2 Kampus Ul Depok, West Java, Indonesia. Email: endang.christine@yahoo.com Nova Maulidina ASHURI, Biology Department, Pascasarjana Program, Faculty of Mathematics and Natural Science, Universitas Indonesia, Gedung E lantai 2 Kampus Ul Depok, West Java, Indonesia. Email: dina_ta_its@yahoo.com Eka HESDIANTI, Biology Department, Pascasarjana Program, Faculty of Mathematics and Natural Science, Universitas Indonesia, Gedung E lantai 2 Kampus Ul Depok, West Java, Indonesia. Email: eka.hesdianti@ymail.com Jatna SUPRIATNA, Research Center for Climate Change, Universitas Indonesia, Gedung PAU Ul lantai 8,5 Kampus Ul Depok 16424, West Java, Indonesia. Email: jsupriatna@conservation.org Appendix Birds recorded on Obi, North Moluccas, Indonesia, from 5 July-27 August, 201 2. Locations: (1) Jikotamo town, (2) Cabang Kiri River area including Cabang Kuning and Cabang-Sumbali river confluence, (3) Tanjung Rijang logging camp, (4) Danau Sagu and Kawasi Town, (5) montane forest north of the Fluk and (6) coastal and marine areas around the island. Relative abundance letters are: (C) greater than or equal to 10 individuals per day, (F) 4-10 individuals per day, (U) 1-3 individuals per day, and (R) less than 1 per day. Legend: (p) species photographed, (s) sound-recorded, (m) mist-netted; others were seen or heard only. Species 1 2 3 Locations 4 5 6 Legend Altitudinal range (m) Dusky Scrubfowl Megapodius freycinet R U R s 0-1,100 Spotted Whistling Duck Dendrocygna guttata F R p,s,m 0-35 Papuan Hornbill Aceros plicatus F F F R R P.5 0-1,100 Dollarbird Eurystomus orientalis R R P 0-30 Common Kingfisher Alcedoatthis U U U U P 0-160 Little Kingfisher Alcedo pusilla R R 0-160 Variable Dwarf Kingfisher teyx lepidus U U U P.5 30-200 Blue-and-white Kingfisher Todiramphus diops U F F U p,s,m 0-200 Beach Kingfisher Todiramphus saurophaga R U p 0 Sacred Kingfisher Todiramphus sanctus F U U u p 0-200 Common Paradise-kingfisher Tanysiptera galatea U F F P.5 30-300 Rainbow Bee-eater Merops ornatus C F F c P.5 0-350 Brush Cuckoo Cacomantis variolosus F U U P.s 0-40 Moluccan Cuckoo Cacomantis heinrichi U U U u P.s 0-1,150 Violet-necked Lory Eos squamata C F C u F P.5 0-1,550 Chattering Lory Lorius garrulus F C P.5 0-1,150 Red-flanked Lorikeet Charmosyna placentis C C C c C p,s,m 0-1,100 Red-breasted Pygmy Parrot Micropsitta bruijnii R 1,350-1,550 136 JOHN C. MITTERMEIER et al. Forktail 29 (201 3) Species 1 2 3 Locations 4 5 6 Legend Altitudinal range (m) Red-cheeked Parrot Geoffroyus geoffroyi F F U U P.s 30-1,550 Great-billed Parrot Tanygnathus megalorynchos U R 0-300 Eclectus Parrot Eclectus roratus U R R 0-35 Glossy Swiftlet Collocalia esculenta C C C C C p,m 0-1,150 Moluccan Swiftlet Collocalia infuscata F F p.s 0-1,000 Uniform Swiftlet Collocalia vanikorensis C C C 0-160 Dark swiftlet sp. Collocalia infuscata/vanikorensis F F C C u 0-1,100 Moustached Treeswift Hemiprocne mystacea C F R p.s 0-300 Moluccan Scops Owl Otusmagicus U U U u p.s 30-1,100 Barking Owl Ninoxconnivens R U F s 0-300 Large-tailed Nightjar Caprimulgus macrurus U U U U s 0-300 Metallic Pigeon Columba vitiensis R 30 Spotted Dove Stceptopelia chinensis R p 20 Brown Cuckoo Dove Macropygia amboinensis U U C F C p,s,m 0-1,150 Great Cuckoo Dove Reinwardtoena reinwardtii R R R s 30-1,100 Emerald Do\eChalcophapsindica C F F R p,s,m 0-1,100 Nicobar Pigeon Caloenasnicobarica R p 150 Scarlet-breasted Fruit Dove Ptilinopus bernsteinii R U p,m 30-1,550 Superb Fruit Dove Ptilinopus superbus U p,m 30-200 Carunculated Fruit Dove Ptilinopus granulifrons U U R p 30-1,150 Black-naped Fruit Dove Ptilinopus melanospila R 300 White-spectacled Imperial Pigeon Ducula perspicillata F C C R p.s 0-300 Cinnamon-bellied Imperial Pigeon Ducula basilica U F U R C p.s 30-1,550 Pied Imperial Pigeon Ducula bicolor U U U p.s 0-150 Red-necked Crake Rallina tricolor R U s 30-35 Bare-eyed Rail Gymnocrexplumbeiventris U s 30 Rufous-tailed Bush-hen Amaurornismoluccanus F F R R s 0-1,150 White-browed Crake Porzana cinerea R U s 30-160 Drummer Rail Habroptila wallacii R s 30 Purple Swamphen Porphyrio porphyrio U p.s 160 Moluccan Woodcock Scolopax rochussenii U U U U p.s 15—1,150 Common Greenshank Tringanebularia R s 30 Wood Sandpiper Tringa glareola R R s 2-30 Common Sandpiper Actitis hypoleucos R R 0-160 Red-necked Phalarope Phalaropuslobatus U 0 Great Crested Tern Sterna bergii U C 0 Black-naped Tern Sterna sumatrana F 0 Bridled Tern Sterna anaethetus F 0 Osprey Pandion haliaetus U U U p.s 0-300 Pacific Baza Aviceda subcristata R R R R p 0-200 Brahminy Kite Haliastur Indus C F F U R p 0-1,100 White-bellied Sea Eagle Haliaeetus leucogaster R R 0-30 Variable Goshawk Acdpiternovaehollandiae U U U U p.s 0-200 Rufous-necked Sparrowhawk Accipiter erythrauchen R 30 Spotted Kestrel Falco moluccensis U U U U R p 0-1,050 Oriental Hobby Falco severus R 350 Australian Hobby Falco longipennis R p 0 Brown Booby Sula leucogaster U 0 Little Pied Cormorant Phalacrocorax melanoleucos F p 0-160 Forktail 29(2013) A survey of the avifauna of Obi island, North Moluccas, Indonesia 137 Species 1 2 3 Locations 4 5 6 Legend Altitudinal range (m) Little Black Cormorant Phalacrocoraxsulcirostris R 30 Little Egret Egretta garzetta U U U 35 Pacific Reef Egret Egretta sacra R 0 Great-billed Heron Ardeasumatrana R 0 Great Egret Casmerodius albus F U U P 0-160 Intermediate Egret Mesophoyx intermedia R P 30 Rufous Night Heron Nycticoraxcaledonicus U U u P 0-30 Black Bittern Dupetorflavicollis R R 0 Great Frigatebird Eregata minor R 0 Lesser Frigatebird Fregata ariel F F 0 Red-bellied Pitta Pitta erythrogaster F F F U p,s,m 0-1,100 Dusky Myzomela Myzomela obscura U U U U U p 0-1,550 Golden Whistler Pachycephala pectoralis C U F c p,s,m 30-1,550 Drab Whistler Pachycephala griseonota F F F F F p,s,m 30-1,550 Torresian Crow Corvusorru U U U P.s 0-200 Paradise-crow Lycocorax pyrrhopterus F C C U C p,s,m 30-1,550 White-bellied Cuckooshrike Coracina papuensis U F F F P.s 0-350 Slender-billed Cicadabird Coracina tenuirostris U U P-5 30-150 Pale Cicadabird Coracina ceramensis R F P 30-1,550 Rufous-bellied Triller Lalage aurea F F C F R P.S 0-1,100 Willie-wagtail Rhipidura leucophrys F U U F P.S 0-350 Northern Fantail Rhipidura rufiventris U F F F C p,s,m 0-1,350 Rufous Fantail Rhipidura rufifrons F s 860-1,550 Hair-crested Drongo Dicrurus hottentottus F F F U F s 0-1,550 Island Monarch Monarcha cinerascens R 200 Spectacled Monarch Monarcha trivirgatus U F F U F P.S 0-1,100 Slaty Monarch Myiagra galeata U F F F F P.S 0-1,150 Shining Monarch Myiagra alecto F F F R P.S 0-950 Island Flycatcher Eumyias panayensis R F P.S 300-1,150 Island Starling Aplonis mysolensis U U F C P.S 0-350 Shining Starling Aplonis metallica C C C p,s,m 0-30 Barn Swallow Hirundo rustica U U u 0-350 Pacific Swallow Hirundo tahitica F F F 0-160 Golden Bulbul Alophoixus affinis F C C U C p,s;m 0-1,150 Mountain White-eye Zosterops montanus F s 930-1,150 Cream-throated White-eye Zosterops atriceps F P.S 500-1,150 Mountain Tailorbird Orthotomus (Phyllergates) cucullatus U P.S 950-1,250 Island Leaf Warbler Phylloscopus poliocephala C P.S 500-1,550 Flame-breasted Flowerpecker Dicaeum erythrothorax U F U u P.S 0-1,350 Black Sunbird Nectarinia aspasia C C C C c P.S 0-1,150 Olive-backed Sunbird Nectariniajugularis C U F F P.S 0-350 Eurasian Tree Sparrow Passer montanus C C p 0-35 Black-faced Munia Lonchura molucca U C C P.S 0-35 138 SHORT NOTES Forktail 29(2013) The seasonality of mixed-species bird flocks in a Sri Lankan rainforest in relation to the breeding of the nuclear species. Orange-billed Babbler Turdoides rufescens ASHOKA JAYARATHNA, SARATH W. KOTAGAMA & EBEN GOODALE Introduction The seasonality of mixed-species bird flocks varies dramatically across the world. At one extreme are temperate flock systems that only form during the winter months or during migration (Morse 1 970), and even some flocks in the subtropics appear to be formed mostly of migrant species (Ewert & Askins 1 991 , Gram 1 998). At the other extreme are tropical systems that occur throughout the year and in which migrant species play a small role (Kotagama & Goodale 2004). Interestingly, however, there are some tropical systems that do show seasonal fluctuations in flock size and composition (Davis 1 946, Develey & Peres 2000). These fluctuations could be influenced by the underlying density of arthropods and/or by the breeding season of the species involved (Develey & Peres 2000). One open question is how flock systems are affected by the breeding season of a 'nuclear species', defined as a species that is important to the formation and/or maintenance of flocks (Moynihan 1962). Munn (1984) studied this question in Peru (see also Munn & Terborgh 1979) and found that flocks continue to function throughout the breeding season of the nuclear species — Bluish-slate Antshrike Thamnomanes schistogynus — with breeding individuals flying far from their nests in order to join flocks. Munn's system was somewhat atypical, however, in that the nuclear species was not particularly gregarious; in most flock systems, nuclear species are highly gregarious (Goodale & Beauchamp 201 0), and in some Asian systems one species can compose a large percentage of the flock (Chen & Hsieh 2002). What happens to flocks that form around such gregarious species when these species breed? In previous work on a flock system of a tropical rainforest in Sri Lanka, we have shown that flock size is seasonally stable, with only a few migrant species joining flocks in the winter months (Kotagama & Goodale 2004). However, we never measured seasonal changes in the density of flocks, so it is possible that flocks might actually still show seasonal fluctuations. This flock system is led by the nuclear species, the Orange-billed Babbler Turdoides rufescens, for which little breeding information is available. Therefore, we had two objectives in this study: (a) to measure seasonal changes in the density of flocks, and (b) to see if that seasonality was related to the breeding of the Orange-billed Babbler. We also aimed to chronicle some aspects of the nesting of this little-studied babbler (Henry 1998). Study site This study was conducted in the north-western sector of the Sinharaja World Heritage Reserve (6.433°N 80.350°E), Sri Lanka's largest remaining patch of lowland rainforest (450-600 m). This sector of the reserve was logged in the 1970s and the effects, including large gaps, are still visible near the main logging road, along which we walked. Annual rainfall is about 4 m with distinct dry (January to March) and wet seasons (April to December) (Kotagama & Goodale 2004). Methods To determine whether the density of flocks changes seasonally, we walked along 8 km of the main logging road that leads from the town of Kudava towards the mountain of Sinhagala. From November 2004 to December 2006, we walked this route three times a month at07h30-1 1 hOO.The months of May and June were only sampled in one year each, because of extremely wet conditions. We recorded all flocks seen or heard within 50 m of this transect, and estimated the number of Orange-billed Babblers present in the flock. Returning along the same route, we watched for any indication of breeding by Orange-billed Babblers, including mating, nesting or feeding of fledged chicks. In conducting two-tailed t-tests, we used a method that does not assume equal variances (Ruxton 2006). Means are given ± one standard deviation. Results Seasonal density of flocks We recorded 492 flocks during the sampling period. In all of these, Orange-billed Babblers were present. Apparently, our transect method, which required relatively fast walking, failed to detect the approximately 8% of flocks that do not contain Orange-billed Babblers (Kotagama & Goodale 2004) — these flocks that do not include the noisy Orange-billed Babblers can be quite cryptic and difficult to detect. The density of mixed-species flocks did not differ seasonally (Figure 1). The highest density of flocks (1.04 flocks per km) was recorded in June, and the lowest (0.52 flocks per km) was recorded in March. There was no significant difference between the density of flocks in the dry season from January to March (1 8 days sampled) and the rest of the year from April to December (54 days sampled), (t-test, t2737 = 0.71, P = 0.48). Figure 1 . Seasonal variation in the density of mixed-species flocks and single-species Orange-billed Babbler flocks. Numbers in parentheses show the number of days sampled for each month during the study. Error bars indicate +1 standard deviation (data from November 2004 to December 2006). No. flocks per km ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ * -^VVV aVVV j? ^ G>> 0> ^